in ARIZONA RESOURCES the and

in ARIZONA RESOURCES the and

VOLUME 9

HYDROLOGY and WATER

RESOURCES in ARIZONA

and the

SOUTHWEST

PROCEEDINGS OF THE 1979 MEETINGS

OF THE

ARIZONA SECTION

AMERICAN WATER RESOURCES ASSN.

AND THE

HYDROLOGY SECTION

ARIZONA - NEVADA ACADEMY OF SCIENCE

APRIL 13, 1979, TEMPE, ARIZONA

VOLUME 9

HYDROLOGY and WATER

RESOURCES in ARIZONA and the

SOUTHWEST

PROCEEDINGS OF THE 1979 MEETINGS

OF THE

ARIZONA SECTION

-

AMERICAN WATER RESOURCES ASSN.

AND THE

HYDROLOGY SECTION

-

ARIZONA - NEVADA ACADEMY OF SCIENCE

APRIL 13, 1979, TEMPE, ARIZONA

Previous Volumes

Preface

TABLE OF CONTENTS

Officers of the Arizona Section AWRA v vi vi

HYDROLOGY AND WATER RESOURCE TECHNOLOGY

Augmenting Water Supply for Home Irrigation (Poster Session)

Barney P. Popkin

Winter Precipitation on a Southeastern Arizona Rangeland Watershed

H B

Osborn, R.B. Koehler, and J.R. Simanton

A Multiattribute Approach to the Reclamation of Stripmined Lands

Fritz H. Brinck, Lucien Duckstein, and John L. Thames

Sediment Production from a Chaparral Watershed in Central Arizona

Thomas E. Hook and Alden R. Hibbert

An Exchange System for Precise Measurements of Temperature and

Humidity Gradients in the Air Near the Ground

L.W. Gay and L.J. Fritschen

An Interactive Model of Suspended Sediment Yield on Forested

Watersheds in Central Arizona

William O. Rasmussen and Peter F. Ffolliott

A Water Budget for a Semiarid Watershed

Severo R. Saplaco, Peter F. Ffolliott, and William O. Rasmussen

How to Select Evapotranspiration Models (Abstract only)

T E A van Hylckama, R.M. Turner, and O.M. Grasz

Ground Water in the Santa Cruz Valley

Marshall Flug

Tests on Arizona's New Flood Estimates

.

.

Brian M. Reich, Herbert B. Osborn and Malchus C. Baker, Jr.

Solar Powered Irrigation Pumping Experiment

Dennis L. Larson and C.D. Sands II

15

21

31

37

43

49

57

59

65

75

1 iii

INSTITUTIONAL ASPECTS OF WATER RESOURCES

Health Effects of Application of Wastewater to Land

James D. Goff

Early Public Involvement in Federal Water Resource Projects

Freda Johnson and Michael Thuss

Negotiating the Water Future of Pima County, Arizona

Michael F. Thuss

79

85

91

Hydrologic Investigation of the Dry Lake Region in East Central

Arizona.

. James J. Lemmon, Thomas R.

Schultz, and Don W. Young

97

Visual Impacts:

Perception and Modification of Surface Mining

Operations on the Black Mesa

Jon Rodiek

105

Impact of Development on Stream Flows

Paul D. Trotta, James J. Rodgers, and William B. Vandivere 109

.

.

Trends in Arizona Water Service Organizations: A Comparative Summary

Jacque L. Emel, Michael D. Bradley, and K. James DeCook

119

.

.

.

.

An Examination of the Buckhorn-Mesa Watershed Environmental Impact

Statement (U.S.D.A., S.C.S., 1978):

A Look at State -of -the Art

Reports

Dale A. Altshul

125

Land Use Planning for the San Tiburcio Watershed

Roberto Armijo and Robert Bulfin 133

Evaluation of Water Management Systems for the Sonoita Creek

Watershed

Hugh B. Robotham

145

The Effects of Second -Home and Resort -Town Development on Stream

Discharge in Navajo and Apache Counties, Arizona

T D

Hogan and M.E. Bond

153

Central Arizona Project Concept of Operation

Frank C. Springer, Jr., P.E. and Albert L. Graves

159

The Impact of Socioeconomic Status on Residential Water Use:

A Cross- Section Time -Series Analysis of Tucson, Arizona

R. Bruce Billings and Donald E. Acthe

167 iv

TO ORDER COPIES OF:

HYDROLOGY AND WATER RESOURCES IN ARIZONA AND THE SOUTHWEST

Volume 1, Proceedings of the 1971 Meetings, Tempe, Arizona

Edited by Dr. D.D. Evans (29 papers)

Volume 2, Proceedings of the 1972 Meetings, Prescott, Arizona

Edited by Dr. Martin M. Fogel (28 papers)

Volume 3, Proceedings of the 1973 Meetings, Tucson, Arizona

Edited by Dr. Ron S. Boster (25 papers)

Volume 4, Proceedings of the 1974 Meetings, Flagstaff, Arizona

Edited by Dr. Ron S. Boster (30 papers)

Volume 5, Proceedings of the 1975 Meetings, Tempe, Arizona

Edited by D.L. Chery, Jr.

(27 papers)

Volume 6, Proceedings of the 1976 Meetings, Tucson, Arizona

Edited by Dr. D.L. Chery, Jr. (44 papers)

Volume 7, Proceedings of the 1977 Meetings,

Edited by Linda M. White (32 papers)

Las Vegas, Nevada

Volume 8, Proceedings of the 1978 Meetings,

Edited by Dr. T.R. Verma (29 papers)

Flagstaff, Arizona

Volume 9, Proceedings of the 1979 Meetings,

Edited by Drs. Gerald Harwood and K.J.

Tempe, Arizona

DeCook (24 papers)

Out of print

$8 per copy

$8 per copy

$10 per copy

$10 per copy

$10 per copy

$10 per copy

$10 per copy

$12 per copy

SEND REMITTANCE TO:

WATER RESOURCES RESEARCH CENTER

Attention:

Arizona Section, AWRA

University of Arizona

Tucson, Arizona 85721

MAKE CHECKS PAYABLE TO:

ARIZONA SECTION, AMERICAN WATER RESOURCES ASSOCIATION

1979 members of the Arizona Section, American Water Resources Association, are entitled to a $2.00 discount on the 1979 volume and $1.00 discount on all past volumes.

Information on membership can be obtained from the AWRA Arizona Section, in care of K.J. DeCook, Executive Secretary, Water Resources Research Center,

Tucson, Arizona 85721.

v

PREFACE

Volume 9 of Hydrology and Water Resources in Arizona and the Southwest is a

compilation of the papers presented at the annual meetings of the Arizona Section of the American Water Resources Association.

The 1979 meetings were held at

Arizona State University, Tempe, April 13, 1979. As in previous years, the meetings

were held jointly with the Hydrology Section of the Arizona- Nevada Academy of

Science.

Twenty -four papers are presented in this volume.

The proceedings reflect a

broad range of topics of interest to hydrologists, planners, engineers, water

resource managers, politicians and citizens alike.

The author affiliation is indicated on each paper.

The primary purpose of these meetings is to enhance communication among various groups and individuals involved in water resources related problems, concerns and matters within the Region.

Anyone in the Region interested in being affiliated with this endeavor may contact the Executive Secretary for information on section activities and membership.

The officers of the Arizona Section of the American Water Resources Association express their thanks to Dr. Charles E. Downs of Arizona State University, Tempe,

Arizona, who assisted in setting up the meetings.

Gerald Harwood, Ph.D., Editor

School of Renewable Natural Resources

University of Arizona

Tucson, Arizona

President

Vice President

OFFICERS OF THE ARIZONA SECTION,

AMERICAN WATER RESOURCES ASSOCIATION

1979

Executive Secretary

Editors of Proceedings

Don Young

Arizona Water Commission

222 N. Central Ave.

Phoenix, AZ

85004

Kennith E. Foster

Office of Arid Lands Studies

University of Arizona

Tucson, AZ

85719

K James DeCook

Water Resources Research Center

University of Arizona

Tucson, AZ 85721

Gerald Harwood, University of Arizona

K. James DeCook, University of Arizona vi

AUGMENTING MATER SUPPLY FOR HOME IRRIGATION

Barney P. Popkin

ABSTRACT

Low rainfall and humidity, and high evapotranspiration, make irrigation necessary for domestic plant growth in the American Southwest.

Irrigation supplies are limited.

A large percentage of potable water used in Southwestern homes is used for home irrigation.

Another large percentage Is returned to sewers.

Water and sewer fees are increasing because of rapid urban expansion and increased water -quality standards.

As fees increase, supplemental home irrigation sources become attractive and are sought.

Major supplemental water sources are grey water, harvested runoff, and roof runoff.

The amount of grey water depends on family size and habits.

The amount of harvested runoff depends on land size and slope, soil's and material's properties, and rainfall.

The amount of roof runoff depends on roof size and geometry, and rainfall.

The quality of these sources is generally suitable for home irrigation.

Engineering systems are required to use supplemental home irrigation water.

systems will have low capital expenditure and low energy requirements.

The most preferred

A large and significant reduction in municipal costs and services is possible if supplemental home irrigation water is developed.

Small -scale analysis indicates that costs are favorable for supplemental irrigation systems.

A suggested research program emphasizes field trials and demonstrations which test design, operation, maintenance, and economics, as well as public and institutional acceptance.

INTRODUCTION

This paper is a written summary of a poster presentation prepared for the 23rd Annual Meeting of the Arizona- Nevada Academy of Science.

Essential poster items are presented as figures to this text.

Because the figures are self- explanatory, each is not reviewed, unless additional clarification is required.

OBJECTIVE

The objective of this report is to outline supplemental water sources for home irrigation with respect to water availability (quantity), water quality, engineering systems, and economics.

WATER USE

Home irrigation accounts for about 65 percent of domestic water use in the Southwest.

percent of the domestic wastewater load is grey water.

About 60

Figures 1, 2, and 3 show domestic water use and wastewater load, estimated regional uses and load, and estimated regional grey water, based adaptations from published sources (1, 2, 3).

85721

Research Assistant, Water Resources Research Center, The University of Arizona, Tucson, Arizona.

1

SUPPLEMENTAL SOURCES

Supplemental sources include grey water (non -toilet home wastewaters), harvested runoff (water produced from prepared land surfaces), and roof runoff (water produced by rainfall that drains a roof).

AVAILABLE WATER

Figure 4 shows the irrigation possibilities of usable domestic wastewater.

Figure 5 shows the quantity of available irrigation, and may be used to estimate irrigation depths from all supplemental water sources.

Estimates of water use and water supply can be made from (3), (4), and (5).

WATER QUALITY

Figures 6 and 7 show water -quality characteristics and treatments for selected domestic wastewater.

Harvested runoff will be low in salts, but possibly high in suspended solids and turbidity.

Roof runoff will have the best quality, though it may show some turbidity.

A combination of crop and soil filtration and uptake, dilution, filtration, and storage account for all identified water -quality problems.

Water- quality aspects are reviewed in (6).

ENGINEERING SYSTEMS

Figure 8 summarizes engineering systems required to use supplemental irrigation water.

designs may be found in (7), (8), and (9).

Practical

ECONOMICS

These may be viewed as large scale and small scale.

Large Scale

If potable water usage can be reduced by reducing irrigation use of potable water, a savings in municipal services is achieved.

If 65 percent of the used potable water is for home irrigation, it might represent up to 65 percent of the water bill.

That amount of water, and its charge, can be reduced by 68 to 100 percent, depending on use and rate structure.

If much of this reduction is due to grey -water reuse, sewer -use fees also could be reduced.

And, if much of this reduction is due to harvested and roof runoff, then street flooding and storm -runoff costs may be reduced.

A large reduction in municipal costs and services is possible if supplemental home irrigation water is developed, and water -use and sewer -fee costs are commensurate with water and sewage reductions.

Municipalities could save money from savings in water supply, sewage treatment, and flood damages.

Small Scale

Figure 9 shows minimum repayment periods for integrated three -day storage systems for supplemental irrigation waters.

These are for four illustrative cases, and two water sources, with ten adjustments.

They assume that each trailer has a 12- by 60 -ft roof, that the apartment complex has a 6,250 sq ft roof, and that the double and single family homes each have a 1,500 sq ft roof.

They assume that the trailer park, apartment complex, and each home occupies 7.92, 0.41, and 0.18 ac, and that 12, 40, and 4 percent of each is paved.

It's assumed that a water- harvest efficiency of 50 percent applies to unpaved areas, and rainfall is ten in /yr.

It's also assumed that irrigation will occur otherwise exclusively with fresh water.

calculations allow a ten -percent mortgage interest.

All repayment

2

The value of water is based on 1978 City of Tucson fees of about $3.56/1000 cu ft.

Because of adjustments in sewer -use fees, we assume that grey water and laundry effluent are worth $9.94/1000 cu ft, while harvested and roof runoff are worth only $3.56/1000 cu ft.

Table 1 shows comparative unit costs of supplemental water, for the four illustrative cases, with two water sources, and three storage systems.

1.

A small -scale economic analysis indicates:

The unit cost for all sources increases with decreasing number of people, and land and roof size.

2.

3.

4.

The cheapest source of supplemental water Is most grey water.

The most expensive source is harvested runoff, exceptwhere dispersed trailers produce more expensive roof runoff.

Integrated, short -term storage, considerably reduces unit and total costs.

SUMMARY

It's summarized that:

1.

2.

3.

4.

5.

Grey water, harvested runoff, and roof runoff are supplemented water sources for home irrigation in the Southwest.

Needed engineering systems can be flexible and innovative.

Costs are favorable for future large -scale supplemental water systems.

Costs become more favorable as water -use and sewage -fee charges increase.

Costs can be reduced considerably by: a.

Integrating collection, storage, and distribution facilities.

b.

Providing three -day water storage.

c.

Using second -hand materials.

d.

Exploiting home -owner labor.

However, southwestern municipalities, states, and the federal government have discouraged supplemental water use.

Municipalities, such as Tucson, have minimum high charges for water and sewer use.

(Sewer -use fees are often computed as a flat percentage of water use, plus a minimum charge.)

States, such as Arizona (10), explicitly outlaw domestic wastewater reuse.

provided inadequate research funds to this area.

And the government has heretofore

RESEARCH

1.

2.

3.

4.

5.

A research program is badly needed (11, 12).

It should:

Emphasize field trials and demonstrations.

Test design, operation, and maintenance.

Evaluate economics.

Address public and institutional acceptance.

Review legal problems.

ACKNOWLEDGMENT

I thank Sol

D. Resnick for his support in this project.

K. J. DeCook was also supportive.

C.

3

B. Ciuff provided useful conversation.

Many home owners and maintenance engineers were helpful.

M.

Busse, L. Donald and N. Svacha drafted and typed this material.

REFERENCES CITED

1.

2.

3.

4.

5.

6.

7.

8.

9.

Murray, C. R. & E. B. Reeves.

Estimated use of water in the United States in 1970.

USGS Cir. 676.

Wash., (1972). 37 p.

Todd, D. K.

The Water Encyclopedia:

A compendium of Useful Information on Water Resources.

Information Center, The Maple Press, Port Washington, (1970), 559p.

Water

University of Arizona.

Water Conservation for Domestic Users:

Desert Climates.

City of Tucson, Tucson, (1977), 32p.

with Special Reference to Warm

Frasier, G. W. ed.

Proceedings of the Water Harvesting. Symposium, Phoenix, Ariz.. Mar. 26 -28,

1974.

USARS ARS W -22, (1975), 329p.

Fink, D. H. & W. L. Ehrler.

Salvaging wasted water for desert -household gardening.

Water Resources of AZ and the Southwest, (1978). 8:125 -131.

Hydrology and

Popkin, B. P.

Land application of selected home wastewater.

Presented to the 1978 Summer Meeting,

ASAE Paper No. 78 -2062, UT State Univ., Logan, (1978). 20p.

Farallones Institute.

Grey Water Use in the Nome Garden.

St., Berkeley, CA, 94710, 1978, 22p.

The Integral Urban House, 1516 Fifth

Murray, M.

Residential Water Conservation.

CA, Davis, (1976411$p.

CA Water Resources Center Rept. No. 35, Univ. of

Winneberger, J. H. T.

Manual of Grey Water Treatment Practice.

Ann Arbor Science, Ann Arbor,

(1974), 102p.

10.

Arizona Revised Statutes.

Paragraphs 36 -1854 to 1859 on Public Health and Safety, and Water

Pollution.

11.

U.S. Dept. Interior.

Executive Summery of the Critical Water Problems Facing the Eleven Western

States, Westwide Study.

U.S. Dept. Interior, Wash., (1975), 85p.

12.

Resnick, S. D. Mar. 14, 1979.

Memorandum to Persons Interested in Water Resources Research.

Water Resources Research Center. Univ. of AZ. Tucson. (Includes attachemnts from U.S. Dept.

Interior, Office of Water Research and Technology on Priority Programs for Water Research for

Fiscal Year 1979.]

4

FIGURE 5

QUANTITY OF AVAILABLE IRRIGATION

IN INCHES OF DEPTH

F-

1-L1

10,000

LI:1 5,000w

OD

I,000-

Z 500-

J

1 l

100-

50-

10100

`

500' '1,000

5,600 10,000

50,000 '

AMOUNT OF IRRIGATION WATER APPLIED

IN CUBIC FEET

9

-

F--

CI)

d

X

Q

Ct

W

~

V

I

U

A11a188n1 sallos 030N3dS11S wnlaos

SdVOS

Al1NI1VS

31VHdS0Hd aNVW30 N33AX0

8311V IN OINV92:10

3SV31d0 aNV 110

8o00

3.11/8/IN

831VM lOH

Hd HJIH

aIVH

S310118Vd o003

WVOd

3NI801HO

H0V318'

VI8310V8

Cr

W

W

Q w

V= u op3

á

is=

Ó Ó

Ú cc

aow

F-=

Mtn

<11-o w

>

V

Z z

Ú

Z

z

w

C9

2 8

3a-

N

10

Cr

Q

>

J

CO

Q

A1t0188n1

S011OS 030N3dSnS wnlaos

SddOS

AlINI-lb'S

31dHdS0Hd

Hd

ONdlN30 N30AX0

8311`dIN OINd980

3St/389 ONb 110

8000

31b81IN

831b'M l0H

5310118t/d 000d wvo

3N1801H0

H0b'318

b'12i310b8

IJ

-

H

I i-

Q

ILI

Q

-

-

Z Z

O

-1--17-a

_J

°

2

J d

°

U

m cc

U

-

z z

JQ

J

O8

UJ d

``

Z

U

w

:1

= i

Z z

F~-

_J

``

_.1

O

!n

o co wY

J j

ía-cl

O

Cw

Ñ

TABLE 1

Comparison of Dollar Costs of Supplemental Water

For 1,000 Cubic Feet Over a Ten -Year Period

WATER

SOURCE

A

B

A

B

A

B

TRAILER

PARK

8.44

18.72

0.69

1.41

0.30

0.62

APARTMENT

COMPLEX

8.90

24.37

0.95

2.19

0.42

0.98

DOUBLE

FAMILY

HOUSE

30.50

77.95

2.45

7.10

1.12

3.20

SINGLE

FAMILY

HOUSE

48.73) Seven -day Nonintegrated

103.37

Storage

3.98) Seven -day Integrated

7.74

Storage

1.82) Three -day Integrated

3.52

Storage

A: Most grey water, and harvested and roof runoff.

B: Laundry effluent, and harvested and roof runoff.

14

WINTER PRECIPITATION ON A SOUTHEASTERN ARIZONA RANGELAND WATERSHED1

H. B. Osborn, R. B. Koehler, and J. R. Simanton2

INTRODUCTION

A principal research effort on the USDA Walnut Gulch experimental watershed in southeastern Arizona

(Fig. 1) has been to identify and quantify rainfall and runoff from summer convective storms (Osborn and

Renard, 1970; Osborn and Laursen, 1973).

A dense network of over 90 recording raingages (Fig. 2) has been installed for this purpose (Renard, 1970).

However, winter precipitation as rain or snow is an important source of rangeland moisture for spring growth of many species of grasses, shrubs, and forbs that are grazed by livestock.

The growth of many of these species is controlled by the amount and timing of winter precipitation (Cable and Martin, 1975).

I

In this paper, winter precipitation data from selected Walnut Gulch gages were analyzed using varying durations and gage spacing.

The effects of elevation were investigated.

Also, the long term

ARIZONA;

precipitation record for Tombstone, centrally located on the Walnut Gulch watershed, was analyzed for intercorrelations of seasonal precipitation, trends, representativeness of the shorter Walnut Gulch record, and occurrence of maximums and minimums within the long term record.

7<.

WINTER PRECIPITATION i'COCHI S E

)

.1

WALNUT GULCH EXPERIMENTAL SATERICEN

PERIPHERAL AREA

) iR»,,...

COUNTY

WATERSHED A EA OETAILEO

Annual precipitation data were divided by season

-winter and summer.

months of

Winter includes the

November and December of one year, and

January through March of the next year.

Winter precipitation in the area is characterized by long duration, low intensity, frontal storms covering large areas (Sellers, 1960).

Figure 2 is an ìsohyetal map of 1965 -1966 winter season precipitation.

December and January are the wettest winter months, and March is the driest.

Less than 10% of winter precipitation is snowfall.

Summer precipitation includes May through September.

It is usually of short duration, high intensity, and limited areal extent (Sellers,

1960; Osborn and Reynolds, 1963).

July and August are the wettest months, and May the driest

(Table

1).

Figure 1.

Location of Walnut Gulch Watershed.

Table 1.

Monthly and seasonal precipitation amounts for Tombstone, AL (1904-

1978) (mm).

May Jun

Jul

Aug Sep] Oct

4.8 12.2

93.2

87.9

37.3

1

19.3

Nov

16.3

Dec Jan

20.8

20.1

Feb

19.6

Mar Apr

16.0

6.9

Seasonal

Mean 235.5

6.9

Range

110.0 -419.6

19.3

-90.2

92.7

19.3 -287.5

0 -4b.0

1.

Contribution of USDA, SEA, AR, Tucson, Arizona 85705.

2.

Supervisory Hydraulic Engineer, Hydraulic Engineering Aid, and Hydrologist, respectively, USDA Southwest Rangeland Watershed Research Center, Tucson, Arizona 85705.

15

0

,-- a

,.e-

----,

° f

ó

I12' 9

( 68

ó

o a

'~2

\

..,/

-o--

0 o

14

° o

0 o

° o a

0

14 o a o o o o o

J o

'

9(

,

; iv 13

/

¡r/ o Ralnpape

Painpope owed in correlOtion analysis

IN lem 0

11un

Note -Depth are in mm

Figure 2.

Raingage network and typical winter precipitation pattern (mm) at Walnut Gulch.

Walnut Gulch

Raingage spacing for representative sampling has been shown to be critical for summer precipitation in areas where thunderstorms are dominant.

Osborn, Lane, and Myers (1979) showed that for storm rainfall, correlation coefficients (r) between gages dropped below 0.9 at a distance of around 1 km; thus, a dense network of gages is needed to quantify storm precipitation.

termine necessary spacing for winter storms.

A similar study was initiated to de-

Correlation coefficients between gages for winter precipitation were determined for pairs of 9 selected gages for 1 -hr, 6 -hr, 24 -hr, and seasonal depths.

The relationships between correlation and distance are shown in Fig. 3 -4.

for 1 -hr depths (Fig. 3) is difficult to define.

The relationship between gages

The most likely explanation is that, since manly of the winter precipitation amounts are quite small, errors in measurement become relatively large.

The correlation for 6 -hr precipitation (Fig. 4) is better defined, indicating a rapid drop in correlation up to about

1 km, with the curve relatively flatter thereafter.

The relationship for 24 -hr depths suggests a steady decay in correlation with distance.

For seasonal estimates, one gage on the 150 -km probably would be satisfactory.

All curves were fitted by eye.

watershed

The 6 -hr, 24 -hr, and seasonal curves for winter events are compared directly with summer curves for the same durations (Fig. 4).

All winter event curves are significantly higher than their summer counterparts.

Four Walnut Gulch gages at different elevations were selected for an analysis of elevation effects on winter precipitation depths.

The maximum 1 -, 6 -, and 24 -hr, 100 -yr winter storms were predicted from

20 yrs of record at these gages using the Gumbel extreme value procedure (Gumbel,1958)(Table 2).

There

Table 2.

100 -yr,

1 -, 6 -, and 24 -hr winter precipitation amounts for four selected gages on Walnut Gulch.

Gage (elevation)

Duration Statistic

2

(1261 m)

1 -hr

X

S

100-yr

5.0

2.4

14.2

6 -hr

X

S

100-yr

11.9

4.8

30.2

36

(1466 m)

(nm)

5.1

2.9

16.5

11.9

4.6

30.0

16

60

(1522 m)

5.0

2.6

15.0

10.7

3.0

22.4

68

(1580 m)

4.6

2.4

14.0

10.2

3.6

23.9

Table 2.

(Continued).

Duration

24 -hr

Statistic

2

(1261 m)

X

S

100-yr

19.1

8.6

53.1

Gage (elevation)

36

(1466 m)

60

(1522 m)

(mm)

18.0

6.4

42.4

16.5

5.3

37.1

10

68

(1580 m)

16.3

3.8

31.5

3

5 10 15

Distance between raingages (km)

20 25

Figure 3.

Correlation between gages and distance for 1 -hr rainfall depths for winter precipitation on Walnut Gulch.

Seasonal

24-hr

5 10 15

Distance between raingoges (km)

20 25

Figure 4.

Correlation between gages and distance for 6 -hr, 24 -hr, and seasonal rainfall depths for winter and summer precipitation at

Walnut Gulch.

17

were no significant differences in 1 -hr, 100 -yr values among the four stations.

There was some indication of a decrease with increasing elevation for the 6 -hr predictions, and a significant decrease with increasing elevation for the 24 -hr predictions.

In the Gumbel distribution, predictions are influenced by variance, and the variance was less at higher elevations.

This could be misleading.

However, mean values also decreased with increasing elevation.

At the least, we have no evidence that extreme events have greater depths at higher elevations.

Also, similar tests of summer total storm rainfall amounts indicated no increase with increasing elevation.

Tombstone

Cooperative National Weather Service daily rainfall records have been kept for Tombstone since the early 1890's.

However, in early years, the record was irregular, and only 75 yrs of data (1904-

1978) were used in this study.

During the period, summer and winter precipitation were 66% and 26%, respectively, of annual precipitation, with April and October rainfall 8% of the annual mean.

Annual summer rainfall ranged from 110 to 420 mm, whereas annual winter rainfall precipitation ranged from 20 to

290 mm.

Five -year moving means of summer, winter, and annual precipitation are shown in Fig.

5 (dashed lines from 1904 through 1907 represent means of less than 5 yrs).

Partial records before 1904 suggest heavy winter precipitation around 1900, with 1904 -1905 the last winter of a reported "wet" period.

Winter precipitation in 1904 -1905 was 290 mm, which is the largest during the record period.

The moving mean for winter precipitation was well above average in the middle 1910's, slightly above average in the

1930's, and has been below average since 1942.

Since 1942, about 1/3 of the winter seasons have been above average (although none exceptionally high), but not for enough consecutive years to raise the moving mean above the average.

In contrast, moving means for summer and annual precipitation were above average as recently as the late 1950's and have been below average since 1961.

Intensive studies were initiated on small Walnut Gulch watersheds and plots in 1962, and moving means for winter, summer, and annual precipitation have been below average since then.

However, just over 150 mm of winter precipitation were recorded in 1977 -78, which was the highest since 1919.

25

20

1

- Annual

-- Summer (Moy -Sept)

-'--

Winler (Nov- Moroh) t 15

Al()

`

/

~ /

/

\

--`,/

5

600

500

---

V- -

200

400

Ê

E

300 ñ

100

O

1900

1910 1920 1930

1940

Year

1950 1960

1970 1980

0

Figure 5.

Five -year moving means for seasonal and annual rainfall for Tombstone,

Arizona (1904 -1908).

18

Correlation coefficients (r) among seasonal precipitation amounts for Tombstone (1904 -1978) are shown in Tables 3 and 4.

In Table 3, winter follows summer; in Table 4, summer follows winter.

The underlined values are within the 95% confidence limits for random selection, and indicate that predicting either summer from winter or winter from summer precipitation is impossible.

Also, there is no suggestion of serial correlation (summer, r = .05; winter, r = .17, and annual, r = .06) from year to year, although wetter and drier periods are indicated by the running means.

Table 3.

Correlation coefficients (r) among seasonal precipitation amounts for Tombstone (1904- 1978); winter follows summer.*

Summer

1.0

October Winter April Annual

Sumner

.01

.09

-.02

.62

October

1.0

.30

.02

.22

Winter

April

1.0

.28

.44

1.0

.16

Annual

*Underlined values are within random error of selection.

1.0

Table 4.

Correlation coefficients (r) among seasonal precipitation amounts for Tombstone (1904- 1978); summer follows winter.*

Summer

1.0

October Winter

April Annual

Summer

October

.01

-.12

-.04

-.11

-.05

.73

Winter

April

1.0

1.0

.29

.22

.54

1.0

.19

Annual

*Underlined values are within random error of selection.

1.0

OBSERVATIONS tion.

Annual winter precipitation seems to be at least as variable from year to year as summer precipita-

Because of the larger areal extent of most winter storms, one raingage probably could be used to give an accurate estimate of annual winter precipitation on the 150 km4 Walnut Gulch experimental watershed.

The data are not as clear cut for winter storm analysis, but we did decide that a well- spaced 8gage network plus several gages at locations where we were carrying out intensive studies on very small areas would be satisfactory (Fig.

2).

Part of our decision, however, was because there have been very few runoff- producing winter storms on Walnut Gulch.

Analysis of gage correlation versus elevation suggested a possible decrease in precipitation with increasing elevation for major winter events.

However, a range of 320 m in elevation may be insignificant as compared with other climatic and topographic features.

The moving means for both summer and winter precipitation have remained below average since 1962, when our intensive small watershed and plot studies were initiated.

Several "lows" have been recorded since 1962, but no "highs."

Low winter precipitation probably has increased the stress on palatable spring -grazed vegetation and further improved the chances for survival of more deep- rooted unpalatable shrubs.

Studies carried out since 1962 are biased by the lower average seasonal precipitation, but only a sustained "wet period" will allow us to determine just how biased.

SUMMARY

Twenty years of winter precipitation data from the Walnut Gulch experimental watershed were analyzed.

Winter precipitation varied widely over the watershed for storm durations up to 6 hr, indicating that a network of gages, rather than a single gage, might be needed for winter as well as summer storms.

However, 24 -hr and seasonal precipitation among raingages were highly correlated, indicating that

1 raingage on the 150 km2 watershed probably would give a fairly accurate estimate of winter precipitation.

19

When 100 -yr return frequencies for Walnut Gulch winter precipitation were plotted against elevation, 1 -hr duration depths showed no trend, but 6- and 24hr storm totals suggested a decrease with increasing elevation.

A 5 -yr moving mean, based on 75 yr of record at Tombstone, Arizona (located on the Walnut Gulch watershed), showed the variability of winter, summer, and annual precipitation from year to year as well as major periods of above and below average precipitation.

This moving mean also indicated that both mean summer and mean winter precipitation have remained below average since 1962, when our intensive small watershed and plot studies were initiated.

There was no evidence of correlation between summer and winter precipitation.

However, there was a good correlation between summer and annual precipitation, and a weaker correlation between winter and annual precipitation.

There was no significant serial correlation between consecutive years of summer or winter precipitation.

REFERENCES CITED

Cable, D. R., and S. C. Martin.

Vegetation response to grazing, rainfall, site condition, and mesquite control on a semidesert rangeland.

USDA Forest Service Research Paper RM -149, Rocky Mt. Forest and

Range Expr.

Sta., Ft. Collins, Colo., July, 1975.

Gumbel, E. J.

Statistics of extremes.

Columbia University Press, 1958, 375 p.

Osborn, H,. B., L. J. Lane, and V. A. Myers.

Rainfall /watershed relationships for southwestern thunderstorms.

Trans. ASAE, 1979.

(In press)

Osborn, H. B., and K. G. Renard.

Thunderstorm runoff of the Walnut Gulch experimental watershed, Arizona, U.S.A.

Proc. IASH- UNESCO Symp. on Results of Research on Representative and Experimental Basins, New Zealand, IASH 96: 455 -464, 1970.

Osborn, H. B., and E. M. Laursen.

Thunderstorm runoff in southeastern Arizona.

J. Hydr. Div., Proc.

ASCE 99(HY7):1129 -1145, 1973.

Osborn, H. B., and W. N. Reynolds.

Convective storm patterns in the southwestern United States.

IASH 8(3):71 -83, 1963.

Bull.

Renard. K. G.

The hydrology of semiarid rangeland watersheds.

USDA -ARS 41 -162, 1970.

Sellers, W. D.

The climate of Arizona.

In Arizona Climate, by C. R. Greene and W. D. Sellers, Univ. of

Arizona Press, pp. 5 -64, 1960.

20

A MULTIATTRIBUTE APPROACH TO THE RECLAMATION OF STRIPMINED LANDS by

Fritz H.

Brinck (University of Arizona, Tucson, Arizona)

Lucien Duckstein (University of Arizona, Tucson, Arizona)

John L. Thames (University of Arizona, Tucson, Arizona)

ABSTRACT

A multiattribute utility function is used to model preferences on outcomes of alternative reclamation schemes for stripmined lands, using Arizona and Wyoming examples.

Each scheme should at least help restore land to its premining value, and is composed of three sets of actions: mining operations, preparations for postmining land use, and mitigating actions.

Grazing and runoff augmentation are examples of postmining land use goals, and mitigating actions may be measures to protect the environment like pollution control in runoff or infiltration.

Conflicting objectives are involved, including the maintenance of sufficient coal production, the alleviation of detrimental environmental effects, and the minimization of loss.

Since the environmental effects are fraught with uncertainty, a multiobjective decision -making scheme under uncertainty is set up to analyze the problem.

The decision model ranks alternative reclamation schemes on the basis of the preference function of a group decision maker, each member of which assessing a separate subset of single attribute utility functions.

INTRODUCTION

The purpose of this paper is to present a framework for applying multiattribute utility theory to the ranking of alternative schemes for reclaiming stripmined lands.

Stripmining,which is regulated by the federal Surface Mining Control and Reclamation Act of 1977, is not acceptable without actions to reduce negative impacts and actions to restore the lands to at least premining value (Imes and Wali,

1977).

Uncertain operating conditions that the decision maker cannot influence are called states of nature, and due to these uncertainties several outcomes of some realized scheme are random.

Accordingly, the evaluation of alternatives has to account for the probability distribution of outcomes rather than fixed outcomes.

The fact that a stochastic process may produce unwelcome outcomes under a distribution with otherwise acceptable expectation introduces risk into the decision making, and hence the need to account for risk attitudes.

The choice of a reclamation scheme may therefore provide a good example of an application of utility theory that accounts for probability distributions of the states of nature as well as for the risk attitude of the decision maker.

The research reported here on defining a multiattribute utility criterion is thought of as part of a comprehensive systems analysis, i.e. the type of trade -off analysis that ideally should be the basis for deciding on mitigating actions and postmining land use.

Much effort goes into the environmental impact statements to make it likely that the chosen scheme satisfies the regulations.

In the analysis reported here the regulations are not treated as strict constraints because the achievement of a certain attribute or performance level is subject to chance.

ulations on the basis of benefits of granting exemptions.

Operations may also be exempted from reg-

The paper is organized as follows.

In the next section, the problem is defined.

Then the multiple objectives are examined more closely, and in the following section the method is presented.

Then follows a description of the assessment of utility functions, and finally comes a discussion of the utility theory framework.

THE PROBLEM

A reclamation scheme has to be found that strikes a balance between several objectives.

Two fundamental objectives are to provide for sufficient coal supplies, and to afford acceptable protection of the water resources.

These objectives are conflicting because a desired level of coal production may have to be traded off against unwelcome environmental impacts.

An accepted reclamation scheme is therefore a result of trade -offs between various attributes or criteria related to the multiple objectives.

What is needed then is a scalar measure of preferability combining all attributes.

Such a measure will allow a ranking of alternative schemes.

21

An analysis to support decisions concerning reclamation of stripmined lands may be bisected into establishing a decision model and a preference model.

The relationship between the two models is illustrated in Figure 1.

This paper deals mainly with the preference model.

A set A of alternative reclamation schemes are considered for ranking a2,...) (1)

A scheme ai consists of three sets of actions; the set M1 of interim land use actions, the set Ci of mitigating actions, and the set Ri of postmining land use preparations.

Thus a1 = {M1, Ci, Rii where the subsets of ai each have specific actions as members.

(2)

Wstates of nature natural resources systems, local production, mining operations

X= F[Ax W outcomes group decision maker, multiple objectives, preference structure

G 0 preference function selection algorithm

A alternative schemes

Decision model best scheme

Ak

Preference model

Figure 1 Models dichotomy

Interim land use actions are obviously the surface mining operations, perhaps along with other activities.

Examples of mitigating actions are watering and chemical stabilization with respect to dust control during topsoil and overburden handling, utilization of emission control equipment, and disposal systems for solid and liquid wastes (Doyle, 1976).

Examples of postmining land use goals are livestock mine in Wyoming (USES 1977), and water harvesting on the Black Mesa in

To some degree the surface mining action set Mi is the same for different schemes

, which is also the case for the set Ci.

There will also be a certain amount of overlapping in that the design of a mining plan, i.e. the choice of mining method, equipment, and sequence of operations is a function of regulations and planned postmining land use.

For example the interfingering of surface mining operations and end use preparations is an important optimization problem as shown by Kirk (1978).

The ranking of alternative schemes, however, assumes that each alternative (M1, Ci, Rj) has an optimal representation as far as possible before ranking takes place.

Another example of an intra- scheme optimization study is given by Brinck 8 al. (1976) relating grading and furrowing to conditions for livestock grazing.

The output of the decision model in Figure 1 is the set X of system outcomes, which are the realized levels of the attributes of the reclamation scheme

22

X =(xl, x2,...)

(3) where xl is the realized level on attribute 1, outcome x2 is the realized level on attribute or performance criterion 2, etc.

Examples of outcomes are: coal production in tonnes per day, area in hectares over which aquifers are destroyed to a specified degree, wildlife habitat in no. of head of certain species, monetary loss, and livestock herd size grazing on the reclaimed lands.

The decision model comprises the model F of the technological and physical relations defining the state transitions.

Then X is the set of system states resulting from the transformation of the Cartesian product of A and W due to the state transition model F

(4)

X =F[Axi 1 where W is the set of states of nature

W =(w1, w2,...)

(5)

Obvious states of nature to be accounted for are random conditions affecting the pollution in runoff and infiltration, like precipitation and infiltration rates on strip mine spoils.

But also the enforcement of environmental protection rules in uncertain, and thus a member of the set of states of nature.

The preferences of the decision -maker are used to rank the schemes ai, i= 1,2,...

These preferences are given by the output of the preference model G as illustrated in Figure 1.

G is a preference mapping of outcomes or consequences of actions into the real numbers R (Bertsekas, 1976)

G: X-R

(6)

The decision on the best reclamation scheme a* thus has to be reached under uncertainty and by trading off multiple attributes.

The task is now to establish the scalar measure of preferability G to find a* among the set A according to a decision rule.

This decision rule says that the best alternative a* is that alternative ai which maximizes the preference function G a * =ai, iff G(ai) =max G(xl(a),x2(a),..) aeA

(7)

In our case G is the multiattribute utility function u

(8) to be explained in a separate section.

GROUP DECISION MAKER AND MULTIPLE OBJECTIVES

In this section the decision maker is defined as a group decision maker (GDM) comprising several groups, and the objectives and related attributes are looked into more closely.

This paper does not treat the problem in its full complexity, hence the sets of attributes, interest groups etc. might have to be expanded for a full case study.

GROUP DECISION MAKER

The opening of a coal field for stripmining affects many groups of people as well as the nation at large.

Typical interest groups are mining companies and public utilities, regulatory agencies, people supporting local postmining production activities, and environmental groups.

The group decision maker

(GDM) will be composed of representatives from interest groups, and decision analysts as well as other experts.

Following Krzysztofowicz (1978), the GDM divides into subgroups related to specific fields of expertise or responsibilities.

The GDM accepts the single attribute utility functions derived by the subgroups as the GDM's own.

In the multiattribute acceptance of a reclamation scheme the nation is represented by the regulatory agencies (Hipel & al., 1974, 1976).

Thq enforce rules that are the results of ongoing trade -offs between national objectives like low inflation, independence from oil imports, and reduced impacts on humans and nature.

The reasons for formalizing the decision analysts' position as a subgroup within the GDM are twofold.

It is often impossible to complete an analysis without judgement on the part of the analysts, not least in the choice of analytical tools.

Judgements in turn may unintentionally imply preferences for outcomes.

Hence the decision makers ought to be made aware of this in a formal manner reminding them of the need that they understand the assumptions along with the conclusions.

Secondly, the model development, data compilation, and sensitivity analyses the analysts carry out all build up systems Insight beyond presentation of quantified answers to questions posed.

Their inclusion as decision maker makes explicit their responsibilities for making it all available.

The next step is to describe the objectives which the reclamation shall serve.

The interest groups and the GDM were discussed first for the reason that only those objectives are sure to be

23

included for which there are people prepared to stand up.

THE OBJECTIVES

The alternative reclamation schemes have the following objectives:

1) to produce sufficient amount of coal

2) to effect a postmining land use reaching at least premining value

3) to alleviate pressures on the environment

4) to minimize costs and losses

Keeney & Raiffa (1976) develop the objectives' hierarchy to the point of a one -to-one relationship between a lowest level of objectives and attributes.

This is not done here, instead the next step within the framework of a cost -effectiveness analysis is applied, which is to transform the objectives into specifications (Duckstein and Kisiel, 1977).

These in turn lead to the definition of attributes.

THE SPECIFICATIONS

Specifications in their full description are the translations of objectives into technological, economic, social, and environmental sub -objectives.

Here also standards and regulations are included.

To objective 1) then correspons the following specifications or sub -objectives

1) minimize reductions from desired production goals

2) do at least as well as required by mining health and safety standards to objective 2)

3) maximize the profit of the postmining land use

4) maximize the protection of already established local production objective 3)

5)

6) landscape as esthetically as possible maximize the protection of renewable and non -renewable resources other than coal

7) meet the air, water, and reclamation standards as well as possible and to objective 4)

8) maximize mining productivity

9) minimize reclamation cost

Spec. 9), i.e. the cost of reclamation is kept separate resulting from the reclamation, spec. 3).

from the profit of local production activities

This is done because reclamation costs including preparations for the local end use are tagged on to the coal price whereas the proceeds from the postmining land use because mining profits is a go elsewhere.

The profitability of the mining is not specified explicitly function of national priorities and regulations in energy matters.

But this study does not treat national trade -offs, and thus, profit is accordingly not specified as an objective.

But given that the productivity is maximized (spec. 8), the profit is maximal for any coal price level.

THE ATTRIBUTES

The attributes are now simply the subobjectives or specifications imperatives maximize, minimize, etc.

of last section stripped of the

Also seemirglystrict regulations are listed as attributes.

When strict regulations relate to uncertain outcomes the strictness becomes muted, but alternatives, although flexible, should be kept as close as possible to the regulations.

The resulting attributes derived from the specifications are xl x2 x3 coal production attainment of mining health and safety standards profit of postmining land use

24

x4 x5 x6 x7 x8 xg x10 xii level of already established local production esthetics of the landscaping non -destruction of renewable and non -renewable resources not to be mined air quality water quality attainment of reclamation standards mining productivity reclamation cost

INTER -SCHEME RANKING AND INTRA- SCHEME OPTIMIZATION

If a reclamation scheme contains actions or control variables that only bear upon a single objective like profit, they can be jointly optimized w.r. to maximum profit separately from any multiattribute ranking or optimization.

But when two or more conflicting objectives are in force for some or all actions, their optimization will have to be handled by a multiattribute preference procedure.

Brinck

& al. (1976) optimize schemes with essentially two types of actions; livestock species for grazing, and furrowing graded spoil piles.

The optimization problem was to combine the two actions to achieve the single objective of maximizing expected profit.

The action to let livestock on the reclaimed range, however, is also the most prominent attribute subject, to strong subjective preferences.

Thus if cattle grazing should turn out to have a smaller expected profit than sheep grazing, a trade -off based on a biattribute preference function would be appropriate.

Multiattribute trade -off procedures do not necessarily contain algorithmic feedback to the optimization of actions.

A continuous action variable that interacts with conflicting objectives has to be discretized, and each level of al thus defined enters into a combination with other actions to make yet another alternative scheme to be ranked under the multiattribute preference procedure.

THE METHOD

In this chapter the choice ofa multiattribute utility function for the preference function G, eq.

(6), is explained and its derivation is outlined.

Once G. has been estimated, the decision rule for finding the best scheme is given by eq. (7).

When outcomes are deterministic, the preference function is called a value function (Keeney & Raiffa, 1976); in the uncertainty case such as treated here, it is called a utility function.

A MULTIATTRIBUTE UTILITY FUNCTION

The scalar index of preferability G is here u, the multiattribute utility function derived on the basis of the von Neumann - Morgenstern utility theory expounded in Raiffa (1970), Keeney and Raiffa

(1976), Krzysztofowicz (1978), and several other references.

The concept of utility theory and its application and alternative forms of multiattribute utility functions are detailed in the literature and will not be repeated here, but for brief comments, (Gros, 1975; Richard, 1975; Keeney and Wood,

1977; Duckstein & Krzysztofowicz, 1978).

Given assumptions about preferential and utility independence certain functional forms of the utility function can be derived.

A fundamental concept of multiattribute utility theory is the assumption of utility independence.

Its role in multiattribute utility theory is similar to that of probabilistic independence in multivariate probability theory.

The existence of utility independence means that there also exist utility functions over attributes individually, i.e.

single attribute utility functions in a multiattribute environment.

Under the condition of mutual utility independence the multiattribute utility theory leads to the conclusion that a utility function is either multiplicative or additive (Keeney and Raiffa, 1976, p. 288).

A multiplicative form for multiple attributes is for K k 0 l +ku(xl,x2,.)= fl(1 +kkiui(xi))

1 =1

(9) where ui are the single- attribute utility functions for the i'th attribute, i= 1,2,..,n.

scaling constants.

k and ki are

The latter can also be called corner utilities (Duckstein and Krzysztofowicz, 1978).

Some authors like Gros, (1975) prefer the phrase "preference function" to "utility function" to avoid confusion with public utilities.

Here,preference function has been used as the general name for the mapping defined by G, in eq. (6), and utility function is understood in the von Neumann -

Morgenstern meaning.

THE SINGLE ATTRIBUTE UTILITY FUNCTION

The utility function allows ranking of probability distributions of the outcome on the basis of expected utility u.

The outcome whose utility is the expected utility of the outcome - distribution is called the certainty equivalent x as the decision maker in his judgement is indifferent between this

25

outcome for certain and the game given by the probability distribution.

Thus instead of ranking alternatives by expected outcome x they are ranked by expected utility ú

ú = E[u(x)]

(10)

If z is equal to z the decision maker is indifferent to any particular realization of the random x, i.e. he is globaly risk neutral.

function.

This is the case when u is a monotonicaly increasingAlinear

Other utility functions may possess local non-neutral risk attituges, even if x equals z.

It is also possible to see the ranking in terms of the certainty equivalent x being a shifted expectation of x.

The derivation of single -attribute utility functions are often framed in the probabilistic notion of two -pronged lotteries with a most preferred reward and a, least preferred

(Halter & Dean, 1971; Lavalle, 1970).

for each attribute

The two- pronged even -chance lottery spanning the whole stochastic range of the outcome, i.e. with the two extreme outcomes as rewards produces one utility point whick defines the indifference between the lottery and the estimated certainty equivalent of the lottery x1.

As the utility function is calibrated by the utility of the most preferred outcome being 1, and the utility of the least preferred outcome xo being 0, the utility of said lottery is h.

Because utility functions are unique up to a positive linear transformation, this technique can be continued for each of the new outcome intervals resulting from the mid -utility splitting, linking the interval utility functions together by said property.

In this way enough points can be estimated to draw a curve.

The technique of finding points on the utility curve by estimating certainty equivalents y of fixed utility lotteries may not work for discrete outcomes.

If outcomes do not exist between the discrete realizations the closest one may have to approximate xxt.

A procedure which may be more accurate keeps the lottery over the extreme and realizable outcomes xo and x *, and varies the odds instead.

The lottery now will produce x* with probability pi and x with probability 1 -pi.

The subject is asked to estimate the probability pi for which he is indifferént between the lottery and a realizable outcome xi for certain.

Because the cardinal utility functions of the von Niemann- Morgenstern type with which we are working are unique up to a positive linear transformation, and secondly, because the scaling convention for the utility function Is the same as for the probabilities pi, the pi-value is also a utility and in fact equal to the utility u.

Thus the pi -point is then a point on thé u -curve for xi.

SCALING FACTORS

The general form of the multiattribute utility function based on mutual utility independence is eq. (11), from which eq. (9) is derived for k # 0, u = n fkiui +k L kik uiuj +k2 E kikjkkuiujuk +.,, +kn- lkik2..knulu2..un

j

1 =1

1=1

Jv ,ksj j >1

The risk neutral form of eq. (11) written for the

It is seen that for k =0 u is (multi -) linear.

combination of best outcomes is

(12) k1 =1

Thus k = 0 and eq. (12) are both expressions of risk neutrality.

and written for the combination of best outcomes n

1 +k= Tf(T +kki)

1 =1

For k 4 0 eq. (11) can be reformulated

These expressions are needed for determining numerically the scaling factors k and k1.

(13)

ASSESSMENT

In this chapter the multiattribute utility function is assessed for a case of three attributes related to multiple land use alternatives covering different grazing schemes x3 profit to the locals of postmining grazing measured by the range carrying capacity in animal units per hectare and year.

Worst outcome is 0 and best outcome is set to

2 AU /ha, year (Brinck et al., 1976) x6 wildlife habitat measured in mule deer, from 0 to 400 deer (USGS, 1977) xB ground water quality measured subjectively in % of excellent (Keeney and Wood, 1977)

26

The procedure follows Keeney and Wood (1977) and can be applied straightforward to cases with many attributes.

The order is

1)

Checks of preferential and utility independence

2)

Derivation of the single- attribute utility functions within the subgroups

3)

Assessment of the GDM's risk attitude

4)

Estimation of the GDM's scaling factors

THE GDM'S QUALITATIVE PREFERENCE STRUCTURE

In eq. (11) the form of the multiattribute utility function was introduced on the assumption of preferential and utility independence.

The present assessment does not treat the true case in terms of decision makers and a full spectrum of consequences, and hence proceeds without verifying these assumptions.

However, e.g. Keeney and Wood (1977) show a way of doing this.

SINGLE ATTRIBUTE UTILITY FUNCTIONS

The utility functions for x respectively.

x and x used in this example are shown in Figure 2 a,b, and c

For x3 a risk prone attitude s explained by unfavorable grazing conditions being manageable, although at a cost.

The carrying capacity may come out worse than planned, and water may have to be piped from farther away, but damage is not irreparable.

However, to restore wildlife habitat, x , takes cooperation from the beasts, and may be impossible or difficult to manage.

Therefore the pBrsistent risk averse utility function for the attribute x,.

Ground water, x below a certain quality level ceases to be useable,,and the decision maker cgn then better gambe that the outcome will be good.

For low quality levels therefore it shows a risk prone attitude, but changes to a risk averse behavior at higher levels.

ESTIMATION OF SCALING FACTORS

Equations for determining the scaling factors k and k are found by considering that eq. (11) for xi at its best level and all other attributes at their wort, yields u(0,..,O,x *i3O..,O) =ki

(14)

This is by the way the reason for calling the k corner utilities.

By taking two attributes say x3 and x6 and keeping the other at its worst level, An indifference relation between two points

(x3,x )' and (x3, x6) " makes an equation for k, k3, and k6 by applying eq. (11) to both sides of this equat 'Ton.

The simplest way to go may be to decide which of the two attributes is the least important, say x6, then fix the point (x,=0, x =400), and ask for an estimate of the x'3 value which makes (x' x =0) indifferent to the fist pont.

The combination with worst levels supresses k, and the resdltg of two indifference relations framed in this way is k6 = .1k3

and k8 =

.3k3

(15)

The utilities are read off the estimated utility functions in Fig. 2.

equations can be had in this way.

With n attributes only n -1

The complementing equation is found by estimating the probability p for which the decision maker is indifferent between a lottery over worst and best combinations, say

(x* ,x *8) and (x xopp), and (x* ,,x for certain, where p is the probability of getting the best

Again using rewgrd.

x aná'it *i mean the wort áña the best outcome on attribute xi respectively.

i eq. (11) fóf this gamble gives k3 = p(k3 +k6 +kk3k6) (16) p was estimated to .8.

Eq. (16), however, inroduced k for which eq. (13) can be solved

1

= k3+ k6 +k8 +k(k3k6 +k3k8 +k6k8) +k2k3k6k8

Eq.s (15), (16), and (17) give the four relations needed.

The solution is k3=.753, k6 =.075, k8=.226, k = -.221

(17)

(18)

Eventually

27

u

.5

u=.753u3+.075u6+,226u6- .013u3u6 .038u3uß .0038u6u8+.00063u3u6u8

u6

.5

1 u8

.5

(19)

0

1.

x

- range carrying dpacity in AU/ha-year

2.

a)

0

0 200

400 x6 - mule deer in head b)

0

0

50 100 x8 - ground water quality in % of excellent c)

Figure 2 Single attribute utility functions

CONCLUSION AND DISCUSSION

Because of the uncertainties in the states of nature and the risks involved the ranking of reclamation schemes for disturbed lands is a good case for applying the multiattribute utility theory.

This paper establishes a framework for a real case application which would have to elaborate the derivation of utility functions and preference structures more carefully than has been done here.

The important step of verifying preferential and utility independence has not been carried out because this analysis does not treat the full case.

The true group decision maker and the full spectrum of consequences would have to be present to carry out the independence checks.

Although the decision maker is thought of as a group decision maker nothing has been said about how the group goes about agreeing on scaling factor derivations.

For suggestions of how to model alternative group decision makers reference is made to the references cited.

The existence of a model for treating the trade -offs between conflicting attributes can be used for sensitivity analyses to assist in solving differences between GDM members.

Finally the dynamic character of the problem has not been touched upon, but it lies close at hand that the reclamation of an area takes place in stages, and calls for framing the development of reclamation plans in a stochastic and dynamic framework.

ACKNOWLEDGEMENT

The research leading to this paper has been supported in part by a grant from the National Science

Foundation ENG 76 -20280 "Uses of Modern System Theory in Water Resources ", and a grant from the U.S.

Bureau of Mines for "Comprehensive Planning for Strip Mine Reclamation in Dry Regions with Emphasis on

Water Harvesting ".

REFERENCES CITED

Bertsekas, D. P.

1976.

Dynamic programming and stochastic control.

Academic Press, New York

Brinck, F. H., Fogel, M. M., and Duckstein, L.

1976.

Optimal livestock production of rehabilitated mine lands.

Hydrology and Water Resources in Arizona and the Southwest, Vol. 6, Proc. 1976 meeting Arizona Section of the American Water Resources Ass. and the Hydrology Section of the Arizona

Academy of Science.

Doyle, W. S.

1976.

Strip mining of coal, Environmental solutions.

New Jersey.

28

Noyes Data Corp., Park Ridge,

Duckstein, L.. and Kisiel, C. C.

1977.

Alternative water reuse systems: A cost- effectiveness approach.

Water renovation and reuse.

Academic Press, New York.

Duckstein, L., and Krzysztofowicz, R.

1977.

A utility criterion for real -time reservoir operation.

Hydrology and Water Resources in Arizona and the Southwest, Vol. 7, p. 207 -218.

Proc. 1977 meeting Arizona Section of the American Water Resources Ass. and the Hydrology Section of the

Arizona Academy of Science, Las Vegas, Nevada, April 15 -16.

Gros, J.

1975.

Power plant siting:

A paretian environmental approach.IIASA res. mem.

Schloss Laienburg, A -2361 Austria.

RR- 75 -44.

Halter, A.N., and Dean, G. W.

1971.

Decisions under uncertainty with research applications.

Western Publishing Co., Chicago, 266 p.

South -

Hipel, K. W., R. K. Ragade and T. E. Unny.

1974.

Metagame analysis of water resources conflicts.

J. of the Hydraulics Division. ASCE, Vol. 100, No. HY10, Proc. Paper 10861, October 1974, pp.

1437 -1455.

Hipel, K. W., R.

K. Ragade and T. E. Unny, 1976.

Political resolution of environmental conflicts,

Water Resources Bulletin, Vol. 12, No. 4, August 1976, pp. 813 -827

Imes, A. S., and Wall, M.

K.

1977.

An ecological -legal assessment of mined land reclamation laws.

North Dakota Law Review, Vol. 53, No. 3

Keeney, R. L., and Raiffa, H. 1976.

Decisions with multiple objective..J. Wiley et Sons, New York.

Keeney, R. L., and Wood, E.

F.

1977.

An illustrative example of the use of multiattribute utility theory for water resources planning.

Water Res. Res., Vol. 13, No. 4, p. 705 -712.

Kirk, S. J.

1978.

The evaluation of surface mining as an interim land use.

Master's thesis, Dept.

of Mining and Geological Engineering University of Arizona, Tucson, Arizona.

Krzysztofowicz, R.

1978.

uncertainty.

Preference criterion and group utility mode for reservoir control under

Reports on natural resource systems, No. 30.

University of Arizona Departments of

Hydrology and Water Resources, and of Systems and Industrial Eng.

Tucson, Arizona

LaValle, I. H.

1970.

An introduction to probability, decision, and inference.

Winston, Inc., New York, 767 p.

Holt, Rinehart and

Raiffa, H.

1968.

Decision analysis. Introductory lectures on choices under uncertainty.

Wesley, Reading, Massachusetts.

Addison -

Richard, S.

F.

1975.

Management Science.

Multivariate risk aversion, utility independence and separable utility functions.

22(1).

School of Renewable Natural Resources.

University of

Thames, J. L.

1979.

Personal Communication.

Arizona, Tucson, Arizona.

U.S.G.S., 1977.

Final Environmental Statement, FES 77 -33.

Eagle Butte mine, Campbell County, Wyoming.

Proposed mining and reclamation plan,

DOI, Geological Survey, Reston, Virginia.

29

SEDIMENT PRODUCTION FROM A CHAPARRAL WATERSHED

IN CENTRAL ARIZONA

Thomas E. Hook and Alden R. Hibbert

ABSTRACT

Sediment production from two chaparral watersheds in central Arizona during a period of heavy winter rainfall in 1978 was compared with sediment production over a 14 -year period (1964 -78).

Results indicate sediment production from chaparral is primarily the result of seasonal periods of heavy precipitation and runoff and not from ephemeral summer rainstorms.

Sediments from 300 acres (122 ha) above a newly constructed stock watering tank we.e produced within a few days time in the late winter of

1978 at an accelerated annual rate of 41.1 ft /acre (2.9 m /ha).

The sediments came mostly from cutting in channel alluvium in upstream tributaries where the sediments are presumed to have accumulated from downslope creep, dry ravel, and overland flow produced by ephemeral, convective rainstorms.

The accelerated rae of sediment production was more than 4 times the average annual rate of

9.8 ft /acre (0.7 m /ha) determined from 14 years of cumulative sediment deposits in a stock tank constructed in 1964.

INTRODUCTION

Sediment production and transport from chaparral watersheds in central Arizona is a complex process, primarily dependent on periods of heavy precipitation.

The ephemeral rainstorm in central

Arizona is an individual precipitation event, usually in summer and early fall, and generally convective in nature.

The ephemeral rainstorm is often intense, of short duration, and may generate sufficient overland flow to dislodge and transport soil from chaparral covered slopes to small upstream tributaries.

The flows dissipate, dropping their sediment loads upon entering the flatter, dry channels.

Only unusually heavy storms of this type produce enough runoff to carry sediment for any appreciable distance downstream.

Heavy seasonal precipitation, on the other hand, will exceed the storage capacity of the soil and produce flow rates capable of transporting large amounts of sediment.

In central Arizona this seasonal precipitation usually is in winter and early spring in one or more cyclonic storms, often occurring in rapid succession.

The exact impact of these two precipitation regimes on sedimentation rates and processes is not well understood for chaparral watersheds in central Arizona.

This paper examines the hypothesis that sediment production from chaparral watersheds in central Arizona is primarily the result of seasonal periods of heavy precipitation and runoff and is not primarily from ephemeral rainfall.

While this hypothesis may appear to be in conflict with results of some investigators, discrepancies may be due to differences in studies.

soils and vegetation and the interpretation of results of small catchment and plot

For example, some early research from the Sierra Ancha Experimental Watersheds (USDA

Forest Service 1953) indicates most soil loss from chaparral slopes occurs during summer months.

It was not clear in the research report what happened to the sediments after they reached the channels, although it was determined that "the surface runoff waters are absorbed by the many small drainage channels and largely evaporate before the next storm occurs and therefore are an ineffective source of streamflow."

The authors are, respectively, Graduate Associate in Research, Department of Geography, Arizona

State University, Tempe, and Principal Hydrologist, USDA Forest Service, Rocky Mountain Forest and

Range Experiment Station, Research Work Unit at Tempe, in cooperation with Arizona State University; station's central headquarters are maintained at Fort Collins, in cooperation with Colorado State

University.

31

Boater and Davis (1972), citing other research in and near the Sycamore Creek Watershed, state that 85$ of the on -site soil movement is in response to summer rainstorms, while runoff occurs mostly in response to winter rains.

Other evidence to support the hypothesis is available from research in chaparral watersheds in southern California.

Measurements of sediment production in the San Dimas

Experimental Forest indicate sediment transport and deposition result from large, infrequent stream flows following heavy winter because of precipitation (Rice 1974).

the wide year -to -year variation in

5 years or less are responsible chaparral watersheds.

rainfall,

Wolman and Miller

(1960) discovered that large winter storms occurring once every for moving approximately 40 -60$ of the sediments produced on

STUDY OBJECTIVES

Study objectives included the measure and evaluation of relatively long term (14 years) and short term (seasonal) sediment production from two small chaparral covered watersheds.

Additional information was sought on particle size distribution and mechanism of transport (suspended or bed load).

To the extent possible, the source of sediment was traced to the channels and slopes.

However, no quantitative measurements were made of channel cutting and soil losses from the slopes.

RESEARCH AREA

The study area lies within the Battle Flat Watershed approximately in the center of the Bradshaw

Mountains in central Arizona.

The Battle Flat Watershed has a median elevation of 5,355 feet (1,633 m) and consists of several contiguous subwatersheds.

The area drains into Turkey Creek and the Agua

Fria

River.

Water yield is generally intermittent with streamflow beginning in early winter during wet years and continuing through May or June.

In dry years little or no flow occurs.

Mean annual precipitati9n in the Battle Flat watershed is approximately 23 inches (584 mm) with 15 -20$ occurring as snowPrecipitation generally results from cyclonic storms in winter and local convective storms in the summer.

Mean daily temperature is approximately 59° F (15° C) with an annual maximum range from -21° F to 103° F ( -29° C to 39° C).

The two subwatersheds under examination together occupy 435 acres (176 ha) in the northwest portion of Battle Flat.

They are hereafter referred to as the south watershed (300 acres) (122 ha) and the north watershed (135 acres) (55 ha).

The general geologic composition of these watersheds is described by Anderson and Blacet (1972) as massive bedded crystallne tuff with recent gravels along stream beds.

Aspect is generally southeast, and slopes range from 15 -40$ with some as high as 60$.

Elevations range from 5,200 to 5,800 feet (1,586 -1,769 m) above sea level.

Soils on the two watersheds are similar and include Moano gravelly loam and Moa -Lynx association in the areas of lower elevation and Moano very rocky loam on the upper slopes.-

Vegetation is dense with shrub crown cover approximately 75 -80$.

The three separate vegetative associations include mixed chaparral of localized emory oak (Quercus emoryi) with alligator juniper ( Juniperus deppeana) overstory, manzanita (Arctostaphylos punqens)

dominated chaparral, and a mixed fire succession association locally dominated by shrub live oak (Quercus turbinella), apache plume (Fallugia paradoxa), and occasionally yerba santa (Eriodictyon angustifolium).

The study watersheds are estimated to yield on the average about 1 inch of water (5$ of the precipitation).

This amounts to an annual production of 36 acre -feet (44,400 m3) of water.

In wet

years much larger amounts are yielded, and in dry years very little.

The water enters the stream by subsurface flow through the soil mantle and as runoff over the soil surface.

In years of above normal precipitation, subsurface flow predominates over surface runoff (overland flow), and stream flow may extend into summer.

Even after surface flows in channels cease, subsurface drainage may continue for a time through deep channel alluvium that may exceed 9 feet in depth.

Loss of ground water in the study watersheds is prevented, probably by bedrock of low permeability.

Two stock watering tanks at the base of the watersheds were constructed by placing earth dams across the channels (figures I and 2).

These small reservoirs are usually dry by mid summer.

The lower stock tank was constructed in

1964 and, until

1977, received all sediment yields from both watersheds.

in watershed.

In 1978 it received only sediment from the north watershed.

The new tank, constructed

1977 approximately 700 feet upstream from the old tank, receives sediment only from the south

Environmental Analysis Report for the Battle Flat Chaparral Pilot Application Project.

Prescott

Nat. For., Ariz., Unpubl. Rep. 59 p.

1978.

' 32

North watershed

South watershed

Upper stock tank -1977

t

200 feet

-; Flow direction

Coarse sediment

Fine sediment

Lower stock tank -1964

Figure 1. -- Stock tanks below the study watersheds.

Water surface

Coarse channel sediments

Fine sediments

Figure 2.-- Idealized cross section of stock tank.

Two distinct sediment deposits are associated with each of the stock tanks (figures 1 and 2).

An elongated deposit of course grained material was deposited in the stream channels at the upper ends of the tanks, while fine grained sediments were deposited in the bottoms of the tanks.

METHODS

The volume of sediment deposited in the stream channels and stock tanks was determined from surface area measurements and depth samples.

The width of the sediment deposits was measured at cross sections 25 feet apart.

The average depth of sediment was measured at several places along each of the cross sections using soil augers, shovels, and soil coring devices.

The base of the sediment deposits was identified by a distinct stratigraphic contact, either an immediate change in lithology from unconsolidated fine and /or coarse grained sediments to the consolidated alluvial material of the old stream bed or a dark layer consisting of decomposed organic material accumulated prior to the deposition of sediments.

The total volume of sediment deposited in the stream channels and stock tank bottoms was determined by multiplying the depth of the sediment layer by the surface area of the deposit.

Volume of the deep, narrow stream channel deposits was calculated separately from the volume of the shallow deposits of the stock tank bottoms to increase accuracy.

The dates of construction of the two stock tanks and the total volume of sediment deposition provide the rate of sediment yield for the study watersheds.

33

Sediment samples from the new stock tank and its immediate upstream channel were collected at the same 25 -foot intervals as the sediment width and depth measurements.

These samples were oven

dried for a period of 24 hours at 105° C.

The samples were then placed in 8 -inch, U.S.

Standard sieves on the Ro -Tap shaker in order to classify them according to the Udden- Wentworth grade scale.

RESULTS AND DISCUSSION

The total volume of sediments deposited in the upper stock tank was 12,325 ft3 (349 m3).

Eighty percent of the sediment was deposited along the immediate upstream channel where water entered the pond, and 20% was distributed more uniformly over the bottom of the tank.

The sediment originated from the 300 acres in the south watershed during a short period of heavy rainfall i the winter of,

1978.

m

/ha).

The accelerated or short term annual rate of sediment production was 41.1 ft /acre (2.87

The3volume of 3sediments deposited in the lower stock tank and its immediate upstream channel was

47,405 ft

(1,342 m ), including the sediment produced by the north watershed in 1978.

Adding the

1978 deposit from the upper Sank, the3sediment production from both watersheds from 1964 through the spring of 1978 is 59,730 ft

(1,692 m ), yielding a mean annual rate of 9.8 ft /acre (0.7 m /ha).

Seventy -three percent of the sediment sampled from the stream channel deposits immediately above the upper stock tank was coarse sand or larger with cobbles and pebbles dominating (figure 3); 69% of the sediment in the thin layer over the bottom of the upper stock tank was medium sand or smaller with fine sand dominating.

Precipitation 2/

on the study area was nearly three times normal for the four months ending

March 31, 1978 (table 1).

The previous year was dry, with little or no runoff.

Soil water recharge began with the first significant rains in late December and continued until the watersheds began to yield water sometime between February 27 and March 4.

Although 17.5 inches (444 mm) of rain had fallen by mid February, no overland flow or streamflow resulted from these earlier storms.

However, both stock tanks were full and overflowing when observed on March 4, after 4 days of nearly continuous rainfall.

These rains added 12.0 inches (305 mm) of water, and rains on March 5 -7 added another 1..9 inches (48 mm).

On March 9 combined flow from the two watersheds was estimated at 1 cfs (0.03 m3 /sec) by measuring the stream cross section ably occurred on March

1 or and flow velocity where it enters

2 during the heaviest part of the the lower storm.

From tank.

Peak flows probhigh water marks visible

50

L

Sediment on bottom of

--I tank

>, 25

c er c

Sediment in channel at inflow

O\ei

0

5

Al" y

ocò

`J

5a

5ca oò

G0 co\

GO

.41.°

GO

Figure 3.-- Particle size distribution of sediment from a chaparral covered watershed by location of deposit within the stock tank.

4/ Precipitation data are for Crown King 7 miles south at 6,000 feet elevation.

The study area receives an estimated 10 -20% less precipitation than the amounts shown here.

34

Table 1. -- Inches of daily precipitation at Crown King, Ariz.

The study watersheds are located 7 miles north and at slightly lower elevations, where precipitation is lower by an estimated 10 -20$.

Day Dec. 1977 Jan. 1978 Feb. 1978

20

21

22

23

24

25

26

27

28

29

30

31

11

12

13

14

15

16

17

18

19

1

2

9

10

7

8

5

6

3

4

Monthly totals

Cumulative totals

.02

.02

.30

.25

1.10

.65

2.34

2.34

.36

1.20

1.05

2.90

1.70

.30

.50

.10

.01

1.59

9.71

12.05

.10

.50

.34

.20

.02

.29

2.80

.95

.05

.15

.23

1.50

7.13

19.18

Mar. 1978

4.87

3.38

2.03

1.43

.37

.08

.08

1.05

.01

1.11

.05

_06

14.52

33.70

on March 9, peak inflow to the lower stock tank was estimated to have exceeded 11 cfs

(0.31 m3 /sec).

On April 27 combined flow from the two watersheds was estimated at 0.2 cfs (0.006

m3 /sec), and by

May 22 it was less than one -half this amount.

By mid June the channels were dry.

Although the exact time of sediment deposition in the upper tank could not be determined,

it is

assumed that transport and deposition took place primarily during periods of high flow, which lasted a few days at most.

It is known that the drainage channels were dry until after February 27, when the last of the big storms began.

By March 9, flows in these channels were only slightly turbid, and little bed material was moving.

The sediment deposits were visible at the upper ends of both stock

tanks, but no measurements were possible until the water subsided.

None of the storms after March 9, and prior to sediment measurement in June, were large enough to cause significant new erosion on the slopes or in the channels.

Therefore, it was concluded that the bulk of sediment transport and deposition took place within a few days, most likely March 1 -3 when rainfall was heaviest.

Movement of sediments from chaparral slopes to upstream tributary channels and eventually out

of the watershed is precipitation.

Soil

a complex process, which is primarily dependent on amounts and intensities of particles are initially transported from slopes to upstream tributaries through the

processes of downslope soil

creep, dry ravel, and overland flow (Hibbert et

al.

1974).

Sediments accumulate in the upstream tributaries until a streamflow of sufficient magnitude occurs to transport the sediments to downstream channels and out of the area.

Limited streamflows from most summer convec-

tive storms are not large enough to do this.

Only rarely are summer storms big enough to generate

and sustain the quantity of overland flow required to transport large amounts of sediment far downstream (the exception is immediately after chaparral wildfires, when overland flow occurs much more

readily as a result of a fire -induced nonwettable layer near the soil surface) (DeBano 1971, Scholl 1975).

A streamflow of sufficient magnitude and duration to transport the sediments to downstream channels and

35

out of the area is more likely to be produced by a series of intense winter storms that result in above normal precipitation.

Sediments In the upstream tributaries of the study watersheds had been accumulating for an unknown period of time prior to the initiation of streamflow on about March 1, 1978.

The high volume of this flow cut and transported the accumulated sediments from the upstream channels to the stock tanks.

These high flows originated primarily from subsurface flows after saturation of the soil mantle, and continued into June.

Observations made on the watershed after the early March activity indicated

that rifting caused by overland flows occurred only at a few locations during periods of intense

precipitation.

The coarse sediments, including gravel carried as bed load, were deposited in the channels

immediately above the stock tanks.

The finer grained sediments were transported as suspended load and were carried into the two stock tanks as they rapidly filled with water.

The fine material was deposited in thin layers on the bottom of the tanks.

The large amount of sediments deposited in the stock tanks below the study watersheds in

response to the heavy winter rains in 1978 supports the hypothesis that sediment production from chaparral watersheds in central Arizona is primarily the result of heavy winter precipitation and runoff and not from ephemeral rainfall.

The accelerated rate of sediment production was 4.2 times the mean annual rate for 14 years, including 1978.

Fourteen years is too short a period to establish a reliable long -term erosion rate for these watersheds when yearly sediment production is known to vary from

nothing in some years to at least the amount (41.1 ft /acre) observed in 1978.

The expected frequency of seasonal precipitation of the magnitude experienced in the winter of 1977 -78 on Battle Flat is not known.

However, precipitation records at Crown King show four winter periods between 1964 and 1977 when precipitation exceeded the approximately 20 inches (508 mm) required to initiate heavy flows by March 1 in the 1977 -78 period:

1.

2.

November 1964 -April 1965

November 1965- February 1966

3.

November 1967 -March 1968

4.

October 1972 -March 1973

24.1 inches (612 mm)

24.4 inches (620 mm)

23.6 inches (599 mm)

27.1 inches (688 mm)

How many of these wet periods produced high flows is not known; nor can we tell how much runoff might have been caused by summer storms on these watersheds.

The Crown King rain gage is too far away to use for estimating summer storm intensities and amounts.

However, since the accelerated rate of sediment production was more than four times the mean annual rate for the entire period, it seems reasonable to attribute the bulk of the production prior to 1978 to the four known wet periods.

REFERENCES CITED

Anderson, C. A., and P. M.

Blacet.

1972.

Geologic map of the Mount Union Quadrangle, Yavapai

County, Arizona.

U. S. Geological Survey.

Boater, Ron S., and Lester R. Davis.

1972.

Soil -loss considerations in chaparral -to -grass conversion.

p.

243 -250.

In Vol.

14, Natl.

Symp. on Watersheds in Transition, Proc. Am. Water Resour.

Conf., Am. Water Resour. Assoc. and Colo. State Univ.

[Fort Collins, Colo., June 19 -22, 1972]

DeBano, Leonard F.

1971.

tion.

The effect of hydrophobic substances on water movement during infiltra-

Soil Sci. Soc. Amer. Proc. 35:340 -343.

Hibbert, Alden R., Edwin A. Davis, and David G. Scholl.

1974.

Chaparral conversion potential in

Arizona, Part

I:

Water yield response and effects on other resources. USDA For. Serv. Res.

Pap. RM -126, 36 p.

Rocky Mt. For. and Range Exp. Stn., Fort Collins, Colo.

Rice, Raymond M.

1974.

The hydrology of chaparral watersheds.

p.

27 -34.

with the Chaparral Proc. [March 30 -31, 1973, Riverside, Calif.].

Sierra Club.

In Symp. on Living

Scholl, David G.

1975.

39:356 -361.

Soil wettability and fire in Arizona chaparral.

Soil Sci. Soc. Amer. Proc.

U.S. Department of Agriculture, Forest Service.

1953.

The Sierra Ancha Experimental Watersheds,

32 p.

Southwest For. and Range Exp. Stn., Tucson, Ariz.

Wolman, M. Gordon, and John P. Miller.

1960.

Magnitude and frequency of forces in geomorphic processes.

J. Geol. 68:54 -76.

36

AN EXCHANGE SYSTEM FOR PRECISE MEASUREMENTS OF TEMPERATURE

AND HUMIDITY GRADIENTS IN THE AIR NEAR THE GROUND

L. W. Gay and L. J. Fritschen

ABSTRACT

Small differences can be very accurately measured with two sensors if precautions are taken to periodically interchange the sensors between observations.

The Bowen ratio model of evapotranspiration requires measurements of air temperature and humidity gradients near the evaporating surface.

The gradients are in the order of only 0.1 °C /m or 10 Pa /m.

Precision in excess of 0.01 °C /m or 1 Pa /m can be obtained only through laborious calibration and replication of instruments, or through periodic interchange.

The design of a simple system for interchanging psychrometers is described.

The system will exchange sensors between two levels one meter apart at selected time intervals.

The vertical exchange path of this design maintains a constant orientation of the sensors and has important advantages over rotating systems used elsewhere.

The principles apply to a variety of measurement problems.

INTRODUCTION

The precise determination of the difference between two quantities is a problem commonly encountered in environmental measurements.

Such measurements are difficult to obtain if the differences are based upon subtraction of readings made with two separate instruments, since errors in either instrument are incorporated into the difference.

Errors become particularly troublesome when the desired differences are small with respect to the absolute value of the quantity.

The errors involved may be either random or fixed.

The random errors of measurement can be minimized or even eliminated with appropriate sampling.

Fixed errors, or biases, are not eliminated by frequent sampling but they can be removed by an exchange technique.

This note reports the design of a simple system that successfully eliminates fixed errors from measurements of temperature and humidity gradients in the atmosphere near the ground.

REMOVING BIASES FROM DIFFERENCE MEASUREMENTS

The effects of fixed errors can be eliminated from difference measurements in three ways.

The first is by using one sensor as a "rover" between the two measurement locations, and obtaining the difference by subtraction of successive readings.

Since the difference sought is that existing between readings at two different points, the variable being measured must not change until after the sensor has moved to the new location and the new sample taken.

The second technique is to subtract the means obtained from a number of sensors at each location.

This requires that the calibrations, number of sensors and distributions of biases combine to yield either a zero bias at each location, or a mean bias that is equal at each location.

Careful calibration is needed.

A third technique requires that two sensors be interchanged between each set of measurements.

The mean value (say of temperature) is then determined at each position over a time interval that contains several exchanges of sensors.

Subtraction of the means obtained by successive measurements at the two locations will give the differences free of systematic errors.

For example, consider measurements of true temperature T1 at level z1, and true temperature T2 at level z2.

The measurements are made with sensor 2 containing systematic error r, and sensor 1 free of systematic error.

The temperature measurements are made at successive intervals (time 1, time 2, etc.) with sensor 1 initially at level 1 and sensor 2 at level 2, but the sensors are interchanged between measurements.

The temperature difference over a period long enough for a single exchange of sensors

(Tanner, 1963) becomes:

The authors are respectively:

Professor of Watershed Management, School of Renewable Natural Resources, University of Arizona, Tucson, AZ, and Professor of Forest Meteorology, College of Forest Resources, University of Washington, Seattle, WA.

Approved for publication as Journal Paper No. 2986,

Arizona Agricultural Experiment Station.

37

Level time 1 time 2

T2

Mean z2

T2 + e 1 /2(T2 +e+T2) zl

T1

'

T1 + e 1 /2(T1 +T1 +e)

Difference

Az T2 -T1 +e T2 -T1 -e

T2 -T1

The interchange technique, in contrast to the multiple sample approach, sets no restriction upon the bias or systematic error a other than that it indeed be systematic, i.e., not change during the observation period.

In contrast to the first technique the temperatures need not be steady state, as measurements are made at both points (21, z2) at the same time.

The difference finally obtained by the interchange technique is actually a mean over the period that elapsed between successive measurements.

Thus some consideration must be given to the frequency of interchange and measurement with respect to the time resolution desired in the data.

FIXED ERRORS AND THE BOWEN RATIO

The Bowen ratio, 8, or ratio of sensible to latent energy, can be written in terms of measured temperature (tT) and humidity (ee) differences between two levels in the atmosphere close to an evaporating surface.

The form given in most standard derivations (see Webb, 1965) is

B

= C

AT where C is a coefficient derived from properties of the air and dependent upon units chosen for the measurement.

The temperature and vapor gradients are rather small (in the order of 0.1 °C or 10 Pa per meter) over most vegetative canopies.

Small gradients are difficult to measure precisely.

The vertical distance between the sensors affects the magnitude of the gradients, as a larger distance develops a larger signal.

If the distance is too great, however, the upper sensor may extend beyond the surface boundary layer into air with properties that are unaffected by the underlying surface.

The gradient measurements will not then represent the surface -exchange processes.

If the distance between sensors is too small, the signal is smaller and the effects of systematic measurement errors become more pronounced.

Gradients over vegetation are commonly measured with a vertical distance of about 1 m, and with the bottom sensor positioned at the approximate level of the tips of the canopy.

The gradient measurements needed in the Bowen ratio are exceptionally well suited to the interchange technique.

The turbulent nature of the atmosphere and corresponding temporal variations in temperature and humidity restrict the usefulness of a single "rover" probe, and the use of many sensors is laborious.

AN AUTOMATIC INTERCHANGE MECHANISM

The mechanism pictured in Figure 1 was designed to move the psychrometers vertically during the exchanging cycle.

It is similar in purpose to several rotary action designs (Sergeant and Tanner,1967;

Black and McNaughton, 1971), and the mechanism of McNeil and Shuttleworth (1975).

Vertical movement maintains the psychrometer water reservoirs in an upright position, eliminating possible spillage and water loss.

The basic unit consists of a support frame, a reversible gearmotor, a timer, and a drive chain that carries the psychrometers.

Two interchange units were contructed for testing.

Although both were controlled by a single timer, each could be manually controlled, independently of the other.

The support frame holds the psychrometers 1 meter apart when fully separated.

After a preselected time increment, the psychrometers are transferred up (or down) to the other level.

A reversible gear driven motor powers the exchanging mechanism.

A pair of travel -limiting microswitches, one at each end of the support frame, opens when the sensors reach the desired position, thus stopping the motor until the timer triggers another exchange.

An additional pair of microswitches near the limiting travel switches are connected to a battery to provide a position signal.

The magnitude of this signal can be used to determine the position of either psychrometer.

The support frame was fabricated with about 12 feet (3.7 m) of 1.25 x 1.25 x 1/8 inch (32 x 32 x

3.2 mm) steel angle iron, arc -welded together with 6013 welding rod.

Two horizontal tower support struts (36 inches or 0.914 m long) were welded 36 inches (0.914 m) apart and perpendicular to the vertical main support member, 42.25 inches (1.07 m) long.

Mandrels for the drive chains were mounted on opposite ends of the vertical support member.

The mandrel support pads contained mounting slots for adjusting the position of mandrels and the tension of the chains.

Two ball- bearing mandrels (Dayton

2X625), two 18 -tooth chain sprockets (Martin #41818) on each mandrel, and two roller chains (Dayton #41) composed the exchange assembly.

Two horizontal aluminum instrument supports (19 x 1.75 x 0.25 in, or

483 x 44 x 6 mm) were bolted to the chains 1 meter apart.

The psychrometers were attached to these

38

supports with 1.25 -inch (32 mm) U- bolts.

directly.

The 4 -RPM geared motor (Dayton 2Z813) drove the upper mandrel

A rubber spider (universal joint) eliminated alignment problems in the drive connection.

Figure 1.

The exchange mechanism with two aspirated psychrometers.

The control system circuit is shown in schematic form in Figure 2, and the components of the circuits are listed in Table 1.

The 115 -volt clock (Dayton Time Switch ( #2E130) in Figure 2C initiates an electrical impulse to the rotary relay switch (Potter & Brumfield #19AP11A), in Figure 2A, which closes the circuit through the oil- filled capacitor (Dayton #4)(426) and energizes the gear motor.

The sensors are driven to different levels until the chain stabilizing support depresses the travel -limiting switch (Micro #(3Z7651).

This opens the power circuit to the gear motor and stops the psychrometer.

After an appropriate period has elapsed, the clock initiates another impulse to the rotary relay switch, which energizes the reverse power to the motor via the capacitor, and the cycle is repeated.

2B.

The position -indicator circuit is wired to two "normally closed" microswitches, as shown in Figure

Whenever the "reference" psychrometer is in the top position, the upper switch is turned to the normally open position, and a positive signal appears on the signal leads.

Whenever the psychrometer is in transit, both switches are in the normally closed position and zero signal appears.

When the psychrometer "parks" at the bottom position, the lower switch is turned to the normally open position, and a negative signal appears on the signal leads.

The signal polarity indicates whether the "reference" sensor is up or down.

The magnitude of the signal is established by the simple voltage divider; the output of the divider shown yields ±0.0089 V.

The exchange system is controlled by an override switch at each driving motor.

To deactivate one of the units, simply turn the manual override switch (DPDT) to the central off position and the psychrometers can be stopped at any position during the cycle.

While the override switch is off on one unit, the clock continues to operate and the cycling continues on the other unit.

Upon restarting, the instruments can be easily synchronized in position with each other by manipulating the override switch in either the up or down position.

This switch will reverse the direction of the gear motor on that particular unit.

39

Reh+p

Swikh

LU Micro- 8wi/0t,

Normally Closed l+.-) eik+rsc

Clock

LD

Mkro-Smi%h,

Normally Closed

4

/Sb/(

1.35-V.

Figure 2.

Exchange mechanism schematics.

B.

A. The timing and control circuits.

The position indicator circuit.

C.

The clock circuit.

SOME SAMPLE CALCULATIONS

The application of the exchange principle to the evaluation of temperature and humidity gradients can be illustrated with data collected by the authors over a stand of saltcedar (Tamarix chinensis).

The exchange mechanism was mounted on a lightweight mast with the lower measurement level at the height of the canopy tips, and the psychrometers were exchanged immediately following each measurement.

sensors were permitted to come into equilibrium at the new position before the measurements

The and subsequent exchanges were repeated.

The choice of time interval between measurement and interchange is not well defined for field experiments.

hr).

In the sample discussed here, the instruments were read and exchanged each 6 minutes (0.1

The 0.1 hr readings were averaged for 0.5 hr, giving 6 readings (3 with each sensor) at each of the two levels for each half -hour average.

The half -hour averages were then subtracted to obtain the mean differences for temperature and vapor pressure.

The data required for the sample calculation of the mean temperature and vapor gradients for 1330-

1400 hrs are tabulated in Table 2.

The temperature and vapor data are listed in the table in the order that they were recorded, i.e., the results from psychrometer 1, followed by the results from psychrometer 2.

The location signal defines the position of each psychrometer at the time of sampling.

In this experiment, psychrometer 1 is at the upper level when the signal reads +8 mV, and at the lower level when the signal is -8 mV.

40

2

1

1

1

1

1

1

Table 1.

Parts list for one controller and one exchange mechanism.

Quantity Parts

Description

4

Microswitches

1

4

1

1

1

2

1

8

2

2

1

1

12 ft

(3.7 mm)

Capacitor, 4.0 MFD, 370VAC

Rotary switching relay

6" (152 mm) L. arm oil- filled

DPDT, 5A, 125VAC fuse holders center position female male male female

4 RPM, high- torque

#BZ- 2RW863 -A2 7651

Dayton #4X426

Potter & Brumfield

#19AP11A lamp fuses

Chassis -mounted buss

DPDT switch

8 -pole socket

8 -pole plug

4 -pole plug

4 -pole socket

Permanent split capacitor motor

Aluminum chassis box

Dayton #22813

Sprockets, 18 -tooth for

#41 chain

Electric clock

U -bolts

Roller chains

Ball- bearing mandrels

Coupling body

Coupling body

Rubber spider

Aluminum instrument supports

Angle iron, steel or aluminum

13 "L x 5 "W x 3 "D

(330 x 127 x 76 mm)

3 5/8" (92 mm) dia.,

5/8" (16 mm) shaft dia., keyed & dual setscrews

Martin #41818 time switch to fit 1 1/4" (31.8 mm) pipe

Dayton #2E130

10 ft (3.05 m), with master links

5/8" (16 mm) dia.,

10" (254 mm) L. shaft

5/8" (16 mm) bore dia.

1/4" (6.4 mm) bore dia.

Dayton #41

Dayton #2X625

Dayton #4X177

Dayton #4X236

Dayton #1X409

19 "L x 1 3/4 "W x

1 /4 "T

(483 x 44 x 6 mm)

1.25" x 1.25' x 1/8"

(32 x 32 x 3.2 mm)

Table 2.

Data collected with a pair of psychrometers mounted on the interchange mechanism.

T1 and el are temperature and vapor pressure measured with psychrometer 1, while T2 and e2 are measured with psychrometer 2.

Time

Location

T1( °C) el Pa

T2( °C) e2(Pa)

1330

1336

1342

1348

1354

1400

+8 mV

-8 mV

+8 mV

-8 mV

+8 mV

-8 mV

33.51

33.92

33.68

33.81

33.69

34.33

570

647

646

648

622

716

33.59

33.86

33.82

33.86

33.72

34.31

693

727

759

728

733

781

41

The mean temperature at level

1 (T1 at level 1) is thus 11 = 1/6(33.59 + 33.92 + ... + 33.72 +

34.33) - 33.87 °C.

Similarly, T2 = 33.82 °C, el = 699 Pa and e2 = 679 Pa.

The gradients over the 30minute period and the 1 m difference in height are thus aT /ez = -0.05 °C /m and to /ez

= -20 Pa /m.

The gradients are ultimately incorporated into the Bowen ratio model and form the basis for estimating evapotranspiration.

The results of these and other calculations have been tabulated by Davenport et al., 1978.

CONCLUSIONS

The basic design is well adapted for obtaining precise temperature and vapor pressure gradients.

The control system operated flawlessly throughout experimental periods totaling 15 days.

Future systems could be lightened by 25 to 30 pounds by fabricating the main support from aluminum angle, and by using bicycle sprockets, chains, and nylon or bronze bearing blocks.

These refinements, however, are not necessary for field use.

REFERENCES CITED

Black, T. A., and K.

B. McNaughton.

1971.

Psychrometric apparatus for Bowen -ratio determination over forests.

Bound. Layer Meteor. 2:246 -254.

Davenport, D. C., R. M. Hagan, L. W. Gay, B. E. Kynard, E.

K. Bonde, F. Kreith, and J.

E. Anderson.

1978.

Factors influencing usefulness of antitranspirants applied on phreatophytes to increase water supplies.

OWRT Completion Report C -6030.

Cont. No. 176, California Water Resources Center,

Davis, CA.

181 pp.

McNeil, D.D., and W.J. Shuttleworth, 1975.

Comparative measurements of the energy fluxes over a pine forest.

Bound. -Layer Meteor. 9:297 -313

Sergeant, D.

H., and C.

B. Tanner.

1967.

A simple psychrometric apparatus for Bowen ratio determinations.

Jour. Applied Meteor. 6:414 -418.

Tanner, C.

B.

1963.

Basic instrumentation and measurements for plant environment and micrometeorology.

Dept. Soils Science, Soils Bull. 6, Univ. Wisconsin, Madison.

Webb, E.

K.

2.

1965.

Aerial microclimate.

In:

Waggoner, P. (Ed.).

Agricultural Meteorology.

Chapter

Meteor. Monographs, Volume 5.

American Meteorology Society, Boston, MA.

ACKNOWLEDGMENTS

The work upon which this publication is based was supported in part by funds provided by the

United States Department of the Interior (Project C -6030) as authorized under the Water Resources

Research Act of 1964, as amended, and in part by the Arizona Agricultural Experiment Station, Hatch

Project 04.

The authors wish to thank Mr. Charles L. Constant for his assistance in the construction of the exchange mechanism.

42

AN INTERACTIVE MODEL OF SUSPENDED SEDIMENT YIELD ON

FORESTED WATERSHEDS IN CENTRAL ARIZONA by

William O. Rasmussen and Peter F. Ffolliott

INTRODUCTION

A prototypical computer simulation model which predicts suspended sediment concentrations in streamflow runoff has been developed to aid watershed management specialists and land use planners estimate the impacts of alternative management practices on suspended sediment yield.

The model allows users at remote locations to readily obtain predictions of sediment yields with modest computer equipment and commonly available data.

The model, called SED, is structured in an interactive format to facilitate operation by persons not familiar with computer operations.

The model is written in ANSI

Standard FORTRAN, requires approximately 5,000 words of core, and is currently operative on a DEC -10 computer at the University of Arizona.

The prototypical version of SED has been developed to represent southwestern ponderosa pine forest and pinyon -juniper woodland ecosystems in central Arizona; however, the conceptual framework is considered applicable to other ecosystems.

CONCEPTUAL SIMULATION MODEL

Instantaneous suspended sediment concentration in surface runoff from a watershed has been represented by functions of several variables (Anderson, 1949, 1954; Rosa and Tigerman, 1951; Ursic and

Dandy, 1963; Hansen, 1966).

The time related variable changing most appreciably for a given storm event is discharge.

Other variables often included in these representations are expressions of vegetative density, accumulations of organic material, etc.

Parameters other than discharge are variable in a longer temporal sense, but are considered constant for a single runoff event.

Concentration of suspended sediment is represented by a function of discharge for individual storm events.

Discharge is considered a function of time.

With this in mind, a runoff event with a total runoff,

Q, flowing for a time T is given as:

Q =1 Tq(t)dt o

The volume of water, dQ

, leaving a watershed in time element dt at time t after initiation of

1 surface runoff is given as: i

(1) dQ1

= q(ti)dt

(2)

This expression, multiplied by the instantaneous concentration of sediment, f(q(t )), at time interval dt, gives the weight of suspended sediment discharged, dWs, in that interval.

l

This is expressed as: dWs = f(q)d0

= f(q)q(t)dt

The integration of this relationship over the duration of the surface runoff event gives the total weight of suspended sediment carried by water as:

(3)

(4)

The authors are, respectively, Assistant Research Professor and Professor, School of Renewable Natural

Resources, University of Arizona, Tucson.

43

A surface event hydrograph is composed of both a rising and a receding component.

If the crossover from one to the other component occurs at time T1 after initiation of surface runoff, the weight of suspended sediment produced throughout the event, W5, is given as:

Ws jT1 fl(q)gl(t)dt

T

+1

f

2

(q)q

2

(t)dt

Ti where: fl(q) concentration of suspended sediment for rising stage; ql(t) discharge function for rising stage; concentration of suspended sediment for receding stage; and f2(q) q2(t) = discharge function for receding stage.

(5)

Here, the rising and receding sediment concentration functions are different for a given discharge.

As an approximation to a single surface runoff event, a hydrograph can be assumed to be of a triangular shape (Figure 1).

Peak runoff rate, q max, is reached midway through the event, at time T12.

The rising and receding discharge functions are linear in the form: q(t) bt + c

(6)

Figure 1. -- Plot of triangular hydrograph.

If the total volume of surface runoff is Q and the duration of the event is T, the discharge function may be written as: jT/2

T gl(t)dt +fT/2g2(t)dt

Q

(7) or

Q = Q/2 + Q/2

(8)

Equation (8) is apparent by inspection of Figure 1.

The discharge function q and q tained either by applying boundary conditions at time t =0, t /2, and t, or through 'use may be ob-

2of trigometric relationships illustrated in Figure 1.

Using trigometric relationships results in:

Q

(9) or, on rearranging:

2Q max =

T

(10)

For the rising component of the hydrograph, the following relationships hold: ql = gmax = L

T/2 t

T/2

44

or, on rearranging:

91 = 4t,

O=t4T/2

Similarily, the descending portion of the curve results in: q2 = 5 T -t), T /2,t4

(12)

(13)

The use of equations (12) and (13) in equation (14) yields a function which predicts the total suspended sediment weight by watershed and runoff event.

This is expressed as:

¡T /2f o i

(4Qt)(

T2

2

T-

)dt +

1T f ( ±T -t))(

2

T

2

T

2

T -t))dt

(14)

SED has been written to numerically integrate equation (14) to yield total weight of suspended sediment discharged from a watershed in a surface runoff event of Q area inches, over T hours.

The above relationships apply to water release during the summer.

Winter release is approximated by a constant value throughout the day.

A daily release of Q cubic feet would yield a constant flow, qc, of: q = ocfs (15)

Using this constant flow in the suspended sediment function, total daily suspended sediment yield is represented by:

Ws ° f(gc)Q

(16)

Equations for instantaneous suspended sediment concentrations have been developed to represent ponderosa pine forests and pinyon -juniper woodlands in central Arizona (Hansen, 1966).

One set of equations for rising and receding components of a hydrograph, respectively, is given as: log(f1(q)) = 2.48 - .02L + .661og(q) - .651og(ga)

(17) and log(f2(q)) = 2.48 .02L + .471og(q) - .65log(ga) (18) where:

L = percent litter cover; qa = average annual streamflow (in inches); f1(q) = suspended sediment concentration for the rising portion of the hydrograph (in ppm); f2(q) = suspended sediment concentration for the receding portion of the hydrograph (in ppm); and q = discharge (in cfs).

These relationships resulted from analysis of source data from watersheds ranging from less than

100 to over 2000 acres in size.

These watersheds are covered with volcanic material, (primarily basalts, although there are smaller areas of cinders), agglomerates, and tuffs.

The heavy clay soils derived from the surface material have a high rock content (Williams et al., 1967).

Relationships similar to the above can be assembled from existing source data from other locations, with resultant equations readily incorporated into SED.

The ability to introduce other sediment response functions into the prototypical model has been one of the main design criteria in the development of this simulator.

This feature should be useful in predicting suspended sediment concentration in locations other than Arizona.

For example, source data from Colorado and Minnesota are currently being analyzed for development of appropriate functions for those locations.

APPLICATION OF MODEL

Perhaps the best way to illustrate the application of SED in simulating concentrations of suspended sediment in surface runoff is through an example.

For illustration, a hypothetical 2,500 acre watershed of southwestern ponderosa pine forest is examined to estimate effects of vegetation manipulation on suspended sediment concentration.

Operation of the model begins with a question as to which type of output, daily or instantaneous, is requested (Figure 2).

In the hypothetical example, daily values will be used.

is answered by entering a

'D'.

As such, the question

Next, the simulator queries the user as to the nature of the surface runoff event, i.e., summer or winter.

A summer event is to be simulated in the example, requiring a

45

'YES' in response to SUMMER STORM (YES /NO, CR GIVES YES) ?1 /

SEDI

DO YOU WANT DAILY (D) OR INSTANTANEOUS (I) VALUES ? D

SUMMER STORM (YES /NOr CR GIVES: YES) ?

WATERSHED AREA (ACRES) ? 2500

ENTER 'L' TO INPUT LITTER COVER AND STREAMFLOW DATA,

OR 'V' TO INPUT OVERSTORY VOLUME DATA. L

LITTER COVER IN PERCENT (CR GIVES: 60) ? 76

ANNUAL STREAMFLOW IN IN. /YR. (CR GIVES: 2.9) ?

DURATION OF RUNOFF IN HOURS (MAX = 24) ? 20

ENTER DAILY STREAMFLOW.

STREAMFLOW IN INCHES (ENTER -1 TO EXIT) ?

.1

TOTAL WEIGHT OF SUSPENDED SEDIMENT = 1285. POUNDS.

MAX. SED. CON.=

MAX. DISCHARGE=

38. PPM.

25.1 CFS.

STREAMFLOW IN INCHES (ENTER -1 TO EXIT) ? .25

TOTAL WEIGHT OF SUSPENDED SEDIMENT 5539. POUNDS.

MAX. SED. CON.= 70. PPM.

MAX. DISCHARGE= 62.8 CFS.

STREAMFLOW IN INCHES (ENTER -1 TO EXIT) 7 -1

ANOTHER VEGETATION DENSITY (YES /NO, CR GIVES: NO) ? YES

LITTER COVER IN PERCENT? 44

ENTER DAILY STREAMFLOW.

STREAMFLOW IN INCHES (ENTER -1 TO EXIT) ?

.1

TOTAL WEIGHT OF SUSPENDED SEDIMENT = 5609. POUNDS.

MAX. SED. CON.=

MAX. DISCHARGE=

167. PPM.

25.1 U.S.

STREAMFLOW IN INCHES (ENTER -1 TO EXIT) ? .25

TOTAL WEIGHT OF SUSPENDED SEDIMENT =

24181. FOUNDS.

MAX. SED. CON.=

MAX. DISCHARGE=

306. PPM.

62.8 CFS.

STREAMFLOW IN INCHES (ENTER -1 TO EXIT) ? -1

ANOTHER VEGETATION DENSITY (YES /NO, CR GIVES: NO) ?

Figure 2. -- Hypothetical example of SED.

The area of the watershed is requested by WATERSHED AREA (ACRES)?

(Figure 2).

A response of 2500 is given

The next question posed is which response function is to be used in the prediction of suspended sediment concentration.

Two alternative response functions are available; both have about the same

"goodness" of fit, as measured by the coefficient of determination.

able, the user selects the appropriate function.

Depending on the input data avail-

For example, it is assumed that percent litter cover and annual streamflow data are available.

The choice of using this data set is conveyed to the simulator by a response of 'L' to ENTER 'L' TO INPUT LITTER COVER AND ANNUAL STREAMFLOW DATA, OR 'V' TO INPUT

OVERSTORY VOLUME DATA.

The values for these two parameters are2yhen requested.

value is requested, with a default value of 60 percent offered._

First the litter cover

In the example, however, a value of

76 percent is considered appropriate; this corresponds to a ponderosa pine forest overstory density of

60 square feet of basal area per acre.

Next, average annual streamflow is requested.

The default value of 2.9 inches is selected.

Two streamflow discharges will be examined in the example, with the duration of both events held at

20 hours.

This value is input in response to DURATION OF RUNOFF EVENT IN HOURS (MAX =24)?

1.

The response of YES may be input to the computer either through depressing the carriage return,

CR, key or by typing 'YES'.

For the example, the carriage return key was depressed.

2.

In many instances, a default value representing the "best" or "most frequent" response to a statement or question is offered to a user in the version of SED.

Acceptance of a default value allows a simulation exercise to continue without explicit knowledge of the input requested.

It should be noted, however, that the user has the option of overriding any default value offered, if desired.

46

The simulator requests the daily streamflow in inches.

For the example a value of 0.1 is input.

At this point, SED outputs total weight of suspended sediment produced from the watershed in one day (pounds), maximum sediment concentration (ppm), and the maximum discharge (cfs).

Following this, a streamflow of 0.25 is conveyed to the simulator, and outputs representing this example are generated (Figure 2).

These are the only discharges we wish to evaluate; therefore, the question of STREAMFLOW IN INCHES

(ENTER -1 to EXIT)? is answered with a

' -l'.

The simulator then asks if the forest overstory density is to be modified.

For the example, we will reduce density by 50 percent.

The simulator is informed of our intentions by a

'YES' response to the question of ANOTHER VEGETATION DENSITY?

Litter cover is input at 44 percent, corresponding to a basal area of 30 square feet per acre, while the annual streamflow is held constant.

The simulation is carried out using the same discharge values as before (i.e., 0.1 and

0.25).

As expected, suspended sediment yield increased due to the reduction of the vegetation density.

Computations for the hypothetical example are completed; therefore, the response to the question

ANOTHER VEGETATION DENSITY? was NO, causing the program to terminate.

INTERACTIONS WITH OTHER MODELS

While this simulator has been designed to operate alone, it has also been structured to be linked with other simulators, if desired.

SED is part of a family of computer models being developed to help watershed management specialists and land use planners estimate impacts of alternative land management practices.

This family, called ECOSIM (Ecosystem COmponents Simulation Models), includes three general modules: FLORA for estimating responses of forest overstory,Tierbaceous understory, and organic material; FAUNA for evaluating animal habitats, carrying capacities, and population dynamics; and WATER for assessing streamflow yield, sedimentation, and chemical quality (Larson et al., 1978).

SED is a component of the WATER module, with interfaces to many of the other modules within ECOSIM.

REFERENCES CITED

Anderson, Henry W.

1949.

Flood frequencies and sedimentation from forest watersheds.

Geophys. Union 30:567 -584.

Trans. Amer.

Anderson, Henry W.

1954.

Suspended sediment discharge as related to streamflow, topography, soil, and land use.

Trans. Amer. Geophys. Union 35:268 -281.

Hansen, Edward A.

1966.

Suspended sediment concentrations as related to watershed variables in central

Arizona.

Paper presented at American Society of Civil Engineers, Hydraulics Dividison Conference,

Madison, Wisconsin, Aug. 16 -18, 1966.

13 p.

Larson, Frederic R., Peter F. Ffolliott, William O. Rasmussen, and D. Ross Carder.

1978.

Estimating impacts of silvicultural management practices on forest ecosystems.

Paper presented at 10th

Annual Waste Management Conference entitled Best Management Practices for Agriculture and Silviculture, Rochester, New York, April 26 -28, 1977.

24 p.

Rosa, J. M., and M. H. Tigérman.

1951.

Some methods for relating sediment production to watershed conditions.

U.S. Forest Service Intermountain For. and Range Expt. Station.

Research Paper No.

28.

Intermountain For. and Range Expt. Station, Ogden, Utah.

9 p.

Ursic, S. J. and Farris E. Dendy.

1963.

Sediment yields from small watersheds under various land uses and forest covers.

Proceedings of the Federal Inter -Agency Sedimentation Conference.

U.S. Dept.

Agr. Misch. Publ. No. 970.

pp. 47 -52.

Williams, John A., and Truman C. Anderson, Jr.

1967.

Soil Survey Beaver Creek Area, Arizona.

USDA

Forest Service and Soil Conservation Service, Superintendent of Documents, U.S. Government Printing

Office, Washington, D.C.

75 p.

47

A WATER BUDGET FOR A SEMIARID WATERSHED by

Severo R. Saplaco, Peter F. Ffolliott, and William O. Rasmussen

INTRODUCTION

In the semiarid Southwest, water supplies are becoming scarce as water tables fall.

Groundwater wells have been lowered about 20 meters to pump water in Tucson; while in Phoenix, wells must be drilled, on the average, 35 meters lower than they were originally to obtain water (Wright, 1966).

Continuous use of the groundwater supply could lower water tables to a point that is uneconomical to pump for domestic uses.

Once the groundwater supply is exhausted, alternative water sources must be tapped to minimize, if not prevent, productive lands from becoming unusable deserts.

To understand the development of water supplies, baseline hydrologic information is needed.

In this study, the water budget on a small semiarid watershed was evaluated and models of surface runoff, soil moisture content, and suspended sediment were developed.

In addition, a chemical analysis of runoff waters was performed and a water balance simulation model was evaluated.

DESCRIPTION OF STUDY

OBJECTIVES

The primary objective of this study was to evaluate the water budget of a small semiarid watershed.

This objective was accomplished by identifying and quantifying the components of a water budget associated with a small instrumented watershed.

The primary components considered were rainfall, surface runoff, soil moisture content, and (collectively) other water losses.

The water budget was conceptualized as:

P = SRO + SMC + L where:

P = rainfall (cm);

SRO = surface runoff (cm);

StN

= soil moisture content (cm); and

L = other water losses (cm).

Secondary and tertiary objectives of the study were to develop regression models of surface runoff, soil moisture content, and suspended sediment, and to analyze the chemical quality of surface runoff, respectively.

THE STUDY AREA

The study site, a 6.5 hectare subwatershed which lies on the southeastern portion of the Atterbury

Watershed, is located 16 km east of Tucson, Arizona.

The watershed drains into the upper Santa Cruz

River, in the geologic area of Arizona known as the basin and range province (Fogel, 1968).

The study site has an elevation of about 3200 feet (975 m).

A relatively flat, single -slope terrain characterizes the entire watershed.

Average slope is 3 percent, ranging from 1 to 6 percent.

The aspect is northwest.

The authors are respectively, Assistant Professor, College of Forestry, University of the Philippines at Los Banos, College, Laguna, Professor and Assistant Research Professor, School of Renewable Natural

Resources, University of Arizona, Tucson.

49

Youngs et al. (1931) identified the major soil series in the Atterbury basin as Tubac, Mohave,

Laveen, Tucson, and Pinal.

The soil textural classification ranges from sandy loam to clay loam.

Gelderman (1964) reported sandy and gravelly soils on ridges of the western and central parts of the watershed.

A cemented zone of lime accumulation occurs 15 to 60 cm below the surface.

On the eastern side of the basin, soils are characterized by sandy loam and loam surfaces underlain by accumulation of clay.

A zone of lime accumulation is found 50 to 75 cm below the soil surface.

Soils along the virtually level drainage channels are characterized by loamy surfaces.

These bottomlands are underlain by clay loams.

Soil infiltration characteristics vary.

The sandy and gravelly sandy soils on the ridges of the eastern and central parts of the watershed have good infiltration rates.

Good infiltration also occurs along the drainage channels; however, infiltration rates on the eastern side of the area are restricted by the clay subsoil (Gelderman, 1964).

The Atterbury Watershed has a semiarid climate.

Mean annual precipitation is 28 cm, occurring during two distinct seasons: summer and winter.

Approximately 55 percent of the mean annual precipitation falls during the summer months, primarily July and August (Fogel, 1968).

Winter precipitation occurs mostly from December to February.

Sumner precipitation is usually characterized by short duration, high intensity thunderstorms.

These summer thunderstorms are the major source of surface runoff from small semiarid watersheds.

Fogel

(1968) reported that summer thunderstorms in southern Arizona are produced by the condensation of water vapor, slowly advected into the area by a light but steady flow from the Gulf of Mexico and the Atlantic

Ocean.

Winter precipitation events are normally of low intensity and long duration.

These winter events are the most important source of runoff and the resulting streamflow contributes largely to groundwater recharge.

Winter precipitation results from the condensation of water vapor carried over the area by migratory cyclones as they move in a west to east direction (Fogel, 1968).

The watershed has sparse vegetation which offers little protection to the soil.

The leaves of the dominant plant species are usually widely dispersed, thin, and small, which makes them ineffective in intercepting precipitation.

Major plant species are creosote bush, white thorn, jumping cholla, catclaw, mesquite, and a few palo verde.

Few forbs, shrubs, and grasses, such as fluffgrass, gray thorn and zinnias are also present.

FIELD PROCEDURES

Rainfall data were collected in a Belfort weighing -type recording gauge from January 1972 to

December 1976.

In addition, rainfall was measured by three plastic Tru -check non -recording rain gauges from January through December 1976.

The recording rain gauge was installed by the Office of the Water Resources Research Center, University of Arizona, to obtain a long term record of rainfall.

The three non -recording rain gauges were systematically installed to index variability of rainfall on the study site.

three non -recording gauges were made following a rainfall event.

Rainfall readings on the

Surface runoff was measured on an HL flume (Belfort Portable Liquid Level Recorder FW -I, which was installed by the Office of Water Resources Center, University of Arizona) from January 1972 through

December 1976.

Surface runoff was also measured on three 4 x 6 feet (1.22 x 1.83 m) runoff plots from

January through December 1976.

The runoff plots were installed adjacent to the non -recording rain gauges to obtain an index of on -site surface runoff.

Runoff was collected after each storm in a 18.6 liter plastic bucket installed at the lower ends of the runoff plots.

A sheet of plywood was used to prevent rainfall entry.

The runoff plots were bounded by aluminum sheets buried 12 cm deep to prevent water inflow and outflow from the plots.

Soil moisture content was determined by the gravimetric method.

Soil samples were collected at points along four transects established systematically on the study area.

The interval between points was approximately 25 meters.

Three soil samples were gathered on each sampling point using a soil collector tube at depths of

0 - 5.1, 5.1

- 10.2, and 10.2 - 15.2 cm.

for 24 hours, and reweighed.

The samples were placed in cans, weighed, oven dried at 110'C

Soil samples were collected before and after each rainfall event from

February through December 1976 to index the changes in soil moisture content.

ured at times of soil sampling.

Rainfall was also meas-

Other water losses of the water budget were assumed to be a combination of evaporation, interception, and transpiration.

Direct measurements of these losses were not made because of the lack of instrumentation.

This component of the water budget was evaluated as a residual term in the total water budget.

50

Water samples used to analyze the water quality were collected from March through December 1976 in the 18.6 liter plastic bucket previously described.

Before the samples were collected, water in the bucket was thoroughly stirred manually.

Then, plastic bottles were submerged in the bucket until they were filled with water.

The water samples were analyzed at the Soils, Water and Plant Tissue Testing

Laboratory, Department of Soils, Water and Engineering, University of Arizona.

ANALYTIC PROCEDURES

Synthesis of the water budget; regression models of surface runoff, soil moisture content, and suspended sediment; and a water quality analysis were made.

All the regression models were evaluated at the 0.05 alpha level.

The inventory method was used to evaluate the water budget for the entire watershed (Gray, 1973).

Magnitude of the major components of the total water budget were evaluated.

Rainfall and surface runoff were evaluated directly from field data; while, soil moisture content was determined using the soil moisture accounting procedure.

In addition, the soil moisture content (in percent) was converted to cm of water (Blake, 1965).

It was assumed that the difference in the soil moisture content before and after a rainfall event is the rainfall amount absorbed by the soil for that time interval.

Regression models of the HL flume and plot surface runoff were developed using stepwise regression analyses (Nie et al., 1975).

In the stepwise procedure, one variable was entered into the regression at each stage, from best to poorest to explain the variance in the dependent variable.

Regression models of soil moisture content and suspended sediment were also developed using a stepwise multiple linear regression analysis.

Water samples were analyzed for electrical conductivity, soluble salts, calcium, magnesium, sodium, chloride, sulfate, nitrate, fluoride, bicarbonate, and carbonate.

The concentrations of all the chemical elements were expressed in mg /l.

RESULTS AND DISCUSSION

SYNTHESIS OF THE WATER BUDGET

A water budget for the 6.5 -ha subwatershed during 1976, the year of detailed on -site measurement, was evaluated.

The total annual rainfall for the study year was 20.2 cm, 6 cm lower than the 5 -year

(1972 -1976) mean annual precipitation on the Atterbury Watershed.

Approximately 55 percent of the total annual rainfall occurred during the summer months.

A preponderance of summer rainfall on the Atterbury

Watershed has been reported by Fogel (1968).

Out of 34 rainfall events, 10 events produced measurable runoff.

The runoff -producing rainfall events occurred during the summer months (primarily July and August) and resulted in intense and localized rainfall.

cent.

Total surface runoff for the study year was 0.4 cm.

The runoff -rainfall ratio was about 2 per-

Ratios for individual events ranged from 1 to 8 percent and closely resembled the runoff precipitation ratio for the entire Tucson drainage (Sellers, 1965).

However, the lower range of the runoff -rainfall ratio (1 percent) was smaller than the lower range (3 percent) for the entire Tucson drainage.

This finding suggests that a lesser amount of rainfall was needed to initiate surface runoff from the smaller drainage of the study area (compared to a larger area of similar hydrologic characteristics).

The runoff -rainfall ratio for the study area was low in comparison with other vegetation zones in

Arizona.

The low ratio may be attributed to high evaporation potential and low precipitation input on the area.

High evaporation potential and low precipitation in the desert shrub vegetation zone of

Arizona has been reported by Ffolliott and Thorud (1975).

The total amount of rainfall absorbed by the soil (soil moisture content) within the 0 -

15.2 cm upper soil zone was 11.1 cm, or 55 percent of the total rainfall for the study year.

The high amount of rainfall absorbed by the soil suggests good infiltration and percolation on the study site.

A good infiltration characteristic of the soil up to a depth of about 0 - 20 cm at the eastern part of the

Atterbury Watershed (where the study site is located) was observed by Gelderman (1964).

An evaluation of the amount of moisture absorbed by the soil at the three different soil depths

(0 - 5.1, 5.1

- 10.2, 10.2 - 15.2 cm) indicated a decreased soil moisture content at the deeper soils immediately following rainfall.

This may be due to a decreasing trend of moisture infiltration into the deeper soils caused by the increasing occurrence of lime accumulation and clay subsoils in addition to the lack of precipitation.

The other water losses component of the water budget was determined as a residual term.

51

The total

amount of rainfall attributed to these losses was 8.7 cm, or about 43 percent of the total rainfall for the study year.

The high amount of the other water losses component may be due to the high potential rate of evaporation on the study site.

The high potential rate of evaporation on the area may be enhanced by the almost complete absence of grass cover during the study period.

In addition to evaporation on the soil surface, intercepted rainfall contributes to the total water supply that would be available for evaporation.

Another potential factor for other water losses was moisture percolation below the 15.2 cm soil depth, the deepest soil depth considered in the study.

However, it was assumed that rainfall percolation below the 15.2 an soil depth was negligible.

This assumption was based on field observations of impermeable lime accumulation on the study site at depths of about 10 - 15 cm.

SURFACE RUNOFF REGRESSION MODEL

Two regression models designed to predict surface runoff from the study area were developed.

One model involves summer runoff -producing storms recorded at the HL flume gauging station from 1972 through

1976.

This model may serve to evaluate surface runoff from the entire watershed.

Another model consists of on -site surface runoff measured on the runoff plots.

The plot runoff model may be used to evaluate the extent of on -site runoff contribution to the surface runoff from the entire watershed.

The surface runoff regression equation for the entire watershed is:

SRO = -0.24 + 0.26 P + 0.19 MXTR r2 = 0.76

where:

SRO = surface runoff (cm);

P = rainfall (cm); and

MXTR = maximum ten -minute rainfall (cm).

Schreiber and Kincaid (1967) reported a similar finding in on -site surface runoff from short -duration convective storms at the Walnut Gulch Experimental Watershed, Arizona; while, Osborn and Lane (1969) found comparable results in precipitation- runoff relations on small semiarid rangeland watersheds in

Arizona.

The average value of the maximum ten -minute rainfall for the study year was used so that the equation may be solved to determine the minimum amount of rain needed to initiate runoff from the entire watershed.

The equation was equated to zero and solved, with the solution indicating that at least 0.65

cm of rain was needed before surface runoff was generated at the HL flume gauging station.

Initiation of surface runoff with the minimum rain of 0.65 cm may be true with the assumption that rain fell as suggested by the average ten -minute rainfall.

This minimum rainfall is expected inasmuch as runoff occurs only with sufficient rainfall.

The needed minimum rainfall to generate runoff from the watershed may help explain why only about 30 percent of the total rainfall events for the study year produced measurable runoff.

The regression equation of surface runoff from the runoff plot is:

SRO = -0.35 + 0.81 P r2 = 0.87

where SRO and P were as defined previously.

To determine the minimum rainfall needed to generate runoff from the runoff plots, the equation was equated to zero and solved.

The solution indicated that at least 0.43 cm of rain was needed before runoff was initiated from the runoff plots.

This minimum rainfall was 30 percent smaller than what was needed to generate runoff from the entire watershed.

It is generally assumed that the smaller the drainage area, the lesser amount of rainfall is needed to initiate surface runoff.

SOIL MOISTURE CONTENT REGRESSION MODEL

Regression equations to predict soil moisture content of soil samples gathered from the study sites after rainfall at three soil depths (0 - 5.1, 5.1

- 10.2 and 10.2 - 15.2 cm) were developed.

The equation of soil moisture content at a soil depth of 0 - 5.1 cm is:

SMC = 3.55 + 0.04 RDUR r2 = 0.74

52

where:

SMC = soil moisture content (percent); and

RDUR = rainfall duration (minutes).

At the soil depth of 0 - 5.1 an, minimum soil moisture content was 3.6 percent.

moisture content approaches the wilting point level for sandy loam soils (Brady, 1974).

This minimum soil

A low soil moisture content was expected because of the high evaporation ratio and low precipitation input on the study area.

The low minimum soil moisture content is perhaps one of the major factors that limits the occurrence and growth of grass species on the study site.

At a soil depth of 5.1

- 10.2 an, the equation of soil moisture content is:

SMC = 2.75 + 0.02 RDUR r2 = 0.75

The minimum soil moisture content at this soil depth (5.1

- 10.2 cm) was about 2.7 percent.

This minimum soil moisture content was about 0.8 percent lower than the minimum soil moisture content at the

0 5.1 cm soil depth.

This lower soil moisture content may be attributed to the delayed redistribution of moisture into the lower soil depth (Gelderman, 1964).

In addition, soil moisture infiltration to recharge the soil moisture reservoir at the 5.1 - 10.2 cm soil depth may not take place when rainfall is limiting.

At the third soil depth (10.2 - 15.2 cm), the regression equation of soil moisture content is given as:

SMC = 2.5 + 3.13 P r2 = 0.85

where:

SMC = soil moisture content (percent); and

P = rainfall (cm).

At this soil depth (10.2 - 15.2 cm), minimum soil moisture content was 1.05 and 0.24 percent lower than the minimum soil moisture content at the 0 - 5.1 and 5.1

- 10.2 cm soil depths, respectively.

This result may be due to the combined effects of redistribution of moisture and limited amount of rainfall.

SUSPENDED SEDIMENT REGRESSION MODEL

A regression equation to predict suspended sediment of the runoff waters from the watershed was developed and is given as:

S = 1066.5 + 755.1 P r2 = 0.77

where:

S

P

= suspended sediment (mg /1); and

= rainfall (cm).

The equation of suspended sediment indicated that surface runoff from the watershed carries over a thousand mg /l of suspended sediment.

The suspended sediment from the study area was over 10 times greater than the acceptable level for aquatic life proposed by the Environmental Protection Agency

(Campbell, 1976).

WATER QUALITY ANALYSIS

The mean electrical conductivity of the water samples was three times lower than the required amount in irrigation (Dutt and McCreary, 1970).

Total soluble salts were 50 percent less than the total dissolved solids recommended by the Public Health Service in domestic water supplies; the recommended upper limit of total dissolved units is 500 mg /l.

The pH of the water samples was within the Environmental Protection Agency levels of acceptability for aquatic life, irrigation, and public water supply (Campbell, 1976).

The water sampled had a low calcium content compared to the standard value set by the World Health Organization (Dutt and McCreary,

1970).

Cobleigh (1934) reported a limiting range for magnesium concentration of 100 to 200 mg /1 for

53

domestic water supplies.

Compared to this, magnesium concentration of the water samples was 50 and 100 times lower than the reported lower and upper limit, respectively.

The sodium concentration of the water samples was also low, about 65 times lower than 200 mg /l, which is considered harmful to people with cardiac, renal, or circulatory diseases (Laubusch and McCammon, 1955).

mg /1

The recommended Public Health Service limit for chloride concentration in drinking water is 250

(Dutt and McCreary, 1970), which is 55 times higher than the chloride concentration of the water samples collected in the study.

The chloride concentration of the runoff waters was 50 times lower than the recommended Environmental Protection Agency limit in public water supply (Campbell, 1976).

For sulfate concentration in drinking water, the recommended Environmental Protection Agency limit is 250 mg /1, which is 10 times higher than the sulfate concentration of the water samples.

The bicarbonate concentration of the water samples was over three times higher than the recommended maximum level of 50 to 60 mg /1 (Dutt and McCreary, 1970).

trial water used for drinking, ice -making, and boiling.

This maximum level applies in indus-

In terms of carbonate ion, the water samples had a zero concentration.

The fluoride concentration of the water samples was 10 to 15 times lower than the recommended

Public Health Service level for domestic water supplies and for irrigation, as proposed by the Environmental Protection Agency (Campbell, 1976).

Mean nitrate content of the water samples was 150 times lower than the concentration recommended by the Public Health Service in domestic waters (Dutt and

McCreary, 1970).

Except for bicarbonate concentration, the concentration of all the other ions were within or lower than the limits set by the Public Health Service, the Environmental Protection Agency, or the World

Health Organization for various water uses.

CONCLUSIONS

The results of the study lead to the following conclusions:

1.

Over 30 percent of the total rainfall events produced measurable runoff.

These runoff producing rainfall events occurred during the summer months, primarily July and August.

2.

The runoff -rainfall ratio was approximately 2 percent.

The ratios ranged from 1 to 8 percent and were low in comparison with other vegetation zones in Arizona.

3.

Soil moisture content accounted for about 55 percent of the total rainfall for the study year.

Soil moisture content immediately following rainfall decreased with increasing soil depths.

4.

About 43 percent of the total rainfall was attributed to other water losses.

The high value of this component of the water budget may have been due to the high potential rate of evaporation on the study site.

5.

Rainfall and maximum ten -minute rainfall were significant variables in a regression model of surface runoff from the entire watershed.

Rainfall,alone, was the significant variable in the regression model of plot surface runoff.

A greater amount of rainfall was needed to generate surface runoff from the entire watershed than from the runoff plots.

6.

Rainfall amount and duration were significant variables in regression models of soil moisture content.

The minimum soil moisture content decreased at lower soil depths.

7.

Rainfall was the significant variable in a regression model of suspended sediment.

The suspended yield from the study site was largely higher than those from other vegetation zones in Arizona.

8.

Except for bicarbonate ion, concentration of the chemical constituents present in the runoff waters were within or largely lower than the limits set for various uses of water.

REFERENCES CITED

Blake, G.

R.

1965.

Bulk density.

In Methods of soil analysis.

Wisconsin.

No. 9, pp. 374 -380.

Amer. Soc. of Agron., Inc., Madison,

Brady, N. C.

1974.

The nature and properties of soils.

McMillan Publishing Co., Inc., New York.

pp. 175 -179.

Campbell, R. E.

1976.

Water quality as affected by forest management treatment in the Beaver Creek pine watersheds.

USDA Forest Service, Rocky Mountain Forest and Range Experiment Station, Study

Plan FS -RM -1611, 27 p.

54

Cobleigh, W.M.

1934.

Utilization of alkali water.

Journal AWWA 26:

1063 -1064.

Dutt, G. R., and T. W. McCreary.

1970.

The quality of Arizona's domestic, agricultural, and industrial watersheds.

Arizona Agricultural Experiment Station, Report 256, 83 p.

Ffolliott, P.

F., and D. B. Thorud.

1975.

Water yield improvement by vegetation management:

Arizona.

National Technical Information Service, PB -246 055, 1,094 p.

Focus on

Fogel, M. M.

1968.

The effect of spatial and temporal variations of rainfall and runoff from small watersheds.

Ph.D. Dissertation, University of Arizona, 90 p.

Gelderman, F. W.

1964.

A soil survey report of the Atterbury Watershed, Pima County, Arizona.

Soil Conservation Service, Special Report, pp.

1 -4.

USDA

Gray, D. M.

1973.

Handbook on the principles of hydrology:

A general text with special emphasis on

Canadian conditions.

Water Information Center, Inc., 1,357 p.

Laubusch, E. J., and C. S. McCammon.

1955.

Water as a sodium source and its relation to sodium restriction therapy patient response.

Amer. Journal of Public Health and the Nation's Health

45:1337 -1338.

Nie, N. H., C.

H. Jenkins, S.

G. Steinbrenner, and D. H. Bent.

1975.

Statistical package for the social sciences.

McGraw -Hill Book Co., New York.

pp. 320 -360.

Osborn, H. B., and L. Lane.

1969.

Precipitation -runoff relations for very small semiarid rangeland watersheds.

Water Resources Research

5:419 -425.

Schreiber, H. A., and D.

R. Kincaid.

duration convective storms.

1967.

Regression models for predicting on -site runoff from short

Water Resources Research

3:389 -393.

Sellers, W.

D.

1965.

Physical climatology.

University of Chicago Press, Chicago.

272 p.

Wright, J.

1966.

The coming water famine.

Coward McCann, Inc., 247 p.

Youngs, F. 0., A. T. Sweet, A. T. Strahorn, T. W. Glassey, and E. N. Poulson.

1931.

the Tucson area, Arizona.

USDA Bureau of Chemistry and Soil, Report 19, pp.

1 -16.

Soil survey of

55

ABSTRACT

HOW TO SELECT EVAPOTRANSPIRATION MODELS

T. E. A. van Hylckama, R. M. Turner, and O. M. Grasz

There are many equations proposed to estimate potential evapotranspiration (E.).

Soil- moisture depletion being some sort of function of Et, the most useful equation can be selected by comparing known values with estimated values of soil moisture.

Once selected, such an equation can reliably be used to predict changes with any given shift in climatic variables.

Studies were made at the Santa Rita Experimental Range on the bajada west of the Santa Rita Mountains south of Tucson, Ariz.

Eleven observation stations were used, spanning more than 500 vertical meters and representing Sonoran Desert, Desert Grassland, and Oak Woodland -Grass ecotones.

The studies showed that models developed by Budyko, Penman, and Olivier were well suited to predict soil- moisture conditions in that 250 km area and probably over most of the southwest desert of the U.S. outside flood plains.

Methods of extrapolating climatic data, gathered in the Tucson area (about 50 km north of the mountains) to the Santa Rita slopes are discussed and graphs are presented showing by statistical analysis, that the correlation coefficient between measured and predicted soil- moisture values is on the order of 0.75 and significant at the 0.01 level.

57

GROUND WATER IN THE SANTA CRUZ VALLEY by

Marshall Flug

INTRODUCTION

Ground water in the Santa Cruz Valley of Southern Arizona represents the only dependable supply of water for municipal, industrial and agricultural needs.

The increasing popularity of the sunbelt coupled with an expanding mineral industry are placing greater demands upon an already overstressed water supply.

Ground water levels in the Valley have been declining steadily for several years with more rapid declines resulting from recent increases in withdrawal.

This paper presents a summary of water use for municipal, industrial and agricultural purposes in the Santa Cruz Valley along with estimates of annual recharge in a section of Rillito Creek.

These data are essential to proper management of the water resources within the drainage basin of the Upper

Santa Cruz River.

Since most wells in the Santa Cruz Valley do not monitor discharge, water levels, or pump performance, extensive ground water data are collected by the Department of

Soils, Water and Engineering, College of Agriculture, University of Arizona.

Annual water level measurements are taken in over 500 wells in the Valley.

In addition, annual data on water use are compiled in the following manner.

Municipal water use is deter ined from pumpage records supplied by the City of Tucson Department of Water and Sewers, private water companies, or estimated from population and per capita water use rates in areas not served by metered water supplies.

Industrial water use is either obtained from individual and corporate pumpage records or estimated from annual production data and associated water demands for each particular industry.

Agricultural consumptive water use is estimated by multiplying a crop water use requirement by the total crop acreage, as determined by a semiannual field survey.

To provide specific data on water use, the Santa Cruz Valley has been subdivided into four districts (see Figure 1) on the basis of aquifer characteristics (Schwalen and Shaw, 1957; Matlock et al, 1965; and Matlock and Davis, 1972).

The Cortaro -Canada

Del Oro District occupies the northern end of the basin lying north of Rillito Creek between the Santa Catalina Mountains and the Tucson Mountain foothills.

The City of

Tucson is located in the center of the Tucson District which extends from the Tucson

Mountain Foothills on the west to the base of the Rincons on the east and from the

Catalina Foothills on the north to San Xavier Mission on the south, including slopes of the Santa Rita and Rincon Mountains southeast of the City.

The Sahuarita -Continental

District, bounded on the north by the Tucson District and by the Pima -Santa Cruz County line on the south, occupies the bottom lands along the Santa Cruz River and adjacent valley slopes, or bajadas, between the Santa Rita Mountains on the east and the Sierrita

Mountains on the west.

The Santa Cruz County District is comprised primarily of narrow bottomlands along the Santa Cruz River from the Pima County line south to the Mexican boundary.

The author is Assistant Professor, Soils, Water and Engineering Department, University of Arizona, Tucson, AZ.

Paper No.

287 of the Arizona Agricultural Experiment Station

59

T115

T125

T 135

T14

TISS

T169

T17S

TIES

T 19 S

T 20 S

RILE

FIGURE 1.

RISE

KEY MAP

KEY MAP

SANTA CRUZ VALLEY

FROM RILLITO STATION TO

THE MEXICAN BOUNDARY

DRAINAGE DIVIDE

GROUNDWATER DISTRICTS lìÿi

- - - -COUNTY LINES

0 S

MILES

IO

ELEVATIONS

SODD-OOd

4000=3000'

5000'

DR 11111

T 21 S

T225

T

LOCATION y

.,

ARIZONA pock

RITE

RI2E RISE RITE RISE RICE RITE

LAND USE

The Santa Cruz Valley drainage area contains about 2240 square miles or two percent of Arizona land area; whereas Pima and Santa Cruz Counties combined comprise over nine percent (Arizona Crop and Livestock Reporting Service, 1977).

The Valley occupies twenty -one percent of the area in these two counties.

Land ownership by individual or private groups is 14% of Pima County land, 37% of Santa Cruz County land, and 17% for the two counties combined as compared to 18% for the state.

The remainder of Arizona land is either publicly owned and administered by county, state and federal governments or by various Indian Reservations.

The population densities in Santa Cruz and Pima Counties are, respectively, 14.1

and 50.7 persons per square mile (Valley National Bank, 1977).

Considering the relatively small proportion and spatial distribution of the private lands in Pima and Santa

Cruz Counties, over 70% of the population resides within the Santa Cruz Valley and particularly in and around Tucson.

The historical and projected populations for this area are given in Table 1.

The population data reflect the migration of people from the northeast and midwest to the sunbelt states and emphasize the need for water management.

60

County

Pima

Santa Cruz

TOTAL

County

Pima

Santa Cruz

TOTAL

SOURCE:

1940

72,838

A,1112

82,320

TABLE 1.

1950

141,216

9,344

150,560

1985

584,800

21,800

606,600

POPULATION DATA

1960

265,660

10,808

276,468

1990

651,000

24,400

675,400

1970

351,667

13,966

365,633

1995

1980

520,000

19,300

539,300

700,700

27,200

727,900

Bureau of the Census and Arizona Department of Economic Security

1977

468,100

17,600

485,700

2000

746,000

30,400

776,400

The Santa Cruz Valley was settled as an agricultural area with diversions for irrigation initially coming from the surface flows of the Santa Cruz River and later from an extensive network of wells.

Irrigated acreage in each district of the Valley varies from year to year as influenced by agricultural prices and federal acreage regulations.

The cropped acreages for each district for the years 1970 and 1977 are given in Table 2 below along with the average for this time span and the percentage in each district of the total average acreage.

A decrease in irrigated acreage from 1970 to 1977 reflects the increasing cost of ground water pumpage due to water level declines and higher energy costs.

These two factors can create an unfavorable economic situation for irrigated agriculture in the Valley.

The bulk of the irrigated acreage is composed of various grains and cotton.

It should be noted that a considerably larger irrigated area, within Pima County but outside the confines of the Santa Cruz Valley, exists west of Tucson in the Avra -Altar Valley.

District

Cortaro -Canada Del Oro

Tucson

Sahuarita- Continental

Santa Cruz Valley in

Pima County

Santa Cruz County

TOTAL -

Santa Cruz Valley

TABLE

2.

IRRIGATED ACRES

1970

2620

2620

10080

15330

5230

20560

1977

3720

2620

7040

13370

2830

16200

Average

1970 -1977

2500

2700

8300

13400

4800

18200

Percent of Total

14

15

45

74

26

100

WATER USE

Ground water in the Santa Cruz Valley is used in increasing quantities to supply municipal, industrial and agricultural demands.

A water use summary is given in Table 3 and includes data compiled for 1970 and 1977.

Water use for domestic and industrial purposes show an increase which is associated with the increase in population as discussed previously.

Domestic water use in the Tucson District showed a significant decline in 1977 when compared to 1976 data, which is attributed to a substantial water price increase and the City of Tucson's "Beat the Peak" program.

Of particular note is the distribution of water use within each district.

Water use in the Tucson District shows the heavily urbanized character of this district whereas in Santa Cruz County almost the entire quantity of water used is for agricultural purposes.

Comparing the total water use in the Santa Cruz Valley for 1970 and 1977 shows a definite trend away from agriculture toward an increased domestic water demand in agreement with the population statistics and irrigated acreage data presented in the previous section.

61

District

Cortaro-

Canada

Del

Oro

Tucson

Sahuarita-

Continental

Santa Cruz

Valley in

Pima

County

Santa

Cruz

County

TOTAL

Santa

Cruz

Valley

D

I

A

T

D

I

A

T

D

I

A

T

D

I

A

T

D

I

A

T

Use

D

I

A

T

TABLE 3.

Water Use, acre -feet

WATER USE SUMMARY

1970

% of % of

District Valley

Total Total

Water Use, acre -feet

1977

% of

District

Total

1740

450

7720

9910

50670

6650

9500

66820

13680

30910

32780

TTYTU

66340

41390

50000

157730

80

0

17790

17870

66420

41390

67790

175600

17

5

78

100

76

10

14

100

1

0

99

100

38

24

38

100

18

40

42

100

42

26

32

100

<1

1

4

6

29

4

5

38

8

17

19

44

38

24

38

100

<1

0

10

10

38

24

28

90

5500

940

11290

17730

56660

8660

9260

74580

15880

29340

24000

69220

88930

41840

44560

1750

1

D

=

Domestic;

I

=

Industrial; A

=

Irrigation;

T =

Total.

2

Includes some non -district data.

300

0

9950

89230

41840

54510

185580

31

5

64

100

76

12

12

100

23

42

35

100

51

24

25

100

3

0

97

100

48

23

29

100

% of

Valley

Total

<1

0

5

48

23

29

100

3

1

6

10

30

5

5

40

48

23

24

95

8

16

13

37

WATER RESOURCES

Although the Santa Cruz Valley is underlain by an aquifer extending to a depth of

1,000 feet in at least part of the basin, the ground water supply continues to be depleted.

Additionally, there is evidence that the quality of ground water deteriorates at increasing depth in the central portion of the Tucson District as evidenced by some wells drilled to over 1,000 feet (Schwalen and Shaw, 1957).

In general, however, the ground water elevations are declining on the average at the rate of about two feet per year with five feet per year not uncommon.

The particular decline in a given well increases as the distance to one of the ephemeral streams increases.

Only the major streams provide natural recharge during periods of extended streamflow.

A ground water mound is created under the streams by the recharge.

The Arizona Water Commission (1975) has estimated recharge to the Upper Santa Cruz

Basin in addition to estimating the quantity of ground water in storage within 700 and

1,200 feet of the ground surface.

These estimates are shown in Table 4 and are used in conjunction with the 1977 water use data, presented in the previous section, to determine an annual water balance (i.e. deficit) for the Santa Cruz Valley.

The water use and estimated recharge data indicate that water is being used at approximately three times the recharge rate.

This fact shows the serious need for water management for the

Santa Cruz Valley and explains the decline in aquifer levels.

62

Use1

Santa

Cruz

Valley

D

A

T

1977

Water Use

89

42

55

TgT

TABLE 4.

WATER BALANCE

1,000 acre -feet

Estimated

Recharge

Annual

Deficit

65 121

Estimated Ground

Water in Storage

2

<1200 ft <700 ft

56,000

28,000

1

D

= Domestic;

I

= Industrial; A = Irrigation; T = Total.

2 Estimates from Arizona Water Commission (1975).

ANNUAL RECHARGE

Annual ground water level measurements have been used in combination with annual water pumpage data to compute recharge along a 14.5 mile reach of Rillito Creek in the

Tucson District (Flug et al, 1978).

These data were used in a complete mass balance of flow along the stream section as follows:

(VF -

VI)

S =

R + Q + pI where

S

VF = volume of saturated porous media at the end of a one -year period;

VI

= volume of saturated porous media at the beginning of the period;

= storage coefficient;

R

Q

= recharge volume within a one -year period;

= volume of water pumped; and dI

= net inflow volume.

The convention taken is that R, Q, 4I are positive if fluid is being added to the system.

Using annual water level data, VF and VI are estimated by a modified use of the

Thiessen polygon method (Matlock and Davis, 1971; Diskin, 1970); data on Q are available from the water use information and S is assumed to be 0.1 for this alluvial section of Rillito Creek (Matlock et al, 1965).

The net inflow was estimated by using

Darcy's law in a discretized form across small sections of the closed region, and adopting transmissivity values from a calibrated digital model of the Tucson Basin (Fogg,

1978).

Results of the available estimated recharge are summarized in Table 5.

Year

Recharge

Ac -Ft

TABLE 5.

1961

4830

1962

ESTIMATED RECHARGE IN 14.5 MILES OF RILLITO CREEK

1963 1964

1965 1966 1967 1968

18670

11120 15200 21590

18710

67750 44960

1971 1974

35000 22540

The average annual recharge for the ten years was 26,000 acre -feet or about 1,800 acre feet per mile of channel.

(Due to missing data recharge could not be calculated for

1969,

'70, '72, and '73.) These results have been shown to be in good agreement with other estimates of recharge for Rillito Creek.

A detailed analysis of recharge has not, however, been performed for all streams within the Santa Cruz Valley.

The potential for recharge in the basin is great with an average annual rainfall of 1,800,000 acre -feet falling within the drainage area boundary (Schwalen and Shaw, 1957).

63

SUMMARY AND CONCLUSIONS

The Santa Cruz Valley, which depends entirely on ground water for supplying municipal, industrial and agricultural water demands, is presently in a serious overdraft situation.

Ground water use is three times the estimated annual recharge with population projections indicating a continuous increase in domestic and associated industrial water use.

As ground water levels continue to decline, most agricultural uses of water will become economically prohibitive.

However, the present domestic water demands alone surpass estimated annual recharge.

Many of the predictions about water use and supply in the Santa Cruz Valley depend upon estimated parameters including population projections, water withdrawals, water use efficiencies, storage coefficients, transmissivities, and stochastic events such as the distribution and intensity of rainfall.

Although many of these parameters have been studied in detail, a comprehensive analysis of their effects on the entire Santa

Cruz Valley has not been performed.

However, the data clearly indicate that a long range solution for supplying the water needs of the Santa Cruz Valley has not been found.

Although temporary relief of the water deficit problem will be provided by water deliveries from the Central Arizona Project when completed, some areas in the

Valley will continue to exhibit water level declines.

REFERENCES CITED

Arizona Crop and Livestock Reporting Service.

1976.

1977.

Arizona Agricultural Statistics,

Arizona Water Commission.

1975.

Summary

..

Phase

I

- Arizona State Water Plan

-

Inventory of Resource and Uses.

Arizona Water Commission, Phoenix, Arizona.

Diskin, M.

H.

Hydrology.

1970.

On the Computer Evaluation of Thiessen Weights.

11(1):135 -144.

Journal of

Flug, M., Abi- Ghanem, G.

V., and Duckstein, L.

1979.

An Event Based Model of Recharge

From an Ephemeral Stream.

Paper No. 2975 of the College of Agriculture, University of Arizona, Tucson, Arizona.

Fogg, G.

1978.

A Groundwater Modelling Study in the Tucson Basin.

Technical Report

32, Hydrology and Water Resources Department, University of Arizona, Tucson,

Arizona.

Matlock, W.

G., and Davis, P.

R.

1971.

Aquifer Specific Yield and Recharge in the

Tucson, Arizona Area.

Paper PR 71 -07 presented at the ASAE Pacific Regions Annual

Meeting, Las Vegas, Nevada.

Matlock, W.

G., and Davis, P.

R.

1972.

Groundwater In the Santa Cruz Valley, Arizona.

Technical Bulletin 194, Agricultural Experiment Station, University of Arizona,

Tucson, Arizona.

Matlock, W.

G., Schwalen, H. C., and Shaw, R.

J.

1965.

Progress Report on Water In the Santa Cruz Valley, Arizona.

Report No. 233, Department of Agricultural Engineering, Agricultural Experiment Station, University of Arizona, Tucson, Arizona.

Schwalen, H.

C., and Shaw, R.

J.

1957.

Water In the Santa Cruz Valley.

Bulletin 288,

Agricultural Experiment Station, University of Arizona, Tucson, Arizona.

Valley National Bank.

1977.

Arizona Statistical Review.

33rd Annual Edition, Valley

National Bank, Phoenix, Arizona.

64

TESTS ON ARIZONA'S NEW FLOOD ESTIMATES

Brian M. Reich, Herbert B. Osborn and Malchus C. Baker, Jr." 1

ABSTRACT

A method for estimating regional flood frequency was prepared by R.

H. Roeske of the U.S. Geological Survey (USGS) in 1978 for the Arizona Department of Transportation.

Hydrologists may wish to use these regression equations for estimating flood peaks or for other purposes in development or flood trol engineering.

con-

Many of those needs are for watersheds smaller than 10 sq. mi., however, for which

USGS measurements are scarce.

Records from two groups of small experimental watersheds near Tombstone and Flagstaff, one gaged by the U.S. Department of Agriculture's (USDA) Science and Education Administration and the other by the Forest Service, were used to independently evaluate the generalized Arizona relationships in specific applications to small watershed work.

The new design floods for each experimental watershed were compared with estimates made using the USGS equation for two of the six flood frequency regions (FFR) in Arizona.

The study showed that use of the generalized regional curve may underestimate flood peaks.

Deviations from the curve can be caused by land use changes, differences in analytical methods, and use of short records.

INTRODUCTION

The users of an equation need to have a feel for the accuracy within which that equation simulates the natural behavior that it attempts to model.

This is particularly true with flood peaks, whose magnitudes are the result of the joint action of many random hydrologic processes upon infinitely complex groupings of soils, land use, and watershed geometry.

In the Southwest, flood peaks per unit area decrease with increasing watershed size because of the limited areal extent of the storms and the flow abstractions in the normally dry channels.

Keppel and Renard

(1962) examined hydrographs at progressive stations and verified that these combined hydrologic processes become particularly important for desert watersheds of 20 sq. mi. or more.

Thus it may be beneficial to stratify Arizona watersheds into large and small subgroups before trying to fit flood peak (Q) area curves.

The U.S. Geological Survey (USGS) collects flood data in

Arizona with financial support from groups within the state such as the Arizona Department of Transportation (ADOT).

Roeske (1978) of USGS prepared a method for estimating regional flood frequency in six flood frequency regions

(FFR) for

Arizona.

The present streamflow network is composed of 245 unregulated sites along Arizona's rivers and streams; the state -wide average is 463 sq. mi. per gage.

This sparse network, particularly in the desert environment where the spatial and temporal variations of the rainfall input far exceed those in more humid parts of the U.S.A., results in a large error component when the data are used to prepare regional flood frequency prediction equations.

Moreover, 79 of Arizona's gages (32 %) are of the crest stage type, which estimate flood peaks less accurately than the other continuous recording stations.

Photographs of two such installations are shown in Fig.

1.

Of 32 USGS stations used by Roeske (1978) in the San Pedro and Santa Cruz basins alone, 65% were equiped with crest -stage gages.

Fig. 1.

USGS crest -gage station.

The hydrologic processes at work on small areas are different from those operating upon the larger desert watersheds,

1/ The authors are respectively:

District, Tucson; Supervisory

USDA, Tucson; and Hydrologist,

Manager, Planning and Resources Development, Pima County Flood Control

Hydraulic Engineer,

Southwest Rangeland Watershed

Research

Center,

Rocky Mountain Forest and Range Experimental Station, Flagstaff.

65

but the larger watersheds have overwhelming weight in the USGS data base.

In the Santa Cruz and San

Pedro basins, all but one of the 19 stations on watersheds smaller than 100 sq. mi. had crest -stage in congages.

Furthermore, at these small watersheds the average length of record was only 12 years, trast with the 32 years for larger watersheds.

Short records and questions about whether such records represent longer climatic extremes may further reduce confidence in the accuracy of predictions for small watersheds.

Reich (1979) examined in detail the prerequisite statistical from each stream gage.

analysis of individual flood series

We combined his graphical analyses with the latest theory (Cunnane, 1978) on plotting positions and applied them to small watersheds in Arizona.

Using the two clusters of small research watersheds in Arizona where land use and hydrologic data are carefully monitored -- one in rangeland in southeastern Arizona near Tombstone

(Walnut Gulch) the other in timber at higher elevations near Flagstaff (Beaver Creek) -- we examined Roeske's (1978) method to determine how well it applies to small watersheds in two of the six flood frequency regions (FFR) that he identified.

DESCRIPTION OF ROESKE'S EVALUATION

The six flood frequency regions (FFR) Roeske (1978) developed are shown in Fig. 2.

The end product of his report was a set of seven regression equations for each FFR, each for a return interval of 2, 5,

10, 25, 50, 100, or 500 years.

'Table

1 presents the 100 -year flood peaks Nino) in each region for a theoretical watershed having a

1 -sq. mi. drainage, which were calculated using Roeske's method.

The table also lists the size range of gaged watersheds upon which Roeske's equation was based and the standard error of estimate of the equa-

4 tion.

Of the 200 watersheds used,

12% were less than

1 sq.

mi., 24% between

1 and 10 sq. mi., 8% between 10 and 20 sq.

l l

\1

1

I_---e-.

L/BEAVER

CREEK ""Vz,

4

F`i

LAGSTAFF

\

1\

PHOENIX

.

3 i i mi., 8% between 20 and 50 sq. mi., 6% between 50 and 100 sq. mi., 8% between 100 and 200 sq. mi., and 13% between 200 and

1000 sq. mi.

About two -thirds of the data used to develop these regression equations were from watersheds larger than 10 sq. mi.

On the small watersheds included in the data base, the record length averaged less than 14 years, whereas for watersheds larger than 100 sq.

mi.

the records averaged about 26 years.

In FFR

3 and 5, watersheds larger than 1,000 sq.

mi. had average records of 38 years.

The relative importance of differing storm types and channel abstractions lead to dichotomous flood behavior between large and small watersheds.

5

-.,\

\ /

Fig. 2.

2

PIMA rCOUNTY

- --FRR AREA

BOUNDARY

I TUCSOÑ ;WAL

UT

GULCH

Location map showing USGS flood frequency regions and the Walnut Gulch and Beaver Creek watersheds.

The practical meaning of the standard error of estimate is significant.

In

FFR 5, for example, Q108

1,230 cfs

(Table 1).

However, the 86% standard error of estimate should be interpreted to mean Q100 at one -third of similar watersheds could actually be outside 1,230 +

(86% of 1,230).

The bounds -172 and

2,288 cfs

-represent a range within which two -thirds of the values lie.

Moreover, the error around the estimate increases as one considers the very large or very small watersheds for which the regression relationship was derived.

In addition to the warning implied by the published standard errors of estimate, one must consider how much of the data came from small or large watersheds.

One should be sure that the lower range of data upon which the equations were developed lend credence to the actual application.

For example, the smallest watershed used in FFR

1 was 1.84 sq. mi.

Similarly, for the high elevation (HE) FFR, where mean basin elevations are above 7,500 ft m.s.l., the smallest watershed was 1.61

sq. mi.

The amount of data used to develop the regression relations for peak discharges are summarized in Table 2.

The table shows again that the analysis of the HE region involved no watershed smaller than

1 only seven of the 60 watersheds were within this smallest size class.

sq. mi.

In FFR 5,

Even more important, these seven very small watersheds added only 98 station -years of data, in comparison with the 339 station -years that came from nine watersheds larger than 1,000 sq. mi.

Therefore, it would be unwise to apply these

66

Table 1.

Estimates of Q100 (Roeske, 1978) with the standard error of estimate and size range of watersheds used to develop regression equations in each flood frequency region.

Q100 cfs

Std. error of estimate

Watersheds used in developing the equations minimum maximum

(sq. mi.)

Q100 from

1 hypothetical sq. mile

(cfs)

5

HE

3

4

1

2

584A'490

1,100A'499

553A'610E-1.30p.915

0.0188A'369E6.09

1,230A'447

72.9A'795

91

83

66

91

86

45

1.84

0.09

0.065

0.17

0.15

1.61

A = Drainage area, in square miles

P = Mean basin elevation, in thousand feet above m.s.l.

E = Mean annual precipitation, inches

5,090

1,810

5,499

3,300

3,610

747

584

1,100

1,183

1,068

1,230

73

Table 2.

Watershed site distributions and length of records used in Roeske's study.

FFR

1

FFR 2 FFR 3 FFR 4 FFR 5

HE areas in square miles

Less than

1

1

- 9.99

10 - 19.99

20 - 49.99

50 - 99.99

100 - 199.9

200 - 999.9

More than 1,000

TOTALS

0

5

2

2

0

1

2

5

22

80

0

50

0

66

23

26

,O

4

0 w

a

g

Z t v t

L

4Y

4-

O

IA

S.

., m v

a

E r t

4,1

J-

3

N

4-

O

ó

4,

v

L q

2

r d

a

E

ó

A

VI

N

0

4-

O

L

r

a

E

S.

v

L

N

L w a

3

4-

0 q

Á

4-

O

L

=

+

2

N

N

3

4-

0

v

a in

L

a

E

Z

ó ro

In q

Á

4

O

L

3 an

L w

4,

N

4-

0

a

E r

t v

0

ó

+Drainage '

4-

O

N

L

JE

4-

O tir a

Ol

E iO

>

!

RI

4+

C

O

.e.

4 ro

I-

°Yó

4o

34

1

3

2

3

2

4

8

4

53

98

51

28

53

25

12

50

24

10

11

17

2

6

6

11

152

246

26

84

94

173

591

386

2

0

2

3

3

5

6

0

66

77 18

26

0

7

6

5

26

0

1

11

2 34

301 140 12

57

9

339

98

266

104

80

0

5

0

2

2

4

2

3

43

51

78

101

0

0

18

18

13

8

24 15

8 6

8

6

8

6

20

28

13

21

5

8

17 257

27

370

87 1752 21 392 60 1233

18

309

100 100

67

equations on watersheds smaller than

1 sq. mi.

in FFR

1 other sized watersheds.

or HE, and they should be used with care on

Consideration of the hydrologic processes at work during flood genesis in various regions and on different size watersheds should supplement the statistical results.

Floods on large Arizona watersheds generally result from widespread frontal storms that occur in the cooler season.

In contrast, extreme flood events on small semiarid watersheds result primarily from air -mass thunderstorms, which dominate the precipitation input in July, August, and September (Lane and Osborn, 1972; Osborn and Laursen,

1973).

Flood peak reduction through channel transmission losses also has considerable impact in the desert (Keppel and Renard, 1962; Lane et al., 1971; and Lane 1972).

In addition, the influence of heavy convective storms that typically cover less than 10 sq. mi. should be considered.

Table 2 shows then that small watershed processes are virtually absent from the input data for FFR

1 and HE.

The USGS equations for those regions are for large watershed relationships.

It is not surprising, therefore, that the hypothetical

1 sq. mi. Q100's in Table

1 for FFR

1 and HE appear to be much smaller than for the other four FFR's.

DESCRIPTION OF STUDY AREAS

Walnut Gulch Research Watersheds

To test Roeske's equation for FFR 5 on the smaller basins, we selected a cluster of experimental rangeland watersheds at Walnut Gulch in southeastern Arizona (Fig.

2).

Each subwatershed is equiped with rain gages and a laboratory calibrated flume -weir constructed and operated by the USDA -SEA -AR,

Southwest Rangeland Watershed Research Center.

There are 11 flume -weirs at Walnut Gulch which measure runoff from subwatersheds of

1 to 58 sq.

mi.

Eight of the subwatersheds are smaller than 10 sq. mi.

typical station is shown in Fig. 3.

A

The Walnut Gulch subwatersheds are in brush -grass rangeland typical

Southeastern Arizona, of much of

Southwestern New

Mexico, and Northern Sonora, Mexico.

Major channels with slopes of about 1% are incised.

All channels abstract large amounts of surface runoff.

The average annual rainfall is

13 to

14 inches, and

70% occurs in the summer as intense, short -duration thunderstorms.

All major flood peaks on Walnut Gulch have resulted from thunderstorm rainfall between April and October.

Runoff records from the eight subwatersheds smaller than

10 sq.

mi.

range from

13 to

22 years.

Flood peaks of up to 1,400 cfs per sq. mi. have been recorded on the smaller subwatersheds.

runoff -measuring

Beaver Creek Research Watersheds

Fig. 3.

Flow at a station, Walnut Gulch,

Arizona.

To test Roeske's FFR

3 equation on small basins, we used watersheds with a wide variety of soils under various land uses.

From 1957 to

1962, the U.S. Forest Service built stream gages on 18 watersheds near Flagstaff ranging in size from 66 to 2,036 acres.

Vegetation representing the Utah and alligator juniper types as well as the ponderosa pine type were included.

A high plateau, sloping mesas and breaks, steep canyons, and valleys characterize the topography of Beaver Creek.

Because of a layer which impedes downward movement of water, the infiltration rate is slow -- from 0.05 to 0.2 inches/ hour when the soil is thoroughly wetted.

Average precipitation varies from 12 to 25 inches across the research area, with 64% falling from October through April.

Average peak snow water equivalent on Beaver Creek is about 3.4 inches.

Fig.

an example of a Beaver Creek streamgage.

4 shows

In testing the FFR 3 equation, we used 14 of the watersheds at Beaver Creek.

They ranged in size from 0.1 to 3.2 sq. mi.

On four of these, forest overstory was either thinned or clear cut so that data could be separated into before and after treatment series.

These effects of forest treatment are indicated in Fig. 5.

In some cases a combined series for the watershed was also subjected to flood frequency analyses.

We studied the effects of various cover types and of thinning or clear cutting because any of these land uses can occur, alone or combined, at some future time on a watershed of interest to a hydraulic designer.

COMPARISON OF ESTIMATED FLOOD PEAKS

Walnut Gulch

The "best fit" in Fig.

6 represent the Q100 that we derived from the very best flood frequency analyses for 10 precisely gaged Walnut Gulch watersheds.

The equation for this least - squares -fitted

68

7000

6500

a- PONDEROSA PINE

.o- JUNIPER o- THINNED PP

6000

o

CLEAR CUT

5000

o

BLANK-SEPT 1970 FLOOD ALSO

SOLID-SEPT. 1970 FLOOD»

OTHER FLOODS o

Fig.

4.

A streamgaging station at Beaver Creek,

Arizona.

20

10 o-WALNUT GULCH

°- FFR

5

I"

1

I I

O

¡

0

U o

o

0

2 pi o

,

°

o-

o

.2

.2

1

2

AREA (mi`)

10 20

50

C3 o o_

4000

3000

o

USGS

/FFR 3

CURVE

2000

,

1000

00

A

//

2

I

AREA (mil) o\4

3

Fig. 6.

Best fit (A) for Q70p at 10 samll watersheds at Walnut Gucñ as compared with

Fig. 5.

scatter found by Roeske (1978) in Pima

County on the basis of his equation for FFR 5.

curve derived at Walnut Gulch is:

2360 A(.688 -.128 log A)

Q100

Effects of forest treatment on run off for selected Beaver Creek water sheds.

where A is the drainage area in square miles.

The deviations about the small.

curve in Fig. 6 are unusually

These 10 Walnut Gulch subwatersheds are very close together, and similarity of climate and geomorphology might be partly responsible for the scatter -free peak area relationship.

We also suspect that there is interstation correlation of the flood values in these separate series.

This property is found among other clustered experimental subwatersheds in the U.S.A., where runoff results from large area storms.

Cross correlation of Walnut Gulch annual flood series showed less dependence than might be expected, however, presumably because of the small southeastern Arizona.

size and random positioning of intense summer storms in

Correlation between concurrent peak discharges for 10 Walnut Gulch subwatershed stations are shown in Table 3.

As might be expected, successive stations on the same drainage -- subwatersheds 11 and 8, 6 and 2, 9 and 6, 2 and

1 in Fig.

7

-- show the highest correlations (r = 0.76 to

0.88) between runoff peaks (Table 3).

Also, runoff peaks from some adjacent watersheds -- 8 and 10,

9 and 10, and 10 and 11 -- are correlated (r = 0.66 to 0.77).

On the other hand, for 95% confidence limits, an r of 0.2 or less can result from random selection.

Therefore at least stations 7,

11, and

15 may be considered as independent sampling points, or stations 7 and 3, 6, 8, 9, 10, 11, and 15, respectively, are independent sampling points.

Stations 7, 11, and 15 have a combined record of 46 years, or an equivalent 46 -year record, if the period of record on Walnut Gulch can be considered as a good sample of a longer period.

If pairs of stations with correlations of less than 0.3 are considered independent sampling points, the record would be the equivalent of about 65 years.

69

Fig.

7.

Walnut Gulch subwatersheds used in best fit analysis.

Table 3.

Correlation (r) between concurrent peak discharges for 10 Walnut Gulch subwatershed stations.

Location of subwatersheds is shown in Fig. 2.

Subwatershed station

1

2

3

6

7

1

2 3 6 7

8 9 10

1.0

.76

1.0

.41

.69

1.0

.57

.63

.47

.69

.46

.54

.88

.29

.77

.72

.55

1.0

.03* .55

-.04

1.0

.80

.42

.85

.39

-.12

-.11

.87

-.09

8

1.0

.59

.76

9

1.0

.78

10

1.0

11

15

11

.57

-.11

.77

.41

.34

.37

.06

.66

1.0

15

.03

.30

.29

.29

.27

.52

.46

.55

.02

1.0

*Underlined values indicate station pairs with no statistical correlation.

The data and amounts for annual maximum peak discharges (cfs) for the same 10 Walnut Gulch subwatershed stations for 1963 -1974 are shown in Table 4.

Many maxima are recorded on the same dates on Walnut

Gulch which indicates strong interdependence between some stations.

However, annual maxima occur only occasionally on the same date at other stations -

(7 and 8, and 11 and 15, for example.

Also, examination of isohyetal maps of major thunderstorm rains (producing the major peaks) shows large differences in amount and areal variabilities between watersheds.

The storm of September 10, 1967, is shown in Fig. 8.

The storm's maximum of 3.45 inches was recorded in 45 minutes.

Annual maximum peak discharges were recorded on watersheds 8, 9, 10, and

11 during this storm, but there were large differences on the three watersheds.

Also, in spite of the magnitude of this "record" event, maximum annual peak discharges were recorded on other dates in

1967 at three stations on Walnut Gulch.

Walnut Gulch represents one locality within FFR 5 where the Roeske method could be tested.

In FFR

5 the curve for Walnut Gulch (Fig. 6) is about 70% higher than the USGS curve over the range from two to

10 sq. mi.

Thus, for small watershed work, the FFR 5 equation should be multiplied by a safety factor of about 1.7.

In other parts of the region, for example the steep -rock faces of the Catalina Mountains in Pima County a safety factor greater than 1.7 might be needed.

70

Fig. 8.

Isohyetal map of storm rainfall on 10 Sept. 1967, Walnut Gulch, Arizona.

Pima County

The flood frequency plots that result from short records at two creststage gages in Pima County are shown in Fig. 9.

The 13 observations on Geronimo Wash gave an inadequate but general indication that the flood which has a probability of 0.01 of being exceeded each year is about 3000 cfs.

The best that could be derived from the Anklam Wash data was even less reliable.

Two frequency lines on the graph suggested inadequate estimates of 3,000 or 6,000 cfs.

6000

1000

500

100

50

10

5

°

2.11 SQ. MILES

ANKLAM WASH

1965-1977

Separating variations from types of installation and difficult.

the short record is

Data were obtained from 45

USGS gages in and around the county.

A complete flood frequency analysis was performed on each of these using three types of probability paper (Reich,

1979).

The best estimates of goo for each time series for the 45 watersheds were plotted in Fig. 6.

The huge scatter of these estimates may be partially due to physical differences between these diverse watersheds.

If it were possible to separate them into sets of recognizable similar hydrologic systems, one would hope to distinguish a family of curves, somewhat concentric to the Walnut Gulch curve, fitting through each of these subsamples.

Attempts to do this by means of four descriptive flood producing features

(channel roughness, channel type, watershed slope, and runoff potential) failed.

Therefore, much of the scatter in Fig. 6 may be due to inadequate data.

1

235 10 2550

Tr

(YRS)

2 3

10 25 50

T (YRS)

Fig.

9.

Flood frequency plots for short records at two selected crest -gage stations in Pima County.

Beaver Creek

An indication of how much land use may alter peak flows within one FFR would be most valuable to hydrologists.

A general equation like one of Roeske's in Table

1 represents only average conditions.

Features like soil, cover, tributary configuration, etc. make peak discharges on a specific watershed higher or lower than average.

The effects of the forest treatment on runoff at Beaver Creek are shown in Fig.

5.

where the ponderosa pine was thinned, Q100

In both cases increase significantly after treatment in the subseries of floods shown by the line and arrow in Fig. 5.

The fact that Q 100 for the intermediate size watershed

72

before the thinning treatment is far below that for the slightly larger watershed emphasizes the expected range of estimation from short records.

Perhaps some intermediate value before thinning of about

1,000 cfs is a more reliable indicator. Similarly, although Q

0 values at the intermediate sized and larger watershed after thinning differ -- over 2,000 and over 3,000 cfs

-- indicating a general after treatment Q100 of about 2,7000 cfs, which means in increase in Q100 of about a factor of 2.7 times.

On the 0.7 sq. mi. watershed, the smallest, clear cutting ponderosa pine changed the Q 100 from 235 cfs to 1,700 cfs.

However, because of the relatively short period of record, the flood frequency plots for these values are uncertain.

More correct pre -and post- treatment Q100's might be closer together, and the treatment may be responsible for less than the large change indicated in predicted peaks.

All of the data after clear cutting included the flood of September 5, 1970, but this flood did not plot off the general trend of other floods.

Very similar QlOO 's would have been estimated even without the 1970 event.

The 1970 event, the largest in the 20 year record for this 2.8 sq. mi. watershed, lies along the line established by the rest of the peak discharges (Fig.

6).

Notice also that peak discharges in ponderosa pine or thinned ponderosa pine make up four of the six largest from the entire ries.

se-

The average of the two larger thinned ponderosa pine, 2,600 cfs, is about 40% greater than would be estimated with the FFR 3 equation.

Four watersheds, (solid symbols) in Fig. 5 did show that the 1970 flood plotted markedly above the general trend in each flood frequency plot (not shown).

In such cases, the 1970 event had to be treated somewhat as an outlier when eye- fitting a line on each frequency plot.

The pion estimates were not inflated as much as they would have been with computer analysis, and as a result these four points lie fairly close to the USGS curve for FFR 3 shown on Fig. 6.

Because of future development, many watersheds will be cleared of vegetation for harvest, roads, or urbanization.

Design values might therefore be expected to be closer to the top of the arrows in Fig. 6 than to the generalized FFR 3 curve.

In fact, a 1.4 sq. mi. watershed in ponderosa pine gaged for 20 years gave a Q100 almost 50% higher than the USGS curve.

Another watershed in ponderosa pine, this one 3.2 sq. mi. with 16 years of record, indicated a Q1p0 of 6,400 cfs from a very good flood frequency plot.

the USGS equation.

This is three times the value from

It appears from Fig. 6, therefore, that small watershed designers in FFR 3 should multiply the tion by a safety factor of about 3.

equa-

This value is compatible with the example we presented on Roeske's published errors of estimate.

Structural engineers often do use such safety factors and their calculations are much more deterministic than those computed using indadequate hydrologic data.

SUMMARY

The regional flood frequency method developed by Roeske (1978) appears to underestimate flood peaks for small rural watersheds.

Evaluation of runoff records from controlled measuring structures on the

USDA -SEA Walnut Gulch experimental Watershed and the USDA Forest Service Beaver Creek Watersheds in Arizona suggested that the Roeske method may underestimate flood peaks.

Although of excellent quality, the

USDA runoff records in Arizona are relatively short, and the Walnut Gulch watersheds are spaced quite closely.

Therefore, flood peak predictions based on these records are also uncertain.

However, the study suggested that Roeske 's method should be applied to small watersheds with caution, because for small watersheds in Arizona, the curves are based on a few short records of questionable the method is used, a resonable safety factor should be applied until accuracy.

If longer and more accurate records are available.

73

REFERENCES

Brown, H.E., Baker, M.B., and Rogers, J.J.

1974.

Opportunities for Increasing Water Yields and Other

Multiple Use Values on Ponderosa Pine Forest Lands.

USDA Forest Service Research Paper RM -129,

Fort Collins, CO, 36 p.

Clary, W.P., Baker, M.B., and O'Connel, P.F.

1974.

Effect of Pinyon -Juniper Removal on Natural Resource Products and Uses in Arizona.

USDA Forest Service Research Paper RM -128, Fort Collins, CO,

28 p.

Cunnane, C.

205 -222.

1978.

Unbiased Plotting Positions - A Review.

Journal of Hydrology.

Vol. 37.

No. 3, p.

Keppel, R.V., and Renard, K.G.

1962.

Div., Proc. ASCE 88(HY3):59 -68.

Transmission Losses in Ephemeral Stream Beds.

J.

Hydraulics

Lane, L.J.

1972.

A Proposed Model for Flood Routing in Abstracting Ephemeral Channels.

Hydrology and

Water Resources in Arizona and the Southwest, Am. Water Res. Assoc., Ariz. Sec. Ariz. Acad.

Sci.,

Hydrology Sec. Proc. 2: 439 -453.

Lane, L.J., Diskin, M.H., and Renard, K.G.

Channel System.

J. Hydrology 13:22 -40.

1971.

Input -Output Relationships for an Ephemeral Stream

Lane, L.J., Osborn, H.B.

Southeastern Arizona.

83 -94.

1972.

Proc.

Hypotheses on the Seasonal Distribution of Thunderstorm Rainfall in

Second International Symposium in Hydrology, Fort Collins, CO, p.

Osborn, H.B., and Laursen, E.M.

1973.

Div., Proc. ASCE 98(HY7):1129 -2245.

Thunderstorm Runoff in Southeastern Arizona.

J.

Hydraulics

Reich, B.M.

1976.

Magnitude and Frequency of Floods.

6, Issue 4, p. 297 -348.

Critical Reviews in Environmental Control, Vol.

Reich, B.M.

put.

1979.

Rainfall Intensity- Duration -Frequency Curves Developed From, Not By, Computer Out -

Transportation Research Record. 9 p.

Roeske, R.H.

1978.

Methods for Estimating the Magnitude Frequency of Floods in Arizona.

ADOT -RS -15 -121 U.S. Geological Survey, Tucson, AZ.

82 p.

Final Report

Williams, G., Gifford, G.F., and Coltharp, G.B.

ed Pinyon -Juniper Sites in Central Utah.

1969.

Infiltrometer Studies on Treated Versus Untreat-

J. Range Manage. 22(110 -114).

Williams, G., Gifford, G.F., and Coltharp, G.B.

chained Pinyon- Juniper Sites in Utah.

1972.

Factors Influencing Infiltration and Erosion on

J. Range Manage. 25(201 -205).

74

SOLAR POWERED IRRIGATION PUMPING EXPERIMENT by

Dennis L. Larson and C.

D. Sands II

INTRODUCTION

More than 50 million acres or about 13 percent of all farmland receives some irrigation (Sloggett, 1977).

Approximately 85 percent of the irrigated land is concentrated in 17 western states.

Underground aquifers provide water for about two- thirds of the irrigated area.

Irrigation pumping accounts for approximately 13 percent of the energy used onfarm in the U.S.

(ERS /USDA, 1977).

However, on farms with deep subsurface water sources, 70 -90 percent of crop production energy may be used for pumping water (Larson and Fangmeier, 1978).

Nationally, electricity drives pumps serving about 45 percent of the irrigated farmland.

Natural gas is used to supply water to 30 percent of the area.

Diesel fuel and LP gas meet most of the remaining pumping energy requirements.

Pumping energy costs vary with energy costs, pumping efficiency, water depth and delivery pressure.

In Arizona, deep well pumping energy costs range from less than

$20 to over $60 per acre foot of water (Hathorn, 1978).

The magnitude of the irrigation energy requirement, potential natural gas shortages, and increased pumping energy costs have led to an examination of solar energy

(Larson, et al, 1978e).

In the past three years in the U.S., four known experiments have been initiated to evaluate the use of solar powered pumping plants on farms.

Photovoltaic cells are used to provide up to 25 KW of electricity for pumpingin an experiment at Meade, Nebraska (Twersky and Fischback, 1978).

Rankine cycle turbine engines and parabolic trough type solar collectors are principal components of the three solar thermal systems.

Two of the installations, a

50 KW unit located near

Gila Bend, Arizona and a 25 KW plant located near Willard, New Mexico, are directly coupled to pumps and have electrical motor backups.

However, only the Willard installation includes energy storage.

This paper describes the fourth experiment, a 150 KW solar thermal power plant presently being constructed on a farm near Coolidge, Arizona.

OBJECTIVES

Objectives of the Arizona solar powered pumping experiment were design, construction, and evaluation of a solar thermal power plant to meet the deep well pumping requirements of Arizona farmers.

The plant size, 150 KW, was selected to meet the peak pumping energy requirements for irrigation of a quarter section or 160 acres.

plant will generate electricity, the principal pumping energy form in Arizona.

The tive.

Unattended or minimally attended operation on an actual farm was another objec-

Potential test sites were evaluated and a farm located six miles southwest of

Coolidge, Arizona, and owned and operated by Dalton Cole was selected.

The University of Arizona will assist by monitoring solar plant operation to determine operational and maintenance requirements.

Full utilization of solar plant output is economically desirable; backup energy is required to assure continuous pump operation.

The local utility company, Electric

District Number Two, agreed to participate in the experiment and will supply or purchase energy as required.

The authors are Assistant Professor and Research Assistant respectively in the Soils,

Water and Engineering Department, The University of Arizona, Tucson, AZ

85721.

Approved for publication as Paper No.

286

,

Arizona Agricultural Experiment Station.

75

DESIGN

In phase one of the experiment, competing conceptual designs were developed to meet specific operational requirements.

The requirements included cost effectiveness, use of currently existing hardware and technology, 20 year lifetime, operation in the

Arizona environment and 150 KW electrical output when solar insolation is greater than

600 watts per square meter.

The U.S. Department of Energy funded and Sandia Laboratories supervised development of alternative concepts by three companies: Honeywell, Black and Veatch, and

Acurex Corporation.

Honeywell proposed use of a series of parabolic, two -axis tracking dishes to concentrate solar energy at the focal point where Brayton cycle turbine engines and generators were mounted on each unit.

Black and Veatch proposed a field of heliostats, or individually tracking reflectors, central receiver, or collector, and Rankine cycle engine.

Acurex proposed use of a field of single axis tracking, parabolic trough type solar collectors and Rankine cycle engine.

The Acurex design was selected for construction, appearing to have the fewest technical unknowns and be most economical at this time.

The solar power plant consists of collector, energy storage and energy conversion subsystems, Figure 1.

The collectors, six feet wide by ten feet long, have reflective parabolic surfaces of aluminum

(Coilzak) and concentration ratios of 35 to 1.

Collector receiver tube is coated with a selective black chrome surface and surrounded by a pyrex tube.

The collectors, manufactured by Acurex, are arranged in a series of north -south oriented rows.

Total collector area will be about 49,000 square feet.

A heat transfer oil, Caloria HT43, is being used as the collector fluid.

This gil remains stable and in fluid state at the temperature to which it will be heated, 550'F.

Therefore, low pressure flow can be utilized.

For simplicity, energy storage is a

30,000 gallon tank of hot oil.

On a sunny June day, enough energy can be collected and stored to provide about 20 hours of operation.

Energy conversion is accomplished by a Rankine cycle engine using a turbine and organic working fluid, toluene.

The theoretical conversion efficiency, about 25 percent, is higher than for alternative types of thermal engines when peak cycle temperature is 550'F.

The engine, made by Sunstrand Corporation, is a scaled down version of one developed for other relatively low temperature applications such as conversion of power plant waste heat.

Net engine efficiency is expected to be 17 percent; overall system energy conversion efficiency is expected to average 7 percent annually.

CONSTRUCTION AND OPERATION

The solar power plant is presently being constructed.

Collector foundations have been poured, collector components are being delivered, and the engine is being tested.

Collector and engine assembly at the site are expected to begin soon.

Plant operational check -out tests are scheduled for summer and to terminate in September.

The solar installation then will be operated and operational data obtained and evaluated by the University of Arizona.

The data will include environmental and operational parameters as well as a record of significant operational events and problems.

Associated pumping, irrigation, and cropping experiments, begun in 1978, simultaneously will evaluate alternative energy use scheduling and conservation techniques.

ECONOMICS

Commercialization of solar power plants is dependent on the relative cost of energy.

The energy costs depend on equipment performance and capital investment, fuel and operational costs.

Solar power plants are presently being developed and evaluated.

Thus, investment cost estimates are of doubtful quality and operational costs are almost nonexistent.

The experimental 150 KW solar power plant under construction in Arizona is costing over $10,000 per kilowatt of net peak output.

Solar collector costs account for about half of this total.

Improvements in collector and engine energy conversion efficiencies can reduce collector requirements.

New design innovations, design for production and mass production can further reduce costs.

Estimates of future costs for solar power plants manufactured in quantities of 1000 have ranged from less than $3,000 per KW

Cost competitiveness will depend on such factors as the cost of competing fuels, operational costs and interest rates (Larson, 1977; Katzman and Matlin, 1978).

76

Small, independent solar power plants might produce energy as inexpensively as larger plants.

However, full utilization of produced energy is needed to minimize energy costs.

Full utilization of a solar power plant, with variable solar input as well as variable demands, is difficult to achieve on a single farm (Larson, et al,

1978b).

Interconnection with the utility company is a logical method to assure energy supply and full use of a solar power plant.

CONCLUSIONS

The Arizona 150 KW solar powered pumping experiment will provide an evaluation of a medium sized solar power plant using distributed collectors and much new technology.

The application is most appropriate since irrigation demand peaks during the period of maximum solar insolation and a farm can provide a good environment for solar devices including adequate land area.

The experiment is expected to provide information on design, operation, and maintenance which will lead to equipment improvements.

The experiment also is expected to indicate the practicality of locating such a solar power plant on a farm and provide comparative cost data.

Thus, the experiment is an important step in the development of solar power plants.

REFERENCES CITED

Acurex Corp.

1978.

150 KWE solar -powered irrigation facility.

Technical Status Review Report to the U.S. Department of Energy.

Acurex, Mountain View, CA.

Barber,

R.

E.

1978.

Current costs of solar powered organic Rankine cycle engines.

Solar Energy 21:1 -6.

ERS, USDA.

1977.

Energy and U.S. Agriculture:1974 Data Base, Volume 2.

Energy Admin. Publ. FEA /D- 77/140.

Federal

Hathorn, Jr., Scott.

1978.

Arizona pump water budgets.

Bulletin, Dept. of Ag. Econ.,

Univ. of Arizona, Tucson.

Katzman, M.

T., and Matlin, R.

W.

1978.

The economics of adopting solar energy systems for crop irrigation.

Amer. J. Agr. Econ. 60:648 -654.

Larson,

D.

L.

1977.

Solar energy for irrigation pumping.

Paper No.

77 -4021 presented at the Annual Meeting of the Am. Soc. of Agr. Engr., Raleigh, NC.

Larson, D.

L., Sands II,

C.

D., Towle, Jr., Charles, and Fangmeier, D.

Evaluation of solar energy for irrigation pumping.

Trans. Am.

D.

1978a.

Soc. of Agr. Engr.

21(1):110 -115, 118.

Larson, D.

L., Williams, D.

W., Altobello, M.

A., McAniff, R.

J., and Gum, S.

1978b.

Agricultural practices which could enhance solar powered irrigation plant utility.

Research Reports, Agr. Exp. Sta., Univ. of AZ, Tucson.

Oct. 1978.

429 p., May 1978; 72 p.,

Larson, D.

L., and Fangmeier, D.

ASAE 21(6):1075 -1080.

D.

1978.

Energy in irrigated crop production.

Tran.

Sloggett, G.

Report No.

1977.

Energy and U.S. Agriculture: Irrigation pumping, 1974.

376, ERS, USDA, Washington, DC.

Agr. Econ.

Twersky, M., and Fischback, P.

E.

1978.

Consideration of irrigation systems in the solar photovoltaic energy program.

Paper No.

78 -2551 presented at the Annual

Meeting of the Am. Soc. of Agr. Engr., Chicago, IL.

77

COLLECTOR SUBSYSTEM

THERMAL STORAGE

SUBSYSTEM

POWER GENERATION

SUBSYSTEM

POWER DISTRIBUTION

SUBSYSTEM

Turb/gen

..1.5.0

EWe nerator

Cooling

Water

Power grid

Vapor Condenser

Pump

Figure 1.

Schematic diagram of the 150 KWe solar -powered pumping facility (Acurex,

1978).

78

HEALTH EFFECTS OF APPLICATION OF WASTEWATER TO LAND by

James D. Goff

ABSTRACT

There is a renewed interest in land application of treated effluent in both the states of Arizona and Nevada.

Conservation of water and energy can be obtained by this treatment method.

Data generated by the design engineer includes health effects related to heavy metals, bacteria, and aerosal spray.

Examples of recent nuisance and consequences are noted.

The application of this practice requires a case by case engineering and management analysis.

INTRODUCTION

In the arid regions of the states of Arizona and Nevada, land application of wastewater effluent should be considered in all wastewater system management plans.

Land application of effluent is ideally suited for reasons of being:

(1) low in energy requirements, (2) generally less expensive than other wastewater treatment processes, and

(3) an indirect method of groundwater recharge.

These treatment processes are consistent with water resources planning in that it encourages the reuse and recycle of wastewater (Goff and Horsefield, 1976).

The three basic types of land application systems are: (1) overland flow,

(2) rapid infiltration, and (3) slow rate process (Boyle Engineering Corporation, 1977).

The health effects of applying wastewater to land has received very little attention in the past; therefore, it is of great interest to public health officials.

PRELIMINARY SITE ASSESSMENT

The process of selecting sites for land -application systems is an iterative process starting with the evaluation of very broad criteria and refining the selection on the basis of more restrictive criteria as it relates to specific systems and sites.

The initial evaluating criteria should consider general location as it relates to the existing wastewater system and its compatibility with proposed land usage in the general proximity (Goff and Ewing, 1977).

Other aspects that should be analyzed in the overall site selection process are the general environmental setting of the potential sites which emcompass such things as climate, topography, soil characteristics, as well as the identification of historically and archaeologically sensitive areas (Boyle Engineering Corporation, 1976e).

Groundwater conditions must be considered if such a system is undertaken by a community utilizing groundwater as a water supply source.

SLOW RATE PROCESS

The most common method of land application of treated wastewater is the slow rate process, also called crop irrigation (Pound and Crites, 1976; Sullivan, et al, 1973).

This method applies wastewater to the land by the controlled discharge of effluent by spraying or surface spreading to both support plant growth and provide additional treatment of the applied wastewater.

The effluent is lost to evapotranspiration and to the groundwater by percolation.

Treatment of the applied wastewater is accomplished by physical, chemical, and biological processes as the liquid passes through or over the soil.

This type of a system serves several purposes:

(1) avoids surface water pollution by the reduction of nutrients, (2) recycles water and nutrients and obtains an economic return by producing marketable crops, and (3) conserves water when lawns, parks, or golf courses are irrigated (Boyle Engineering Corporation, 1976b).

Senior Engineer, Boyle Engineering Corporation

Newport Beach, California

79

RAPID INFILTRATION SYSTEMS

Rapid infiltration systems apply wastewater to the soil by discharge to specific basins.

The treatment of the effluent is similar to that of the slow rate process except that it is allowed to infiltrate at a relatively high rate, thereby reducing land area requirements.

while a

The major portion of the wastewater percolates to the groundwater smaller quantity is lost to evaporation.

The purposes of rapid infiltration systems are to: recharge the groundwater, renovate effluent, and often to recover the renovated water by the use of wells or underdrains.

OVERLAND FLOW SYSTEMS

The third method of land application is overland flow.

This technique is probably the least used.

It consists of applying wastewater, usually by spraying, on the upper reaches of sloped terraces of relatively impermeable soils and allowing it to flow through vegetation.

The wastewater is ultimately collected as runoff at the bottom of the slope minus that amount lost to evaporation and percolation.

plished by sheet flow through the vegetation.

Renovation is accom-

The purposes of this type of a system are to provide additional treatment prior to discharge to surface waters and the reuse of collected runoff.

GROUNDWATER

Land treatment of wastewater can have a significant effect on both groundwater quality and quantity.

Land treated wastewater can either be a pollutant or a means of improving groundwater quality, depending on the quality of the natural groundwater.

areas of extremely high total dissolved solids content, wastewater can improve this

In factor sufficient to make the groundwater suitable for irrigation and other purposes

(Goff, 1977

).

Conversely, high quality groundwater used for domestic purposes could become contaminated by wastewater application if adequate controls are not exercised

(Berry and Shaffer, 1978).

TRACE METALS

Although some metals, usually in small quantities, are essential for plant growth, most are not.

Indeed, many metals are acutely toxic to plant life and microorganisms at even low levels, particularly citrus crops.

The major problem resulting from heavy metals is their long -term accumulation in soil.

Heavy metals are retained in the soil matrix by absorption, chemical precipitation, and ion exchange.

In general, copper, nickel, and zinc comprise the largest portion of total heavy metal content in effluent.

Boron is an essential plant nutrient at low levels, but is toxic to many plants at dosages of 2 mg /1.

To a limited extent, boron can be removed in the rapid infiltration treatment method by fixation to the soil in the presence of iron and aluminum oxides.

Sand and gravel will not remove boron; clay, however, is an excellent fixant (Ragone, et al, 1975).

Results from the rapid infiltration system at Flushing Meadows, Arizona, show that copper and zinc are removed by the soil to a great extent.

Cadmium and lead appeared to be relatively unaffected by rapid infiltration.

Mercury removal was about 40 percent.

One observation was that extended underground travel of the applied effluent does not result in additional removal of heavy metals (Boyle Engineering Corporation,

1977).

The slow rate process appears to be the most effective means of metal removal, with plant uptake comprising the primary removal mechanism.

Irrigation removes 95 percent of applied trace metals; overland runoff removes 60 percent to 90 percent; and rapid infiltration removes 50 percent to 90 percent, depending upon the type of soil.

The important metals appear to be cadmium, copper, molybdenum, nickel, and zinc.

The accumulation of these metals may pose a hazard to plants, animals, or humans under certain circumstances (Wolman, 1977).

Experience already shows that a more successful control of heavy metals lies in major reduction of these objectional materials at their source.

PATHOGENIC ORGANISMS

As increasing use is being made of land application treatment methods, more and

80

more attention is being paid to the public health effects of such action in relation to bacteria and virus spreading.

In general, the potential hazard associated with the application of wastewaters to the land is low, and in fact, is less than that associated with their discharge to surface waters (Benarde, 1973).

However, certain potential public risks associated with the use of wastewater for land application do exist.

Therefore, as a matter of good public health practice, effluent applied to land should first receive secondary treatment.

Rapid infiltration on land offers an excellent means of virus destruction.

Studies indicate that no viruses are present in effluent after infiltration through 200 feet (61m) of soil.

California authorities permit such water to be used not only for irrigation, but for recreational purposes as well.

At the Flushing Meadows project in

Arizona, total coliforms decreased to a level of two organisms per 100 ml at a distance of 30 feet (9m) from the point of application.

The application method consisted of flooding the basins for one day, followed by a drying period of three days (Gilbert, et al, 1976).

Crops receiving wastewater effluent applied as irrigation water may, in turn, become contaminated with bacteria and viruses.

In spite of reductions due to desiccation and die -off in the ground, large numbers of pathogens may still survive to constitute a health hazard in crops.

Bacteria do not appear to enter healthy and unbroken vegetables; they may, however, enter broken, bruised or damaged plants (Sorber, 1974).

Pathogens may travel nozzle.

in aerosols a distance of up to 200 feet (61m) from a sprinkler spray

If wastewater applied to land passes through 5 to 10 feet (1.5 to 3m) of continuous fine soil before entering a groundwater system, then microbiological contamination of the groundwater should not usually occur (Kocerba, 1973).

Even if bacteria do enter groundwater, they can only travel a few hundred feet horizontally.

However, in the case of gravels and coarse soils, viruses and bacteria may travel long distances underground.

Viruses in wastewater infiltrating through the soil are removed by absorption to the soil particles.

There is, however, much information that is lacking concerning the effects upon public health of direct or indirect groundwater recharge (Lance, 1975).

Aerosols are small particles of liquid having a size range of 0.01 to 50 microns and suspended in air.

As of 1975, specific studies of biological aerosols emitted by spray irrigation of secondary wastewater effluents had not been found in literature.

It appears that destruction of bacteria in aerosols occurs through evaporation of the droplet and desiccation of the bacterial particle.

Investigations of aerosol evaporation rates show that the majority of bacteria die within three seconds.

The remaining bacteria, protected by chemical additives which inhibit evaporation, die at a decreasing rate over time (McNabb and Dunlap, 1975).

The bacteria of primary concern in wastewater, Escherichia Coli, generally has a very short life span in aerosol form.

On the other hand, Klebsiella, a pathogen affecting the respiratory tract, can form a protective capsule to prevent desiccation while in the air (Bryan, 1975).

Love reported that pathogens can travel over fairly large distances in aerosols from sewage or sludge spraying, but seldom travel very far in soil.

He stated, "The greatest hazard would probably result from disrupting good treatment or sanitary practices.

Despite associated disease outbreaks, the practice of land application of effluent can function safely.

The limited quantities of essential chemicals and water available to produce food makes it necessary to find a safe use for wastewater and sludge." (Love, et al, 1975).

Rapid infiltration systems also show very good removal efficiencies for viruses.

The Santee, California system was tested with polio virus; none were found after the water traveled 200 feet (61m) through the alluvial stream bed.

Virus studies have also been conducted at St. Petersburg, Florida.

In one test, over 12 million virus units were applied per day; the viruses were detected infrequently at both 10 -foot (3m) and 20 -foot (6m) depths.(Dugan, et al, 1975).

Buffer zones of 50 to 200 feet (15 to 61m) in width should be provided around any land application site.

In the case of spray application, prevailing winds have a great effect upon the distance aerosols may travel.

For example, experiments have shown that a

7 mile (11km) per hour wind can cause aerosol -born bacteria to travel 500 feet (152m) from the spray nozzle.

The maximum distance of travel has been estimated to be 1,000 to 1,300 feet (305 to 396m) in the case of an 11 mile (18km) per hour wind (Burbank,

1975).

81

The various land application methods provide a very high degree of bacteria and virus removal (90 to 99 percent).

Removal efficiencies of Escherichia Coli by infiltration percolation are usually excellent (Pound and Crites, 1973).

Table 1 presents a comparison of Escherichia Coli concentrations in applied effluent and the resulting renovated water for three such systems.

TABLE 1

ESCHERICHIA COLI REMOVAL EFFICIENCIES FOR

RAPID INFILTRATION

Fecal Coli Concentrations

(MPN /100 ml)

Applied Effluent

Renovated Water

1,000,000

60,000

Distance of

Travel (feet)

Location

Flushing Meadows,

Arizona

Hemet, California

Type

Sand

Sand

10

11

100

8

Santee, California Gravel

130,000

10

200

According to the Corps of Engineers Cold Regions Research and Engineering Laboratory, "Virus, pathogenic bacteria, and coliform are completely removed by percolation through a tions.

5 -foot (1.5m) depth of soil under controlled slow infiltration design condi-

Survival times are typically a maximum of three months for organisms retained on soil particles.

It is important that pathogens be removed, for the wastewater reuse cycle may be shorter with land application than with a point source river discharge.

Consideration must be given to die -away rates in storage ponds and to the removal capacity of the receiving land." (Uiga, 1976).

CONCLUSIONS

In the arid areas of Arizona and Nevada, effluent applied to a properly designed land application system should cause no adverse health effects provided health regulations are followed, and normal operations are provided.

As a matter of good public health policy, all public water supplies should receive proper disinfection to inactivate any pathogens that might be present.

Although there are questions still unanswered regarding health effects, land application of effluent and reuse of the water are far better methods to employ than the practice of discharging partially treated wastewater into stream beds that may serve as recharge areas to groundwater supplies.

The engineering capability is available today to properly plan, operate and monitor on a site specific basis land application of wastewater in a manner as to cause no adverse health effects.

82

REFERENCES CITED

Benarde, M.

A.

1973.

"Land Disposal and Sewage Effluent: Appraisal of Health Effects of Pathogenic Organism ", Journal of the American Water Works Assoc., 65(6):432 -440.

Berry, James W.

and Shaffer, T.

B.

Wastes Engineering, p.

14 -17.

1978.

"Viruses:

Monitoring is the Key ", Water and

Boyle Engineering Corporation.

1977.

"Large Array of Wastewater Land Treatment

Alternatives ", U.S. Army Corps of Engineers Phoenix Urban Study.

Boyle Engineering Corporation.

1976a.

"Preliminary Assessment of the Potential Land

Application Sites ", U.S. Army Corps of Engineers Phoenix Urban Study.

Boyle Engineering Corporation.

1976b.

"Potential Reuse Options for Municipal Wastewater Effluent and Residual Solids ", U.S. Army Corps of Engineers Phoenix Urban

Study.

Bryan, E.

H.

1975.

"Management of Municipal Wastewater Treatment Residuals ",

Proceedings of the 1975 National Conference on Municipal Sludge Management and

Disposal.

Burbank, N.

C.

1975.

"Mililani Town Wastewater Reclamation and Recycling Project ",

Journal of the American Water Works Assoc., p 67, 487.

Dugan, G.

L., et al.

1975.

"Land Disposal of Wastewater in Hawaii ", Journal Water

Pollution Control Fed., 47, 2067.

Gilbert, R.

G., et al, 1976.

"Wastewater Renovation and Reuse:

Soil Filtration ", Science, 192(4243):1004 -1005.

Virus Removal by

Goff, James D.

1977

.

"Municipal Wastewater Reuse Priorities for Metropolitan

Phoenix ", American Society of Civil Engineers, Reprint No. 2943.

Goff, James D., and Horsefield, Dave.

1976.

Texas ",

"Wastewater Resources in North Central

Journal of the American Water Works Association.

Goff, James D.

and Ewing, Ronald E.

1978.

"Modeling for Land Application of Wastewater", Modeling, Identification and Control in Environmental Systems, Proceedings of the IFIP Working Conference on Modeling and Simulation of Land, Air and

Water Resources Systems, North -Holland Publishing Company, Amsterdam.

Kocerba, B.

A.

1973.

"Land Application of Wastewater at Colorado Springs ", APWA

International Congress and Equipment Show, Denver, CO.

Lance, J.

C.

1975.

"Fate of Nitrogen in Sewage Effluent Applied to Soil ", Journal

Irrigation and Drain Div., Proc. American Society Civil Engr., p

101, 131.

Love, G.

J., et al.

1975.

"Potential Health Impacts of Sludge Disposal on Land ",

Proceedings of the 1975 National Conference on Municipal Sludge Management and

Disposal.

McNabb, J.

F., and Dunlap, W.

J.

1975.

"Subsurface Biological Activity in Relation to

Groundwater Pollution ", Groundwater, p

13,

33.

Pound, C.

E., and Crites, R.

Application ",

Vol

W.

1973.

"Wastewater Treatment and Reuse by Land

II, U.S. Environmental Protection Agency, Technoloah Transfer

Series, Report EPA 660/2 -73 -0066.

Pound, C.

E., and Crites, R.

W.

1975.

"Design Seminar for Land Treatment of Municipal

Wastewater Effluents: Design Factors, Part I ", U.S. Environmental Protection

Agency, Technology Transfer Program.

Ragone, S.

E., et al.

1975.

"Mobilization of Iron in Water in the Magothy Aquifer

During Long Term Recharge with Tertiary Treated Sewage, Bay Park, N.Y. ", Journal of Research of the U.S.G.S., p

3, 93.

Sorber, C.

A.

1974.

"Public Health Aspects of Land Application of Wastewater Effluents", Land Application of Wastewater, Newark, Delaware.

Sullivan,

R.

H., et al.

1973.

"Survey of Facilities Using Land Application Wastewater"

U.S. Environment Protection Agency, Office of Water Program Operations, Report

83

EPA -430/9 -73 -006.

Uiga, A.

1976.

"Let's Consider Land Treatment -

Not Land Disposal ", Civil Engineering,

ASCE, 46(3):60 -62.

Wolman, Abel.

1977.

"Public Health Aspects of Land Utilization of Wastewater

Effluents and Sludges ", Journal Water Pollution Control Federation, p 2211 -2218.

84

EARLY PUBLIC INVOLVEMENT IN FEDERAL WATER RESOURCE PROJECTS by

Freda Johnson and Michael Thuss

INTRODUCTION

Recognizing that the development of a public participation plan calls for time and effort on the part of federal agencies, it is maintained that developing a productive relationship between the agency and private citizens in the early planning stages can result not only in better projects but also in public support for the resulting projects.

A case study approach is used to describe how and when the U.S. Army Corps of Engineers (in its

Tucson Urban Study) has achieved positive results in dealing with an active public in a community where water resource issues evoke extensive public controversy.

It is critical to the success of this early, productive relationship that an agency identifies the public's role in the planning process.

Clarifying the agency's expectations of the public, by defining what decisions can be made by citizens, results in public willingness to participate and lays the groundwork for meeting federal public involvement requirements in later stages of the planning effort.

How the interested publics were identified, brought together, and organized into an Urban Study oriented mechanism for public participation and approach taken to resolve conflicting preferences between local governments and the public are described.

BACKGROUND

THE URBAN STUDIES PROGRAM

An urban study is a federally sponsored program that identifies a range of water resource plans addressing water -related problems in an identified urban area.

At the request of local governments, urban studies are authorized by the U.S. Congress and carried out by the U.S. Army Corps of Engineers.

Congressional authority for the Urban Studies Program was established in 1974.

A three -stage planning process, taking from 36 to 48 months, results in recommended plans developed in sufficient detail to be adopted and carried out by local governmental agencies or by the

U.S. Army Corps of Engineers if a federal interest has been identified.

Corps planning objectives state that the recommended plans, or solutions, must be socially and environmentally acceptable, have local support, and be implementable.

Other important aspects of the Urban Studies Program are that the efforts recommended not duplicate existing local or regional planning activities and that the overall study be consistent with local, regional, State and Federal objectives.

Since 1974, about 36 urban studies have been authorized by Congress.

THE TUCSON URBAN STUDY

1978.

The Urban Study for the metropolitan area of Tucson, Arizona, was officially initiated in February

Prior to the initiation announcement, the Pima Association of Governments passed a Resolution petitioning Congress to authorize and appropriate funds for the study (April 1976); the U.S. House of

Representatives' Subcommittee on Public Works and Transportation approved the study (May 1977); and

President Carter signed the bill authorizing the study (August 1977).

Conducted by the Urban Study Section of the Los Angeles District of the U.S. Army Corps of

Engineers, the Tucson Urban Study is managed locally in Tucson by a Study Manager, an Assistant Study

The authors are, respectively, a public involvement consultant, Rillito Consulting Group, Tucson, and

Study Manager, Tucson Urban Study, U.S. Army Corps of Engineers.

85

Manager, a secretary, and temporary employees for editorial assistance and engineering planning.

public involvement consultant has been employed as a Technical Writer since March 1977.

A

Stage 1, the development of a Plan of Study, has been completed and a Plan of Study document, telling what the study will do and how it will be conducted, has been reviewed and approved by elected officials of all local governments in Pima County, and by the interested Federal and State agencies.

Major problems being addressed in the study are: 1) flooding along and degradation of the area's watercourses, and 2) depletion of the area's natural water supply.

CORPS REQUIREMENTS, STAGE 1, PUBLIC INVOLVEMENT

According to the Corps' requirements for the Plan of Study document, there should be, at a minimum, a study initiation announcement, the development of a public participation plan for Stages 2 and 3, and a public meeting to present and hear public responses to the Plan of Study.

The Corps also identifies an overall goal for public involvement as the establishment of continuous two -way communication between the Corps and the public throughout the life of the study.

Later, during Stages 2 and 3, it is expected that the Corps will work with a body of informed people, known as the "participating public," while conducting the Tucson Urban Study.

This group assists in tradeoff analysis by assessing impacts and evaluating effects of developed alternative solutions to water resource problems identified in the

Plan of Study.

The Corps' definition of "public" is any non -Corps entity and, therefore, includes local, regional, State, and Federal governmental agencies as well as organized interest groups, community associations, and any interested individuals.

The use of the word "public" in the following sections refers to citizens acting on their own beliefs, not as representatives of governmental agencies.

PUBLIC PERCEPTIONS IDENTIFIED AT BEGINNING OF STAGE

1

The Corps set out to learn whether the public in eastern Pima County was interested in the subjects to be addressed in the Tucson Urban Study and, if so, what aspects were of greatest concern, and how much was the public willing to be involved during development of the Plan of Study.

Information on the Tucson Urban Study provided by the Corps at the beginning of Stage form of a Study Initiation Announcement mailed to 4,500 individuals in December 1977.

1 took the

A mailing list developed by the 208 Water Quality staff of the Pima Association of Governments was used.

A broad cross -section of community interests was represented on the list, including environmental, business, construction, and civic groups, as well as governmental agencies at the local, regional, and State levels.

The initiation of the study was also announced through television and newspapers.

From responses to this announcement and informal discussion with individuals and representatives of community organizations, it was learned that there was significant public interest in the Corps'

Urban Study.

Over 200 people contacted the Urban Study office by mail or phone to tell the Corps what kinds of problems should be addressed and what aspects of water resource planning should be emphasized.

Many of these people, fearing a traditional construction- oriented approach by the Corps, expressed uneasiness at the Corps' presence in Tucson.

However, they were also willing to learn more about the urban study concept and asked to be kept informed of study progress.

The Corps also learned, through means of clipping and reading the area's newspaper and attendance at meetings of local governments, that a vocal and active public appeared to represent a variety of conflicting opinions, values, and goals on water -related issues.

Finally, the Corps needed to gauge how much the public was willing to be involved.

This was accomplished by learning about other mandated public involvement programs and evaluating their effectiveness.

This was not attempted in a formal way but through observation and personal contacts with members of the public and with agency staff responsible for conducting other public participation programs.

It was known that a variety of public involvement activities competed for the public's attention.

In Pine County and the City of Tucson, there were, in 1977, over 60 active citizen boards, commissions, and committees in existence.

Numerous other local, regional, State, and Federal agency programs required public attention and involvement on a regular basis.

It appeared that the public perceived that most of these public participation programs were token gestures designed to fulfill the agencies' legal requirements.

Often, it was not clear to the public exactly what was expected of them.

Also, individuals were skeptical about how their responses or participation would affect decisions being made.

From this general, informal survey of the community, the Corps concluded that:

1.

2.

The public expected to be listened to and to have influence on decisions made during the Urban

Study planning process.

It was not enough to contrive to be

The Corps must achieve a genuine openness from the start.

open or be patronizing.

86

3.

The Corps' expectations of the public must be clearly defined.

The following objectives for the public involvement'program of Stage 1, Plan of Study Development, were established as a result:

1.

2.

Provide detailed information on the scope of the study.

Encourage involvement of any interested individuals or organized groups.

3.

Provide opportunities for public discussion of preferences on the identified parts of the Plan of Study.

4.

Analyze public preferences and report decisions made based on public response.

5.

Resolve any remaining conflicting issues between the participating public and local governments.

DEFINING THE PUBLIC ROLE

If the Corps was to avoid the pitfall of confusing and frustrating potential participants in Urban

Study activities, it appeared that the public should be brought into the planning process as soon as possible.

Therefore, meeting the Corps' requirement of presenting the draft Plan of Study at a public meeting was identified not as the first formal public involvement activity but as the last step in a public involvement process for Stage 1.

Based on what was known about public dissatisfaction with citizen involvement programs, the Corps determined that defining the role of the public was the major task to be addressed in Stage 1.

DECISION -MAKING OR ADVISORY

Consideration of the concept of a decision- making role for the public was acceptable to the Corps.

While most urban studies had described the general public's role as advisory (review and comment), early contacts with representatives of special interest groups and community organizations in the

Tucson area revealed a strong desire to have a responsible, informed, decision- making role for citizens during the planning process.

It was apparent that citizens wanted to participate in formulating decisions rather than simply responding to them.

Knowing this, the Corps' Urban Study staff considered decisions related to the Plan of Study development and addressed themselves to the following questions: decide?

What are we not willing to let the public decide?

What are we willing to let the public

(Subjects identified as not appropriate for public decision -making were those required or mandated by authorities beyond Corps control.)

Four elements of the Plan of Study were identified as being appropriate subjects for public preferences and decisions.

These were Problem Identification, Study Area, Study Management, and Public

Participation for Stages 2 and 3.

Those elements identified as not pertinent to the public's interest in Stage

1 were Justification of the Study, Statement of Planning Objectives, Institutional Considerations, Study Effort Allocation, and Schedule of Work Tasks and Costs.

MEANS TO IDENTIFY PUBLIC PREFERENCES

An Option Paper (with a response form) was prepared for each of these four parts of the Plan of

Study.

The papers were distributed to points of contact in each of the five local governments (City of

Tucson, Pima County, City of South Tucson, and the towns of Marana and Oro Valley).

A preliminary governmental consensus was developed based on the local governments' preferences.

The Option Papers were then distributed to any interested individuals and all identified special interest groups.

Availability of this means to solicit public preferences was announced in the press, radio and television, and in special mailings to civic organizations in eastern Pima County.

The governmental consensus was reported to the public recipients of Option Papers, not as a limiting factor or indication that the Corps had already made a decision, but as information needed by the public to make realistic comments and report their preferences.

The next step or method in verifying public preferences was the sponsorhip of an all -day workshop in June 1978.

Preceded by an informational meeting several days earlier, the workshop was designed to test reported preferences and to identify any conflicts between the public and the local governments' preferences.

87

CONFLICT IDENTIFICATION AND RESOLUTION

This process revealed that the public concurred with local governments on all but one issue.

That issue was the question of citizen representation on the Steering Committee, the management mechanism that would provide policy direction to the Urban Study staff for the duration of the study.

Local governments had proposed a 5- member Steering Committee made up of city managers (Tucson, Pima County,

South Tucson) and locally elected officials (Marana and Oro Valley).

The public reponse to that proposal, derived from Option Paper responses and the results of the workshop, was that there should be general public voting representation on the Steering Committee.

Suggestions stated in discussion groups at the workshop had ranged from appointing to the proposed committee one citizen voting member to establishing a Citizen Executive Committee.

Reasons expressed at the workshop to justify citizens in a decision-making role included:

"give citizens respect and motivation to become knowledgeable "; "advisory role (only) not preferred, has no more power "; "temper politicians "; "lack of trust in elected officials to carry out public wishes "; public acceptance of actions of Corps."

In addition to learning that the public desired representation on the decision -making mechanism for the Urban Study, the Corps also learned that both agency and general public response to the concept of an advisory committee was positive and specific.

Characteristics of an advisory committee agreed to were that it have broad representation, be Urban Study- oriented, oversee educational outreach efforts from the Corps to the general public, have a funded staff, and that there be no political appointments.

For nearly three months at the end of Stage 1, the Corps worked with local governments to achieve a compromise solution to the question of makeup of the Steering Committee for the Tucson Urban Study.

The Urban Study staff developed a proposal for Steering Committee makeup that would meet the requirements of local governments and also provide the desired citizen representation.

The Corps' plan was for an 11- member Steering Committee, adding five citizens to the committee of three managers and two elected officials, plus one Corps of Engineers representative.

Initially, contacts in the five local governments agreed to the fact that including citizens in a steering capacity was a reasonable idea.

However, the local governmental representatives were not willing to have citizens sit with them as equal voting members on Urban Study agenda items.

It appeared to this committee that, by including citizen members, more time would be required to deal with issued addressed and that there was no precedent for such an arrangement.

After consideration and discussion of many options, the local governments accepted the Corps' recommendation that the Urban

Study be under the direction of a Citizen Steering Committee.

The Steering Committee was proposed to have nine members: one from each of the five local governmental jurisdictions in eastern Pima County, three at- large, and one from the Papago Indian Reservation.

The Corps and the Urban Study's Citizen

Advisory Committee would jointly nominate the nine members to the PAG Regional Council (the regional council of governments) for appointment.

After the local governments agreed to this arrangement, the result was reported publicly.

RESULTS OF THE STAGE

1

PUBLIC INVOLVEMENT PROGRAM

The Citizen Steering Committee for the Tucson Urban Study has been formed and meets regularly to provide direction and control for Stages 2 and 3 planning activities.

Responsive to preferences of an ongoing Citizen Advisory Committee, the Steering Committee will formally recommend study results to the

PAG Regional Council.

The Advisory Committee is open to any interested people and currently has nearly two hundred members, of which approximately 80 are eligible to vote under rules established by the committee itself.

Currently being developed by the Urban Study staff is a range of alternative plans addressing solutions to flooding problems on the area's watercourses.

Also under study are wastewater reuse; urban runoff; groundwater recharge; and enhancement of recreation, wildlife, historical, archaeological, and cultural resources.

In conclusion, it can be stated that all parties have a clear understanding of what role each will play in the course of the Tucson Urban Study.

The Corps has accepted the responsibility of maintaining a productive, if time- consuming, relationship with the public in eastern Pima County.

The responsibility undertaken requires a genuine commitment on the part of the Urban Study staff, administrative assistance and support for the citizen committees, and budgeted funds to monitor and maintain the program.

The participating public has accepted the responsibility to inform itself on Urban Study matters, attend frequent meetings, and take into account legal and institutional limitations as alternative plans are evaluated.

88

SUMMARY OF PUBLIC INVOLVEMENT ACTIVITIES DURING STAGE

1

A -- General Public

B -- Participating Public

C -- Governments /Agencies

X

X

X

X

X

X

A

X

X

X

X

X

X

X

X

X

X

X

B

X

X

X

X

X

C

X

X

1.

"Urban Study Program," article in bulletin of Southern Arizona Environmental Council;

2 pp., November 1977.

2.

Study Initiation Announcement, December 1977.

Mailed to 4,500 individuals in eastern

Pima County.

3.

Meeting with individuals and representatives of prominent community organizations to determine criteria for a successful public involvement program, 28 December 1977.

4.

Preparation and distribution on request of a Briefing Paper on the Tucson Urban Study,

3 pp., 16 January 1978.

5.

Option Paper, Public Participation Program, Stage 1; 13 pp., 10 March 1978.

Distribution to local governments and individuals expressing interest in the Tucson Urban

Study.

X

X

X

X

X

X

X 6.

Coordination meeting with Option Paper reviewers and other interested persons to discuss and seek public agreement on the Public Participation Element of the Plan of

Study, Stage 1, 30 March 1978.

7.

Display of the Tucson Urban Study Planning Process (8' x 40 ") exhibited at the Parade of Homes, April -May 1978.

8.

Announcement of Public Participation Program, Stage 1, 14 April 1978.

Tucson Urban Study and press release provided to:

Fact sheet on a) presidents or executive directors of ten community organizations, b) seven editors of widely read organizational newsletters and bulletins in southern Arizona, c) sixteen homeowners associations in the Tucson metropolitan area, and d) eight newspapers in the Tucson area.

9.

Established a sequence of events for option paper review by local governments represented on the Pima Association of Governments (PAG), April -May 1978.

Informal meetings with representatives of the five local governments were held to discuss contents of each option paper.

10.

11.

Distribution of option papers on elements of the Plan of Study (Problem Identification, Study Area, Management, Public Participation, Stages 2 and 3) to individuals and representatives of community organizations, April -May 1978.

Local governments' preferences were reported to these people.

Various appearances by Urban Study staff on television and radio, April -June 1978.

12.

Mailed 4,000 flyers to individuals in eastern Pima County announcing activities #13 and #14.

13.

Informational meeting on the Tucson Urban Study for the general public, 20 June 1978.

Cragin Elementary School.

70 individuals attended.

14.

15.

Tucson Urban Study Workshop, 24 June 1978.

El

Rio Neighborhood Center.

65 participants indicated preferences on water resources problem identification, study area, management steering mechanism, and future public involvement.

Announcement of results of Workshop mailed to 150 individuals (workshop participants plus other interested people), 7 July 1978.

16.

Workshop results provided to five local governments, 10 July 1978.

17.

18.

Meeting with local governments to discuss workshop results on the management steering mechanism for the Tucson Urban Study, 20 July 1978.

Public Information Brochure (summary of the draft Plan of Study) and announcement of

Public Meetings mailed to 433 people in August 1978.

Copies were also provided to 12 branches of the Tucson Public Library.

89

A

X

B C

X X X

19.

20.

Radio and television appearances by Urban Study Manager to promote public awareness of and participation in the scheduled Public Meetings, July- August 1978.

Public Meetings, 22 -23 August 1978.

90

NEGOTIATING THE WATER FUTURE OF PIMA COUNTY, ARIZONA by

Michael F. Thuss

INTRODUCTION

Because current water consumption in the eastern Pima County region exceeds the average annual long -term natural supply by a rate of over 3 to 1, the region's major water users have come together to discuss the problem and develop possible solutions.

With the support of the elected officials of the local governments, and the representatives of the interested Federal and State agencies; the major agricultural, mining, municipal, and tribal water users have requested that the U.S. Army Corps of

Engineers provide technical advice and assistance in presenting the facts regarding the water needs facing the region over the next 50 years, and in preparing alternative solutions to the anticipated problems.

The Corps has designed the Eastern Pima County Regional Water Resources Study so that the major water users can make decisions regarding the wise use of water throughout the planning period.

The objectives of the study are to assess and project current and future water demands and supplies, facilitate the major water users' negotiation process, and quantify and test an array of alternative solutions involving groundwater, wastewater, and Central Arizona Project water.

The major water users have organized themselves into a group called the Water Resources Coordination Committee, and the U.S. Army Corps of Engineers provides staff support to this committee.

The

Corps is assisted by a Multipurpose Technical Committee made up of representatives of Federal, State, and local agencies; a Citizen Advisory Committee; and consultants.

The Corps efforts have been integrated with those of the Arizona State Legislature's Groundwater

Management Study Commission, the City of Tucson, and the U.S. Geological Survey and are being conducted over a 10 -month period in three phases.

Phase I is devoted to developing the facts associated with the problem and the possible solutions.

In Phase II, the major water users use the results from Phase

I to search for acceptable and feasible solutions, and in Phase III the data are refined so that a preferred plan may be selected.

The total cost of this effort is $480,500.

PROBLEM DESCRIPTION

This all came about because the metropolitan area of Tucson, with its almost one -half million inhabitants, is one of the largest urban areas in the world totally dependent on groundwater for its water supply.

It is also one of the fastest growing communities in the United States.

Based on projections made by the Pima Association of Governments (1977) for the period 1975 to 2000:

1.

Population is expected to increase from 475,000 to 760,000.

2.

Agriculture is expected to either remain the same at 54,000 cropped acres or decline to about

10,000 cropped acres.

3.

Manufacturing and mining employment is expected to increase from 21,000 to 38,000.

In 1975, these water consumers exceeded the natural average annual long -term water supplies by a rate of over 3 to 1.

Approximately 340,000 acre -feet of water were consumed while nature provided only about 110,000 acre -feet of recharge.

Consumption by these major water users in 1975 is estimated as follows (Thuss, 1978):

1.

Agricultural irrigation -- 74 %.

2.

Industrial consumption, which includes manufacturing, mining, and electrical power generation

-- 15 %.

3.

Municipal and recreation consumption -- 11 %.

The author is Study Manager, Tucson Urban Study, U.S. Army Corps of Engineers.

91

The numbers shown here may differ slightly from those of other sources, but the difference is caused by definition rather than fact, and they are in complete agreement with and fully support the estimates of water use and supply presented by the Arizona State Water Commission in its state water plan.

This overdraft is confirmed in that groundwater tables in the region have dropped an average of two feet each year since 1930 according to Professor Sol Resnick of the University of Arizona.

Even with the introduction of Central Arizona Project water in 1987, the overdraft is expected to continue.

These facts mean that the natural legacy of groundwater saved over thousands of years is being used up.

The potential economic, social, and environmental costs of this overdraft are enormous.

The economic costs are for the exploration, energy needs, and transportation requirements of getting water supplies from greater and greater depths as well as from distant basins and regions.

Social costs take the form of increased hostility and conflicts between agricultural interests, municipalities, industrial and mining interests, recreational activities, and the Papago Indian Tribe.

from the reduction of wildlife habitat and the threat of subsidence.

Environmental costs stem

Currently, there is no single authority in the region which has control over water quality, water supply, and wastewater management systems.

Current management practices call for separate control of small portions of the water supply, use, and wastewater generation systems in the region.

In the 1960's and 1970's, many of the region's major water users have been involved in legal actions against one another over rights to the groundwater supply.

The Papago Indian Tribe has brought suit based on the Winters Doctrine and the Cappaert decision, against the non -Indian water users in the

Tucson Basin.

COMMITTEE DEVELOPMENT

In January of 1978, Mr. William E. Strickland, representing the Papago Indian Tribe, invited the major water users of eastern Pima County to meet together to discuss means of resolving the various legal suits out of court.

Representatives of the region's mining, agricultural, municipal, and other interested parties began to meet on a weekly basis.

This group was known as the Negotiating Committee.

The U.S. Army Corps of Engineers was invited by the City of Tucson and the Papago Indian Tribe to attend these meetings as an observer in late January 1978.

The initial task of this group was to prepare a scope of work directed towards evaluating the water needs and the sources of supplies available to eastern Pima County, and determine who should perform the scope of work.

Mr. Will Worthington, the Chief of the Urban Study Section for the Los Angeles

District, U.S. Army Corps of Engineers, gave approval for the internal development of a Corps of

Engineers- sponsored proposal reflecting how a study might be conducted.

A draft proposal was prepared and circulated at the 8 March 1978 meeting of the Negotiating

Committee.

This proposal provided an overview of the work that needed to be done, how it would be done, what management arrangements would be required, how the Corps fits in, how much the study would cost, and how it could be funded.

The Negotiating Committee reviewed the document and approved the concept.

Their support for a Corps of Engineers managed program was transmitted to Brig. General

Hugh G. Robinson (then Colonel), the District Engineer, who immediately recognized the importance and necessity of this effort, not only for the local residents, but also in carrying out the President's water resources policy guidelines.

Although there was some initial reluctance on the part of some

Corps officials to undertake this project, Gen. Robinson authorized the development of a detailed plan of study to accomplish the task.

The major advantage to the local water users for having the U.S. Army Corps of Engineers conduct this study was because the Corps is an outside agency.

It had the confidence of all the interested parties to conduct an open and objective program, and it had the capabilities to bring together the necessary resources.

In the fall of 1978, the elected officials of the five local governments (Pima County, the cities of Tucson and South Tucson, and the towns of Marana and Oro Valley), as well as representatives of the interested Federal and State agencies, approved the concept.

In December 1978, the Plan of Study was approved by the major water users and work began.

formed itself into the Water Resources Coordination Committee.

on

Figure

1.

The informal committee organized by Mr. Strickland

The members of the committee are shown

CORPS AUTHORITY

The authority for the Corps to participate in this program is an Urban Study Resolution sponsored by Congressman Morris K. Udall and adopted by the Committee on Public Works and Transportation of the

U.S. House of Representatives on 23 September 1976.

It is the latest in a series of authorizations for studies of the Gila River and tributaries.

This program is in keeping with President Carter's Water

Resources Policy Reform Message of 6 June 1978 and his Memorandum to the Secretary of the Army on

92

U.S.

Army

Corps of Engineers

Management

Organization

District

Engineers

Los Angeles

Chief, Urban

Study Section

Water Resources

Coordination

Committee

U.S. Army Corps of Engineers

Study Team

Citizens

Advisory

Committee

(wastewater reuse)

E-

Technical

Committee

H

Consultants

Pima County

City of Tucson

Mining Interests

Agricultural Interests

The Papago Indian Tribe

Papago Allottees

Tucson Electric Co.

University of Arizona

Bureau of Indian Affairs

Corps of Engineers

Davis -Monthan AFB

Private Citizen

y

Study

Participants'

Own Technical

Staffs

Figure 1.

Management Organization.

12 July 1978 concerning Federal and Indian Reserved Water Rights, Environmental Quality, and Water

Resources Management.

THE REGIONAL WATER RESOURCES STUDY

PURPOSE AND OBJECTIVES

The purpose of this program is to develop the facts regarding the water needs facing eastern Pima

County for the next 50 years, and the facts associated with alternative water use patterns.

Neither the determination of the legal right to a specific water resource nor the establishment of a water resource management organization is within the scope or purpose of this program.

It is within the role of the U.S. Army Corps of Engineers to provide technical assistance and advice to the major water users in the region so that, collectively, they can make decisions regarding the wise use of water throughout the planning period.

The objectives of this program are to:

1.

Inventory historical water use in eastern Pima County.

2.

Project the water quality and quantity needs of eastern Pima County through the year 2030.

3.

Inventory, assess, and project future water supplies for eastern Pima County through the year 2030.

4.

Develop water budgets and other materials comparing supply to demand, indicating the nature of the problem (technical and economic) which faces all water users in eastern Pima County.

5.

Provide a forum for open discussion and communication of needs and wants among water users regarding:

93

a.

The nature of the water situation.

b.

The goals of a program to provide alternative solutions.

c.

The impacts of any proposed solutions on major water user groups and the groundwater supply.

6.

Provide the facts associated with various water supply and conservation alternatives, and assist the group in.selecting the most feasible and acceptable alternatives.

7.

8.

Determine capital and annual operation and maintenance costs of alternative delivery systems which meet the requirements of the supply scheme.

Investigate the environmental impacts and the institutional arrangements necessary for implementation of the alternative solutions.

9.

Formulate recommendations for other necessary studies, design, and construction.

STUDY AREA

The political boundaries of the study area are defined by the limits of Pima County on the east, north, and south, and by the eastern boundary of the Papago Indian Reservation on the west.

The hydrologic unit chosen as a basis for delineation of the working study area is the surface water drainage subunit as defined by the United States Geological Survey.

These include those portions of the Altar Valley /Avra Valley subunit, the Santa Cruz subunit, and the San Pedro subunit within Pima

County.

This boundary indicates the physical area of intensive effort, but it does not limit the staff's ability to investigate alternative solutions both inside and outside the area described.

SCHEDULE

This program is being conducted over a 10 -month period in three phases.

Phase

I

-- Problem identification.

This phase includes the collection, assessment, and projection of water demand and water supply information.

Water budgets and modeling results are used to depict the magnitude of the existing problem.

During this phase, the goals of the remaining two phases will be established.

This phase will be completed with the issuance of a Problem Identification Report which will include:

1.

A report which shows depth to groundwater for the Avra /Altar basin and the Tucson Basin projected over the planning period.

This includes reports describing the Arizona Water

Commission's computer model which is utilized in the process, and the water use data of the participants.

This will be done primarily by the Arizona Water Commission.

The contract has been executed and the report will be available in May 1979.

2.

A report considering the feasibility of exotic water supplies, including importation from the

Yukon and the Columbia rivers system, weather manipulation, vegetation manipulation and soil conditioning, desalinization of sea water, and even icebergs.

University of Arizona under contract to the Corps.

This is being done by the

3.

A report on the feasibility of utilizing wastewater as a water supply north of the treatment plants for agricultural irrigation in Marana and /or the Avra Valley area; and south of the treatment plants for agricultural irrigation and mining process water, and as one source of water for the Papago Indian Tribe.

4.

A report on the projected unit costs of power.

5.

A report on the preliminary feasibility of alternative alignment and termination sites for

Central Arizona Project water.

This will be done by the Bureau of Reclamation.

Phase II -- Alternative plan development.

This phase includes the support of the Water Resources

Coordination Committee's search for acceptable and feasible solutions to the problems defined in

Phase I.

Various solutions proposed by staff and other study participants, and approved by the Water

Resources Coordination Committee, will be examined to determine their preliminary engineering, economic, and environmental feasibility; their impact on the existing problems; and their impact on each study participant.

This phase will be completed with the Water Resources Coordination Committee's selection of the most viable solutions (2 -3) for detailed analysis.

Phase III -- Plan selection.

This phase takes the viable solutions (a small array) and develops the information necessary for selection and implementation of a preferred plan.

Costs, benefits, environmental impacts, and engineering feasibility are considered.

This information is provided to the

Water Resources Coordination Committee for final plan selection.

Once the decision is made, final

94

details are developed for the selected solution.

This phase is completed with the issuance of a

Preferred Plan Report.

MANAGEMENT ARRANGEMENTS

The District Engineer is charged with final authority for the administration and management of the study.

He relies upon various cdnmittees and his staff to discharge these responsibilities.

Figure 1 shows the structure through which the study will be conducted.

Various responsibilities are outlined as follows.

The Water Resources Coordination Committee.

The formation of this committee was explained earlier.

Its purpose is to provide a forum for open discussion regarding water supply and use pattern problems and their possible solution.

This committee has local review and approval authority over the study.

Voting by this group provides general policy guidance, approval of study methods, and approval of interim technical reports.

This voting procedure will not necessarily apply to the negotiation process utilized during Phase II and III.

The committee will determine the actual negotiation process.

The

Corps will facilitate the negotiation process, but recognizes that each study participant must develop an individual position or plan of action during negotiations.

Technical Committee.

The Technical Committee is composed of representatives of Federal, State, and local agencies; private groups; and required experts.

Members were selected by the U.S. Army Corps of Engineers so as to provide adequate technical expertise for the area of concern.

This committee will not necessarily meet together at any single time, but is a more informal arrangement with continuous interaction between its members.

The duties are to:

1.

Identify specific planning problems and conflicts.

2.

Recommend study methodologies.

3.

Provide agency representation and agency policy.

4.

Review, assess, and comment on program results.

5.

Provide a communication link between the U.S. Army Corps of Engineers and the represented agency.

6.

Provide access to technical information available within the respective agencies which may contribute to the study.

The U.S. Army Corps of Engineers Study Team.

The day -to -day study management is carried out by the U.S. Army Corps of Engineers' Study Team.

This staff is supplemented by specialists from within the Los Angeles District, other Corps agencies, and firms under contract.

The Study Team is located in

Tucson.

The Citizens Advisory Committee.

The Citizens Advisory Committee is utilized as an advisory committee for wastewater reuse matters.

Membership in this committee is open to everyone, and the committee has been asked to nominate a citizen to the Water Resources Coordination Committee.

OTHER PLANNING EFFORTS

There are several other ongoing planning efforts into which the results of this program should be integrated.

The Arizona State Legislature's Groundwater Management Study Commission will address water management options in the state's critical groundwater basins from the State Legislature's point of view.

Their report is to be completed in June 1979, and the water management strategies developed as a result will be administrative in nature.

The results of the Corps program will be the basis for the technical implementation of the Groundwater Management Study Commission's work.

The U.S. Geological Survey is conducting a regional basin study over a five -year period.

The

Corps' work will be of considerable value to the U.S. Geological Survey's study.

The City of Tucson and the U.S. Geological Survey are conducting a study of subsidence in the region.

The work in the Corps' program will provide projection of groundwater depletion at a site specific level for their use.

95

STUDY COSTS

The total cost of the study is $480,500.

This estimate reflects a Federal effort amounting to

$293,000 and a non -Federal effort of $187,500.

Expected funding is as follows:

1.

2.

Major water users (in -kind service contributions)

Environmental Protection Agency

3.

Bureau of Indian Affairs

$187,500

$101,000

4.

U.S. Army Corps of Engineers

$ 87,000

$105,000

SUMMARY

In conclusion, the U.S. Army Corps of Engineers is working with private citizens, State and

Federal agencies, local governments, and the major water users of eastern Pima County to develop acceptable and feasible solutions to the water supply problems of the region.

This program is now in the middle of the problem identification phase, and should be completed this year.

Almost 40% of the cost is being borne by the local study participants.

The results of this program are being coordinated with other Federal, State, and local planning efforts.

A community has only so much energy and time to expend during any one period on any one problem.

The Tucson region now has the resources available to address its water supply problems.

Now is the time to resolve this issue so that in the near future the community can transfer its attention to the region's other pressing social and environmental needs.

REFERENCES CITED

Pima Association of Governments.

1977.

Economic base and employment; population; and land use change by drainage area reports.

Water Quality Management Study (PAG -208), Tucson, Arizona.

An overview.

Pima Association of

Thuss, Michael F.

1978.

Water resources for Pima County, Arizona:

Governments, Tucson, Arizona.

Six volumes.

559 pp.

96

HYDROLOGIC INVESTIGATION OF THE DRY LAKE REGION

IN EAST CENTRAL ARIZONA

James J. Lemmon, Thomas R. Schultz and Don W. Young

ABSTRACT

The Dry Lake Region is located in Navajo County, Arizona, near the southern margin of the Colorado Plateau.

The region's internal drainage basin of 160 mi2 is further augmented by 50 mi2 of the Phoenix Park Wash drainage.

The dominate surface water inflow to the playa is the 12 to 13 MGD of paper pulp mill effluent from Southwest Forest Industries near Snowflake, Arizona.

As a result, the playa surface water is now covering several thousand acres.

Dry Lake water quality is relatively poor by

Arizona Department of Health Services (ADHS) drinking water standards.

Ground water in the region is produced from the Coconino Aquifer which is comprised of the Coconino

Sandstone and the Kaibab Limestone.

The depth to ground water is 400 feet with a saturated zone 100 -175 feet thick.

Wells in the region yield from 0 to 500 gpm.

The presence of the Holbrook Anticline and the Dry Lake Syncline influence both ground water flow direction and artesian conditions.

There is concern that the playa may not be suited as an evaporative disposal basin because of the potential influence that karst topography and linear surface features may have on the water balance of the region.

INTRODUCTION

This report is a combination of on -site field investigations and the review of available literature and aerial photography.

The investigation was undertaken in connection with a proposed land exchange between the Arizona State Land Department and

Southwest Forest Industries, Inc.

(SWFI) during the late summer and fall of 1978.

PHYSIOLOGIC SETTING

The Dry Lake playa lies in the southern portion (Township 14 North, Range 18

East) of Navajo County, Arizona, approximately 126 miles east, northeast of Phoenix,

Arizona and 20 miles west, northwest of Snowflake, Arizona (see Figure 1).

The region is historically made up of small farming, ranching and recreational communities such as the Town of Snowflake (population 2600) which date back to the late 1800's.

Several small abandoned towns dot the region, such as the Town of Zeniff (Barnes,

1960).

The largest single employer within the region is Southwest Forest Industries,

(SWFI), who owns and operates a pulping and paper mill located in Section 21, Township

13 North, Range 19 East west of Snowflake, Arizona (Figure 1).

Liquid wastes from their operation are discharged to Dry Lake (Section 1, Township 14 North, Range 18

East) for evaporation and settling.

The project area lies in the COLORADO PLATEAUS PROVINCE (Heindle, 1960) between

5800 and 6000 feet elevation (MSL).

Vegetation types trend from desert grassland to juniper -pinon woodland.

The alluvial soils found in the Dry Lake region are predominantly derived from decomposed Moenkopi sandstones, silts and clays of the Moenkopi

Formation as well as some igneous materials.

Precipitation in the region averages 12.28 inches per annum, and the mean annual air temperature is 52.1 °F, with extremes from below freezing to 100 °F (Johnson, 1962).

Very windy conditions (25 -45 mph) prevail throughout the year.

The authors were all members of the Water Rights Division staff of the Arizona State

Land Department, Phoenix, Arizona.

Their positions were respectively Technician,

Hydrologist, and Hydrologist.

97

1.ì; ,..

.

.n1

is

.

SWFI PAPER MILL

_L

1-

/___._,.

-4

!sE row u s. csoiocol sUMV[Y

FIiF FsF wP i'.100000

'A,

,o

1.

;

s

++11

w

..-.

~

.

\

., io

"

1LL

I o r

¡

Ì - -i

-S. `

{

1. .....

AREA OF s

,.-

1,21

,---".

`

REPORT

10 0

10

F 322L-E--í7.-.'__.-__].-:

0

10

70 MILES

20 KILOMETRES

FIGURE 1

Area of Report Showing Dry Lake, SWFI Mill and

Effluent Discharge Channel

(Taken from New Mexico Geological Society, 13th Field Conference, 1962)

98

GEOLOGIC SETTING

Dry Lake lies equidistant between the surface expression of the Holbrook Anticline to the north and the Dry Lake Syncline to the south (see Figure 2).

The Lake overlies approximately 60 feet of Quaternary alluvium.

Beneath the alluvium is a thin veneer of Kaibab Limestone (Arizona Bureau of Mines, 1960).

The Kaibab and the underlying

Coconino Sandstone are the principal aquifer in the area (Mann, 1976).

The Coconino is 729 feet thick beneath Dry Lake (California Oil Co., 2 test holes).

Beneath the

Coconino is the Supai Formation, an interbedded sequency of shales, siltstones, and evaporites comprising some 1400 feet of thickness.

The Dry Lake drainage area comprises about 160 square miles of internal drainage and 50+ square miles of drainage from Phoenix Park Wash, which breeches the internal drainage basin along its southern edge.

The origin of the Dry Lake depression appears

(Bahr, 1962) to be the result of collapse from dissolution of the evaporites within the Supai.

Geologic evidence indicates the collapse was pre -Quaternary (greater than one million years ago).

The drainage patterns illustrated in Figure 2 are indicative of the geologic time frame.

A less dramatic geomorphic feature is the karst topography east of Dry Lake (Township 13 North, Range 20 East) and northwest of Dry Lake

(Township 16 North, Range 17 East). The sinkholes surrounding Dry Lake are locatable on the ground and easily identified on aerial photography.

It is thought that the origin of the smaller sinkholes located in and around Dry Lake are a result of tension

- induced features at the time Dry Lake depression collapsed.

Figure 3 is an interpretation of

U -2 aerial photography indicating many of the sinkholes located in and around Dry Lake.

Field work at Dry Lake has disclosed at least one set of sinkholes along the southern edge of the lake, enclosed by a dike to prevent movement of water into those features.

The deepest portions of the sinkholes have a liquid residue purported to be the remnants from a filling of the features several years ago.

of seepage through the dike.

below the elevation of the lake surface.

There was no evidence

The elevation of the residue surface was about ten feet

Reportedly, several sinkholes opened up several years ago as a result of a rapid rise in lake level.

Bahr (1962) found alluvium- filled older sinks, which indicates the appearance of sinks is an on -going process.

Proper monitoring of the potential for new sinkholes and subsequent diking should prevent movement of Dry Lake water into any new sinkholes.

HYDROLOGIC SETTING

GROUND WATER

The following ground water conditions are summarized from Mann (1976).

The Kaibab and Coconino, combined, comprise the aquifer in the region surrounding Dry Lake.

Well yields are from 0 to 500 gpm and are a function of the saturated thickness, artesian conditions, and well design.

Wells are designed for the intended use, that is, any yield up to 500 gpm is generally possible.

In the near vicinity of Dry Lake, depth to water is approximately 400 feet.

of the lake.

Water is generally absent several miles to the north

The thickness of the saturated zone will vary in the region south of the lake, but rough estimates of 100 -175 feet would coincide with regional trends.

Volume of recoverable water is highly variable.

In general, ground water moves perpendicularly to the structural contours of the

Coconino.

However, near Dry Lake, regional flow is northeast.

From an interpretation of Mann (1976) ground water probably moves northwest and southeast at Dry Lake as a result of the Holbrook Anticline.

Recharge to the aquifer occurs along the Mogollon

Rim through the more permeable materials.

occurs along the northward flowing streams.

Some percentage of the recharge undoubtedly

For example, in Lone Pine Dam south of

Show Low, seepage losses into the Kaibab are of such magnitude that almost all of the water is lost.

SURFACE WATER

Internal drainage patterns developed within the subsiding Dry Lake basin.

Some older drainage channels show evidence of reversal as the gradient reversed (refer back to Figure 2).

The major drainage pattern of the basin does not appear to be fault aligned; however, Phoenix Park Wash, which is a major drainage within the basin, exhibits a dendritic pattern.

There is evidence, though, that numerous minor drainages are influenced by shallow sub -surface fractures.

The small sinkholes inspected on the south side of the lake exhibited a closely spaced linearity.

Although not substantiated by previous geologic data, it is believed that these smaller (12 -25 foot diameter) sinkholes are not

99

1mile

SHORELINE

18 MAY 72

15 MAR 78

SINKHOLES

UNDERWATER

MAR 78

FIGURE 3

Interpretation of Aerial Photography

Showing Dry Lake, Sinkholes and Other Changing Features

D o

101

associated with dissolution subsidence of the Supai Formation as previously eluded to, but are, in reality, associated with dissolution phenomena occurring within the Moen kopi Formation (or within a thin veneer of Kaibab Limestone) which underlies the alluvium in the region.

The Moenkopi Formation consists of "predominantly sandstone, siltstone, claystone, mudstone, limestone, and gypsum beds" (Johnson, 1962).

As mentioned previously, fracturing of the Moenkopi Formation probably occurred as a result of compression stresses on the upper strata as the Dry Lake basin subsided.

Surface drainage flowing through such fractures, and subsequent dissolution of shallow lenses of limestone and /or gypsum, could account for formation of localized karst formations, which periodically exhibit themselves through collapse of the alluvial overburden.

Present day internal drainage to the basin most likely is continuing to play an active role in the formation of this localized karst formation.

The possibility of leakage from Dry Lake and /or the effluent channel to the Coconino aquifer increases significantly if active karst formation is taking place within the Moenkopi and /or Kaibab Formations.

WATER QUALITY

Other than Dry Lake playa, the quality of the surface water in the area is assumed to be similar to other regions in the Colorado Plateau.

There are no known mineral or salt deposits that would add large amounts of dissolved or suspended material to the surface runoff.

Phoenix Park Wash and Sheepskin Draw both carry large amounts of sediments during peak runoff events as is the nature of ephemeral streams in Northern

Arizona.

The water in Dry Lake playa had a noxious taste and odor when sampled during the field investigation.

The dark color of the water sample submitted to the Arizona

Department of Health Services Laboratory has hampered the analysis procedure.

Ground water samples from two wells were tested by Arizona Testing Laboratories for tannin /lignin content and compared to the basin surface water.

One well

[A(14- 18)5adb] was two (2) miles west from Dry Lake and the other well [A(14- 20)3odad] was seven (7) miles to the east.

The results of the chemical analysis for these wells shows the quality to be only marginal for domestic purposes at this time.

The surface water was found to have 170 milligrams per litre (mg /1) tannin /ligfi,n -like substances while the ground water contained less than 0.2 mg /1 tannin /lignin.

There is an abandoned well within several thousand feet of the basin

[A(14- 19)7ccc] which was recommended to SWFI as a potential monitor well for the Coconino aquifer with respect to potential contamination from Dry Lake.

This well was recently refurbished by SWFI and pump tests and sampling are presently under way.

No chemical results are available for this well at the time of this report, however, verbal reports of the pump test results would indicate that the Coconino aquifer in this region may have a low specific capacity (0.5 gallons /minute /foot of drawdown).

CONCLUSIONS

ENVIRONMENTAL CONCERNS

The playa is currently being used as the main disposal basin of a watery pulp mill effluent (average discharge is 12 -13 million gallons per day).

The toxicity and longevity of the discharge substance is unknown; however, McKee and Wolf (1967) report that tannin /lignin levels of 100 mg /1 will kill goldfish (carp) in nine

(9) hours.

The amount of chlorine needed to disinfect a public supply would greatly increase if there is lignin in the raw supply.

The levels of tannin /lignin in the basin will increase over time regardless of the amount of effluent disposed into the basin due to evaporation.

The photographic interpretation of linear features and sinkholes in the Dry Lake playa leads to a conclusion that there are geologic conditions that could allow contamination of the ground water system by the surface disposal of effluent.

There are

1.

The .2 mg /1 would be the approximate lower limit of test accuracy performed by ATL, however, there are measurements in the range of .05 mg /1 accuracy (McKee and

Wolf, 1967).

102

no less than 25 sinkhole -like structures in the present wetted portion of the lake

(Figure 3).

While the ground verification or truthing of the aerial photographic interpretation cannot be done under the water in the playa, there are some areas on the southern and western margins of the basin that could be reached.

The geologic conditions that are evident in the area include sinkholes of varying sizes from several feet across to several hundred feet wide.

The linear features, or surface expressions of fractures, faults or cracks, show that there are subsurface geologic conditions that are influencing surface topography and possibly ground water movement.

Percolation rates may be on the order of several inches /day to several feet /year; however, little is known about the Dry Lake region's Coconino Sandstone permeability.

Transmissivity coefficients of Coconino Sandstone have been developed for other areas of the Colorado Plateau, but they may not be applicable to the hydrologic situation here.

The regional direction of ground water flow trends to the northeast; however, the Holbrook Anticline may channel the flow to the east and southeast at Dry Lake.

The sealing clays of the playa floor and the silts and mudstones of the Moenkopi

Formation may not bridge the wider cracks and voids resulting from sinkholes and fracturing.

This means that even if the majority of the playa floor is solid clay, the sinkholes and cracks could provide a direct pathway from the surface to the Coconino

Sandstone.

If there is a direct link of the surface and ground water systems, then the horizontal rate of ground water movement in the Dry Lake region becomes important.

RECOMMENDATIONS

1)

The need for further hydrologic investigation into the exact nature of ground water movement and possible contamination in the Dry Lake basin is evident from the inconclusive data collected from the area to date.

The data that was collected and analyzed indicates that, while at the present time there is no apparent contamination outside the basin proper, little is known about the site in terms of geology and hydrologic systems as they relate to water quality and environmental hazards.

2)

The interpretation of the U -2 aerial photography resulted in a number of sinkholes being identified in the wetted portions of the lake.

Field investigations of the area did not confirm or deny any observable downward flow of surface water.

The use of currently available geophysical methods should be employed to determine the extent of the openings and the rate of downward water flow, if any, for each sinkhole and fracture site.

3)

A water balance of the inflow and evaporation rate from the surface of the lake should be recomputed (Dry Lake is equipped with a stage recorder; however, this was inoperative at the time of this investigation).

The rapid build -up of a 40 -acre deltaic deposit at the lake head should be examined to determine the exact nature of the delta.

It is conceivable that the surface runoff into the basin was underestimated by several thousand acre feet (personal communication with L. Mann, September,

1978).

This additional inflow combined with a smaller surface area for evaporation could result in higher lake elevations and greater downward pressures on the basin floor.

New and larger collapse features might develop beneath the water, yet remain undetected by visual inspection.

4)

Although a potential exists for contamination of the ground water in the Dry

Lake area by pulp -mill wastes, there is little evidence to substantiate whether influx to the Coconino Sandstone aquifer is actually occurring.

Continued monitoring and sampling is recommended.

Possible consideration might be given to drilling additional monitoring wells in the regions east and northwest of the lake.

5)

The contamination of regional ground water by pulp mill wastes cannot be discerned at this time.

Further data collection and analysis would be required to either substantiate or disprove the theory that waste materials are entering the

Coconino Sandstone aquifer from Dry Lake or the waste transport channel through karst dissolution cavities.

103

REFERENCES CITED

Aley, Thomas, 1972.

Vol. 14, No.

3.

Groundwater Contamination from Sinkhole Dumps, Caves and Karst:

Arizona Bureau of Mines, 1960.

Geologic Map of Navajo and Apache Counties:

Bureau of Mines, Tucson, Arizona.

Arizona

Guide Book of the

Bahr, C. W., 1962.

The Holbrook Anticline, Navajo County, Arizona:

Mogollon Rim Region, New Mexico Geological Society.

Barnes, Will C., 1960.

Arizona.

Arizona Place Names:

The University of Arizona Press, Tucson,

Fuller, W. H. and J. R. Erickson, 1961.

Area West of Snowflake, Arizona:

Soil and Land Classification of the Dry Lake

Sante Fe, New Mexico.

Guyton, William F.

& Associates, 1975.

Compilation of Data and Information for the

Coconino Sandstone Aquifer in the Snowflake Area, Arizona:

Houston, Texas.

Heindle, L. A. and J. F. Lance, 1960.

divisions of Arizona:

Topographic,Physiographic, and Structural Sub-

Arizona Geological Society Digest, Vol. 3.

Johnson, P. W., 1962.

Water in the Coconino Sandstone for the Snowflake -Hay

Area, Navajo County, Arizona:

Hollow

U.S. Geological Survey, Water Supply Paper 1539 -S.

Mann, Larry J., 1976.

Arizona:

Ground Water Resources and Water Use in Southern Navajo County,

Arizona Water Commission Bulletin 10, Phoenix, Arizona.

State Water

McKee and Wolfe, 1967.

Water Quality Criteria California Standards:

Quality Control Board, Sacramento, California.

McGraw -Hill Book

Morisawa, Marie, 1968.

Streams, Their Dynamics and Morphology:

Company, New York, New York.

Sweet, R. and R. H. Fetrow, 1975.

Ground Water, Vol. 13, No. 2.

Ground -Water Pollution by Wood Waste Disposal:

Valley National Bank, 1977.

Arizona Statistical Review:

Phoenix, Arizona.

Valley National Bank,

Young, D. W. and R.

B. Clark, 1978.

Evaluation of Potential Effects of in situ -

Copper Leaching on Ground Water Quality in the Miami, Arizona Region - Special

Project Report, Arizona State Land Department, Phoenix, Arizona (Unpublished).

104

Visual Impacts:

Perception and Modification of Surface Mining Operations on the Black Mesa

Jon Rodiek

ABSTRACT

Scientists and industrialists are now seeking from their reclamation brethren an explanation of the new criteria and standards which will enable all of us to recognize a visually reclaimed site when we see one.

To do so may require more of the viewer than the ability to perceive visually.

One must understand the larger game in which reclamationis played.

INTRODUCTION

The title of this paper implies a relationship between visual impacts, perception and the resource base of a given area.

The discussion of such a lofty issue will be reduced for practical purposes to:

1.) a discussion of technical treatments regarding the visual resource on surface mined lands and; 2.) the relationship of visual resource modifications and perception of the environment.

The situation on the Black Mesa is a classic example of the resource development- conservation dilemma.

We speak of this dilemma in contemporary terms as the energy development environmental protection issue.

Any discussion of this subject requires some fairly definite meanings of the word

"development" and "conservation ".

I shall forego the development of background material and simply propose two definitions.

into potential capital.

Resource development has to do with the conversion of inert natural processes

Resource conservation consists of modifying or reducing a people's standard of consumption in the direction of the future.

Resource development on the Black Mesa is designated as the production of coal.

Resource conservation of the visual resource involves the elemental reconstruction of the landscape; the rehabilitation of its productive processes; and the creation of a compatible land use for its people.

BACKGROUND

The Navajo and Hopi Nations have chosen to convert their coal resources into social capital.

That decision sets into motion the concerns for the rehabilitation of the residual landscape.

Guidelines for beginning this task are outlined in the Surface Reclamation Act of 1977.

Its goals simply stated are to encourage the development of positive procedures to stimulate the simultaneous achievement of energy supply and environmental objectives.

mining is not possible on the Black Mesa.

Immediate restoration of the visual resource after disturbance by

There are three reasons for this.

WATER SUPPLY

The semi -arid conditions found within the juniper -grassland ecosystem of northeast Arizona make it quite impossible to realize any biomass productivity without the acquisition and manipulation of water.

Water is too much a limiting factor here to permit us to develop vegetative cover in its absence.

Ground water supplies are not available nor reliable enough to be considered for supplimental use.

Surface water from offsite sources are not feasibly available either.

On site surface water in the form of precipitation, runoff, and snow melt are the single largest sources of usable water.

OVERBURDEN

The low sulfur content coal lies in a seam some sixty feet below the land surface.

105

Surface

RODIEK /Visual Impacts

PAGE 2 mining techniques have large volumes of overburden that connot be removed or reshaped beyond some minimal level.

Large scale recontouring and subgrade reconditioning treatments will in all likelyhood not be cost effective.

UNIQUE CULTURE grazing.

The present land management practices of the resident indian population is oriented around sheep

Productivity on these lands is inherently low.

Therefore successful grazing requires the

Indian land manager to utilize available forage over large areas of the landscape.

complicates and worsens the condition of the available range resource.

Overgrazing simply

The reorientation of an existing culture through the availability of new capital, new values, and new life styles has not converted these people's lives, it has complicated them.

The problem of land reclamation on the Black Mesa forces the decision makers to bring not only technical, but ethical and cultural strategies into the planning process.

equally involved in this web of values.

Concerns for rehabilitating the visual landscape are

The existing visual image is very much fixed by the dyiamics of ecosystem productivity, land use production or yield and the intangible values of a culture.

Overtime the activities for engaging man with land take on identifiable patternsand characteristics.

These resource use practices formed by a peoples own mode of reasoning are consistent with other values and behavior within their culture.

people may see in their landscape may very well reflect theirsystem of involvement with it.

What a

Perception and modification of the visualresource therefore transcends the reconstitution of the physical structure.

TECHNICAL TREATMENTS APPLIED TO THE VISUAL RESOURCE

Landscape treatments involve hydrologists, biologists, and landscape planners and centers on substrate treatments, recontouring treatments and revegetation treatments.

The cast overburden and parent materials makeup the physical conditions in which plant succession takes place.

The fundamental purpose of substrate treatments is to modify the negative growing conditions found there.

Presently chemical additives (lime carbonate, organic fertilizers, soil binders) and organic topsoils are being tested.

A series of mulches (rock, stone, slay material, bark mulch) used in combination with organic soil additives may offer realistic alternatives to a 100% topsoil covering.

Recontouring the substrate -topsoil layer is designed to influence water regime balances in terms of water retention, drainage, and soil -water availability relationships.

The purpose of the recontour plan is to capture and harvest water and deliver it to designated vegetative areas on site.

The recountour plan must be designed with microclimatic, visual modifications and land use strategies in mind.

The substrate which is set into a predetermined configuration with proper consideration given to water retention and drainage problems provides a fixed set of growing conditions for selected plant species.

These conditions set standards for selecting a proper collection of suitable plant material.

Our best rehabilitation efforts rely heavily on the reintroduction of vegetation so as to establish the process of self- repair (plant succession) and self -adjustment (species tolerance)on the site itself.

Through these preliminary treatments a visual modification strategy can be developed.

VISUAL ACCESSIBILITY

Visual accessibility is a term used to describe the kind of vantage points, roads or avenues that afford the incidental viewer or traveler a view of the scar.

When applied skillfully and strategically topographical screening, vegetative buffers and scenic enrichment through textural and tonal additives can greatly enhance the picture plane the viewer perceives.

The designer is essentially working with position and vantage point characteristics to locate treatment activities that enhance the visual image.

ORIENTATION AND CONFIGURATION OF THE SCARRED AREA

Surface mining operations are designed to respond to the position and location of the ore bearing seam, the climatic conditions and the existing topography.

Visual rehabilitation schemes must work with these residual landforms and within recontouring cost constraints.

The goal is to retrofit this residual land form back into its surroundings.

The designer must manipulate the physical form of the

106

RODIEK /Visual Impacts

PAGE 3 overburden piles so that they simulate characteristics of the adjacent landscape.

manipulation is further enhanced by fine line form manipulation.

This gross form

This is accomplished by treating the tonal character of the exposed surface (i.e. soil. rock, vegetational) so as to yield a desired reflectance characteristic in both color and light quality.

Textural characteristics of the natural elements can be manipulated in a similar manner.

The land planner is essentially employing the principles of biology (succession, growth form etc.) and the principles of three dimensional perception to integrate the landscape into one visual image.

CONTEXTUAL INTEGRITY

The undisturbed visual image of the landscape is a product of cultural and natural forces interacting together.

When the physical scene is changed so is our perception of it.

The culture of a people consists of a set of activities and a fund of beliefs and techniques for manipulating their resources.

One fact is universal to all cultures.

By shaping their environment they shape themselves and their future as well.

The visual resource like its mineral and biological counterparts does not exist, within our perceptions at least, until we define it.

Like its counterparts the visual resource is not separately defined until we extract it from its surroundings.

The dismanteling of our environment reveals to us its systematic workings and organizational levels if we look critically at it.

Reintegrating the landscape contexturally back into its setting both structurally and functionally is a task of a much different nature.

This requires of us the highest order of understanding of ourselves and our relationship with our environment.

RESOURCES AND PERCEPTION

The problems of contemporary life in any setting are insoluble.

Our actions are mediated by values through which we balance opposing impulses.

We shape, or in the case of visual modifications to mined lands, reshape the landscape by finding appropriate principles to guide us.

The natural environment is perceived and understood at its most fundamental level by the science of ecology.

Natural resources are defined by our values, techniques and activities.

The perception of the two in the dynamic state of alteration and change transcends simple physical reconstruction techniques.

apply these to the resolution of our problem.

Somehow we must

Visual resource restoration within surface mined lands on the Black Mesa is three things.

Its technical.

We attempt to integrate biological and perceptual elements.

It is also ethnological in nature.

We realize the vague line of dependency between culture and environment is particular to each setting. To ignore this reality is to invite failure. Finally it is ethical in character.

We have unleashed powerful, inventive and exploratory urges upon ourselves and the landscape.

They seem to be part of our inheritance here in the western world.

The legislative and technological constraints associated with visual modification treatments simply demonstrate the increasing price we have to pay for indulging them and the ingenious ways in which we contrive to meet that price, no matter how steep it becomes.

107

IMPACT OF DEVELOPMENT ON STREAM FLOWS by

1

2

Paul D. Trotta, James J. Rodgers, William B. Vandivere

3

ACKNOWLEDGEMENT

The research reported in this publication was supported by the Rocky Mountain Forest and Range

Experiment Station, U. S. Department of Agriculature, Forest Service, through the Eisenhower Consortium for Western Environmental Forestry Research (EC 270).

INTRODUCTION

In the arid regions of the southwest and west forests provide significant quantities of the water for the life style chosen by its human inhabitants.

The forests not only provide the "catchment" for renewing surface and ground water resources but also have a significant role in regulating the supply of water to these sources.

It is essential, therefore, to the forest manager as well as the city planner to come to a qualitative as well as quantitative understanding of the impacts of development on stream flows and how the regulating effect of the watershed on stream flow is influenced by changes in land use. In a recent survey of municipal planners one of the currently unsatisfied informational needs of planners was reported as being the relationship between planning decisions and impacts on the water resources of the region, and all the resultant feedback effects.

A hydrologic simulation model was chosen in this study to help answer some of the above questions.

LITERATURE REVIEW

URBANIZATION

Inquiry into the hydrologic characteristics of urbanizing watersheds has concentrated on the alteration of: shape and timing of runoff hydrographe, flood magnitudes for events of varying recurrence intervals; and, quantitative evaluation of runoff volume and stream sediment loading. In a

California study,

McBride

(1976) observed that urbanization of a watershed in the Berkeley Hills region produced values of peak daily streamflow, ten times those of a similar non- urbanized watershed.

Another California investigation (James, 1965) contrasted runoff volumes from a watershed in both its pre -urban and urbanized forms.

It was estimated from streamflow records volume from the completely urbanized watershed was 2.3 times that that runoff produced by the same basin in its rural condition.

For a topographically dissimilar area in North Flordla, Turner, et.

al.

(1975) found that urbanization, defined as residential- commercial land use, of a watershed near Lake Jackson increased peak stream discharge and total stream discharge volume.

A review of publications dealing with the effects of urbanization on floods of varying recurrences intervals was compiled by

Hollis (1975).

Conclusions projected from empirical studies by Leopold

(1968) showed that paving of less than 5% of the drainage basin area had a negligible effect on floods with return periods of one year or more, while paving of 30$ of the watershed may double observed discharge resulting from a 100 -year flood. In addition, Herr (North Carolina) was cited as having inferred that a flood with a recurrence interval greater than 150 years is essentially unaffected by urbanization.

Studies involving the urbanization of previously dominant forested watersheds often necessitate consideration of more variable soil and topographic regimes than those customarily identified with low lying urban plains.

An inquiry of this nature was made by Wooldridge (1967) on a small forested watershed near Seattle,

Washington.

Results demonstrated urbanization increased the proportion of rainfall manifesting itself as daily discharge approximately three fold:

From 0.099 to 0.296 c.f.s. per inch of rainfall.

In a similar study conducted in East Africa, Dagg and Pratt (1962) related runoff from the largest observed storms on a 31 -acre forested watershed to less than 1$ of measured precipitation.

In contrast, runoff from a 32 -acre contributing area dominated by a housing and administrative complex amounted to 36% of rainfall when subjected to these same hydrometerologic events.

1)

Asst Professor, College of Engineering and Technology, Northern Arizona University, Flagstaff,

Arizona.

2)

Hydrologist, U.S.

Tucson, Arizona.

Forest Service, Rocky Mountain Forest

&

Range Experimental Sta.,

Tempe, Arizona.

3) Research Ass't, College of Natural & Renewable Resources, University of Arizona,

109

An extensive study directed by the Northeastern Forest Experiment Station (Lull and Sopper,

1969) documented hydrologic response (annual ratios of stormflow to precipitation) and its seasonal variation in progressively urbanizing forested watersheds in the Southern New England region.

The effect of urbanization was evaluated as a net reduction in evapotranspiration.

Runoff from paved surfaces is assumed to enter directly into streamflow.

Increases in the impervious area of 25,

50, and 75 percent resulted in annual increases in runoff of 15, 29, and 41 percent respectively.

It was also found a measureable seasonal variation existed between forest and urban affected stormflows.

A completely forested area produced 80% of its runoff during the dormant season (October through

March) where as runoff for this same period from a heavily urbanized area (75% imperviousness) amounted to just 62$ of total annual streamflow.

Additionally, Lull and Sopper, utilizing data from East Coast watersheds, attempted to gage the effect of urbanization on streamflow records.

Regression analysis of actual annual runoff and predicted runoff, simulated through the use of a computer program (Black, 1967), indicated increasing stormflow, peak flow, and annual runoff for progressively urbanizing watershed.

THE WOODS CANYON STUDY AREA

Within the Beaver Creek study area maintained by the U.S. Forest Service is the Woods Canyon area which was chosen for this study.

There are four basic reasons why the Woods Canyon area was chosen: I. Plausibility of Development; 2. Similarity to other Developed & Developing areas in the region; 3. Data Availability;

4.

Forest Service interest in this Area.

Figures 1 and 2 illustrates the Beaver Creek study area and Woods Canyon within it.

PARTITIONING WOODS CANYON

Within the Woods Canyon area four separate hydrologic response units (HRU) were delineated.

Although eight hydrologic response units appeared initially to be the ideal number, from a purely hydrologic homogeneity etandpont, the realities of computer time and system complexity resulted in a reduction to four.

A variety of parameters were used to delineate the HRUs.

Among those most important from a hydrologic point of view were, drainage, elevation, slope and aspect.

Figure 2 is a sketch of the Woods Canyon area with the delineated HRUs. The area of each HRU is shown along with the major drainage ways and their length.

Table 1 shows additional information for each HRU such as: average slope, aspect, average width, average length, perimeter, length of boarder with adjacent HRUs top elevation, bottom elevation, and length of major definable drainage ways.

Figure 2

Woods Canyon Watershed

(within Beaver Creek)

Delineated HRUs & Linkages

Figure 1

Locator Map

-

Beaver Creek

Watershed (USFS 1977)

ARIZONA

SALT.VERDE

BASIN

PHOENIX spit IVIES

FLAGSTAFF

Beaver Creek

Watershed

110

Soil Types Within Woods Canyon.

within a watershed.

The types of soil and their depths is a significant parameter

Fortunately, a detailed study of the soil types to be found in the Woods Canyon area has been completed by the U. S. Forest Service,

(1967).

This detailed data has permitted for each HRU the calculation of an aggregate weighted average value for the field capacity, wilting point, saturated capacity and depth of soil.

Table 2 shows the relative percents of each soil type within each HRU.

The values of the ground moisture and soil parameters for each type of soil are presented in Table

3 again from the Woods Canyon Soils

Report.

These values were combined by weighted average based on the relative percents presented above.

The results of, this analysis are presented in Table 4, where the averaged field capacity, wilting point, saturated water capacity and depth are listed for each

HRUs.

HYDROLOGIC BACKGROUND

HYDROLOGIC VARIABLES

Hydrologic studies devoted to watershed response generally represent the physical processes occurring at the air surface -sub -surface interface by way of a simplified mathematical model, or water budget.

The outwardly manifest elements in this model are precipitation inputs, either snow, rain, or mixture, and runoff which may appear as surface (overland) or subsurface flow.

Changes in land use affect not only individual parameters, but their dynamic interaction as well. In the following paragraphs the potential impact of development upon these hydrologic variables is discussed.

Interception.

A portion of the precipitation which falls within a watershed is intercepted by vegetation as well as other protrusions from the ground surface.

Additionally, a variable proportion, termed throughfall, will clear the forest canopy uninhibited by obstructions and proceed directly to the forest floor.

Intercepted precipitation will at some point in time, depending on its fctrm, either evaporate from vegetative surfaces or exit downward through the canopy as drip or stemflow and eventually reach the ground surface.

Measurement of intercepted precipitation is usually accomplished by subtracting the value of in- forest precipitation catch from that obtained from a nearby open field gage.

This estimate is necessarily a rough one due to the

neglect of stemflow.

However, this component is generally considered negligible for coniferous stands.

Its significance, however, has been noted for some deciduous species (Eagleson, 1970).

The mechanics of snowfall interception are much more complex than that pertaining to rain.

Variables such as density and structure of snowflakes, wind speed, and air temperature affect the impaction process which creates an immense same factors may also force storage potential in the forest canopy.

a reversal of the process.

Solar radiation and

However, these winds aided by the turbulence produced by rough forest stands are capable of removing accumulated snow from impacted branches.

Reception of solar radiation pulses is also sometimes sufficient to melt the base layer of plastered snow resulting in the sliding of a snow mass from its anchoring branch (Miller,

1966).

Construction on second home sites as well as appurtenant small scale commercial development will certainly reduce interception of precipitation in the Woods Canyon area.

The difference between developed and undeveloped interception could be expected to be more pronounced during rainstorms than snowfall periods due to storage, albeit temporary, of snow by man -made structures.

However, because of the influence of heat conduction from underlying materials, a factor which is essentially negligible in soils of forest stands, the melting of these isolated accumulations would proceed quickly.

1

2

3

4

HRU

Area

(Mi.2)

4.48

6.44

5.02

3.00

Table 1

Ave.

Slope

.017

.037

.037

.030

Hydrologic Response Unit (HRU) Data

General

Aspect

Ave.

Length

(Mi.)

Ave..

Width

(Mi.)

Top

Elev.

(Ft.)

Bottom

Elev.

(Ft.)

315° 4.7

1.4

6370

270° 3.0

2.3

6850

7040 6480

270°

225°

2.8

2.5

1.8

1.2

7560

7480

7000

7080

Perimeter

(Mi.)

12.2

11.7

9.5

8.4

111

Table 2

Soil Types by Percent for

Woods Canyon Hydrologic Response Unite (HRUs)

Soil Type

Friana -seep

Complex

Rocktop Cobbly Loans

Cruice Clays

Bootlegger Cobbly

Clay Loans

Rocktop -Brolliar

Loans

Herrn- siesta Complex

Fain Cobbly Loans

Siesta Lome

Herm- rocktop Complex

Siesta- rocktop Complex

Percent of Total

Unaggregated 1

1.4

16.3

0.3

5.4

22.9

11.9

3.0

8.0

10.8

20.0

Percent of Total by Weighted Ave of HRUS2

1.3

'

19.4

nill

4.8

24.7

14

4.3

5.7

8.9

17

1

Hydrologic Unite

2 3

4

1 1

2 1

8

-

10

29

-

7

21

-

-

13

-

-

22 25

24 29

16

0

4

13

26

8

3

3

6

18

20

7

15

4

7

12

12

14

14

5

1) This was calculated by Forest Service for entire Woods Canyon by use of computer digitiser (U S. Forest Service, 1967).

2)

This was calculated in study by planimeter for each HRU and subsequently aggregated by weighted average using HRU areas.

Detention and Depression Storage.

area, a film of water will

Once the infiltration capacity is reached over a certain build up on permeable and impermeable surfaces.

In addition, natural depressions will store water which is then eventually infiltrated or evaporated.

These abstractions, defined as detention storage and depression storage, affect the timing of runoff to a greater degree than they affect volume of stormflow.

In combination they could be expected to change slightly in response to second home construction.

of

Evaporation and Transpiration.

course, water.

solar radiation,

All evaporative processes require the availability of energy and,

Inputs of energy which drive the evaporative mechanism are sensible heat

(heat used to warm the environment without a change of state), and wind, a potentially important factor in forest stands.

short and long wave

Evaporation from wet surfaces, be it soil, litter, or arboreal is relatively efficient in its utilization of available energy.

Rain moisture also maintains a greater absorptivity of solar radiation than snow which has a high albedo.

The aggregate evaporation from intercepted rainfall is dependent on the frequency of precipitation events.

The more times interception storage is filled, the greater the evaporative opportunities.

Transpiration will fluctuate according to the supply of radiant energy, turbulent mixing by wind currents, nature of leaf surfaces, and the state of internal plant water relations.

Increases in streamflow volume have been attributed primarily to transpiration reduction in a wealth of studies including that by Galbraith (1975) for forested regions of western Montana.

Removal of tree stands in the Woods Canyon watershed would logically be expected to reduce evapotranspiration in the area.

Even with the protection or of the regeneration of low lying vegetation in the harvested areas, the removal deep forest canopy would decrease the possibilities for evapotranspiration potential of forest canopy from that of the low vegetation.

Galbraith concludes that the forest more often exhibits a higher rate of evapotranspiration.

Rauner (1960,

63, 65) has also shown that warm air advection, certainly expectant over north central Arizona, enhances this increased rate.

112

Table 3

Ground Water Parameters for Woods Canyon Soil Types (U.S. Forest Service, 1967)

Soil Type

Total Depth

FLDCAPI

1

(inches)

2

Total Plant

Avail. Water

(inches)

3

Total Fld.

Capacity

(inches)

4

Wilting

Capacity

(inches)

(5 -4 -3) (6 -4/2)

WLTCAP

(5/2)

Friana-seep

Complex

48 7.3

17.2

9.9

.36

.21

Rocktop Cobbly

Loams

30

4.2

10.5

6.3

.35

.21

Cruice Clays

Bootlegger Cobbly

Clay Loams

36

22

5.2

3

12.7

7.3

7.5

4.3

.35

.33

.21

.20

Rocktop -brolliar

Lome

27 3.7

8.9

5.2

.33

.19

Herm- siesta

Complex

Fain Cobbly Loams

Siesta Loams

Herm- rocktop

Complex

Siesta- rocktop

63

32

65

41

43

8.5

3.1

7.6

5.2

5.6

20.3

7.7

17.2

13.5

13.6

11.8

4.6

9.6

8.3

8.0

.32

.24

.26

.33

.32

.19

.14

.15

.20

.19

SATWC

2x(6)

.72

.70

.70

.66

.66

.64

.48.

.52

.66

.64

1)

Field capacity per inch of soil as fraction of 1 inch.

2)

Wilting point capacity per inch of soil as fraction of 1 inch.

3)

Saturated water contents assumed to be approximately twice filed capacity.

Table 4

Averaged Soil Parameters for Woods Canyon

Hydrologic

Unit (HRU)

Soil Depth cm

3

4

1

2

Field Capacity

(FLDCAP)

% of soil depth

32

33

32

31

Wilting Point

( WLTCAP)

X of soil depth

19

19

19

18

Saturated Water

Content (SATWC)

% of soil depth

64

66

64

62

102

91

107

102

Infiltration.

Before the runoff process is initiated, a portion of the water occurring at the ground surface is extracted by the soil system.

This phase in the redistribution of water influx, referred to as infiltration, is dependent upon a variety of soil and ground cover factors (Hewlett and Nutter,

1969).

Infiltration rates are probably the most uncertain of all hydrologic parameters.

Not only do these rates vary seasonally, a result of fluctuations in vegetal activity, but differences are even apparent from storm to storm.

Higher rainfall intensities often result in sealing of entry pores.

This impact- produced effect causes an in- washing of fine soil particles which impedes the infiltration process. However, in forested areas where sufficient ground litter exists to eliminate this

113

disruption, it is a decrease in the capillary potential which accounts for the concurrent decrease in the rate of infiltration (Bodman and Colman, 1944).

shed.

The rates of infiltration capacity exhibited by forest soils are usually greater than prevailing rainfall intensities.

Thus, surface runoff is seldom experienced over an undisturbed forested water-

There are exceptions, though, and the shallow, clayey soils prevalent in Woods Canyon evidently provide relatively low infiltration capacities (Rogers, personal communication). The action of continuous degredatioii of the forest floor by grazing, logging, road building, and human usage

(hunters, recreatfonists, etc.) should also impair infiltration rates in the affected areas (Hewlett and

Nutter, 1969).

This intensified usage

Woods Canyon basin to establishment of a second home, recreation- oriented community.

of this alteration development.

will, may expectedly play a role in the hydrologic response of the

The extent as with previous abstractions, be a function of the nature and degree of

Percolation.

The progressive movement of water through the soil profile into the ground water system is an essential component of the water budget in humid regions.

But in the semi -arid environment of Arizona, most channel systems are ephemeral in nature and are not sustained by extensive groundwater.

Percolated water is thus neglected in the present study.

HYDROLOGIC SIMULATION

ECOWAT.

The watershed simulation model ECOWAT developed by J. Rogers (1975) is the vehicle used in this study to evaluate the effects of second home development on the hydrologic response of Woods

Canyon.

The model was assembled to components follows.

represent ecosystem dynamics indigenous to the coniferous forest biome.

A brief summary of the interaction between altered hydrologic parameters and model

Detailed explanation of ECOWAT formation and procedure may be found in

"Contribution to Woodlands Snythesis, Ecosystem Dynamics Section, Water

Relations and Hydrologic

Cycles ", (Waring, Rogers, and Swank, 1975).

Canopy Water Balance and Snowpack Components.

Water balance in the forest canopy as considered in the model has adequate parametric capacity with which to simulate changes in evaporation.

Since the canopy water holding capacity is described in terms of leaf, branch, and stem area, and vegetative type, storage capacities and their depletion and recharge are satisfactorily expressed.

In addition, evaporation of excess snow in canopy storage is neglected in favor of its redistribution to the ground surface.

ECOWAT's treatment of snowpack energy and water balance realizes the thermo dynamic energetica involved in the snowmelt reaction.

Litter Water and Surface Water Balance Components.

The litter layer found in forest regions is an important source of moisture storage which also increases the evaporative potential of the water influx.

Its displacement will most probably influence soil disruption and therefore infiltration capacities and subsequent runoff.

The litter water balance component successively evaporates moisture from storage, transmits it through the litter, and passes moisture through litter openings.

This litter component is significant in that it may affect the surface water balance even in partially cleared areas depending upon the degree of degradation practiced and the nature of construction at second home sites.

The surface water balance segment handles detention storage, infiltration, and overland flow.

Changes in infiltration capacities due activities are accounted for by this component.

to antecedent conditions, and construction and use

Soil Water Balance Component. This component covering percolation, lateral flow, and non- capillary flow within the soil system models critical soil conductivity and the lateral flow process which contributes to aggregate streamflows in the channel system.

Aside from soil water utilized in the transpiration process (plant -water component), developmental impact upon the soil water balance is expected to be negligible.

Plant Water Relations Component.

The simulation by this component is crucial to the successful assessment of changes in water yield.

Transpiration, as mentioned earlier, is expected to withdraw the greatest percentage of water from the watershed system.

Detailed description of the plant water relations encompasses consideration of water uptake, transpiration, internal storage, moisture stress, leaf resistance, and surface resistance.

Ponderosa pine is predominant at the experimental site; however, there is also a number of deciduous trees (Gambel oak).

Deciduous species must rely totally upon uptake through the soil -root system while coniferous trees are capable of drawing from an internal storage reservoir.

This component is a potentially sensitive indicator of the internal water relations which govern transpiration losses.

Simulated Development.

In order to appraise the long term impacts of development upon stream flow, six levels of development were simulated for the Woods

Canyon watershed.

These ranged from a no development situation to a complete development situation with a fairly high density.

The high density situation was provided to approach a quarter acre subdivision with paved streets.

114

Development progressed from the lowest elevation HRU up to the highest.

Fortunately, for the

Woods Canyon area this also allows the development to proceed from areas in close proximity to the major transportation link to areas more remote, steeper, and less suited for development.

illustrates the progression of development simulated.

Table 5

Table 5

Development Level

Simulation Number

1

4

5

2

3

6

-0--

HRU -2

Tr-o

},

0 0

1

1

} i

1

1

1

}

1

1

1

1

HRU -3

0

0

} i

1

0

HRU -4

0 -Zero Development

}- Moderate Development 1 -3 acre subdivision approximately 50% built

1 -Full Development, } -} acre subdivision approx. 75% built.

The variables which were deemed most sensitive to the simulated development were: a)

The density of the canopy cover (normal, summer and winter); b) Water holding capacity of the litter, c)

The density of foliage on the forest floor, d) The fraction of global radiation attenuated by the vegetative cover; e) The saturated conductivity of the soil, f) The saturated water content, g) and

Average soil depth.

Development results

in the partial clearing of the land.

Regardless of the developers intent to "save trees" the earth moving grading and paving operations inevitably remove significant numbers of trees.

This seems to be proportional to the density of development since more roads are needed to service a more finely divided area.

It has been estimated for the purposes of this study that for the moderate development described above there would be a 25% decrease in the forest canopy cover.

Full development would reduce the canopy cover even in a "good" development by 50% or more.

The litter on the forest floor, on the average, can suffer a significant decrease in its ability to hold water with increasing density of development.

Significant areas are cleared of all litter by the process of developing the land and building the structures.

Within this study it was estimated that a

25% decrease in the total water holding capacity of the litter would result from moderate development and a 50% decrease would result from full development.

It intense land use in other areas over the same time.

is noted that there could be significant recovery of litter in some areas of a developed HRU but this could be offset by an increase in

The density of foliage on the forest floor suffers the same as does the trees and litter.

The same factors of 25% and 50% decrease with moderate and full development respectively were used as estimates of developmental impacts.

As the ground cover and canopy cover are decreased there is a corresponding increase in the percentage of global radiation which reaches the ground.

It is estimated that in the undeveloped condition, only 20% of the total radiation reaches the ground.

With development, the percentage increases to approximately 40% partial development and to 60% with full development.

The hydrologic simulation model used in this study does not include an explicit input parameter describing infiltration.

Infiltration is logic studies

seen to be the

result of the complex interaction of water, ground litter and soil qualities.

To simulate changes to what is commonly referred to in urban hydroas urbanization (the ratio of impervious surfaces to pervious surfaces) the parameters of soil conductivity, saturated water content, and soil depth were used.

The saturated conductivity of the undisturbed soil which is the ultimate limitation upon infiltration was estimated at about 0.7

cm. /min. (1000 cm. /day).

It was further assumed that the average conductivity of each HRU would be affected by the development in that HRU.

In the full development situation one quarter to one third of the land surface could be sealed off from infiltration.

This could result in a net decrease of the average conductivity in that HRU of roughly the same amount (25 -33 %).

Consequently, with full development the average saturated conductivity for the affected HRU was assumed to be 0.5

cm. /min.

(approx. 700 cm. /day).

For moderate development 0.6 cm. /minute was assumed (approx.

860 cm. /day) as an average value for the affected HRU.

Using much the same logic as was used above for the saturated conductivity, parameters describing saturated water content and soil depth were also modified from the original values.

It was realized that development would cut off large areas from infiltration altogether.

It was therefore decided that the total moisture holding content of the soil would be reduced.

The total moisture present in a HRU is a function the average soil depth and saturated water content for developing H RU's were decreased by approximately of soil moisture and soil depth.

To simulate the desired changes,

15 and 30 percent respectively for partial development and full development.

Table 6 summarizes the values chosen for the variables used to reflect changes in the land use of the H RU's.

115

Variable

Name

COVMAX

ELTCAP

ELTCOV

ATNMAX

SANK

SATWC

DEPTH

Table 6

Values of Hydrologic Variables for Various Levels of Development

Hydrologic

Variables

Development

Zero Development

0

Level

Moderate

1/2

Full

1

Canopy Cover

Density (Z)

Water Holding

Capacity of Soil liter (cm)

Foliage Density on Forest Floor (Z)

Fraction of Radiation

Stopped by Canopy (Z)

Saturated Conductivity of Soil (cm /day)

Saturated Water Content

1.00'

1.00

1.00

.80

1000

.52

.75

.75

.75

.60

860

.45

.50

.50

.50

.40

700

.36

Soil Depth

Z of original value

100 80

60

RESULTS and CONCLUSIONS

CALIBRATION

As runoff data was also available for the four year period for which the other necessary hydrologic data was available (1967 -1970 inclusive) a calibration phase of the study was conducted.

Model base line parameters were adjusted to attempt to bring predicted flows in line with actual measurements

Despite this calibration attempt the model overestimated by almost 100% the total flow in 1967.

This is partially explained by the extreme hydrologic conditions found in that year.

Table 7 however, illustrates that the average annual flows in years 1968, 1969 and 1978 were simulated more accurately with the total flow error averaging +8 %.

Table 7

Calibration Results Average Annual Flow (cm.)

Year

Predicted Total

Measured Total

% Error

1967

19.5

10.0

+95.0

1968

11.1

13.0

- 14.6

1969

21.6

20.3

+6.4

1970

7.5

7.7

-2.6

Table 8 illustrates the peak flow month of the predicted annual hydrograph and the actual hydrograph.

The one and two month mismatch experienced in years 1968 and 1969 is viewed as a result of small differences in the individual totals for the two or three most significant months in each year.

These two or three months had total flows which were close and only a slight error in any of these months would cause the shift of peak flow as seen.

Table 8

Calibration

Results

Peak Month

Year

Predicted Max. Mo.*

Actual Max. Mo.*

1967

12

12

1968

12

2

1969

1

2

1970

*Numbers indicate months with January being 1 and so on.

9

9

Table 9 shows the average annual flow for each year simulated and for all six levels of hypothesized development.

Also, shown are the peak months for each year and level of development.

It

116

can be seen that there is a steady increase in average annual flow with development.

Model year

1967 shows a 4041 increase in flow with the full development situation while model year 1970 shows almost as much with just the first level of development and an almost 400% rise in flow with full development.

Years 1968 and 1969 also show marked increases with full development 100% and

50% respectively.

The rise in flow with development appears steady with the level of development.

There appears to be no significance to a regression analysis, however, since the levels of development were defined arbitrarily as simple fractions of existing conditions in a progression of 1 through 6.

The statement that runoff from a previously forested watershed seems to be positively affected over the long run by development of an urban character seems to be the only significant result of the study.

Table 9

Average Annual Flow and Peak Month vs. Development Level

Development

Level

3

4

1

2

5

6

A

67

19.5

21.5

22.6

24.7

25.9

26.9

10

Model

68

11.1

14.3

16.8

19.9

21.8

23.4

13.0

Year

69

21.6

24.3

26.8

29.7

31.7

33.3

20.3

70

7.46

11.6

15.9

21.1

24.4

27.2

7.7

Average Annual

Flow

(cm.)

5

6

A

1

2

3

4

12

12

12

12

12

12

12

12

12

12

1

1

1

2

2

2

2

2

2

2

1

9

9

9

9

9

9

9

Peak Month

CONCLUSIONS AND RECOMMENDATIONS poral.

of the

The long range effects of development upon streamflow seems to be more quantatitive than tem-

That is, although significant changes in total monthly and annual flows (measured by average monthly or average annual flows) were encountered as development increased, the large scale phasing of those flows seemed to be affected little by development.

This seems to in concert with the view urbanization sealing off large quantities of earth and resulting in enhanced streamflow.

It is to be remembered, of course, that even in highly urban settings this increased streamflow is often at net vailing expense of ground waters which could sustain the flow in the stream through drier periods.

So annual increases of flow are not at first expected when due consideration is given all the counter hydrologic factors affecting a forested watershed.

long

As contributing watershed.

(Qualitative aspects may however be effected.) effects effect a planning input, the results of this study seem to indicate that development will have no range adverse quantitative effects upon water systems depending upon the developing area as a tity and

Further, water harvesting seem to predominate, actually increasing the net productivity of the area.

The degree of the seems to vary significantly depending upon not only the degree of development but the quantemporal distribution of precipitation as well.

(Short range changes may, however, be adverse as the watersheds response to particular events may be changed.

This may result in greater flooding threats during extreme and short lived meteorlogical events.) Planned development could have less of an effect on watershed productivity if attention is paid to ground cover and natural forest floor litter as well as simply saving trees and providing open space.

The scope of this work did not permit sensitivity analysis.

The watershed development simulated affected many parameters together in a progressive manner.

Consequently, it is impossible, from, this depth area study to identify which parameter was most important canopy cover, ground cover, litter, soil and quality etc., may all respond to development and in turn may affect the hydrology of the independently of each other.

Further studies in these areas

(separately) could focus the results of this study further into developmental criteria for maintaining or modifying in a planned and coordinated way the productivity of watersheds.

117

REFERENCES CITED

Bodman, G. B., E. A. Colman.

1944.

"Moisture and Energy Conditions During Downward Entry of of Water Into Soils," Proc. Soil Sci. Soc. Am. 8, pp.

116 -122.

Black, P. E.

1967.

"Thornthwaite's Mean Annual Water Balance" Silviculture General Utility Library.

Program 60 -101, 1966.

N.Y. State University College of Forestry at Syracuse Univ., 20 pages.

Dagg, M., M.A.C. Pratt.

1962.

"Relation of Streamflow to Incident Rainfall ", E. African Agr. and

Forestry J., 27: p. 31 -35.

Eagleson, P. S.

1970.

Dynamic Hydrology, McGraw Hill,

Inc., New York, N.Y., 1970, pp. 201.

Galbraith, A. F.

1975.

'Method for Predicting Increases In Water Yield Related to Timber Harvesting and Site Conditions ", In "Watershed Management ", Proceedings of a Symposium, Irrigation and

Drainage Division of the American Society of Civil Engineers, Logan Utah, August 1975, pp.169 -184.

Herr.

(North Carolina)

Cited in Hollis 1975 below.

No other information available.

Hewlett,

J. D.,

i W. L.

Nutter.

Athens, Ga., 1969, p. 35.

1969,

An Outline of Forest Hydrology, Univ. of Georgia Press,

Hollis, G. E.

1975.

"The Effects of Urbanization on Floods of Different Recurrence Interval ".

Resources Research.

Vol. 11, No. 3, Pp. 431 -435, June 1975.

Water

James, L. D.

1965.

"Using a Digital Computer to Estimate the Effects of Urban Development on Flood

Peaks." Water Resources Research.

7

: pp. 223 -234.

Leopold, L. B. 1968.

"Hydrology for Urban Land Planning - A Guidebook on the Hydrologic Effects of Urban Land Use.' U.S. Geological Survey. Circ. SS4, 18 pages.

Lull, H. W., 6 W. E. Sopper.

1969.

"Hydrologic Effects from Urbanization of Forested Watersheds in the Northeast ", U.S.D.A. Forest Service Research Paper, NE -146, 1969,

31 pages.

McBride, J. R. 1976.

"Impact of Urbanization on Water Yield, Flood Peak Sediment Yield, and Water

Quality in the Berkeley Hills, Calif.

port, September 1976.

16 pages.

California Water Resources Center, Davis.

Completion Re-

Miller, D. H. 1966.

"Transport of Intercepted Snow from Trees During Snow Storms." U.S. Forest

Service Research Paper.

PSW -33, Berkeley, 30 pages.

Rauner, Iu. L.

49 -59,

1960, 63, 65.

Teploroi balans lesa, Izv. Akad, Nauk SSSR Ser. Geograf. No. 1, pp.

1960. Izmerenie Teplo -i rlagoobmena mezhdu lesom i atmosferoi pod vliianiem okruzhaiuschikh territoríi. Izv. Akad. Nauk SSR Ser. Geograf. No. 4, pp. 15 -28 1963.

0 gidrologicheskoi roll lesa, Izv. Akad, Nauk SSR Ser. Geograf. No. 4, pp. 40 -53, 1965.

Rogers, J.

1975.

Documentation of ECOWAT

:

Coniferous Forest Biome, Corvallis, Oregon.

A General Water Balance Model. Manuscript in Review.

Southeastern Forest Ex Station 1961.

baseflow, ".

"Watershed Mangement

:

Some ideas about storm runoff and

U.S. Forest Service, SEFES Annual Report 1961, pp. 61 -66.

Turner, R. R., Et.

Al. 1975.

"The Effect of Urban Land Use on Nutrient and Suspended- Solids Export from North Florida Watersheds." In:

"Mineral Cycling in Southeastern Ecosystems." 1975, pp. 868 -888.

U. S.

Forest Service 1977.

"The Beaver Creek Program - Advancing Forest and Range Resource

Management ", U. S. Dept. of Agriculture, p.

1.

U. S.

Forest Service 1967.

"Soil Survey - Beaver Creek Area, Arizona," U.S. Dept. of Agriculture.

Waring, R. H., Rogers, J. J., Swank, W. T.

1975.

:

'Water Relations and Hydrologic Cycles, Ecosystem Dynamics Section, Contribution to Woodlands Synthesis,

Vol.

I.

B.P., Reichle, editor.

Wooldridge, D. D. 1967.

"Water Transport in Soils and Streams." In: American Society of Mechanical

Eng. Proc. 1967, pp. 3 -20.

118

TRENDS IN ARIZONA WATER SERVICE ORGANIZATIONS:

A COMPARATIVE SUMMARY

Jacque L. Emel, Michael D. Bradley, and K. James DeCook

INTRODUCTION

The pattern of water occurrence profoundly influenced the economic and physical development of Arizona.

Water, a vital and elemental resource, was often in short supply, undependable and unevenly distributed in the river basins and valleys.

Early solutions to this problem took the form of physical development and were primarily matters of engineering and technology.

Dams and canals were surveyed and built, distribution systems were constructed, wells were drilled and pumped.

It soon became apparent that technological development also required effective management.

Since water was a public resource, public management in the provision of water supplies became the general rule.

In addition to the problems of water engineering, a full understanding of water development in Arizona also requires an appreciation of the experience gained in managing water supplies through public institutions.

Water policy and institutions can be examined at many levels, from the national perspective of coordination and regional development, to the state level of water rights and planning, to the local level of service delivery.

This study examines water service organizations at the local and district level.

It is here that actual deliveries are made to ultimate water users and here that water policy is directly implemented.

A study done for the Arizona Water Commission (DeCook et al., 1978) provided information on water service districts and organizations, statewide.

This information was used to compare the Arizona experience with that of California, which also has a decentralized district system of water service delivery.

ARIZONA WATER SERVICE ORGANIZATIONS

Five types of organizations distribute water in Arizona.

Four are public districts; the fifth type encompasses private companies.

Table 1 shows the number of each type of organization and cites the enabling legislation.

Irrigation districts are the most numerous public water service organizations in

Arizona.

Approximately one -half of the irrigated acreage within the state is organized into irrigation districts.

The range in district size is from 550 to over 150,000 acres.

Districts are municipal corporations with broad powers: they can purchase or acquire water rights, own or sell property and real estate, construct facilities, generate electricity, appropriate water for irrigation and power generation, tax and charge for service, appropriate money, and exercise eminent domain.

Irrigation or water conservation districts empowered to conduct drainage activities have the word

"drainage" in their titles.

The authors are, respectively, Graduate Assistant in Research, and Associate

Professor, Department of Hydrology and Water Resources, and Associate Hydrologist,

Water Resources Research Center, University of Arizona, Tucson.

119

Type of District

AGRICULTURAL WATER

SERVICE ORGANIZATIONS

Irrigation District or Water Conservation

District

Irrigation Water

Delivery Districts

Agricultural Improvement

District

Multi- County Water

Conservation District

Agricultural Water

Companies

Water Associations

Irrigation Companies

Canal Companies

TABLE I

ARIZONA WATER SERVICE ORGANIZATIONS

General Law or Code

(Arizona Revised Statutes) Code Section

Water

(Title 45)

Water

(Title 45)

Water

(Title 45)

Water

(Title 45)

Corporations and

Associations

(Title 10)

1501 to 1866

1201 to 1396

1900 to 1955

901 to 1041

2601 to

2634

054 to 509

MUNICIPAL AND INDUSTRIAL

WATER SERVICE ORGANIZATIONS

Municipal Water Depts.

and Utilities

Private Companies

Cities and

Towns

(Title 9)

Corporations and

Associations

(Title 10)

511 to 542

054 to

509

Number of

Districts (1978)

40

23

1

1

>40

37

(>2,500 persons served)

30

( >2,500 persons served)

The purposes of an irrigation district are to provide the landowners within the boundaries with water, electricity, and other public conveniences.

A 1971 decision by the Arizona Supreme Court states that a petition for the organization of a district must indicate the purpose of the organizers is to irrigate arid lands and thereby improve agricultural and farming lands.

City of Scottsdale v. McDowell Mountain

Irrigation and Drainage District (1971), 107 Ariz. 117, 483 P. 2d 532.

Irrigation water delivery districts distribute water for lawns and pasture in urbanized areas.

The total acreage organized into water delivery districts is 3,300 and the range in district size is from 10 to 1,300 acres.

As lands within irrigation districts and water company service areas are developed for residences, irrigation water delivery districts are often organized to take over distribution and financial responsibilities.

The majority are in rapidly urbanizing areas, particularly the

Phoenix and Yuma regions.

Irrigation water delivery districts are corporate bodies but not municipal corporations, with perpetual succession and the power to contract; sue; acquire, hold, and sell real and personal property; adopt a seal; incur debts; and contract with the federal government for irrigation services.

Irrigation water delivery districts have the powers of eminent domain and taxation.

Special districts are of two types: the agricultural improvement district and the multi- county water conservation district.

The Salt River Project is the only agricultural improvement district and the Central Arizona Water Conservation District is the only multi- county conservation district in Arizona.

An agricultural improvement district stores and delivers water, provides drainage, improves facilities, sells surplus water, and finances construction.

It may not acquire water rights.

A multi county water conservation district is a tax -levying public improvement district and municipal corporation that taxes to pay its administrative costs and to repay project costs to the federal government.

120

Both types of special district have the power of eminent domain.

The agricultural improvement district is financed by issuing bonds with the approval of the majority of electors, while the multi- county water conservation district levies an ad valorem tax against all taxable property in the district.

Water companies distribute water to both agricultural and municipal users.

Agricultural water companies have many names, including water associations, mutuals, canal companies, and irrigation companies.

Some 40 agricultural water companies were identified; others exist, but the exact number cannot be determined as there is no categorical registration of such companies.

Generally, the companies distribute water to areas ranging from 500 to 5,000 acres, averaging approximately 1,500 acres.

Agricultural water companies are private corporations and associations with perpetual succession.

They do not have the power of eminent domain and are financed by stocks, bonds, and water sales revenues.

About 65 municipal and industrial organizations supply water to communities of more than 2,500 persons.

Municipal water departments or municipal utility districts that supply water are subject to the general provisions of the legislation governing municipal corporations and most are regulated as utilities by the state public utilities commission.

Municipal water service organizations have the power of eminent domain; industrial water service organizations do not.

Revenues and assessments vary according to the license issued by State Public Utilities Commissions.

CALIFORNIA WATER SERVICE ORGANIZATIONS

More than 1,000 public and private agencies share responsibility for management and distribution of California's water.

Nine hundred are special districts, most of which are recently organized.

Table 2 shows the 20 types of California water districts.

In general, these districts are authorized to levy taxes, issue both general obligation and revenue bonds and set rates.

Many of the newer types of districts function as general municipal governments and provide the basic services characteristic of such.

In 1971, several of the Community Services Districts provided fire protection, waste disposal, recreation and park programs, lighting, library services, police, and road and street construction (Goodall, Sullivan and DeYoung, 1978).

TRENDS

Several trends, apparent in the formation and functions of water service institutional models in California, can be compared with the Arizona experience.

In California, fewer districts formed in recent years are related to agricultural water use

(even though 85% of water use in the state is irrigation).

Similarly, fewer irrigation districts are being formed in Arizona.

Between 1920 and 1950, 21 of the present 40

Arizona irrigation districts were formed.

Since 1950, excluding those districts formed for the purpose of obtaining Central Arizona Project water, only 11 relatively small irrigation districts have been organized.

In contrast, Arizona irrigation water delivery districts, whose primary function is to distribute water for lawn irrigation, constitute an increasing proportion of the districts being formed.

three existed whereas 20 have been organized since then.

Prior to 1950 only users.

Currently, agricultural districts are delivering water to domestic and industrial

In California, 75 to 100 percent of the districts were predominantly concerned with irrigation water distribution from 1880 to 1930.

From 1930 to 1970, 50 to 70 percent of the deliveries were for both irrigation and urban use.

At present, over 69 percent of the distribution is for urban users.

The pattern is also found among the

Arizona districts; approximately one -fourth now deliver domestic and lawn or suburban pasture irrigation water.

Another type of trend is evident in terms of voter qualification as related to district formation and operation.

In California, the tendency is toward creation of districts under those enabling acts which require the participation of landowners in formation and which stress an acreage- weighted voting system.

In the earlier part of the century, most California districts were tied to agrarian purpose and voting was generally open to registered resident voters.

Presently, three methods are used to fill principal positions and to decide bond issues: one person- one -vote elections; property -weighted voting based on number of acres or value of acreage; and appointment by the county board of supervisors.

In Arizona, property

121

TABLE 2

CALIFORNIA WATER DISTRICTS BY TYPE OF ENABLING LEGISLATION AND

NUMBER OF DISTRICTS, 1970 -71 AND 1974 -75

Type of District

General Law or Code Code Section No. of Districts

1970 -71 1974 -75

1. Community Servicesa

2. Flood control & water conservation

3. Harbors & ports

4. Municipal Improvement

5. Maintenance

6. Reclamation

7. Recreation & parksc

8. County Service Area

9. Municipal Utility

10. Public Utilityd

11. California Water

12. County Water

13. Metropolitan

14. Municipal Water

15. Water Agency or

Authority

16. Water Conservation

17. Water Replenishment

18. Water Storage

19. County Waterworks

20. Irrigation

Government

General laws

Harbors & Navig.

General laws

Sts. & hwys.

Water

Public resources

Government

Public utilities

Public utilities

Water

Water

General laws

Water

General laws

Water

Water

Water

Water

Water

61000 et seq b

6200 et seq b

5820 et seq

50000 et seq

5780 et seq

25210.1 et seq

11501 et seq

15501 et seq

34000 et seq

30000 et seq e

71000 et seq b

74000 et seq

60000 et seq

39000 et seq

55000 et seq

20500 et seq

103

24

8

1

8

90

107

6

28

3

52

160

192

1

50

4

33

8

7

1

116

4

34

10

27

11

1

8

88

103

8

1

6

43

52

3

162

189

1

47 a.

b.

c.

d.

e.

includes one special -act district special -act districts only includes four special -act districts includes two special -act districts includes Metropolitan Water District only.

Source: M.

R. Goodall, J.

D. Sullivan, and T.

DeYoung, California

Water: A New Political Economy, (Montclair, N.J.: Allanheld,

Osmun /Universe Books, 1978).

ownership is a requirement for voter eligibility in nearly all public districts: the exception is the multi- county water conservation district.

Two- thirds of the irrigation districts in Arizona have the one -person- one -vote participatory system, the remainder employing the property -weighted (usually one vote per acre) method.

The enabling legislation for irrigation water delivery districts stipulates that each landowner is entitled to 1/5 vote per acre.

The agricultural improvement district employs the acreage voting system while the taxing is ad valorem.

All registered voters residing within district boundaries may vote in the multi- county water conservation district elections.

A California study (Goodall, Sullivan and DeYoung, 1978) shows a greater turnout

(45 to 60 percent) for elections in those districts employing the one -person- one -vote system.

Under this system, politics are generally more competitive than under the property- weighted system because there are fewer appointments or uncontested seats.

In districts where property- weighting is used, the average estimated turnout was only

5 percent and elections were fewer.

122

One -person- one -vote districts tend to gain revenue as a function of expenditure, incur relatively less debt and, in general, exhibit relatively modest fiscal policy.

The districts with weighted voting systems commonly incur greater debt and show erratic financial behavior.

SUMMARY AND CONCLUSIONS

Three principal trends have been identified relative to the organization and operation of water districts in both Arizona and California: (1) Proportionately fewer districts are being organized for the primary purpose of serving agricultural irrigation; (2) existing districts are increasingly delivering water to urban domestic and industrial users; and (3) provisions for district voting show a preponderance of property ownership- weighted voting systems.

Since water districts are providing water services not only to irrigators but also to a growing number of municipal and commercial users, their activities are more than physical and economic; districts are inherently legal and political entities, especially when they determine how and when water is delivered and who pays the cost of distribution.

Water service organizations deserve careful and complete study by those interested in revising state water law and developing a rational and effective system of water resources management.

As illustrated by comparisons with California, local water organizations can expand and develop in various ways, sometimes accepting functions beyond the original intention of their institutional charter.

This suggests that a careful examination of the dynamics of these extremely important limited -purpose agencies is in order.

In a time of changing institutional purpose and political responsibility, Arizonans need to appreciate more fully the role of local water service agencies in order to assess the equity, efficiency, and stability of their operations and functions.

Rational water management in an arid state requires a thorough understanding of the institutional arrangements for controlling and allocating water, a limited natural resource.

REFERENCES CITED

DeCook, K. James; Emel, Jacque L.; Mack, Stephen F.; and Bradley, Michael D.

Water Service Organizations in Arizona.

Water Resources Research Center,

Univ. of Arizona, Tucson, 1978.

Goodall, Merrill R.; Sullivan, John D.; and DeYoung, Timothy.

California Water:

A New Political Economy.

Allanheld, Osmun and Co. Publishers, Inc.,

Montclair, N. J., 1978.

123

AN EXAMINATION OF THE BUCKHORN -MESA WATERSHED ENVIRONMENTAL

IMPACT STATEMENT (U.S.D.A., S.C.S., 1978): A LOOK

AT STATE -OF- THE -ART REPORTS

Dale A. Altshul

ABSTRACT

The National Environmental Policy Act of 1969 was written with the intent of fostering a spirit of harmony in the day to day operations of Federal agencies with the environmental concerns voiced by the general populace.

To examine how Federal agencies have assimilated E.I.S. procedures and guidelines a typical report was reviewed.

In general, compliance with environmental law and procedural guidelines was found to be adequate.

In some ways, particularly in assessment of Cultural Resource Impact, the statement was exceptional in its evaluation.

However, the sections of the report detailing the benefits and costs of the alternatives was not up to the standards expected in an E.I.S.

Because the benefits and costs were not calculated in consistent units and the no action alternative was not adequately examined, the entire alternatives section is called to question.

By re- evaluating the data provided in the

E.I.S. in consistent units, it was found that the alternative selected had neither the highest benefit/ cost ratio nor the lowest environmental impact.

It is concluded that alternatives should be as fully evaluated as the project itself in order to integrate environmental considerations into the overall planning process.

INTRODUCTION

The purpose of this study is to review the requirements and guidelines of the National Environmental Policy Act of 1969 (hence -forth referred to as NEPA or the Act) and the Council on Environmental

Quality (CEQ), and to examine how Federal agencies have assimilated these mandated procedures into

Environmental Impact Statements (EIS).

For the purpose of this examination, an EIS filed in 1978 was selected at random; by chance it turned out to be the Final EIS for the Buckhorn -Mesa Watershed

(U.S.D.A., SCS, 1978) in the State of Arizona.

NEPA was written in very general terms, being specifically written by Congress as a mandate to

Federal agencies on environmental policy.

Section 102(C) of the Act detailed what the legislators determined appropriate for the newly required Environmental Impact Statement:

.detailed statements on all major Federal actions significantly affecting the quality of human environment include, (i) the environmental impact of the proposed action, (ii) any adverse environmental effects which cannot be avoided should the proposal be implemented, (iii) alternatives to the proposed action,

(iv) the relationship between local short -term uses of man's environment and the maintenance and enhancement of long -term productivity, and (v) any irreversible and irretrievable commitments of resources which would be involved in the proposed action should it be implemented."

To the agencies that prepare impact statements, Section 102(C) of the Act has been interpreted as delineating a format.

Indeed in the EIS reviewed, this five -fold division was followed.

In terms of CEQ guidelines, this division of reports is not a necessity, but it does supply a means of standardization.

The real purpose in reviewing any EIS is not merely to check that standard format was used, but rather that the CEQ guidelines (40 CFR part 15002 (C)) are followed:

"Section 101 of NEPA sets forth the substantive requirements of the Act, the policy to be implemented by the 'action- forcing' procedures of Section 102.

These procedures must be tied to their intended purposes, otherwise they are indeed useless paperwork and wasted time...

...(i) In securing more accurate, professional documents the lead agencies are responsible for the professional integrity of reports, and care should be taken to keep any possible bias from data prepared by applicants out of

Department of Hydrology and Water Resources, University of Arizona.

125

the environmental analysis.

A list of people who helped prepare documents, and their professional qualification should be included in the

EIS.

(ii) Recording in the decision how the EIS was used.

To this end agencies must also produce a concise public record, indicating how the EIS was used in arriving at the decision.

The record must state what the final decision was; whether the environmentally preferable alternative was selected; and if not, what consideration of national policy led to-another choice.

(iii) Insure follow -up of agency decisions..."

The intent of this study is to check on the degree of compliance with these CEQ guidelines through analysis of a randomly selected EIS.

We want to take the information as presented in the case study EIS to derive:

1) what is the project; where it is located and why it is needed; 2) what are the present physical conditions in the project area; 3) what are the economics of the selected project and its alternatives; 4) what are the environmental impacts of the selected course of action and its alternatives.

The objective of using this open -ended inquisitive approach in review of an EIS is simply to find:

1) if the information as presented in the EIS would lead a trained professional reviewing the statement to the same logical conclusion as that of the lead agency; 2) if the EIS complied with NEPA, CEQ guidelines, and other legal formats which regulate Federal manuscripts; 3) whether the alternative selected by the lead agency had the lowest environmental impact, and if not, was it properly justified in the report.

WHAT IS THE PROJECT (LOCATION AND NEED)

The plan to be implemented in the Buckhorn -Mesa watershed is a watershed protection and flood prevention project located in Maricopa and Pinal Counties, Arizona.

The project is to be carried out by the sponsoring local organizations with Federal assistance under provisions of Public Law 83 -566, 83rd

Congress, 68 Stat. 666, as amended.

The purposes of the project are to reduce flooding and associated flood damages occurring within the flood prone area, to reduce sediment and erosion throughout the watershed, to increase efficiency of irrigation water use, and to allow flood protection to lands now undergoing rapid urbanization.

The actual project consists of five floodwater retarding dams which are designed to trap 823 acre feet of sediment over a 100 year design life of the structures.

urban land, and other developments will be reduced.

Downstream sediment damage to cropland,

The land treatment measures as detailed in the EIS have essentially been installed.

PRESENT PHYSICAL CONDITIONS

The watershed in question covers 69,172 acres in eastern Maricopa and northwestern Pinal Counties,

Arizona.

Included within the watershed are a portion of Apache Junction and the northeastern quadrant of Mesa.

Both of these towns are within the greater Phoenix metropolitan area.

lation in the watershed is classified as urban and 15% as rural.

About 85% of the popu-

Of the total watershed area, nearly

60% is flood prone.

The area that would be innundated by a 100 -year flood is 25% of the watershed.

This flood -prone area is undergoing a tremendous rate of population and development growth.

About 27% of the flood -prone area is irrigated cropland.

Floods are part of the natural scene in the watershed.

Since 1910, an estimated 40 floods have occurred.

The magnitude of damage expected from a storm to occur on the average of once in 100 years

(one percent frequency of event) would seriously affect the local economy for several years.

The flood resulting from a storm of this magnitude would innundate approximately 28,300 acres; of which approximately 60% is in urban or irrigated agricultural uses.

Projections indicate that by the year 2000, even without the flood protection afforded by the proposed project, the land use will change to primarily urban use; crops being reduced to small acreages on "ranchettes."

The economy of the area is based heavily on retirement- recreational type development, and many people are engaged in employment in the service trades.

Residents employed outside the watershed commute to the Mesa -Tempe- Phoenix area, and in the mines in the Superior area.

The native vegetation is Sonoran Desert type.

Wildlife species inhabiting the watershed include a wide variety of mammals, birds, amphibions, and reptiles.

The watershed includes three species on the endangered list, one species on the threatened list, four are peripheral, and two are of undetermined status (including Gila Monster).

Archaeological and historical sites in the watershed were examined and evaluated by Arizona State

University.

In the opinion of the investigators only site AZ U:10:51 (ASU) warranted additional investigation.

No sites in the area were listed in the Federal Register of Historic Places.

The Arizona

State Historic Preservation Officer concurs that none of the cultural resources located and identified meets criteria for inclusion in the National Register.

126

CONSIDERATION OF THE ALTERNATIVES (ECONOMICS AND IMPACTS)

From the data presented in the EIS, there are six listed alternatives that were considered by the lead agency.

Through close review of the material, it was found that in reality there were seven alternatives; as it was found that the actual project to be implemented was not listed in the Alternatives section of the EIS.

In specific the Alternatives as listed are:

No.

1

- No Project

No. 2 - Structural Protection of Existing Urban Development Only, with Further Urban Buildup

No.

Prevented.

3 - Structural Proection for All Flood Prone Areas, but With Further Urban Buildup Prevented on

Prime Irrigated Cropland.

No. 4 - Alternative to The Spook Hill Floodwater Retarding Structure.

No.

5 - Accelerated Land Treatment and Floodways.

No. 6 - Accelerated Land Treatment and Flood Water Retarding Structures.

Recall from page 2 of this paper that the project selected by the lead agency was described as five floodwater retarding structures, floodways, and the already implemented land treatment measures.

Obviously this constitutes a seventh alternative.

It is indeed curious that this option was not listed in the Alternatives section of the EIS.

Upon closer examination it was found that discussion of this seventh alternative was contained within discussion of Alternative No.

1

- the No Project alternative:

"Alternative No.

1

- No Project

This alternative includes the ongoing land treatment program.

Because technology land use, and land ownership change, the land treatment program is a continuous updating process.

The Soil Conservation Service through the Natural

Resource Conservation Districts, will continue to provide technical assistance for installation of this program.

Land use projections for the No Project" alternative are the same as for other alternatives.

As desert land and cropland are taken for urban development, the following are among the impacts that are expected: loss of productive cropland; loss in scenic quality; reduced air and water quality; more energy use; loss in wildlife habitat; and more traffic congestion.

Officials of the community recognize that a flood problem exists.

With or without this project, a flood plain management program will be developed.

The flood plain management program will be developed.

The flood plain management program will encompass proper land -use planning, protective measures for existing developments, and land use regulation.

Specific flood hazard areas will be identified through detailed flood plain information studies.

Common recognition of these hazards will be the key to the action program for flood plain management that will follow.

The first item in the action program is adjustments in existing structures and occupancy in the identified flood hazard areas.

Because of characteristics of the flood plain, the studies may show that most present development is in a flood hazard area.

From studies of aerial photographs, it is estimated that there are 19,940 existing homes or commercial establishments that would need to be floodproofed.

A preliminary cost estimate to floodproof these establishments is $64,000,000.

Flood plain land use will be controlled through the following zoning ordinances; subdivision regulations, including utility extensions; building codes; acquisition and evacuation; building financing and related tax assessment adjustments; flood hazard warning signs and notices; and flood insurance.

The regulations will have two purposes.

One is to maintain regular floodways that will have sufficient cross -section area for passing a specified flood flow through the developed areas without damage.

The second is to regulate development of the floodplain to prevent damage to future development.

Cost of the structural measures would be considerably more than the planned project measures.

They would consist, to a great extent, of floodproofing by diking existing development; maintaining floodways for internal drainage of present and projected developments; and floodproofing of future development through either dikes or landfills, with subsequent increases of the flood problem in unfilled areas.

127

With the planned project measures, only the floodways will be required; and these would be substantially reduced in size.

Under the "No Project" alternative, a total of 823 acre -feet of sediment would move downstream causing damage to roads, bridges, irrigation facilities, urban developments, crops, and other properties over the next 100 years.

Periodic floodwater and erosion damage, consisting of scour damage to cropland and other unprotected land, would occur.

On an average annual basis, the project will provide benefits of $2,808,790 while costs will be $1,122,800.

The net monetary benefits to be foregone by not implementing the project is estimated to be $1,685,990 annually"

(USDA, SCS, 1978).

After the fifth paragraph of the preceding discussion, the lead agency ceases to discuss the

No Project" alternative and in fact begins discussion of the measures, impacts, and economic analysis of the selected project.

This confusing procedure will be discussed later in this analysis.

The confusion of the Alternatives section is compounded by the manner in which the economic data for each specific alternative is presented.

Table 2 -A shows the actual economic data for each alternative as presented in the EIS.

From the number of blank spaces in the table it is apparent that a great deal of pertinent data was omitted in the environmental report.

Through extrapolation of the data from other parts of the report, I was able to obtain a B/C ratio for Alternative No. 2; which is higher than the ratio for the selected program.

The projected environmental impacts of each alternative is contained within Tables 3 -A and B.

Note should be taken that the projected effect of Alternative 2 is the same (or less) than the projected environmental impact of the selected course of action.

From the data presented in the EIS, what conclusions are we able to draw.

need for some form of flood control in the defined watershed.

First, we find obvious

Present conditions, as resulting from floods under the 100 year event, have caused considerable loss of income, productivity, and environmental quality.

Rapid urbanization of flood prone sectors of the watershed will only increase the predicted harmful consequences of flood events.

Second, the watershed in question has been intensively developed over the past 50 years, both in urban and agricultural sectors.

This development has already disrupted the natural desert ecosystem in the lower reaches of the watershed.

Because of site disruptions accompanied by the secondary effects of urbanization, we can conclude that the potential adverse environmental impacts of any of the presented reasonable alternatives would not be excessively "disruptive."

However, that is not to say that some alternatives have less effective disruption than others.

To answer the most important question posed for analysis: are we logically drawn to the same conclusions as the lead agency, we will have to focus specific attention on the Alternative section of the EIS.

Recall that Section 102(0) of NEPA specifically stated:

"...Study, develop, and describe appropriate alternatives to recommend courses of action in any proposal which involved unresolved conflicts concerning alternative uses of available resources."

The intent of the authors of the Act was to generate complete review of all appropriate alternative courses of action, and to move this review as early as possible in the decision- making process.

The legal necessity for open discussion of alternatives was reinforced by the Court in Natural Resources

Defense Council Inc. vs. Morton (43, USCA, 133N; 1970) in which it was stated:

"Alternatives must be explored and discussed thoroughly in order to compare with the intent and requirements of Section 432(2)(c) of NEPA."

The alternative section of the Buckhorn -Mesa EIS was improperly presented.

First, examine the

No Project Alternative.

The description of the No Project section immediately relates that No Project includes:

1) continued land treatment measures; 2) land use projections the same as for the other alternatives," 3) the expected impacts of (a) loss of cropland, (b) loss in scenic quality, (c) reduced air and water quality; (d) more energy use, (e) loss in wildlife habitat; (f) more traffic congestion.

All this is adequately presented data from which reasonable conclusions could be drawn about the No

Project Alternative.

At this point one would expect a statement of benefits and costs for this alternative and then simply move on to discussion of other alternatives.

The lead agency did not organize the section in this manner.

Through close analysis of the economic data in the No Project Alternative, it was found that what has been described is in reality the course of action desired by the lead agency; described as accelerated land treatment, floodwater retarding structures, and floodways.

In other words, the lead agency chose to place the project opted for within the No Project alternative.

In this manner, they failed to properly examine the chosen course of action within the framework of the NEPA requirements for discussion of alternatives.

128

Analysis of the remaining Alternatives is complicated by the poorly presented benefit /cost analysis.

I would like to state my concern that agencies which prepare benefit /cost analysis of proposed actions should attach analysis, or summaries thereof, to the environmental impact statement; and should clearly indicate the extent to which environmental costs have not been reflected in such analysis.

Basic principles of engineering economics emphasise the need for evaluation of appropriate alternatives in sound decision making.

Summarization of six basic principles would read as follows:

1.

2.

3.

4.

5.

6.

Decisions are among alternatives; it is desirable that alternatives be clearly defined and that the merits of all appropriate alternatives be evaluated.

Decisions should be based on the expected consequences of the various alternatives.

Before establishing procedures for project formulation and project evaluation, it is essential to decide which viewpoint is to be adapted.

In comparing alternatives, it is desirable to make consequences commensurable with one another insofar as practicable.

That is, consequences should be expressed in numbers.

In economic decisions, money units are the only units that meet the foregoing specifications.

Only the differences among alternatives are relevant in their comparison.

Insofar as practicable, separable decisions should be made separately.

(Grant and Ireson, 1970).

From the data presented earlier, recall that the course of action opted for by the lead agency included intensive land treatment measures, structural measures (flood control dams), plus the intensive use of floodways.

All structures are designed for control of the 100 year event, and design life is set at 100 years.

This data, as stated in the No Project Alternative and as listed in the EIS appendix as the "Comparison of Benefits and Costs for Structural Measures," has a calculated B/C ratio of 2.5.

Alternative No. 1, the No Project alternative does not have an assigned B/C ratio presented in the report.

Alternative No. 2; described as structural protection for existing urban developments only, with further urban development prevented; has non -consistent data, but through recalculation we derive a

B/C ratio of 2.86.

Alternative No. 3; described as structural protection for all flood -prone areas, but with further urban buildup prevented on prime irrigated cropland; has no presented estimation of benefits, and therefore, a benefit /cost ratio is impossible to derive for this alternative.

Alternative No.

4; the alternative to Spook Hill flood water retarding structure; was dropped.

Alternative No. 5; described as accelerated land treatment and floodways; contains no calculations of potential benefits, therefore it is impossible to derive a benefit /cost ratio for this alternative.

Alternative No. 6; described as accelerated land treatment and floodwater retarding structures; again contains no calculated estimate of benefits of the alternative, therefore there is no benefit /cost ratio for Alternative No. 6.

As stated in both the principles of engineering economics and NEPA, the reliable evaluation of appropriate alternatives is not only a requirement but is fundamental to the decision making process.

Emphasis must be made that the purpose of NEPA is to bring consideration of environmental impact and alternatives into the decision making process as early as possible.

One is forced to ask; with data presented as shown in the case study EIS, how is any reasonable discussion of alternative actions possible?

The close examination of the Alternatives section of the Buckhorn -Mesa EIS has shown that:

1.

2.

3.

4.

5.

Incomplete assessment of benefits and costs were done for perfectly reasonable alternatives, without explanation for the omission of such data; the "No Project" alternative was improperly assessed; the alternative selected by the lead agency was not listed in the Alternatives section of the EIS, but was is reality buried in the No Project Alternative; methods used for evaluation of alternatives were not elaborated upon, or even given some sort of explanation in any section of the EIS; basic concepts of engineering economics were ignored in preparation of the Alternatives section by: a.

b.

notclearly defining the merits of all appropriate alternatives, decisions were not based on the expected consequences of various alternatives, and finally, c.

the units used were non -consistent, in fact, some alternatives did not even include a monetary assessment of one key factor.

Even with these problems, it became readily apparent that the alternative opted for by the lead agency was not the alternative with either the highest benefit /cost ratio nor with the lowest environmental effects.

Alternative No. 2 had a higher benefit /cost ratio (2.86) and a less significant impact on the immediate environment.

Federal laws and guidelines do not specifically require that the alternative with the highest benefit /cost ratio and the lowest environmental impact be selected, but they do require that adequate description be provided to show why an alternative with a lower B/C ratio and /or a higher environmental impact was selected.

Such a description was never presented in the Buckhorn -Mesa EIS.

Unless such data can be produced, it is only a logical conclusion that the lead agency was not using the EIS guidelines in preparation of the impact statement.

Full description of all alternatives, including the No Project alternative and the B/C and environmental impacts are necessary for a proper review of an EIS and should be available if the project

129

planning is to follow the spirit of NEPA, i.e., "...to build into their decision making at the earliest possible point an appropriate and careful consideration of the environmental aspects of the proposed action."

REFERENCES

The National Environmental Policy Act of 1969, 42 U.S.C. & 4321 et seq, 83 Stat. 852, P.L. 91 -190

Council on Environmental Quality Preparation of Environmental Impact Statements:

40CFR Part 1500, 38 -FR 20550 (Aug. 1, 1973).

Guidelines,

Buckhorn -Mesa Final EIS; USDA; SCS; 1978.

Natural Resources Defense'Council Inc. vs. Morton, 43, USCA, 133N; 1970.

Grant, E.

L. and Ireson, G. W.; Principles of Engineering Economy; Ronald Press, 1970.

Alternative

No Project

1

2

3

Spook Hill Dam

4

5

6

Land Treatment

Measures

X

X

X

Alternative dropped

X

X

7 X

Table 1 -A

Structures (Dams)

Urban Rural

X

X X

Floodways

Zoning Measures

X

X

X

X X

X

X

X

X

X

Table 2 -A Data as Presented in the EIS

Alternative Costs (total) Benefits (total)

Costs (annual) Benefits (annual) B/C ratio

5

6

1

2

3

4

$ 72,000,000

$124,000,000 no economic data given

75,000,000

$

26,400,006 alternative dismissed

$1,122,800

$2,808,790

$2,058,000

Table 2 -B Derived Economic Data

Alternative

Costs (total) Benefits (total) Costs (annual) Benefits (annual) B/C ratio

5

6

7

1

2

3

4 no data given

$ 72,000,000

$124,000,000 alternative dismissed

$ 75,000,000'

$ 26 400,000

$11 ,28 000 no data

$2.058(108) no data no data no data

$2.81(108) no data

$7.2(103]

$1.24(106)

$7.5 103

$2.610

$1.12(((106: no data

$2.058(106) no data no data no data

$2.81(106)

2.86

2.50

130

Alternative No.

1

(No Project)

2

Table 3 -A

Projected Environmental Effects

Beneficial Effects

Adverse Effects

-loss of cropland

-loss of scenic quality

-reduced air quality

-reduced water quality

-more energy use

-loss in wildlife habitat

-more traffic congestion

-flooding reduced

-restricted urbanization

-=26000 more acres of agricultural open spaces

-productive cropland preserved

-scenic quality unaffected

-water and air quality improved

-wildlife habitat preserved

-less traffic congestion

-loss of 1316 acres of desert vegetation

-visual impact of structures

-construction impacts

-disturbance of one significant archaeological site

-loss of tax revenue

3

-flood effects reduced

-restricted urbanization

-cropland stays in production

-scenic quality unaffected

-water and air quality improved

-wildlife habitat preserved

-less traffic congestion

-loss of desert vegetation

-visual impact of structures

-construction impacts

-disturbance of archaeological site

-loss of tax revenue

4

5

6

-alternative dismissed

-downstream developments protected from sediment

-flood effects reduced

-less vegetation disturbance

-less visual impact

-flood effects reduced

-sediment would be dumped into RWCD floodways

-more irrigated cropland development would be required

-greater visual impact

-greater impact on negative vegetation

-relocation of an additional four homes

Beneficial Effects

Table 3 -B Projected Environmental Effects of the Selected Project

Adverse Effects

-protection of urban land

-protection of cropland

-stabilization of agricultural industry

-reduction of sediment load

-increased recharge behind retarding structures

-provide protection for Superstition Highway and CAP aqueduct

-reduction of erosion

-loss of desert vegetation

-less recharge downstream from retarding structures

-visual impact of structures

-construction impacts

-relocation of an additional four families

-disturbance of one significant archaeological site

-loss of tax revenues

-accelerated movement of bedload

131

LAND USE PLANNING FOR THE SAN TIBURCIO WATERSHED

Roberto Armijo and Robert Bulfin

ABSTRACT

Land use planning, within the context of socio- economic development, is characterized by many conflicting objectives.

This paper defines ojectives for the San Tiburcio watershed in northern Mexico.

A mixed multiobjective programming model is developed.

The model serves as an aid to a group of decision makers in choosing a

"satisficing" feasible set of non -mutually exclusive land use alternatives.

The paper concludes with a discussion of possible solution techniques.

INTRODUCTION

Land use planning has been of interest to both practitioners and researchers for many years.

Recently, the use of quantitative models has become widespread.

Utility theory (Edwards, 1977), simulation (Baur and Wegener, 1975), and optimization methods (Goicoechea, 1977; Nijkamp and Rietveld, 1978, and Bamni and Bamni, 1979) have been used to aid land use planners.

Multiobjective models seem particularly appealing since many of the goals of land use planning are incommensurate and conflicting.

This paper discusses a multiobjective mixed -integer programming (MMIP) model that will provide insight for land use planning in the San Tiburcio watershed.

This region is located in the state of Zacatecas, Mexico and is a rural area.

Currently, the area is not well developed and lacks much of the infrastructure normally assumed to exist in land use planning studies.

Therefore, land use planning must be considered within the framework of a total regional development program.

While this paper does not explictly discuss the development program for the region, the land use alternatives are compatible with the development goals and in some cases projects considered have impact beyond the specific land use.

The paper contains a description of the San Tiburcio watershed and a discussion of its land use problems, which provides a decision scenario for the study.

This is followed by a group assessment procedure, which gives the value structure for the MMIP land use planning model.

Preliminary solution results and future plans for the model are discussed.

DECISION SCENARIO

CHARACTERISTICS OF THE SAN TIBURCIO WATERSHED

The characteristics of the region have been well documented (Medina, 1973).

This section will summarize those characteristics.

The San Tiburcio Watershed lies in the northeast corner of the state of Zacatecas in the north of` Mexico, comprising approximately 1500 square kilometers.

It is located in the Chihuahuan Desert with an elevation ranging from 1700 meters above mean sea level in the bottomlands to 2500 meters in the mountains.

On the basis of a five -year record, the mean annual temperature is about 12 °C; mean annual evaporation is 2100 millimeters and the mean annual precipitation occurs in a few individual rainfall events during the summer season.

In addition to this, the area suffers from a short

The authors are respectively on the faculty at the Department() de Recursos Naturales,

Universidad Autonoma Agraria Antonio Narro, Saltillo, Mexico and the Department of

Systems and Industrial Engineering, University of Arizona, Tucson.

This work was partially supported by CONACTY of MEXICO.

133

drought period in the middle of the summer, known as the August or Intraestival drought.

The distinctive feature of this period is a striking decrease in the number of rainy days, which accounts for a drastic diminution in the amount of precipitation.

The maximum yearly temperatures occur during this time, which combined with the low moisture conditions, makes this part of the year a critical one for all agricultural activities of the peasants.

Erratic rainfall, short growing season, and shallow and saline soils, have resulted in highly uncertain conditions for the entire range of agricultural activities.

The net effects are low and unstable crop yields, total loss of crops in three out of five years, abandomnent of farming parcels, highly erodable soils and frequent floodings.

in

The risk inherent in these factors results a reduction of income, which for the total rural population averages three U.S.

dollars per capita per month, with a range between one and five U.S. dollars.

The region suffers from an increasing demographic pressure.

This situation tends to persist and is resulting in a highly pernicious phenomenon that is detrimental to the social and economic development of people: the "minifundia" (Manzanilla, 1969).

Here small holdings of less than five hectares per family are used for farming and combined with seasonal part -time labor market activities such as the extracting and selling of native fruits, fiber, wax and other natural products.

The labor structure is traditionally of an agricultural -pastoral- native productsgathering nature.

About 90% of the economically active population is engaged in primary activities, generally for local consumption.

Therefore it is difficult for the tenants to accumulate working capital for husbandry improvements or to finance repairs.

The region's infrastructure exhibits an overall deficiency in such aspects as roads, communications, electricity, health centers, schools, hospitals and potable water.

For example, the educational level of the people is as follows: no single householder have attended secondary school; only 5% have finished elementary school and 44% have no formal education at all.

tions.

Villages and rural communities are widely scattered and most have small popula-

The access to these small communities is through unpave roads that are almost impassable during the rainy season.

In 1970 the total population resided in

21 towns or cities with the following distribution:

62% had less than 100 inhabitants,

27% with 100 -500 people and 11% with more than 500 people.

This situation becomes a factor that limits the introduction of the basic public services within reasonable costs.

This factor, coupled with the low productivity of land, accounts to a large degree for the unfavorable commericalization of products, high prices of basic food items, and monopolization of produced goods.

Moreover, in most instances the local villages present a internal duality with respect to production.

On the one hand, a small number of landholders monopolize most of the local resources, and by exerting economic pressure on the least affluent tenants, control the local economy.

On the other hand, most of the people have small choice in farming systems and are customarily engaged in operations that produce low outputs per hectare, such as goat and cattle raising and gathering native plants.

Maize and beans comprise 95% of the crops.

To a lesser extent, and dependent on the year's climatic conditions, wheat, barley, and vegetables are planted.

Normally, seeds used for planting are "creole" seeds, i.e., seeds harvested and stored from previous years.

In general, mineral fertilizers are not used, natural fertilizers are used on only few parcels.

Land tenure is also a problem in the San Tiburcio watershed.

The lands in the region are either privately owned or are ejidos.

The ejidos, accounting for 87% of the land area, are collectively owned lands which are subdivided and operated on an individual basis.

This creates a diversity of land usage within a small area. The pork will be directed towards ejido lands with the hope of initiating negotiations for better use of these lands.

INSTITUTIONAL SETTING

There are three major agencies involved with the development, and hence land use planning, of the San Tiburcio watershed. They are Secretaria de Programacion y

Presupuesto (SSP), Secretaria de Agricultura y Recursos Hidraulicos (SARH) and the

Comision Nacional de Zonas Aridas (CONAZA).

Through a specific research and development agreement, the Universidad Antonoma Agraria "Antonio Narro" (UAAAN) has been designated the coordinating institution in the structuring of the San Tiburcio development program.

UAAAN has operated an experimental station in the region for the past six years and collected extensive data on potential impact of various land use alter-

134

natives.

Due to the complexity of the issues and the multiplicity of the agencies and people involved, it was deemed necessary to explicitly state the underlying basis for any decision making criteria.

Thus a multidisciplinary group was formed which will provide a first -level value structure.

Using this value structure in conjunction with the MMIP land use model, UAAAN will develop a set of potential land use alternatives to present to a second -level decision body.

This body, composed of government officials, will then choose the course of action to be pursued in the development of the San Tiburcio watershed.

DECISION REQUIREMENTS

The preceding discussion of the characteristics of the region and the institutional setting highlight certain requirements and /or restrictions placed on the decision framework.

They can be summarized as follows:

1.

both factual and subjective information must be considered,

2.

3.

several levels of decision making exits, high natural uncertainty exists,

4.

5.

6.

existing conditions limit the range of alternatives, community participation is desirable, but difficult due to educational levels, location, etc., ecological interdependence among units within the watershed is high.

DEFINITION OF TERMS

In the development of following sections it will be necessary to define the following terms:

Objective In general, it will indicate a "direction" which results in "improvement" (Keeney and Raiffa, 1976).

Goal

-

A specific point in the "direction" of an objective; its achievement is binary, either it is or it is not achieved.

Attribute - Given by the physical or physiological characteristics identifiable with the alternatives under consideration.

The attributes can be viewed as means or information sources availabli to the decision maker for formulating and achieving desired objectives (Starr and

Zeleny, 1977).

Criterion Standard upon which a judgment is based; it implies some measurement and /or scale.

Project Indivisible element that compirses the alternative set for the decision making task.

Note that the projects are not necessarily mutually exclusive.

Program -

.

A subset of the set of projects that satisfies the set of constraints and contributes in the largest extent to the fulfillment of some overt objectives.

In the context of land use planning a project represents a land use alternative.

For example, a project might be to plant maize.

In this case the decision would entail not only whether or not to undertake the activity, but also at what level it will be operated.

Other projects, such as constructing a water catchment, may only involve a go -no -go decision.

A program then would be a particular land use plan for the region.

GROUP DECISION MAKING

The structuring of a development program is viewed as a closed loop procedure requiring the selection of projects which emphasize a particular area of concern.

A group of experts is responsible for ascertaining a set of objectives, establishing

135

the project relationships, and providing the value structure on which the selection procedure is based,

In this sense, the present work strives to set up a framework in which land use planning becomes one aspect in the process of structuring a development plan for a rural community in Mexico.

The next sections deal mostly with the group organization, task, and assessment procedure under this general framework.

CHARACTERISTICS OF THE GROUP

Within the context of the overall problem, the actual decision making transactions constitute only one activity among several others which are required of the group.

Hence, it is more appropriate to speak of a multidisciplinary group whose aim is to define objectives, resolve conflicting viewpoints and confront technical problems.

The presence of mutual influence between group members through open and direct communication distinguishes this approach from other Procedures used to extract factual or subjective information.

Current research is yielding support for the assertation that cooperative task group processes generate a greater collective and member performance, than their competitive counterpart (Dailey, 1978).

Among the factors which contribute more notably to the group's behavior, the literature mentions group size, cohesiveness, task certainty, and task interdependence (O'Keefe, et al, 1975: Wallmark, et al, 1973).

Since group size has been shown to be negatively related to satisfaction, Tévéi of agreement, and personal involvement (Hackman, et al, 1970: Daily, 1978), it is consistent with the stated requirements to contemplate the structuring of this group with no more than fifteen members.

However, care must be taken in the structuring process to include the necessary skills and expertise demanded by the problem.

The importance of the task characteristic and its impact on the group process has been pointed out by several researchers (Frank, 1971: Dailey, 1978).

In addition, task interdepencence pertains to the dependencies among group members to perform their individual jobs.

It should be pointed out that task interdependence is related to the organization and structuring of the group transactions, hence it is an integral part of the planning stage.

Basically, the tasks assigned to the group can be grouped into two categories: problem definition activities and decision making transactions.

In the first case, the specific output sought is a hierarchy of objectives along with a set of measures of effectiveness.

The set of measures of effectiveness (or attributes) will be required to satisfy certain conditions that would insure the existence of an assessable utility function (Keeney and Raiffa, 1976).

The culmination of the decision making transaction is the statement of a group utility function for the set of attributes defined previously.

A period of feedback between the model's output and the actual group preference is also desirable.

A preliminary inquire into the hierarchy of objectives applicable to the San

Tiburcio Watershed have provided a tentative structure which is being applied to the land use alternative case (Figure 1).

GROUP VALUE STRUCTURE ASSESSMENT

Group decision making has neither a well developed theory, nor enough applications to enable the extraction of guidelines regarding imolementability of utility models (Sheridan & Sicherman, 1976 Keeney & Raiffa, 1976).

The main thrust lies in the inadequacy of utility aggregation models which are reliant on interpersonal comparison of utilities.

Although it is evident that such comparisons are often made as Horsanyj (1974) points out, their subtleness is difficult to quantify.

Edwards

(1977) stresses that any utility assessment scheme must consider the actual behavior observed in the decision making process.

Hence, if the assessment framework is too unlike that which decision maker is accustomed, it will introduce unnecessary bias into the group response.

In order to avoid the utility aggregation problem, Kryzsztofowicz (1978) has proposed an alternative scheme which as proven successful in dealing with real time reservoir control problems.

Although it is premature to claim real life success, the proposed procedure does seem to circumvent some of the shortcomings of other schemes.

It is interesting to note that the requirements imposed by the complex nature of the present problem make it the ideal proving ground for such an assessment scheme.

The assessment process proceeds as follows:

Step 1: The group makes a decision in relation to regrouping its members into subgroups or committees.

136

Step 2:

Each subgroup assesses its own utility function and determines its own decision role.

Step 3:

The whole group must establish its tradeoff coefficients among the attributes.

Two assumptions are critical in the above procedure: a.

b.

the group acceptance of the subgroup utility functions as their own, and each group member accepts the Von Newman- Morgenstern (1947) utility axioms (or any other equivalent set of axioms that insure a, weak order on a set).

Each group and subgroup must determine its own decision rules.

An arbitrary but desirable decision rule has been adopted as part of the group structure.

In the case of a dichotomous situation a majority voting rule is adopted and a median rule is used when the situations are related to the value of a continuous variable.

The entire group transactions include a predecision stage, an interchange of points of views stage, and a final decision.

LAND USE PLANNING MODEL

The approach taken in this paper is to develop a multiobjective optimization model for land use planning.

This approach is quite common in the literature (Bamni and

Bamni, 1979; Barber, 1976; Etushenko and Mackinnon, 1976 Nijkamp and Rietveld, 1978.)

The model developed in this paper differs from most previous models in that integer valued variables are included.

These variables must be included since certain infra structural projects are prerequisites for carrying out particular land use alternatives.

Mathematically, the problem can be stated as follows: v.max

n

E m f.(xj) +

E c1 yi

1 =1,2,...,p j =1

1 =1

(1) s.t.

E i =1 aiixj

+ n

E i =1 a111y1 < bi i=1,2, q (2)

E jr y.

<

1 yi-yi

< 0

V1ET

(3)

V(i,j) Et-

(4) yi-y1

= 0 yi +yj +yk <

1

V(i,j) E SC

V(i,i,k)

E

C

Liyj

< xi

< Mlyi

VjEE x1 > 0 yi

E

{0,1}

1 =1,2

1 =1,2 n m

(5)

(6)

(7)

(8)

(9)

The variables, objectives and constraints for this formulation will now be discussed in some detail.

VARIABLES

There are two types of variables considered in this model, continuous variables

(x.) and integer variables (y.).

Each variable is associated with a project.

Projects ari either land use alternatives or infrastructural industrial projects which must be done before certain land use alternatives are feasible or economically viable.

A list of variables used in the model is given in Table i.

OBJECTIVES

The land use planning model has three proxy objectives, economic return, social benefit and environmental impact.

137

The proxy objectives can be derived by use of a subset of the attributes of Figure 1.

The attributes actually used for the land use planning objectives are al

total cost, properly discounted (E).

total revenue, properly discounted (E) a2

a3

04

- employment level

(man -days)

- community acceptance (per cent) a5 - ecological impact (change in carrying capacity).

Calculation of a a a and a are striaghtforward for, a given project.

may be obtained fFom quegtionnaires or interviews.

The values for

Finally, values of a5 can be fund by using regression techniques or available empirical data.

The proxy objectives can then be derived from the attributes.

Economic return can be expressed as the difference of return and cost.

In order to determine social benefit, a utility function assessment over a and a4 using the group utility procedure previously outlined must be carried ouf.

For the purposes of this study, ecological impact will be represented exclusively by the change in carrying capacity of the land.

With the possible exception of the portion of the social benefit objective associated with the continuous variables, all objectives are linear in the decision variables.

The social benefit utility function will be linear if the subgroups have linear utility functions and the entire group is risk -neutral.

If not, additiona zero -one variables can be added and a piecewise- linear approximation can be used.

CONSTRAINTS

The constraints of the land use planning model can be considered to be either resource constraints or logical constraints.

Examples of resource constraints (constraints (2) in the model formulation) include a limited budget and limitations of certain types of land suited for particular land use alternatives.

Logical constraints ((3) -(9)) insure that certain logical conditions must be met.

Constraints

(3) represent mutually exclusive projects, for example projects for water delivery to the greenhouse from wells or from a run -off catchment are mutually exclusive.

Constraints (4) and (5) represent (strict) contingency relations between projects.

An example of this would be wells and water distribution.

It would be useless to build a watering distribution system without first drilling a well,

Other constraints may represent complementary relations: that is doing two projects together may have an effect different than if the projects were done separately.

Constraints (6) handle this case.

An example might be road construction and processing of goat cheese.

Each project has certain characteristics independently of the other, but if both are implemental, new markets are opened up and the impact is greater.

Constraints (7) have several purposes.

They enforce thresholds, account for economies of scale, and provide the mechanism for handling piecewise linear objectives.

An example of this case would be that yucca fiber transformation can take place only if a processing plant is to be built.

Finally, the constraints (8) and (9) required levels to be non -negative and certain projects to be done or not done.

For the sake of brevity the entire formulation isnot presented in this paper.

The prototype problem has more than fifty variables (including slacks and artificials), seventy constraints, and three objectives.

SOLUTION APPROACH

There are three approaches normally taken in solving multiobiective optimization models such as the one discussed in this paper.

They are goal programming, utility approaches, and vector maximization.

Goal programming has been widely used (Lee,1972;

Lee and Moore, 1977; Dover and Krueger, 1977; and Bussey, 1978) but has certain drawbacks.

For objectives expressed as utilities, such as social benefit, it is very hard to set meaningful goals.

In addition, weights must be chosen a priori for the different objectives, which may also prove a difficult task.

from a

Utility approaches suffer similar difficulty: determining a utility function over the different objectives.

Explicitly determining such a utility function requires much time and effort on the decision maker (Edward 1977; Farquhar, 1977) and when decisions are made by a group, as in this case, the problem is magnified.

Implicit derivations of the utility function are also possible through interactive solution procedures.

Vector maximization (Geoffrion, 1967: Kornbluth, 1972; Zeleny, 1974) is an attempt to generate all non -dominated (or efficient) solutions, and is usually computationally intractable.

138

Initial attempts to solve the prototype problem by use of an explicit utility function have not been promising.

The particular decision structure seems to necessitate explicit tradeoffs at the second level.

Thus, the solution approach will be to generate some subset of the efficient solutions.

Previous work (Bitran, 1978; Zionts,

1978; Villarreal and Darwin, 1978; Banker et al., 1978) indicates that generating the entire efficient set for multiobjective mixed - integer problems of reasonable size may be beyond the capabilities of existing algorithms and computers.

For example, the largest problem solved by Banker et al.

(1978) has nine variables, seven constraints and four objectives.

The algoritl used was, for this problem only twice as efficient as total enumeration.

The authors are currently working on an heuristic for this problem, which will generate a subset of the efficient set.

Controlled enumeration over the feasible region examining all objectives simultaneously will form the basis of the algorithm.

Of course, the quality of fathoming rules will determine how well the solutions generated appeal to the second level decision makers.

REFERENCES CITED

Bamni, D. and D.

Bamni.

1979.

Development of a Comprehensive Land Use Plan by Means of a Multiple Objective Mathematical Programming Model,

Interfaces, vol.

9, No.2.

Banker, R; M. Guihurad; S.K. Gupta. 1978.

An Algorithm for Generating Efficient Solutions to the Multiple Objective Integer Programming Problem, RP, Wharton ARC, Penn.

Barber, G.M.

1976.

Land -use plan interactive multiple -objective programming.

vironment & Planning, vol. 8, No.

5, pp. 625 -636.

En-

Baur, V. and M. Wegener.

1975.

urban planning,

Simulation evaluation and conflict analysis in

Proceedings IEEE, vol. 63, No.

3.

Bitran, G.

1978.

Theory and Algorithms for Linear Multiple Objective Programs with

Zero -One Variables,

Tech. Report No.

150, OR Center, MIT.

Bulfin, R.L. and H.L. Weaver.

1977.

Appropriate Technology in Natural Resources

Development: An Overview, Annotated Bibliography and Guide to the Literature,

Office of Interdisciplinary Programs, Office of Arid Lands Studies, University of Arizona, Tucson.

Bussy, L.

1978.

N.J., 491 pp.

The Economic Analysis of Industrial Projects, Prentice Hall,

Inc.

Dailey, R.

1978.

The Role of Team and Task Characteristics in

R &

D Team Collaborative Problem Solving and Productivity,

Management Sci., Vol. 24, No.

15, pp.

1579 -1588.

Dauer, J.P., R.J. Krueger.

1977.

Quarterly, 28(3ii), pp.

383 -417.

An Iterative Approach to Goal Programming,

OR

139

Edwards, W.

1977.

How to use Multiattribute Utility Measurement for Social Decision making, IEEE Transactions, Vol. 5MC -7, No.

5.

Evtushenko, Y., R.D. Mackinnon.

1976.

Nonlinear programming approaches to National

Settlement system-planning, Environment & Planning, Vol. 8, No.

8, pp. 637 -654.

Farquhar,

P.

1977.

A survey of multiattribute utility theory.

In:

Multiple

Criteria Decision Making

(J. Cochrane and M. Zéleny, eds.), pp.

59 -60.

North -

Holland, Amsterdam.

Geoffrion, A.

1967.

Solving Bicriterion Mathematical Programs.

OR 15, 39.

Geoffrion, A.;

J.

Dyer; A. Feinberg.

1972.

An interactive approach for multi criterion optimization with application to the operation of an academic department,

Management Sci.

19, 357.

Goicoechea, A.

Management,

1977.

A Multi- Objective, Stochastic Programming Model in Watershed unpublished Ph.D. Dissertation, University of Arizona, Tucson,

Arizona.

Harsanyi, J.

1974.

Vol.

7, pp.

61 -82.

Nonlinear Social Welfare Functions.

Theory and Decisions,

Helly, Walter.

1975. Urban Systems Model, Academic Press, New York.

Keeney, R.,

R.

Raiffa.

1976.

Decisions with Multiple Objectives:

Value Tradeoff, 569 pp., Wiley, N.Y.

Preferences and

Krzysztofowicz, R.

1978.

Preference Criterion and Group Utility Model for Reservoir

Control Under Uncertainty, pp.

107, Technical Report No.

30, Systems and Industrial Engineering Department.

Lee,

S.

1972.

Goal Programming for Decision Analysis, Manaaement Science, Averback,

Philadelphia, Penn., 389 pp.

Lee, S.,

L.

J. Moore, 1977.

Multicriteria School busing Models, vol. 23, No.

7, pp.

703 -715.

Manzanilla,

5.

1969.

La Reforma Agraria Mexicana,

Management Science.,

Comunidades 12:31, Madrid.

Marglin, S.

1967.

Public Investment Criteria,

MIT Press, Cambridge, Mass.

103pp.

Medina, J.

G.

1978.

Tiburcio Watershed,

Preliminary report on the management alternatives for the San

Department of Renewable Natural Resources UAAAN,

Mexico.

140

Nava, R., R. Armijo, J. Gasto.

hombre,

1978.

Ecosistema:

UAAAN- Press, Mexico ,

365 pp.

La unidad dela naturaleza y el

Nijkamp, P. and P.

Rietveld.

1978.

New Multi- Objective Technqiues in Physical

Planning, paper presented at the NATO Advanced Study Institute on Water Resources and Land

Use Planning, Lonvain-la.Neuve.

Sheridan, T.B., A. Sicherman.

1977.

Estimation of a Group's Multiattribute Utility

Function in Real Time by Anonymous Voting,

IEEE Trans. Sys. Man and Cybern,

SM( -7(5), pp. 392 -394.

Starr, M,

M.

Zeleny.

1977.

Holland Publ., Amsterdam.

Multiple Criteria Decision Making,

326 pp. North -

Villarreal, B., M.

H.

Karwan.

1978.

Dynamic Programming Approaches for Multicriterion

Integer Programming,

Research Report No.

78 -3, May, Department of IE, State

University N.Y. at Buffalo.

Von Neuman, J., O. Morgenstern.

1947.

Theory of Games and Economic Behavior,

2nd

Ed., Princeton University Press, Princeton, N.J.

Wallmark, J. T., S. Eckerstein, B. Langened, and H. Holmovist.

1973.

The Increase in

Efficiency with Size of Research Teams,

IEEE Trans. Engineering Management,

Vol. EM -20, No.

3, pp. 80 -86.

Zeleny, M. 1974a.

N.Y.

220pp.

Linear Multiobjective Programming,

Springer -Verlag, Berlin and

Zionts, S.

1978.

A Survey of Multiple Criteria Integer Programming Methods,

Paper presented at the DO 77 Conference, Univ. of British Columbia.

O'Keefe, R.D., J.

A.

Kernaghan, H.

A. Rubenstein.

1975.

Group Cohesiveness:

A factor in the Adoption of Innovations Among Scientific Work Groups, Small Group Behavior, vol.

6, No.

3, pp.

282 -292.

141

x9 x10 x11 x12 x13 x6 x7 x6 x2 x3 x4 x5 x

14

DECISION

VARIABLE xi x15 x16 x17 x18 x19

Agricultural -NRM:

Rest and shrub control the halophyte shrubland.

Agricultural -NRM: Rest only the Halophyte shrubland (H.S.)

Agricultural -NRM: Rest only the Halophyte grassland (H.G.)

Agricultural -NRM: Do nothing new to the H.G. unit.

Agricultural -EC: land of corn.

Rainfed crop-

Agricultural -EC:

Run -off /supplemented corn cropland

Agricultural -EC: cropland.

Rainfed bean

Agricultural -EC: Run- off /supplemented bean cropland.

Agricultural -EC:

Opuntia plantation in L -F unit 1.

Agricultural -NRM:

Manual shrub control to the L -F unit.

Agricultural -NRM:

Range reseeding the L -F unit.

Agricultural -NRM: Selective shrub control

Y -0 unit

Agricultural -NRM: the Y -0 unit.

Rest only

Agricultural -NRM:

Semimechanical shrub control y -O unit

(selective)

Agricultural -NRM:

Mechanical shrub control Y -0 unit (selective).

Agricultrual -NRM: control y -O unit.

Total shrub

DECISION

VARIABLE x20 x21 x22 x23 x24 y1 ylw y20 y2W y2 y3b y4 y5 y

6

Y7 y8

Agricultural -NRM: Range reseeding & water supplemented the L -F unit.

Agricultural -P: Animal units year of goats (AUYG)

Agricultural -P: Animal units year of cows (AUYC)

Industrial -Guayole production level

(kgs of rubber).

Industrial -,oat milk benefiting

(Lts).

Industrial -Yucca fiber transformation.

Industrial -Agave fiber.

Industrial -Opuntia food distribution level.

Agricultural -C:

Greenhouse

Infrastructural -W: Water -runoff structure for watering points.

Infrastructural -W: Well perforation for greenhouse and community consumption.

Infrastructural -W:

Well perforation and extracting units for watering points.

Infrastructural -L -W: Watering units.

Industrial -Goal milk by -products.

Industrial -small scale guayole plant.

Industrial -Yucca and agave textile facility

Industrial -Opuntia food and candy

Infrastructural - Roak construction to industrial installations.

Infrastructural -Fencing

Table 1.

Decision Variables under consideration for the land -use planning preliminary phase.

143

EVALUATION OF WATER MANAGEMENT SYSTEMS FOR THE SONOITA CREEK WATERSHED by

Hugh B. Robothaml

INTRODUCTION

The purpose of this paper is to present a water management study which is being conducted for the

Sonoita Creek Watershed, and to provide an example of the use of the standardized cost effectiveness methodology in evaluating water resources systems.

The watershed is about 209 square miles in area and is located in Santa Cruz County, southeastern Arizona.

The towns of Sonoita, situated on the northern edge of the basin, and Patagonia situated in the floodplain of Sonoita Creek, are the main population centers in the area.

Previous studies done by Halpenny (1964), Nasseridin (1967) and more recently Ben -Asher, et al.

(1976) have indicated that a steady state condition currently exists in the watershed when viewed on a yearly cycle.

However, seasonal fluctuations in water levels, in response to the winter and summer rainy seasons and the intervening dry season, have caused great concern over the adequacy of the areas water supply in meeting future demands.

A water management study is necessary at this point because of the following reasons:

1.

Most of the water used in the Basin is pumped from the alluvial aquifer which underlies the floodplain.

Although agricultural activities are expected to maintain only a constant level of productivity in the area, municipal consumption is expected to increase substantially over the next twenty years.

This is in anticipation of copper mining operations which could start at Red Mountain, southeast of Patagonia, in the next ten years.

A substantial increase in population could translate into increased pumpage from the aquifer.

2.

Increased pumping in Patagonia could stop the perennial flow of the creek through the Sonoita

Creek Sanctuary, which has long served as a refuge for rare and endangered species of birds and fish.

3.

Continued release of sewage effluent, from the Patagonia sewage treatment plant, a few hundred feet upstream of the sanctuary, could lead to serious deterioration of water quality in the creek and in lake Patagonia, a recreational facility in the area.

The possibility of conflict therefore exists among potential water demands in the basin.

The standardized cost -effectiveness (CE) methodology has been selected to develop and compare alternative system solutions for the Sonoita Creek water management study.

develop and compare alternative systems, has been widespread in recent years.

The use of CE approach, to

Important applications of the methodology to water resources problems include development of the Lower Meking River Basin

(Chaemsaithong et al., 1974), and comparison of alternative water reuse systems in Tucson (Duckstein and Kisiel, 1977).

CE methodology is particularly suited to water resources problems because of the multiplicity of the objectives and the sometimes unquantifiable nature of both costs and effectiveness.

The next section presents the standardized CE methodology as applied to the Sonoita Creek water management study.

The conclusions and recommendations follow.

APPLICATION

The standardized cost -effectiveness approach, as proposed by Kazanowski (196A), consists of the following ten steps:

1.

2.

Define the desired objectives that the systems are to fulfill.

Identify requirements or specifications that are essential for the attainment of the desired goals.

3.

4.

5.

Establish system evaluation criteria that relate system capabilities to requirements.

Select fixed -cost or fixed -effectiveness approach.

Develop alternative systems or solutions for accomplishing the objectives.

1.

Graduate Research Assistant, Dept. of Hydrology and Water Resources, Univ. of Arizona, Tucson.

145

6.

7.

8.

9.

10.

Determine capabilities of the alternative systems in terms of evaluation criteria.

Generate systems - versus - criteria array.

Analyse merits of alternative systems.

Perform sensitivity analysis.

Document the rationale, assumptions, and analyses underlying the previous nine steps.

OBJECTIVES

The objectives that the Sonoita Creek Basin Water Management Plan are to accomplish are grouped under demand, environment, treated sewage effluent disposal, and flexibility.

The planning horizon covers the period 1980 -2000.

Water Demand

The most economic means of fulfilling the water demands of the Basin should be determined.

demands are;

1) municipal, which includes projected use by potential mining employees;

These

2) recreation, namely Lake Patagonia and the Sonoita Creek Sanctuary;

3) agriculture and; 4) rural activities.

Quality requirements for each should be met.

Environment

The water management plan should have no negative effects on the environments of Sonoita Creek

Sanctuary, Lake Patagonia and the National Forest area.

Disposal of Treated Sewage Effluent

Treated sewage effluent from the Patagonia plant should be efficiently reused.

of other waste waters produced in the Basin should be made.

Flexibility

Proper disposal

The proposed systems should be flexible enough to meet a wide variety of future requirements, most of which cannot be foreseen at the present time.

SPECIFICATIONS

The specifications consist of expressing the objectives in quantitative terms, inasmuch as that is possible.

The planning horizon of 20 years was chosen because; 1) uncertainties involved in forecasting population growth and water use makes a longer period unadvisable, which may be introduced would make a shorter period unrealistic.

2) structural changes

The specifications are developed in a one -to -one correspondence with the objectives.

Demand

Municipal and Industrial.

Figure 1 and 2 gives the projected water demand for the communities of Patagonia and Sonoita for the period 1970 to 2000.

Figure 1 (median projection) is based on a constant population growth rate of 3% per year over 1976 estimates and a constant increase in water use rate from .21 acre -feet per person per year in 1976 to .23 acre -feet per person per year in 2000.

Figure 2 (median projection) is based on a 2% per year population growth rate and a steady increase in water use from .083 acre -feet per person per year in 1970 to a maximum of .23 acre -feet per person per year in 2000.

The plan should provide enough water to cover the high estimate in Figure 1 and the median estimate in Figure 2.

Quality standards for water supply in Patagonia and Sonoita are those established by the Environmental Protection Agency for domestic use (EPA, 1920).

Recreation.

The recreational facilities that will be directly affected by a water management plan are the Nature Conservancy's Sonoita Creek Sanctuary and Lake Patagonia.

A minimum flow of 0.5 cubic feet per second should be assured for the perennial segment of the creek near the southern edge of the sanctuary.

This 0.5 cubic feet per second is the minimum flow measured at that point in the creek over the period for which records were available.

Water quality specifications of the creek are those prescribed by the EPA (1972) for fresh water aquatic life.

146

1970

1975

1980 1985

DATE

1990

1995

2000

Figure 1.

Total Water Use in Patagonia, Arizona, 1970 -1976, and Forecast to the Year 2000. -- Based on population forecast, water use records from the Patagonia, Arizona, Town Clerk and assumed rate of water use for the year 2000 equal to 0.23 acre -feet per year.

High, median and low projections are shown.

In order to function adequately as a recreational facility the water level in Lake Patagonia should be kept a minimum of five feet below spillway crest (3,766 feet); a lower level would render the marina and beaches useless.

Quality standards for the water in the lake are those stated by the EPA for recreational uses.

Agriculture.

Agricultural activities are expected to maintain a constant level of activity throughout the planning period.

The amount of water used in 1975, about 1500 acre feet, is expected to remain the same throughout the period.

Water quality standards are those stated by the EPA for agricultural purposes.

Rural use.

Total use by rural residences and cattle ranches was estimated to be about 180 acre feet (Ben -Asher et al., 1977).

Rural water supply is not expected to be affected by activities in the alluvial valley.

-

Environment

The environment status of the Sonoita Creek Sanctuary, Lake Patagonia and the National Forest areas should not be negatively affected.

Requirements for the sanctuary and lake are the same as those stated under water demand for recreation.

In addition, water levels in the vicinity of the sanctuary should not fall to levels which may prove detrimental to the life of the phreatophytes in the tract.

It is difficult to determine this level because phreatophytes generally have very deep root systems which invariably do reach the water table.

Flexibility

Flexibility should be an undispensible property of the Sonoita Creek water management plan, because of the many uncertainties which characterize the problem (Diaz Pena, 1978).

These uncertainties include natural, strategic, technological and informational uncertainties.

147

200-

ò

100 c

90.

<

80-

70

=

6CI

501

201

Cf.

4.:.......--

/

///

/

O/

'/ i./

/

//

..

/

P.

p

//

//

....

//

//

//

'

//

//

/

/

/

/

//o

/O

/A

/

/

/

// '//

//

// 4

/ i

/

/O//

./

1C

1970

1975 1980 1985

DATE

1990

1995

2b00

Figure 2.

Total Water Use in Sonoita, Arizona. -- Projection 1970 -2000 based on population projections and assumed rate of water use equal to 0.23 acre -foot per person per year by the year 2000.

High, median and low projections are shown.

The plan should be able to cope with all of these uncertainties and posses the necessary flexibility to undergo changes that become necessary as new elements are introduced or become more important in the future (Kazanowski, 1972).

SYSTEMS EVALUATION CRITERIA

The measures of effectiveness (MOE) are developed in a one -to -one correspondence with the specifications.

Demand

Municipal and Industrial.

The MOE for municipal and industrial demand is the opportunity loss measured by a specific number of people that cannot live in the community for each acre -foot shortage of water.

Tables 1 and 2 gives the "opportunity loss" functions for the communities of Patagonia and

Sonoita respectively.

These values represent the ratio between the median projection for population and the median projection of total water consumption.

The MOE for water quality takes the qualitative ratings of very good, good, fair and poor.

Recreation.

The MOE for water quantity demand in Sonoita Creek is the probability that the minimum flow falls below the required flow of 0.5 cubic feet per second.

For Lake Patagonia the MOE for water quantity demand is an opportunity loss associated with the number of people that cannot use the facility as a function of water level depths below spillway crest.

The precise form of this function is still to be determined.

The MOE for water quality demand for Sonoita Creek and Patagonia has the ratings very good, good, fair and poor.

148

400

300

Q

200

/

/ i

.i.

.

--

..----° ii/ _',.o

? 100 y

90

80

70

60

50

40

30

20

10

1970 1975 1980

1985

DATE

1990

1995

2000

Figure 3.

Sewage Water Effluent from Patagonia, Arizona. -- Projection to the year 2000 based on ADEPD

(1971) report and estimated volume of domestic water use.

High, median and low projections are shown.

Table 1.

Opportunity Loss for Patagonia.

Year

1980

1985

1990

1995

2000

Population per

Acre- Foot /Year

4.72

4.63

4.52

4.43

4.35

Table 2.

Opportunity Loss for Sonoita.

Year

1980

1985

1990

1995

2000

Population per

Acre- Foot /Year

8.68

7.44

6.15

5.24

4.35

149

Agriculture.

The number of acres of agricultural land that cannot be irrigated due to a shortage of water is the measure used for agricultural water demand.

Using an estimate of five acre -feet of water per acre of land irrigated per year (Ben -Asher et al., 1977), each acre -foot of shortage leaves

0.20 acres unirrigated.

The criterion for water quality demand is; very good, good, fair and poor.

Environment

The environmental impact of the water management plan, on the sanctuary, lake and national forest, is assessed using the qualitative measures; beneficial effects, unaffected and adverse effects.

Disposal of Treated Sewage Effluent

The effectiveness of the systems with respect to sewage water disposal is measured by the fraction of secondary treated sewage effluent not utilized, especially in agriculture.

The exact form of this function is still to be determined.

Flexibility

The flexibility of the systems is measured by their ability to cope with uncertainties (sensitivity) and the ease with which they can be transformed.

Correspondingly, the two M0E's are; not sensitive, sensitive or very sensitive and very good, good, fair or poor.

SELECTION OF FIXED COST OR FIXED EFFECTIVENESS APPROACH

The choice between fixed cost and fixed effectiveness is necessary in cost -effectiveness analyses and is in general not a trivial decision (kazanowski, 1968).

In the fixed -cost approach the alternatives are judged on the basis of the amount of effectiveness obtained for a given expenditure of resources.

In the fixed -effectiveness approach the alternatives are evaluated on the basis of the amount of cost incurred or resources required to obtain a given level of effectiveness.

The nature of the objectives, which require certain minimum specifications to be met, necessitated the selection of a fixed -effectiveness approach.

It is generally recommended that both approaches be used especially when major assumptions are supported only by subjective judgement and sketchy data.

DEVELOPMENT OF ALTERNATIVE SYSTEMS

The importance of this step is stressed by Kazanowski (1968) who stated that the results of the evaluation can be no better than the conception of attractive candidate systems."

A major problem usually encountered in implementing this step is in determining the degree of detail required for the definition of the candidate systems.

Too little system definition usually results in unreliable estimates of system effectiveness and cost while too detail definition would defeat the basic purpose and value of cost -effectiveness.

Suitable candidate systems have been developed by Diaz Pena (1978).

The principal characteristics of these systems are displayed in Table 3.

CAPABILITIES OF ALTERNATIVE SYSTEMS

Once appropriate criteria have been identified and the candidate systems have been adequately defined the next step is to express the abilities of these systems in terms of the criteria.

This should be done in quantitative terms, whenever possible, or qualitatively.

Diaz Pena (1978) established the framework for the present study.

For the purpose of his thesis he expressed dil the criteria in qualitative terms then performed a subjective evaluation of the systems capabilities.

The reader is referred to this thesis or a subsequent paper, published by

Diaz Pena, et al. (1978) for the details.

The main aim of this study is to perform a quantitative evaluation of the alternatives, insomuch as that is possible, and to determine the most suitable plan for the basin.

Although substantial research has been done over the past nine months several problems still remain:

1.

The effects of continued release of treated sewage effluent, at the present site, on the environments of Sonoita Creek Sanctuary and Lake Patagonia have to be researched.

2.

3.

Cost estimates of each alternative have to be determined.

Operating policies for Lake Patagonia have to be researched.

In order to do a quantitative evaluation of the candidate systems the above problems must be resolved.

The author expects work on this study to be completed within the next two to three months.

150

Table 3.

Main Characteristics of Alternative Systems

Alternative

Systems

I

Patagonia Municipal

Water Supply

From existing wells and new wells in the future.

II

V

III

IV

VI

Same as in

Alternative I

From existing wells and from Red Rock

Reservoir.

Same as in

Alternative I.

Same as in

Alternative I.

From existing wells and from Lake

Patagonia.

Sewage Effluent

Agriculture

Water Supply

Released at site upstream from

Sonoita Creek

Sanctuary

Unchanged, from wells and springs.

Base Flow in Sonoita

Creek Sanctuary

No provisions made for maintaining minimum flow through

Sonoita Creek

Sanctuary.

Piped downstream of Lake

Patagonia Dam.

Same as in

Alternative I.

Same as in

Alternative

I.

Same as in

Alternative

I.

Same as in

Alternative I.

Minimum flow assumed through Sanctuary by water from Red Rock

Reservoir.

Pumped 6 miles upstream for use in Rail -X

Ranch for irrigation. Total

Lift: 250 feet.

Pumped 1 mile upstream for use in Box -T

Ranch for irrigation.

Total Life:

25 feet.

Same as in

Alternative I.

Rail -X Supplemented with second ary- treated sewage effluent.

Box -T Supplemented with second ary- treated sewage effluent.

Same as in

Alternative

I.

Same as in

Alternative I.

Same as in

Alternative I.

Minimum flow assumed through Sanctuary by water from Lake

Patagonia

Lake Patagonia

No use for water supply within the basin.

Same as in

Alternative I.

Same as in

Alternative I.

Same as in

Alternative I.

Same as in

Alternative I.

Used for sup plementary water supply to Patagonia and Sonoita

Creek Sanctuary

SYSTEMS VERSUS CRITERIA ARRAY

The system -versus- criteria -array is regarded by many as the core of multicriterion decision making.

This array is easily generated once the abilities of the candidate systems have been expressed in terms of the criteria.

In general, the criteria are identified at the tops of columns and arranged in decreasing order of importance while the alternatives are listed vertically (Kazanowski,

1968).

This arrangement is particularly useful when many alternatives are being evaluated because less likely candidates can easily be eliminated leaving the major contenders.

The ultimate selection is generally based on a judicial evaluation of system capabilities and requirements.

The remaining steps of the cost -effectiveness methodology consists of; 1) analysing the merits of the alternatives,

2) performing sensitivity analysis, and 3) documenting the results, etc.

These steps can be dealt with more appropriately when systems capabilities can be adequately evaluated.

CONCLUSIONS

A water management study for the Sonoita Creek watershed is needed to resolve possible conflicts among potential water demands in the area.

These demands have certain requirements which should be met.

Implementation of a water management plan should have no negative effects on the environments of Sonoita Creek Sanctuary, Lake Patagonia and the National Forests.

Cost -effectiveness is an appropriate approach for this problem since it allows for the consideration of all the important factors, both quantitative and qualitative, involved in the decision -making process.

The approach enables the identification and proper handling of the uncertainties which exist in such complex decision problems.

151

Much work needs to be done before a thorough evaluation of the candidate systems can be made.

The main problems that still remain are:

1.

The impact of treated sewage effluent on the environments of Sonoita Creek Sanctuary, and

Lake Patagonia.

2.

Cost estimates of the alternative systems.

3.

Optimal operating policy for Lake Patagonia.

A qualitative evaluation can still be done, as exemplified by Diaz Pena.

the versatility of the cost -effectiveness methodology.

This further illustrates

REFERENCES CITED

Ben -Asher, J., J. Randall, and S. Resnick.

1976.

Determining Areal Precipitation in the Basin and

Range Province of Southern Arizona Sonoita Creek Basin.

Hydrology and Water Resources in Arizona and the Southwest.

6:161 -167.

Chaemsaithong, K.

1973.

Basin.

Design of Water Resources Systems in Developing Countries:

Ph.D. dissertation, University of Arizona, Tucson.

The Lower Mekong

Duckstein, L., and C. Kisiel.

1973.

Cost -Effectiveness Analyses of Disposal Systems.

Journal of the Environmental Engineering Division, ASCE.

99(EE5):577 -591.

Environmental Protection Agency.

1972.

Water Quality Criteria, Washington, D.C.

Halpenny, L.C., et al.

1964.

Groundwater Supply of Patagonia Area, Santa Cruz County.

Water Development Corporation, Tucson, Arizona.

Kazanowski, A.D.

1968.

A Standardized Approach to Cost -Effectiveness Evaluation, in J. English

(editor), Cost -Effectiveness:

The Economic Evaluation of Engineering Systems.

John Wiley and

Sons, Inc., New York.

Nasseridin, M.T.

1967.

Hydrogeological Analysis of Groundwater Flow in Sonoita Creek Basin.

Santa

Cruz County, Arizona.

M.S. thesis, University of Arizona, Tucson.

Office of Economic Planning and Development.

1974.

Established Natural Areas in Arizona, A Guidebook for Scientists and Educators.

Planning Division, State of Arizona, Pheonix.

Pena, E.D.

1978.

Application of the Cost -Effectiveness Methodology in Water Management: tion of the Sonoita Creek Basin System.

M.S. thesis, University of Arizona, Tucson.

Formula-

Pena, E.D., et al.

1978.

Water Management in the Sonoita Creek Basin, Arizona:

A Multicriterion

Approach.

Working Paper No. 78 -26, Department of Systems and Industrial Engineering, University of Arizona, Tucson.

ACKNOWLEDGEMENT

The research leading to this paper is supported in part by the Office of Water Resources and

Technology Grant No. 14- 34- 0001 -6003.

The help and advice of J. Ben- Asher, D. Davis, M. Diskin, L. Duckstein, E.D. Pena, J. Randall,

S. Resnick and M. Sniedovich is also gratefully acknowledged.

152

THE EFFECTS OF SECOND -HOME AND RESORT -TOWN

DEVELOPMENT ON STREAM DISCHARGE IN

NAVAJO AND APACHE COUNTIES, ARIZONA by

T. D. Hogan and M.

E. Bond

This study brings together water data and information on second -home development to discover whether relationships can be identified between second -home development and water flows.

The research specifically focuses on surface water conditions, as measured by stream discharge readings, and has been restricted to areas where both second -home data and stream discharge data are available.

The methodology for this study has been determined by time and financial constraints.

Limited allocations of both resources precluded a long -term field -data collection effort.

Hence, data from existing secondary sources have been used for the analyses.

Prior investigations recognized that second -home development would predictably affect runoff volumes where the streams were adjacent to the second -home locations.

Construction of dwellings, roads, and related structures heighten stream discharge during rains because of the increased immediate water runoff and lessen the measured flows in nonrain periods because of reduced groundwater absorption and seepage.

The measurability of this relationship through the use of secondary data --has not been tested in the context of second -home and resort -town development.

Thus, without preconceived expectations, this study has been directed toward the relationships between watershed changes associated with development and the resultant impacts on streamflow.

LITERATURE REVIEW

Two subsets of empirical analyses of watershed changes are of particular interest to this project.

The first are investigations of the effects of changes in vegetation upon the runoff of existing watersheds.

In general, results demonstrate that forest removal both increases the volume of runoff and modifies the time pattern of that runoff.

The second body of literature investigates the impact of urban growth upon streamflow.

The available evidence indicates that the land use changes associated with urban development have significant impacts upon hydrologic relations.

Waananen has summarized these effects in the following list: (1) increase in total yield from stormflow and in annual discharge; (2) decrease in base flow of those streams that remain under generally natural conditions;

(3) modification of low flow of streams influenced by the importation of water, the use of which results in discharge of wastewater; (4) decrease in recharge to the underlying groundwater basins; (5) increase in precipitation in urban areas and corresponding increase in yield (Waananen, 1969).

The forested areas of northern and eastern Arizona provide a major portion of the total state surface runoff, which is vital to the State's agricultural sector and to its metropolitan areas.

In the post -World War II period there has been a dramatic trend of second -home development in these same northern and eastern counties providing summer homes for the residents of the Phoenix and Tucson metropolitan areas.

No comprehensive study of second -home development in Arizona has yet been undertaken, but some information on this development can be compiled from a variety of sources.

Thompson and Lewis conducted a U. S. Forest Service study of residential development on private land in the Mogollon Rim area

(including prime second -home areas in Coconino, Gila and Navajo Counties).

They estimated that there were 150 subdivisions containing 16,000 lots within their study area in 1972.

Their analysis also revealed that dwelling units had been constructed on only 3,300 of those lots up to that year (Thompson and Lewis, 1973).

The authors are respectively, Research Associate, Bureau of Business and Economic Research,

College of

'Business Administration, Arizona State University; and Dean, College of Business Administration,

State University.

Memphis

The project on which this paper was based was funded by the Eisenhower Consortium for Western Environmental Forestry Research Grant No, 16- 787- GR(EC261).

153

In 1977, Bond and Dunikoski examined the impact of such developments on water availability in

Arizona and inventoried the number of second homes within the major second -home areas of Coconino, Gila,

Navajo, and Yavapai Counties.

Bond and Dunikoski estimated there were 5,500 second homes in 1967 and

10,500 within the study area in 1975; and based upon these data, together with analysis of additional information, they also projected the number of second homes within their study area would grow to over

21,000 by 1985 (Bond and Dunikoski, 1977).

Gerking, Holmes, and yanBrackle studied second -home development in Navajo and Apache Counties during the 1958 -1977 period.

Un the basis of field work, they concluded that:

(1) the number of second homes in the area had grown dramatically; (2) most of this development has occurred in the Show

Low -Pinetop- Lakeside area; (3) the prospects for future development appear bright; and (4) data on second homes in Navajo and Apache Counties were difficult to obtain but an acceptable proxy measure was available from electric utility records kept by the Navopache Electric Cooperative, Inc. (Gerking, et al., 1979).

The growing magnitude of rural residential development has led to increasing research concerning the possible impacts upon the areas and the nearby communities where such development is occurring.

Looking specifically at water issues, analyses have discussed second -home development in terms of its impacts upon the local demand for water, upon the erosion of nearby lands, and upon water quality in affected streams.

Bond and Dunikoski examined the magnitude of water depletions associated with second home use and concluded that future development in northern Arizona would have only minor impact upon water availability in the affected hydrologic regions.

Bricklen and Utter examined the water quality issue in a study of three lakes and three streams in the White Mountains of northeastern Arizona.

Although vacation homes, campgrounds, and day -use sites were located within the watersheds, no major pollution problems were identified in the lakes and streams studied.

The authors concluded that the potential for problems still does exist unless research or land managers' current endeavors are continued and expanded (Bricklen and Utter, 1975).

DATA

The major objective of this research is to determine whether a statistical relationship can be identified between second -home development and stream discharge.

Such an investigation requires a time series approach.

Gerking's study of second -home development used time series data from the Navopache

Rural Electric Cooperative, a utility serving parts of Navajo and Apache Counties, to develop estimates on second homes.

The Navopache Rural Electric Cooperative data are also used in this study as a proxy for the time pattern of second -home development.

At the time of this study the utility maintained a very minimal charge for seasonal off /on power connection charges, and most seasonal second -home owners could minimize their annual utility outlay by a seasonal connect /disconnect charge rather than pay a monthly minimum charge for periods of nonuse.

Persons at the utility agreed that most of these seasonal connections were for summer cabins; thus, it was determined that these "seasonal hookup" data would provide a time series on second -home activity that would not otherwise be available.

This analysis also relies upon secondary data for measures of stream discharge in the study area.

The United States Geological Survey publishes water data reports in which stream discharge data are reported from each of its gauging stations on streams throughout Arizona.

Using the study area defined by the Navopache Electric Cooperative service area, the watersheds to be analyzed were identified by hydrologic maps of northern Arizona.

Potentially useful gauging stations were identified on these streams, and by field inspection, the study team selected a set of six gauging stations to be utilized in the analysis.

A list of these stations is provided in Table 1.

Analyses of the impact of urban development have generally demonstrated that such development affects both the volume and time pattern of streamflow.

To evaluate these two types of change, monthly stream discharge data measuring both maximum flow (in cubic feet per second) and total discharge (in acre feet) from each of the six stations were compiled from the annual issues of water Resources Data for Arizona (U. S. Geological Survey).

In addition to the second -home development variable, two measures of climatic conditions have also been employed in the analysis as explanatory variables.

Since the magnitude of flow would likely be positively related to the volume of precipitation occurring on the watershed, a measure of total monthly precipitation was therefore incorporated as one of the variables in the estimating equation.

It also seems likely that the pattern of stream discharge would be related to temperature within the watershed area; thus average monthly temperature was also included as the second climatic variable in the explanatory models.

Other factors that might be expected to influence the pattern of stream discharge, such as soil types, topography of the area, and the nature of the vegetation, would not be expected to change in a given location over the relatively short study period without outside action such as land development, cultivation, foresting, fire damage, etc.

For this exploratory analysis, the assumption has been made that no changes in these factors occurred that were not associated with the process of second -home development.

154

The data series employed as measures for the two climatic variables were compiled from U.

S.

Department of Commerce sources.

Series relating to average monthly temperatures and total monthly precipitation from the McNary station were employed in equations for gauging stations 3905, 4910, and

4960, and data from the Alpine recording station were utilized for gauging stations 3834, 3835, and

4890.7.

Gauge Number

3834

3835

3905

4890.7

4910

4960

TABLE 1

STREAM DISCHARGE GAUGES

Location

Little Colorado River at Greer, Ariz.

Nutrioso Creek above

Nelson Reservoir near

Springerville, Ariz.

Show Low Creek near

Lakeside, Ariz.

Period of Record

8/60 to present

6/67 to present

1/59 to present

North Fork of East

Fork Black River near Alpine, Ariz.

North Fork White

River near McNary,

Ariz.

6/65 to present

1/59 to present

Corduroy Creek near mouth, near Show Low,

Ariz.

9/51 to 9/75

EMPIRICAL ANALYSIS

Statistical models of streamflow of the following general form were estimated with ordinary least squares regression procedures for the streamflow series from each of the six gauging stations included in the study sample:

Fit = ai

+ bliTit

+ b2iPit + b3iSit + ui for i

= 1, 2,

3, 4, 5, 6-- corresponding to each of the six gauging station locations and where

F1 = either maximum monthly streamflow or total monthly streamflow at location "i" in month "t" ai

= the intercept of the estimated equation for location "i" incorporating the net impact upon streamflow of all the factors not explicitly included in the equation.

bli, b2i, b3i

= the estimated regression coefficients for the three explanatory variables,

Tit= average monthly temperature at location "i" in month "t ",

Pit= total monthly precipitation at location "i" in month "t ",

Sit= number of seasonal electric hookups (as a proxy for the number of second homes) in the county in which location "i" is situated in month "t ", ui = the stochastic error term for equation "i ".

The initial results obtained by estimating such regression models were disappointing.

The coefficients of determination (R2) were generally very low --in fact, the R2 statistics implied that most of the equations could explain only 1 to 2 percent of the total variance in the streamflow series.

Further, while some of the equations indicated statistically significant relationships between the climatological variables and streamflow, in no case was any association shown between second -home development and the pattern of stream discharge within the study area.

155

Both the streamflow and second -home series demonstrate strong seasonal patterns, and the character of the seasonal patterns were not similar.

The seasonal hookup data have predictable annual cycles with a peak each summer and a nadir in the winter, while the water series have much more complex patterns of variation.

Therefore, all of the data series were seasonally adjusted, and another set of regression equations was estimated employing these seasonally adjusted data.

The results of these regressions are set forth in Table 2 for the maximum flow series and in Table 3 for the total flow series.

These tables set forth the intercept term, the estimated coefficients for each of the three explanatory variables, with the corresponding t- statistic indicating the statistical significance in parentheses below each estimated coefficient, and the coefficient of determination (R2).

After seasonal adjustment of the data, the results are substantially different than the initial regression findings.

The estimated equations computed with seasonally adjusted series were able to explain a much higher proportion of the total variance of the streamflow series.

(0f course, the total variance in these data was much smaller in magnitude than the variance figures in the original series.) The computed R2 statistics imply that the models are able to explain 37 to 67 percent of the total variance in the streamflow series compared with 0 to 25 percent using the initial equations.

Turning to the findings with respect to the seasonal hookup variable --the measure employed in this investigation to assess the impact of second -home development- -the maximum flow equations demonstrate a statistically significant and positive relationship for only two of the six gauging stations.

The t- statistics for the coefficients of the second -home variable did not indicate significant impact of the number of second homes upon peak streamflow in the other four maximum flow equations.

The total flow equations provide more evidence of impacts of second -home development upon the volume of stream discharge.

The coefficients of the second -home variable in four of the six equations were found to be both positive and statistically significant.

In the estimated equation for station

4960, only the coefficient for precipitation was indicated to be statistically different from zero at even the 90 percent level of confidence.

A possible explanation for this result might be that the characteristics of the watershed of this gauging station were such that changes in the volume of precipitation swamped the influences of the other two factors.

With the other equation (station 3834), the regression results imply a significant negative relationship between second -home development and stream discharge.

This empirical finding is contrary to expectations and might be due to the specific nature of the watershed area for that particular station.

It was beyond the scope of the present investigation to become involved in detailed field study of each watershed, but such a case study methodology would be very interesting in gaining a better understanding of the development- streamflow relationship in rural mountain areas.

TABLE 2

Gauging Stations

REGRESSION RESULTS:

SEASONALLY ADJUSTED MAXIMUM FLOWS

Intercept

Coefficients

Average

Temperature

Total

Precipitation

Number of

Seasonal

Hookups

3834

456.9

-10.0

(3.99)

7.78

(1.86)

.04

(1.52)

3835

3905

4890.7

4910

4960

-

5.6

-2132.0

-

27.9

897.2

-2409.4

- 1.48

(4.93)

41.7

(3.99)

- 3.50

(5.58)

19.13

(2.96)

40.53

(1.99)

7.69

(1.75)

112.69

(6.49)

17.62

(1.91)

32.33

(3.01)

288.13

(8.52)

.09

(4.42)

(

.01

.58)

.24

(5.43)

.01

(1.10)

(

.01

.23)

R2

.44

.47

.38

.53

.45

.59

156

TABLE 3

Gauging Stations

REGRESSION RESULTS:

Intercept

3834

3835

-

24422.3

12.7

SEASONALLY ADJUSTED TOTAL FLOWS

Coefficients

Average

Temperature

-434.07

(.2.85)

- 37.23

(4.90)

Precipitation

-

Total

557.49

(2.20)

(

4.08

.03)

Number of

Seasonal

Hookups

-5.51

(3.95)

2.51

(4.66)

3905

7404.9

4890.7

4910

511.0

39863.0

-187.83

(2.52)

- 79.91

(5.22)

-821.15

(4.07)

761.42

(6.17)

-

(

8.37

.03)

316.63

(

.94)

.24

(3.06)

4.66

(4.30)

.37

(1.71)

4960

10421.2

-317.15

(1.55)

2533.16

(7.48)

(

.19

.85)

R2

.44

.37

.67

.38

.38

.67

CONCLUSION

The study results provide positive evidence that second -home development has effects on the immediate watershed.

In particular, it appears that the volume of stream discharge is increased as second home developments reallocate utilization of the land from one production form to another.

This evidence of a positive relationship was found, however, only after the data series were seasonally adjusted.

And it must be recognized that, although the initial hypotheses were confirmed, the estimated magnitude of the impacts and the statistical precision of the estimated relationships were not overly strong.

The Navopache data were the best available; however, these were a second -best proxy of trends in second -home development.

Similarly, since certain gauging stations were geographically away from second -home development, the precision of the empirical tests of the hypothesized relationships was necessarily compromised.

Thus, further study should consider development of alternate measures of both the streamflow and second -home development variables; more careful measurement should strengthen the conclusions.

It is also recognized that these conclusions apply to a narrow period of time and a small area within Arizona.

The time dimension may not be significant, but a broader base of geographical study should be considered for future research.

The popularity of second -home developments, when considered in concert with the public call for better planning of the use of our forested areas, requires this broader examination in other regions with forested areas now undergoing or available for such development.

157

REFERENCES CITED

Bond, M. E. and Dunikoski, R. H. 1977.

North Central Arizona.

The Impact of Second -Home Development on Water Availability in

Eisenhower Consortium Institutional Series, Report No. 1, Bureau of Business and Economic Research, Arizona State University, Tempe, Arizona.

Bricklen, S. K. and Utter, J. G. 1975.

of Arizona.

Impact of Recreation Use on Water Quality in the White Mountains

Final Report on Cooperative Agreement 16- 340 -CA, U. S. Forest Service, Rocky Mountain

Forest and Range Experiment Station, Ft. Collins, Colorado.

Gerking, S. D., Holmes, C. J., and VanBrackle, M. 1979.

in Northeastern Arizona.

A Short -Term Forecasting Model for Second Homes

Final Report, Grant 16- 699 -GR, U., S. Forest Service, Rocky Mountain

Forest and Range Experiment Station, Ft. Collins, Colorado.

Thompson, J.

C. and Lewis, G. D. 1973.

Rural Residential Development on Private Land in the Mogollon

Rim Area of Arizona, U. S. Forest Service, Albuquerque, New Mexico.

U. S. Department of Commerce.

National Oceanic and Atmospheric Administration.

Climatological Data,

Arizona. 63 -80, Washington, D. C.

U. S. Department of Interior.

U. S. Geological Survey. 1959 -1976.

Water Resources for Arizona.

Washington, D.

C.

Waananen, A. 0. 1969.

Urban Effects on Water Yield.

In W. L. Moore and C. W. Morgan (editors), Effects of Watershed Changes on Streamflow.

University of Texas Press, Austin, Texas.

158

CENTRAL ARIZONA PROJECT CONCEPT OF OPERATION

Frank C. Springer, Jr.,P.E. and Albert L. Graves

ABSTRACT

The Central Arizona Project (CAP), presently under construction, will convey Arizona's remaining entitlement of Colorado River water to three central Arizona counties.

As a result of the recently completed CAP Real -Time Operations Study, a concept of operation has been developed.

The concept of operations defines three types of operation beginning with an initial manned operation in 1985, a transition operation, and a permanent operation using a computer assisted remote control system.

Under the permanent operation, computer models will be run in advance to define weekly and daily pumping plant and check gate schedules.

INTRODUCTION

Scheduled to deliver Colorado River water to the Phoenix area in 1985 and Tucson in 1987, The

Central Arizona Project (CAP) is one of the largest water conveyance systems in the United States and it offers a unique opportunity for application of modern operation concepts.

Recognizing this opportunity and the need to select the proper operation concept and associated type of control system the USBR has completed the CAP Real -Time Operation Study.

The CAP concept of operation developed by the Operation Study is similar to operation concepts used for control of large power distribution systems and the California Aqueduct.

The concept centers around the use of computer models off -line in advance to generate pumping plant and check gate schedules which meet water and power operating objectives and constraints.

will be implemented daily by a computer assisted remote control system.

The operating schedules

CAP OVERVIEW

The CAP was authorized by Congress in 1968 after years of controversy, court battles, and legislative maneuvers.

The project's 300 plus miles of aqueducts, will convey Arizona's remaining entitlement of Colorado River water to Central Arizona.

The Colorado River water will be used in

Central Arizona to defer the overdraft of the groundwater supply.

Recipients of CAP water include

Municipal and Industrial users in major metropolitan areas, Indian and Non -Indian agricultural users, mining and power production interests, and wildlife and recreation entities.

Initial construction on the project began in 1971 at the Navajo Power Station in Page, Arizona.

The Federal participation in the Navajo Plant assures CAP of adequate power for operation.

The sale of available excess energy from the Navajo Station will also provide the CAP with additional revenues.

Over half of the Granite Reef Aqueduct from the Colorado River to Phoenix is completed or under construction.

Contracts for relift pumping plants, Salt -Gila Aqueduct, and associated project features will be initiated in the near future.

PHYSICAL DESCRIPTION

Havasu Pumping Plant, located on the rugged shorelin to the mouth of the Buckskin Mountain Tunnel.

of Lake Havasu will lift CAP water 820 feet

Six 500 fti /s pump units in the Havasu Plant will utilize over half of the project's power source when operating at capacity.

After passing through the 6.8 mile concrete lined Buckskin Mountain Tunnel the water will ente, the head of the open channel Granite Reef Aqueduct (GRA).

The initial 18 mile Reach of the GRA will have an operational storage capacity of approximately 4400 af.

The GRA extends 190 miles to the

Phoenix area with three intermediate pumping plants, seven inverted wash siphons, and two intermediate horseshoe section tunnels.

The relift pumping plants all have a capacity of 3000 ft3 /s with lifts of

Frank C. Springer, Jr., Chief, Operations and Maintenance Branch, Arizona Projects Office, U.S.

Bureau of Reclamation, Phoenix, Arizona.

Albert L. Graves, Civil Engineer, Operations and Maintenance Branch, Arizona Projects Office, U.S.

Bureau of Reclamation, Phoenix, Arizona.

159

101, 103, and 180 feet.

Dual radial gate check structures are located at approximately 6.5 mile intervals in the open channel sections to control water levels and flow.

in the open channel section of the GRA.

No wasteways are included

The Salt -Gila Aqueduct (SGA) begins at the terminus of the GRA with the 76 -foot pump lift at the

Salt -Gila Pumping Plant.

The SGA extends 58 miles from the Phoenix area toward Tucson.

Expected to be similar in design to the GRA with an initial capacity of 2,700 ft /s, the SGA will step down in capacity as delivery points are established along its length.

Picacho

Mountain Pumping Plant signifies the terminus of the SGA and the initial point of the

Tucson Aqueduct.

Still in the preliminary planning stages, the Tucson Aqueduct will convey CAP water the remaining 60 miles to the Tucson area.

The Operation Study which developed the operating concepts outlined herein considered only the real -time operation of the Granite Reef and Salt -Gila Aqueducts without seasonal regulatory storage features.

CONCEPTS OF OPERATIONS

Considering the timing and quantity of water delivered, and the limitations involved in completing and accepting the CAP control system, the proposed concept of operation for CAP facilities can be broken into three basic types of operation.

The initial manned operation, commencing almost immediately following completion of the project's physical facilities, should correspond to a period of low initial water deliveries.

Control of the pumping plants and aqueduct check gates will be accomplished by personnel at the pumping plants and along the aqueduct.

During this period, a shakedown of the aqueduct mechanical and electrical systems will take place, as well as "off- line" and "on- line" debugging of the control system.

The duration of the initial operation will be approximately 12 months.

At the conclusion of this phase, "final" acceptance tests will be performed by the control system contractor.

The transition operation will be characterized by increasing water deliveries, and a continuation of the final "on- line" checkout and debug of the remote control system by project personnel.

During the transition period the pumping plants and aqueduct check gates will be controlled remotely by the remote control system with manuel back -up in the field.

the remote control system should become fully functional.

After a transition period of 6 to 12 months,

At that time, most of the major distribution systems could be completed, and permanent operations can be initiated.

On an annual basis, an operating plan will be developed using the Colorado River water allocation for CAP, water and power contracts, and master water delivery schedules for the service area.

The annual plan will outline water diversion requirements at the Colorado River on a monthly basis and will identify monthly allocations of water service contracts and operating policy.

Coordination of CAP water user schedules and water availability in conjunction with Colorado

River operations will produce an annual operating plan which will be coordinated through Western Area

Power Administration (WAPA) to schedule yearly CAP power requirements and market excess Navajo energy.

Within the framework of the annual operating plan, coordination with water users and Colorado

River operations will occur on a monthly basis to generate an operating plan which will distribute each users monthly allocation into weekly delivery schedules.

by WAPA for refined power scheduling and marketing.

This updated information will be used

On a weekly basis, the water users will be required to notify the CAP Operations Center of their weekly water orders by the preceding Wednesday for the following week.

The water scheduling staff will compile and approve the water orders.

The staff will also process scheduling programs on the computer system in the Operations Center to determine a set weekly operation at the Havasu Pumping

Plant, and nominal schedules at the relift plants.

It is expected that the Havasu Pumping Plant will normally operate as scheduled during the weekly period.

This essentially firm Havasu operation will insure that CAP's portion of Navajo power can be utilized effectively, consistent with water delivery and power marketing requirements.

When emergency or unforeseen conditions arise which require CAP to utilize energy in excess of the scheduled amount, power will be obtained through other sources within the integrated Federal power system, or purchases from commercial suppliers.

Under most operating conditions, the water users will be allowed the flexibility to contact the project water scheduling staff and update the water delivery schedules on a daily basis.

The daily changes in delivery schedules must be telephoned in to the Operations Center by 9 a.m. on the preceding day to be approved and incorporated in the following days operating schedules.

160

The updated daily water delivery orders will be input as data into the scheduling computer models.

The models will be run on the computer at the Operations Center, and will define an operating schedule at the pumping plants and check gates for the following day.

The control operators will review, and approve the schedules prior to implementation by the control system.

During actual operations for the following day, conditions which deviate significantly from the projected operations (unewal:or emergengy cpnditton4)will be detected by the control system', and will generally require a response by the control system operators.

Small deviations may trigger the processing of special computer programs which evaluate conditions, and suggest an appropriate response to the operators.

More unusual problems may require operator response based on experience.

This experience will be developed by the control system operators throughout their terms as trainees during the initial and transition operations.

When unusual or emergency conditions have been cleared, the computer scheduling programs will be rerun using the current aqueduct conditions.

The revised schedules will be reviewed, approved, and implemented.

Unusual and emergency conditions will require close coordination with WAPA, and, depending on the severity of the situation, coordination with the water users and Colorado River operations.

During partial or complete power failures of the system, load shedding operations as the result of an outage at Navajo Power Plant, or various other emergency conditions, close coordination will be required with WAPA power schedulers.

user entities.

Curtailment in deliveries will require coordination with water

OPERATIONS CENTER FUNCTIONS

The Operations Center will house the control system, control system operators, water scheduling staff, and administrative, operations and technical staff.

The center will be the main control point for the entire CAP system.

The capability to transfer control to a local site will be available, but during normal conditions, the Operations Center will direct all aqueduct operations.

The control system computer will be an integral part of the Operations Center.

The computer is expected to be dual in nature for reliability.

The dual aspect of the system should extend to all parts of the computer.

The computer system will direct the collection of telemetered aqueduct system status and alarm data, process the information to check for software generated alarms, display appropriate information to the control system operators, and implement control commands.

The computer will utilize the current telemetered information to maintain an updated hydraulic representation of the aqueduct in memory at all times.

PUMPING PLANTS OPERATION

After an appropriate period to phase out pumping plant operators which will be on temporary assignment during the initial and transition periods, the pumping plants are expected to be operated remotely during normal, and most unusual and emergency operating conditions.

However, maintenance personnel will be assigned to the plant, and available for short -term emergency operations if required.

The control system will telemeter status and alarm conditions "continuously" from the plants to the Operations Center, with control commands telemetered back to the plants at the appropriate time.

During long term unusual and emergency conditions, the pumping plants may be manned by roving operator teams and operated locally in coordination with the Operations Center.

During normal operations, the roving teams will visit the pumping plants periodically to monitor the remote operations.

Havasu Pumping Plant will be operated to optimize the extent of offpeak pumping within the limits of available operational storage in Reach 1 and water delivery requirements.

The plant's operation will be on a firm weekly cycle with changes in pumping operations made only when unforeseen circumstances require deviations.

The relift plants will be operated under a variety of water -power operating concepts.

Various computer models will be used at different times to specify in advance the desired relift plant operations.

Three types of operations are currently defined.

Optimal operation (maximize offpeak pumping) of the relift plants may be possible during certain flow ranges using water /power optimization models.

A non -optimal, onpeak, offpeak operation may be specified using the constant volume concept model.

The pumping plants could also be operated on a baseload concept, cycling pumps to match downstream delivery requirements using simplified operating models.

This concept will be used during initial operations.

In addition, the baseload concept would be required at high aqueduct flows, when water deliveries are scheduled on a pro- ration basis.

Regardless of the concept utilized, the operation of the relift plants will be defined nominally on a weekly basis, and updated daily to reflect changes in delivery orders.

161

CHECKGATES OPERATIONS

The check structures will be operated remotely to control pool water levels and flows.

The computer scheduling models will develop a daily schedule of checkgate movements consistent with the water -power concept incorporated in the models.

The computer models will prescribe gate operations which will assure that specified flows or levels are maintained.

As with the pumping schedules, once a daily schedule of gate movements are defined by the models, the control system operators will review the schedule for errors, make changes where necessary, and implement the schedules in the control system.

The following day, the control system will automatically move the gates according to the schedules, unless directed otherwise.

During emergency situations, control of the gates can be made remotely, or locally at the check structures.

The control system operators will make gate changes as suggested by emergency operating programs or by personal experience.

After an emergency condition has been cleared, the scheduling models will be updated and rerun, defining a new pumping plant and gate schedule which will be implemented through the control system.

TURNOUTS OPERATION

The gated turnouts will be operated remotely by the control system.

Flow changes will be made by the control system at the times scheduled by the water users, consistent with operating policy.

Should a change be required in the planned turnout schedule, the control system operators will make the change remotely.

At the non -gated turnouts, flowmeter readings will be telemetered to the

Operations Center and displayed for operator surveillance.

PERSONNEL REQUIRED

The control system operator staff will be large enough to have two men on duty at all times.

The operator staff will also have other duties that can be performed when the system is functioning normally.

day.

At least one member of the water scheduling staff will be on duty 7 days a week, one shift per

This staff will compile monthly, weekly, and daily water orders and coordinate with the water users, WAPA, and with Colorado River operations.

The control system and Water Operations Division will be supported by an engineering technical staff familiar with the control system hardware and software, the aqueduct electrical and mechanical systems, and the operating principals and concepts.

This staff will consist of a number of electronic engineers, civil engineers, and support technicians.

The engineering technical staff will assist the operating staff with major problems, perform on -going studies, and evaluate overall system effectiveness.

This staff will also make software improvements to the control system as necessary.

The two roving operator teams are expected to consist of two men each.

The normal procedure for the teams will be to visit individual pumping plants on a regular basis evaluating and monitoring the remote operation.

The operating personnel required under permanent operations will be somewhat less than under the initial or transition operations.

The staff reduction will be accomplished by returning temporary personnel to their respective projects, transfers to the Tucson Aqueduct, shifting jobs to become control operators, and by attrition in the control system operator staff.

The permanent control system operator staff is expected to consist of twelve system operators who will man the control system at the Operations Center at all times, four system operators who will form roving operating teams and four system operators who will work with the water scheduling staff in developing daily operating schedules.

COMPUTER MODELS

The current concept of computer models fall into three general categories of operation specification programs, operation design programs, and simulation programs.

The operation specification programs are computer models which define target depths or target flows throughout the aqueduct as a function of time over a 24 -hour period.

These programs can define an array of operation specifications.

The two models currently used define an optimum water -power operating concept and a constant volume operating concept.

Other specification programs may be written to define operations under other operating concepts.

162

The operation design programs make up the heart of the "off-line" concept of operation.

The design programs will use the specified operating criteria (flow as a function of time or depth as a function of time in each pool) to design operating schedules for the pumping plants and check structures.

These "designed" operating schedules will reflect operating constraints of limited hydraulic transient generation and operating objectives of limited water level fluctuations in the pools.

The simulation program represents the actual aqueduct.

The simulation program uses approximate mathematical techniques to model gradually varied unsteady state flow in open channels.

This technique has been thoroughly tested over a number of years.

WPOM

The Water -Power Optimization Model (WPOM) falls under the category of operation specification program.

The WPOM is currently under development for use under actual operations.

Using operation research techniques, WPOM optimizes the use of aqueduct prism operational storage to minimize onpeak energy use.

The optimization process is constrained by the operating constraints (primarily downstream rates)(Yeh, 1979).

CVM

The Constant Volume Model (CVM) is also categorized as an operation specification program.

Using turnout demands, initial conditions, and target storage values for each pool, the CVM develops pump schedules and target pool levels for onpeak and offpeak time periods.

Although the constant volume approach does not produce an optimal onpeak, offpeak operation, the approach does provide for a great deal of operational flexibility.

With the aqueduct prisms at a relatively constant target storage throughout the day, flow changes can be made throughout the system almost instantaneously.

In contrast to the traditional operation which requires long lag times for introducing flow changes and no onpeak, offpeak pumping, the constant volume concept allows for pumping and check gate flow changes to be made simultaneously with acceptable pool level fluctuations.

The concept is not workable for medium to large flow changes due to the variation in pool levels during the flow changes.

Under the constant volume concept, low flows tend to have higher checked water levels and high flows tend to have lower checked water levels.

This variation is due to the difference in wedge storage between high and low flows.

GSM

Under the category of operation design programs, one Gate Stroking Model (GSM) was formulated for use in connection with the WPOM (Yeh, 1979) and one GSM was developed for the CVM (Falvey, 1979).

The GSM developed for WPOM uses specifications of flow as a function of time at each check structure and a finite difference -staggered net method to arrive at gate schedules which minimize hydraulic transient activity.

The GSM related to CVM used depth as a function of time as a specification and a characteristic grid method to generate gate schedules which minimize hydraulic transient activity

(Bobley, 1978) (Wylie, 1969).

Do to the nature of the gate stroking concept, an open channel flow problem cannot be completely specified.

One degree of freedom is necessary for the solution procedure to progress.

The degree of freedom allowed within the GSM is depth in each pumping plant afterbay pool.

This pool is allowed to

"float" according to the input specifications for the other pools and initial conditions.

ASM

The Aqueduct Simulation Model (ASM) falls into the category of simulation programs.

The model uses the method of specified time intervals (a particular type of method of characteristics solution) to solve the partial differential equations defining gradually varied unsteady flow (Shand, 1971).

The model has boundary routines which model checkgates, pumping plants, turnouts, and changes in section.

The model can also perform calculations in trapezoidal sections, round sections, and horseshoe sections.

163

USE OF COMPUTER MODELS

Under actual operations, a version of ASM will be running at all times in the computer operating system using telemetered water levels, gate opening, and pump flows from the aqueduct.

The ASM calculated water surface and flow profiles will be used by the control system to generate certain software alarms.

At 10 a.m. each day; the ASM will run on the control system using projected turnout schedules, actual 10 a.m. hydraulic conditions, and the gate and pump schedules for that day.

This run will yield the projected conditions in the system as of midnight that day.

From these projected conditions and the turnout schedules for the next day, an operation specification program

(WPOM or CVM, etc.) will generate the desired operation specifications for the next day.

operation design program (GSM) will then use the projected initial conditions and operation

The specifications to design the gate and pump schedules for the next day.

Initially, the model generated pump and gate schedules will be evaluated by the ASM using the projected initial conditions.

Should these evaluations indicate an undesirable operation has been designed, the operation specification and operation design programs will be rerun with proper biasing at the problem areas.

It is expected that operating experience with the specification and design programs will eliminate the need for a daily ASM evaluation of the operating schedules.

Once the next days schedules have been evaluated by the computer models, the schedules will be made available to the control system operators.

The daily gate and pump schedules for the next day are expected to be available to the control system operators by no later than 4 p.m. each day.

The control system operators will visually inspect the schedules on the control system display screens and authorize the system to implement the schedules at the proper time the next day.

164

REFERENCES CITED

Bobley, W.E. and Wylie, E.B.

1978.

Control of Transients in Series Channel with Gates.

Journal of the Hydraulics Division, ASCE(HY10):1395 -1407.

October.

Falvey, H.T. and Luning, P.C.

1979.

Gate Stroking.

Internal Report (to be submitted for publishing), Engineering and Research Center, Bureau of Reclamation, Denver.

55p.

Shand, M.J.

1971.

Final Report, Automatic Downstream Control Systems for Irrigation Canals,

Technical Report HEL -8 -4.

Hydraulic Engineering Laboratory, University of California at

Berkeley.

159p.

Wylie, E.B.

1969.

Control of Transient Free -Surface Flow.

Journal of the Hydraulics Division,

ASCE 95(HY1):347 -361.

Yeh, W.G., Becker, L., Toy, D., Graves, A.L.

Project.

Unpublished.

1979.

Operations Model for the Central Arizona

165

THE IMPACT OF SOCIOECONOMIC STATUS ON RESIDENTIAL WATER USE:

A CROSS- SECTION TIME -SERIES ANALYSIS OF

TUCSON, ARIZONA

R. BRUCE BILLINGS AND DONALD E. AGTHE

ABSTRACT

The impact of a selected set of socioeconomic variables on residential water consumption per household is examined using a combined cross -section time -series analysis by census tract for Tucson, Arizona for 1974, 1975, 1976 and 1977.

The estimated income elasticity of demand for water is .23, which means that a 10- percent increase in income produces a 2.3 percent increase in water use.

Additionally, the number of persons per household and the percent of households with head age 65 or more also are shown to have a strong positive relationship to water use.

New residential units are shown to have a strong tendency to utilize less water than older units, presumable because of a shift away from water using yards.

Both Black and Spanish- surnamed dominated areas tend to consume a lower than expected amount of water for their income and family size characteristics, but the coefficients on these variables are not sufficiently strong to accept this relationship.

INTRODUCTION

The purpose of this study is to measure the relationship of selected socioeconomic variables to residential water consumption in Tucson, Arizona.

A knowledge of this relationship should help to improve water system planning and policy.

A cross -section time -series analysis is utilized to make estimates of the relationship between household water consumption and personal income, the proportion of relatively new housing units, household size, the proportion of elderly persons, and the racial composition of the population.

Since the data used in this study were collected by census tracts, all of which are served by the City of Tucson, all of the water customers faced the same price structure for water so the influence of price changes on water consumption was not examined.

Although Tucson is a Southwestern city with its own unique temperature and rainfall patters, it is likely that the results found for Tucson will be similar to those in many other arid urban areas, and especially those with relatively hot climates.

Using only one city for analysis tends to eliminate intercity differences in temperature and rainfall which require somewhat tenuous adjustments for multi -city cross -section studies ies.

A few socioeconomic variables have been related to water consumption in past stud-

A study by Darr (1975) for Israel found that income and household size were the only statistically significant socioeconomic variables influencing water consumption in each of the urban areas studied.

Age of the head of household and cultural origin were also significant for about one -half of the cities included in the Israel study which dealt only with Jewish households.

Tucson, like Israel, has citizens with multiple cultural backgrounds and an arid climate.

Morgan's (1973) study of water scarce Santa

Barbara, California found that both income and number of persons per household are statistically significant determinants of household water consumption.

A recent water price elasticity study of Tucson by Young (1973) did not include socioeconomic variables.

The authors are respectively, Lecturer in Economics, University of Arizona and

Assistant Professor of Economics, St. Mary's University of San Antonio.

167

METHODOLOGY

The relationship between residential water consumption and selected socioeconomic variables was examined using cross- section time -series analysis of census tracts served by the Tucson, Arizona Water System for the years 1974, 1975, 1976 and 1977.

The data included annual water consumption per household for all individually metered single family detached dwelling units, townhouses, condiminiums, mobile homes, duplexes and triplexes for this period.

The socioeconomic variables utilized to explain variations in water use included average family income, average household size, the percentage of households with the head age 65 or over, the percentage of Black households and the percentage of Spanish surnamed households.

In addition, the percentage increase in the number of water connections in each census tract was used as a measure of the effect of new residential units.

Census tracts in which less than 25- percent of the estimated population is served by the water utility and those tracts in which total water system connections increased by over 100 percent during the study period were omitted from the analysis.

This was done to avoid a possible bias in the data which could result when the water customers ara not properly represented by the socioeconomic data from the census.

Seven of the

57 tracts served by the city water system were omitted from the study, leaving 50 census tracts for the regression analysis.

The omitted areas are inner city tracts composed of apartment complexes with centrally supplied water, outlying tracts served by private water systems, or tracts marked by very rapid housing development.

In the regression analysis, each census tract was weighed by the number of households served by the water system.

This makes the individual household rather than the census tract the underlying unit of analysis, avoiding any bias toward the water use patterns of lower population tracts which would result from weighing each tract equally.

The number of cases used in determining the statistical significance of results after weighing equals 200.

The following model was tested using multiple regression analysis:

(1)

W= a0+ alY+ a2G+ a3H +a4A +a5B +a6S +a7DA +a8DB +a9DC +u

The model was also formulated in multiplicative form, on which a log transformation was performed to estimate the income elasticity of demand for water.

(2)

W =e(b0

+b7DA +b8DB +b9DC),Yb1 0b2 Hb3 Ab4 13b5 sb6

(3) log W =b0 +bi log Y +b2log G +b3log H +b4log A +b5log B +b6log S +b7DA +b8DB +B9DC +u where:

W

Y

Gi

H

A

B

S

DA

DB

DC is annual average water consumption (100 cubic feet) per household for single residential water connections for years 1974, 1975, 1976 and 1977 by census tract (Tucson 1978).

is average family income (thousands of dollars) (U.S., Table P -4).

is percentage growth in single residential unit water connections from January 1974 to:

1) the 1974 average number of connections;

2) the 1975 average number of connections;

3) the 1976 average number of connections;

4) the 1977 average number of connections (Tucson, 1978).

is average household size (U.S., Table H -1).

is the percent of households whose head is 65 or older (U.S., Fourth Count

Tape).

is the percent of heads of households who are Black in census tracts with over 400 Black population (U.S., Table H -3).

is the percent of heads of households who are Spanish- surnamed in census tracts with over 400 Spanish- surnamed population (U.S., Table H -3).

is the Dummy Variable for 1975 (1975 =1).

is the Dummy Variable for 1976 (1976 =1).

is the Dummy Variable for 1977 (1977 =1).

Water use per residential connection is used as the dependent variable.

This is the total measured water use of all individually metered residential customers in each census tract, divided by the average number of active water customers.

Average annual household water use is shown in Figure 1 for each census tract and each year included in this research.

168

As shown in previous studies, income has a positive impact on residential water use (Gottlieb, 1963; Darr, 1975; Morgan, 1973; Headly, 1963).

Average family income was used as the measure of economic well -being and the ability to consume.

People with higher incomes may use more water due to their greater economic ability and the fact that water is a smaller portion of their budgets.

Higher income families also tend to own larger homes with multiple bathrooms and an above average number of water using appliances, such as dishwashers and clothes washers.

These more expensive homes are often on large lots with swimming pools, extensive irrigated lawns, and other water hungry plantings.

Because of their anticipated impact on water consumption, several additional indicators of income and wealth were tried and discarded because of their high multicollinearity with average income, which explained water use better than any other available measure of economic status.

These indicators included level of education, number of bathrooms, and house value.

Median income was tried as an alternative to average income but it did not show the strength of average income in the multiple regression analysis.

Changes in the characteristics of the housing stock may influence water consumption, particularly if new units are significantly different from existing units.

To test the hypothesis that housing units added during the study period are different from previously existing units, a growth variable defined as the percentage increase in residential water connections from January 1974 to the annual average for each year was computed.

Census tract data was unavailable prior to January 1974.

Average water consumption is expected to decline as the percent of new connections becomes larger because a large proportion of the new connections represent individual mobile homes, town houses, and condominiums, very few of which have extensive outside yards with their associated sprinkling demands.

The number of individually metered town houses and condominiums increased by 18 percent between the third quarter of 1974 and the last quarter of 1976.

In contrast, the number of conventional homes increased by only six percent (Tucson Trends, 1976).

Many of the new homes which were built during this period featured desert landscaping in place of the conventional water hungry lawns and shrubs.

A high proportion of new homes also have refrigerated air conditioning, which consumes little or no water, instead of the traditional evaporative cooling which uses large quantities of water on hot summer days.

The growth variable measures the impact of these changes on aggregated water consumption by census tract.

Average household size is included to measure the impact of the number of individuals served by each water connection.

The coefficient on this variable shows the increase in water use for each additional household member, and is expected to be positive since large families tend to use larger amounts of water.

Household size accounts for only part of all water use, however, since a substantial amount of water is used for purposes such as lawn sprinkling which are more closely related to the existence of the household than to its size.

A significant coefficient on the household size variable would direct planners to consider the anticipated changes in household size by neighborhood in addition to the expected changes in the number of connections in their projections of future water use.

The percentage of households with heads age 65 or over was included as an independent variable since Tucson is a reputed retirement community.

It is possible that these individuals may have different water use patterns than the remainder of the population.

These households may consume less water because their members are older and less active, or they may consume more water since they are home more often, have more time for gardening and may have unchanged water consumption habits which were developed in pre- retirement years in the less arid Eastern areas of the United States.

Tucson has a substantial Spanish- surnamed population and some Black residents.

The proportion of these minority group members in each census tract was included to see if varying cultural heritages influenced water consumption.

The racial composition of each census tract with respect to Black and Spanish- surnamed population was included, without any a priori notion of their probable effect on water consumption, in order to find out if they in fact do have different water use patterns than the remainder of the urban population.

Because there is no theoretical reason to expect their water consump tion to vary from the average, very strong statistical results would be required before any positive hypothesis about their behavior could be accepted.

Since there were changes in water prices, rainfall, and water conservation publicity over the four year period that could not reasonably be accounted for in the annual data, it was necessary to include a set of dummy variables (DA, DB, DC) to account for these intertemporal differences.

Price changes would be difficult to include since they occurred at various times within each year, and it was felt that the use of some averaging technique would misrepresent the price changes because of the presence of

169

only four annual observations.

There is also some evidence of a shift in preferences toward desert landscaping, resulting in a downward shift in the entire demand schedule.

During the course of the study, it was decided to "try on" the percent of the population which had moved into the metropolitan area from outside during the past five years to find out if these people tended to consume more or less water than longer term residents.

It was thought that they might tend to consume more water since many of them came from the wetter Eastern and Great Lakes sections of the country.

In running the equations, however, the coefficient on this variable was not significant, and it was dropped from the model.

RESULTS

(4)t

W= 71.89 + 4.842Y - .2644G + 28.05H + 2.038A - .4884B

(4.79) (9.20)

(-3.61) (4.32) (6.72)

(-2.71)

- .1209S - 11.6ODA - 41.47DB - -70.80DC

( -1.19) ( -3.61)

(- 11.53) (- 19.74)

R2 = .82

(5)

F = 96.06

log W= 4.18 + .2232 logY - .0136 logG + .3987 logli + .1490 logA

(34.47) (7.40) ( -3.24)

(7.42) (7.09)

- .0237 logB - .0204 logS - .0413DA - .1845DB - .3508DC

(-3.26) (-2.63)

(-2.35)

(-10.24) (20.02)

R2 = .82

F = 99.42

The models yeilded the preceeding results.

As expected, income was found to strongly influence water consumption, with consumption increasing about 484 cubic feet per household per year for each $1,000 increase in income.

The estimated income elasticity of demand for water is .22, indicating that for each 10 percent increase in household income, water use can be expected to increase by 2.2 percent.

Using the average values for income and water consumption, the linear equation yeilded an income elasticity of .25, which is very similar to the logarithmic results.

Table 2 provides a comparison of the income elasticity found in this Tucson study with those found by other studies of water scarce areas.

The elasticities found in the

Tucson study are similar to those shown in studies of Israel and San Francisco while being somewhat smaller than those for Santa Barbara and near the low end of the spectrum of elasticities in the Kansas study.

The coefficient on the percentage of new dwelling units connected to the water system is negative and significantly different from zero at the 95 percent level.

This result strongly confirms the hypothesized relationship between the proportion of newly constructed residential housing in an area and its average water use.

TABLE 2

AREA

Tucson

Kansas

Gottlieb, 1963)

San Francisco -Oakland

(Headly, 1963)

Income Elasticity of Demand for Urban Water,

Selected Studies of Water Scarce Areas

INCOME ELASTICITY

TYPE STUDY

.22 to

.25

Cross Section

.28 to

.58

Cross Section

.00 to 40

Time Series

.33 to .50

Cross Section

Santa Barbara

(Morgan, 1973)

Israel

.18 to

.34

Cross Section

YEAR

1974 -77

1963

1950 -59

1973

1975

170

As expected, water use increased strongly with the number of persons per household, and the coefficient for household size was significant at the 99 percent level.

The

Morgan (1973) study found household size elasticity coefficients of .31 and .37 which is similar to the .39 for the logarithmic equation for Tucson.

These household size elasticity coefficients cannot be compared with those found in the study of Israel because that study used water consumption per capita as its dependent variable (Darr,

1975).

The use of per capita water consumption produced negative household size elasticity coefficients because of economies of scale in household water use.

The most interesting result obtained in this research is the positive relationship between the percent of households whose head is at least 65 years old and water consumption.

The coefficient on this variable is significant at the 99.5 percent level, providing strong support for the hypothesis that people it census tracts with a larger proportion of elderly residents do use more water than expected on the basis of their income, family size and other characteristics alone.

This result may be due to the large proportion of the elderly who are retired and therefore spend more time at home, some of which is devoted to gardening and yard maintenance, which requires large amounts of water.

The positive relationship between water use and elderly residents may also result from these households spending on the basis of their permanent incomes rather than their much lower current retirement incomes.

The coefficient on the percent of Black population in each census tract is negative and statistically significant at the 90 percent level.

However, the linear equation indicates a decrease in consumption of only 49 cubic feet per one percent increase in Black households.

Since there is no a priori reason to expect that predominantly

Black neighborhoods consume either more or less water than other areas, no important conclusion about their water use can be drawn from these results.

The coefficients on S, the percent of Spanish -surnamed households, is negative in the log equation and is statistically significant at the 90 percent level.

However, this variable is not significant at this level in the linear equation.

Therefore, no conclusive result concerning the water use patterns of these households can be drawn from this model alone.

The strong result of the intertemporal dummies suggest that annual changes in rainfall, prices and water conservation attitudes are also important.

Since price changes occurred in mid -year, rainfall exhibits great seasonal variation which must be correlated with such factors as evapotranspiration, and water conservation news was varied throughout the year, it is thought best to leave these variable out of the current study until more than four data points become available.

However, the strong results from the dummy variables suggest that research of time -series effects of price, rainfall and conservation publicity should be fruitful.

CONCULSIONS

This study examined the relationship of domestic water consumption with a selected set of socioeconomic variables using a cross- section time series analysis by census tract for Tucson, Arizona for each of four sequential years.

The model employed was able to explain a large part of the variation in average household water use.

Water use is shown to increase by 484 cubic feet for each $1,000 rise in household income.

The estimated income elasticity of demand for water is .22.

The study also shows that newer residential units tend to be more water efficient, primarily because of reduced use of lawns and water consuming plantings.

The strong positive relationship between household size and water use was anticipated.

The highly significant positive coefficient found for the percent of households whose head is age 65 or more was the most interesting finding of this research project.

The regression results for the percent of households with Black or Spanish- surnamed heads of household provide weak support for the hypothesis that these groups tend to consume less water than expected on the basis of their incomes and family size.

171

REFERENCES CITED

Arizona Department of Economic Security, unpublished data, 1977.

Darr, Peretz, Stephen L. Feldman, and Charles S. Kamen, "Socioeconomic Factors Affecting

Domestic Water Demand in Israel," WATER RESOURCES RESEARCH, Vol. 9, No. 6

(December 1975):

805 -809.

Gottlieb, Manual, "Urban Domestic Demand for Water:

A Kansas Study,"

LAND ECONOMICS,

XXXIX, No. 2 (May 1963):

204 -210.

Headly, J. Charles, "The Relation of Family Income and Use of Water for Residential and

Commercial Purposes in the San Francisco -Oakland Metropolitan Area," LAND ECONO-

MICS, XXXIX, (November 1963):

441 -449.

Morgan, W.D., "Residential Water Demand:

The Case from Micro Data," WATER RESOURCES

RESEARCH, Vol. 9, No. 4

(August 1973): 1065 -67.

Tucson, Arizona, Department of Water and Sewers, unpublished data, 1978.

TUCSON TRENDS, 1976, p. 25.

U.S. Bureau of the Census, CENSUS OF POPULATION AND HOUSING: 1970, CENSUS TRACTS,

TUCSON, ARIZONA SMSA.

Young, Robert A., "Price Elasticity of Demand for Municipal Water: A Case Study of

Tucson, Arizona," WATER RESOURCES RESEARCH, Vol. 9, No. 4 (August 1973):

1968 -72.

172

* *

Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement