Variability of Intermittent Headwater Streams in a Boreal Landscape

Variability of Intermittent Headwater Streams in a Boreal Landscape
Variability of Intermittent Headwater
Streams in a Boreal Landscape
Examensarbete vid Institutionen för geovetenskaper
ISSN 1650-6553 Nr 248
Influence of different discharge conditions
Tum Nihm
Dynamic expansions and contractions of stream networks can play an important
role for hydrologic processes as they can connect different parts of the
landscape to the stream channels. However, we know little about the temporal
and spatial variations of stream networks during different flow and wetness
conditions. This study focuses on the contraction and expansion of stream
networks during different flow conditions in the boreal Krycklan catchment,
located in Northern Sweden. The stream network and initiation points were
extracted from a gridded digital elevation model (DEM) of 5-meter resolution,
and then compared with the stream network initiation points (heads) observed
during the spring flood (freshet) period in 2012. From the results of the study, it
was clearly seen that the observed stream heads and the stream heads appearing
in the stream network map extracted from DEM did not agree very well. 49% of
the total observed stream heads (49) fell onto the low order stream branches and
headwater streams derived from the DEM. Only few of them exactly matched
the modeled stream heads. Moreover, the modeled stream network was much
denser than the observed stream network, and so the simple raster based
dynamic model developed could not well represent the dynamic stream network
extension in the real system. Most headwater streams in the study catchment
were man-made ditches, which were dug to drain water wetlands and to
increase forest productivity. The majority of observed stream heads were
formed by seepage from the saturated surrounding soils, while only a few of
them were formed by saturation overland flow. On the other hand, the dynamic
stream network derived from the DEM suggested that the number of streams of
lower order and their lengths was sensitive to change in streamflow, especially
during the high flow episode.
Uppsala universitet, Institutionen för geovetenskaper
Examensarbete E1, Hydrologi, 30 hp
ISSN 1650-6553 Nr 248
Tryckt hos Institutionen för geovetenskaper,
Geotryckeriet, Uppsala universitet, Uppsala, 2012.
Variability of Intermittent Headwater
Streams in a Boreal Landscape
Influence of different discharge conditions
Tum Nihm
Examensarbete vid Institutionen för geovetenskaper
ISSN 1650-6553 Nr 248
Variability of Intermittent Headwater
Streams in a Boreal Landscape
Influence of different discharge conditions
Tum Nihm
Copyright © Tum Nihm and the Department of Earth Sciences Uppsala University
Published at Department of Earth Sciences, Geotryckeriet Uppsala University, Uppsala, 2012
Abstract
Variability of intermittent headwater streams in a boreal landscape:
Influence of different discharge conditions
Tum Nhim
Department of Earth Sciences, Uppsala University
Villavägen 16, SE-752 36 Uppsala, Sweden.
Dynamic expansions and contractions of stream networks can play an important role for
hydrologic processes as they can connect different parts of the landscape to the stream channels.
However, we know little about the temporal and spatial variations of stream networks during
different flow and wetness conditions. This study focuses on the contraction and expansion of
stream networks during different flow conditions in the boreal Krycklan catchment, located in
Northern Sweden. The stream network and initiation points were extracted from a gridded digital
elevation model (DEM) of 5-meter resolution, and then compared with the stream network
initiation points (heads) observed during the spring flood (freshet) period in 2012. From the
results of the study, it was clearly seen that the observed stream heads and the stream heads
appearing in the stream network map extracted from DEM did not agree very well. 49% of the
total observed stream heads (49) fell onto the low order stream branches and headwater streams
derived from the DEM. Only few of them exactly matched the modeled stream heads. Moreover,
the modeled stream network was much denser than the observed stream network, and so the
simple raster based dynamic model developed could not well represent the dynamic stream
network extension in the real system. Most headwater streams in the study catchment were manmade ditches, which were dug to drain water wetlands and to increase forest productivity. The
majority of observed stream heads were formed by seepage from the saturated surrounding soils,
while only a few of them were formed by saturation overland flow. On the other hand, the
dynamic stream network derived from the DEM suggested that the number of streams of lower
order and their lengths was sensitive to change in streamflow, especially during the high flow
episode.
Keywords
Dynamic stream network, intermittent streams, headwater stream, stream heads, wetness
conditions, streamflow
i
Referat
Variabilitet av periodiskt återkommande bäckar i ett borealt
landskap: Betydelse av olika avrinningsnivåer
Tum Nhim
Institutionen för geovetenskaper, Uppsala universitet
Villavägen 16, 752 36 UPPSALA
Utbredningsdynamiken av bäckars flödesnätverk kan spela en viktig roll för hydrologiska
processer eftersom de kan ansluta olika delar av landskapet till bäckar och älvar. Vi vet dock
väldigt lite om de temporala och spatiala variationerna i utbredningunder olika flödes- och
fuktförhållanden. Denna studie fokuserade på utbredningsdynamiken av flödesnätverk under
olika flödesförhållanden i Krycklans avrinningsområde, beläget ca 50 km nordväst om Umeå i
Västerbotten. Bäckarnas nätverk- och initieringspunkter identifierades från en digital höjdmodell
(DHM) med 5-meters upplösning och jämfördes sedan med samma bäckars initieringspunkter
(källor) observerade i fält under vårfloden 2012. Från resultaten av studien var det tydligt att de
observerade och simulerade initierings punkterna inte stämde bra överens med varandra. 49% av
de totalt observerade initieringspunkterna (49) utgjordes av bäckar med låg strömordning och på
källsflöden hämntat från DHMet. Endast ett fåtal av dem överenstämde exakt med de simulerade
initieringspunkterna Dessutom var det simulerade flödesnätverket mycket tätare än det
observerade, och den enkla raster-baserada dynamiska modellen representerde dåligt det
dynamiska flödesnätverket som observerades i fält. De flesta källflöden i avrinningsområdet är
diken som grävts för att dränera våtmarker och för att öka skogsbrukets produktivitet.
Majoriteten av de observerade initieringspunkter har bildats av läckage från den omgivande
mättade marken, medan endast ett fåtal av dem har bildats av ytavrinning på mättad mark. Å
andra sidan visade det simulerade dynamiska flödesnätverket att antalet bäckar av lägre ordning
och deras utbredningslängd var känsliga för förändringar i flödesförhållanden, speciellt under
högflödesepisoder.
Sökord
Dynamiska flödesnätverk, periodiskt återkommande bäckar, källflöden, källor, vätaförhållanden,
avrinning, flöde
ii
Abbreviations
DIC
DOC
DEM
GPS
LDD
m.a.s.l
Dissolve Inorganic Carbon
Dissolved Organic Carbon
Digital Elevation Model
Global Positioning System
Local Drain Direction
Meter Above Sea Level
iLL
Table of contents
1. Introduction ................................................................................................................................. 1 1.1. Research objectives .......................................................................................................... 1 1.2. Key research questions ..................................................................................................... 1 2. Background.............................................................................................................................. 2 3. Materials and Methods ................................................................................................................ 4 3.1. Study location ...................................................................................................................... 4 3.2. Derivation of terrain indices and flow directions ................................................................ 6 3.3. Development of a dynamic stream network model ............................................................. 7 3.4. Field investigation for mapping stream heads ................................................................... 10 3.5. Testing of model parameters’ value ................................................................................... 12 4. Results ....................................................................................................................................... 12 4.1. Stream network and frequency of surface flow generating ............................................... 13 4.2. Stream network and geology of the study catchment ........................................................ 17 4.3 Modeled stream length and stream orders during different flow conditions ...................... 19 4.4. Field observations .............................................................................................................. 23 4.5. Comparison of observed and predicted stream heads ........................................................ 25 4.6. Slope-area relationship........................................................................................................... 34 5. Discussions ............................................................................................................................... 35 5.1. Comparison of observed and simulated stream heads ....................................................... 35 5.2. Variation of stream length and stream orders at various flow conditions ......................... 35 5.3. Model uncertainties ............................................................................................................ 36 6. Concluding remarks .................................................................................................................. 36 7. Suggestions for future work ...................................................................................................... 37 8. Acknowledgements ................................................................................................................... 38 References ..................................................................................................................................... 39 Appendix 1: Field protocol ............................................................................................................. a Appendix 2: Dynamic stream network model’s scripts .................................................................. c Appendix 3: Coordinates and descriptions of observed stream heads ............................................ h Appendix 4: Field data records ........................................................................................................ j List of tables and figures
List of figures
Figure 1: DEM and geology class map of the Krycklan catchment ............................................... 4 Figure 2: Krycklan catchment and flow gaging sites...................................................................... 5 Figure 3: LDD code for direction of runoff .................................................................................... 6 Figure 4: Average daily flow series at site 7................................................................................... 9 Figure 5: Investigated subcatchments of site 9, 7, and 1 .............................................................. 11 Figure 6: >90% frequency map of surface flow generating ......................................................... 13 Figure 7: >90% frequency map of connecting surface flow ......................................................... 14 Figure 8: Stream network map at maximum streamflow.............................................................. 15 Figure 9: Stream network map at minimum streamflow .............................................................. 16 Figure 10: >90% frequency map and geology of site 9 ................................................................ 17 Figure 11: >90% frequency map and geology of site 1 ................................................................ 18 Figure 12: > 90% frequency map and geology of site 7 ............................................................... 19 Figure 13: Relationship between total stream length and flow..................................................... 20 Figure 14: Relation between total stream length and corresponding streamflow ......................... 21 Figure 15: Total stream length versus different stream orders ..................................................... 21 Figure 16: Total stream length versus different stream orders ..................................................... 22 Figure 17: Stream head with seepage erosion............................................................................... 24 Figure 18: Stream head with saturated overland flow .................................................................. 24 Figure 19: Stream head with seepage from saturation .................................................................. 25 Figure 20: Observed stream heads within subcatchment 7 ........................................................... 26 Figure 21: Observed stream heads in catchment of site 9 ............................................................ 27 Figure 22: Observed stream head in catchment of site 1 .............................................................. 29 Figure 23: Observed stream heads in other subcatchments .......................................................... 30 Figure 24: Observed headwater streams ....................................................................................... 32 Figure 25: Relationship of local slope/source area of stream head for man-made ditches........... 34 Figure 26: Relationship of local slope/source area of stream head for natural stream ................. 34 List of tables
Table 1: Flow gaging sites in Krycklan catchment......................................................................... 6 Table 2: Statistical characteristics of the flow series .................................................................... 10 Table 3: Model calibration for b value ......................................................................................... 12 Table 4: Model calibration for Zmin value ..................................................................................... 12 Table 5: Modeled total stream length during highest and lowest flow conditions ....................... 20 Table 6: Observed stream heads ................................................................................................... 23 Table 7: Formation of stream heads.............................................................................................. 23 Table 8: Number of surveyed stream heads falling within subcatchment of site 7 ...................... 26 Table 9: Number of stream heads falling within catchment of site 9 ........................................... 28 Table 10: Number of stream heads falling within catchment of site 1 ......................................... 30 Table 11: Number of stream heads falling within other subcatchments ....................................... 31 Table 12: Characteristics of headwater streams............................................................................ 33 1.Introduction
Stream heads are important transition points since, similar to riparian zones, they are at the
interface between hillslopes and stream channels (Montgomery & Dietrich, 1989). Connecting
the stream heads to the downslope stream networks, headwater streams maintain many
hydrological and biogeochemical processes that influence, for example, stream chemistry, flow
and wetness conditions in the catchment. More importantly, these controlling processes may
vary according to topographic conditions, geological characteristics and different landuse in a
basin. One challenging aspect in the study of where the stream initiates is that little is known
about the dynamic temporal and spatial variations of the stream networks since they often occur
at relatively small scales, and in mapped parts of the stream network that are seldom mapped.
There might be large differences in the number of active zero or first order streams, with varying
flow conditions. For example, during high flow episode, the stream network might considerably
expand resulting in many active low-order streams, which in turn would be likely to increase
transport of sediments, organic matter, or other stream solutes. At low flow conditions, on the
other hand, the stream network would contract and, as a result, the connectivity between
landscape and surface waters would decrease.
Many studies use DEMs to extract and examine the stream networks such as O'Callaghan &
Mark (1984), Mackay & Band (1998), McMaster (2002), and Hancock & Evans (2006).
However, because of different resolutions of DEMs and variable topography and landscape, the
stream initiation points may be inaccurately mapped. In this study, to better understand the
controlling factors in headwater streams and dynamic extension of the stream networks, DEMderived stream heads were compared to stream heads that were mapped during field visits. The
relationship between local slope and source area of the stream heads was also discussed
regarding different formations and geology to which those stream heads belong.
1.1.
Researchobjectives
The spatial extent of stream networks can play an important role in controlling the stream
chemistry as it connects different parts of the landscape to the streams. However, relatively little
is known about temporal variability of the extent of the stream network during different flow and
wetness conditions. The aim of this study was to determine to what extent the stream network
expands at different flow conditions and to investigate if the variable extend could be reproduced
using a raster based stream network model.
To reach this aim, several steps were accomplished:
o Establishing a dynamic stream network model based on flow-groundwater table
relationship, using gridded digital elevation data of 5 meter resolution;
o Mapping stream initiation points from the stream network extracted from the digital
elevation data;
o Determining criteria for selecting stream heads;
o Mapping stream head locations in the terrain during a field survey
1.2.
Keyresearchquestions
In this thesis, some key research questions were addressed upon completing the stated
objectives:
o Would model simulation of the stream network agree with the observed stream network
in the real terrain?
1
o To what extent, would the stream network expand at different flow conditions?
o Could the variable extent of stream network be reproduced using a simple raster based
stream network model?
2. Background
Many studies demonstrate the importance of study on the stream network expansion and the
locations where the streams begin. Lyon et al. (2004) mention that knowing the locations of
source areas is useful for understanding non-point source pollution. O'Callaghan & Mark (1984)
and Hancock & Evans (2006) show the importance of stream network extraction from digital
elevation data on the study of geomorphology and hydrology. Rayburg et al. (2005) draws the
attention to the impact of landuse change on the channel network evolution. Their study further
concludes that better understanding about the evolution of the water delivery networks would
make it possible to make better predictions of the impact of landuse change on the hydrologic
functioning.
Montgomery & Dietrich (1988) define channel initiation points as the points that are located
closest to the upslope drainage divide with the presence of channelized morphology.
Montgomery & Dietrich (1989) also suggest that channel heads can be gradual or abrupt and that
the channel reach may touch the channel network immediately downslope or discontinuously
expand.
Connecting the stream head locations to the downslope stream channels, headwater streams
maintain many inducing processes that control, for example, stream chemistry and flow
conditions in the catchment. Jaeger et al. (2007) presents physical characteristics of the
headwater streams. A headwater stream, for instance, can be continuous or discontinuous.
Usually, it has convergent topography and can be found at the foot of a valley. In addition, to be
considered as a headwater stream, it should have channel length of at least 5 meters.
To get more knowledge on the governing processes in the headwater streams, it is interesting to
understand the relationship between source area and slope. Some studies discuss the source
area/slope relations to better understand the inducing processes in the stream head initiation
points on downslope channel network. For instance Montgomery & Dietrich (1988) suggest that,
to know where the stream channels begin, the relationship of source area/slope should be tested
and that the channel head source areas could be selected based on field accessibility, ranges of
slope values, and climatic and geological conditions. In a more recent study, Jaeger et al. (2007)
investigated the source area/slope relationships of headwater channels in sandstone and basalt
lithology of forested landscape by using the source area of channels, delineated in digital
elevation model (DEM) and comparing DEM derived heads to heads that were mapped in the
field using Global Positioning Systems (GPS). The results of the study showed, however, that
there was a poor relationship between the source area and slope since the source areas vary
lithologically and that the source area derived from DEM and mapped in the field did not agree
well.
Another important factor controlling the headwater stream is the formation of the stream heads.
To understand more about the sources of the stream heads, field observations are very valuable.
Channel heads can be formed by several processes such as overland flow (saturation excess or
return flow) or subsurface flow seepage or landslide failure (Jaeger et al., 2007). In their study
on channel initiation, Montgomery & Dietrich (1988) collected 71 points of stream heads in
forested landscape underlain by folded Paleocene and Eocene sandstones. The results of the
study show that, of all the 71 head locations, the formation of the steep-slope stream heads was
either through subsurface flow seepage or through landslide failure, while that of gentle-slope
channel heads was overland flow. The study done by Montgomery & Dietrich (1988) further
2
shows that the initiation points of the channel heads on steep gradient may be formed by
subsurface flow erosion. The field investigation also depicts that the initiation point of abrupt
channel heads on low slope is formed by seepage erosion, while that of gradual channel head is
controlled by overland flow saturation
Many studies rely on DEMs to extract flow directions of the stream networks. There are,
however, some uncertainties due to either low DEM resolution (McMaster, 2002) or difficulties
in deriving flow directions in gentle slope areas (Mackay & Band, 1998). Unfortunately, there is
general a lack of field data on channel initiation locations and the processes that form and
control those transition points (Montgomery & Dietrich, 1988).
3
3.Matterialsan
ndMetho
ods
3.1.Sttudylocattion
The ressearch was conducted
d in the borreal Kryckllan study catchment,
c
situated ab
bout 50 km
m
northweest of Umeåå, northern Sweden, w
with a catchm
ment area of 67 km2. T
The catchmeent geologyy
(Figure 1) is pooorly weatheered gneissiic bedrock covered with
w
sedimeent depositts at lowerr
elevatioon and moraine (glaciaal till) at higgher elevattion. The caatchment haas a gentle topographyy
with eleevations rannging from
m 126 to 3772 m above sea level (m.a.s.l).
(
Thhe catchmeent receivess
mean annnual preciipitation of 623 mm, aand has the mean annu
ual air tempperature of 1.8 oC. Thee
catchmeent vegetattion cover is dominatted by foreests, coveriing 88% oof the catch
hment areaa
followeed by 8% off wetland, 3% of agricuultural land, and 1% of lake. (Grabbs, 2009)
F
Figure
1: DE
EM and geoloogy class maap of the Kry
ycklan catchm
ment
4
Figure 2: Krycklann catchmentt and flow gaging
g
sites
Flow gaaging sites in Kryckla
an catchmeent
There w
were primarrily 16 flow gaging sitees (Figure 2)) in Krycklaan catchmennt, but later site 11 andd
8 were no longer considered
c
as a flow ggaging sites.. Site 16 is the main ouutlet of the catchment,,
whereass site 7 is thhe site whosse dischargee series from
m April 17,, 2007 to D
December 31
1, 2008 wass
used in the model. Subcatchm
ments of gagging site 1, 7, and 9 arre interestinng study areeas for fieldd
investiggation sincee they repressent differennt types of catchment
c
geology.
g
5
Table 1 below show
ws details about
a
each ssubcatchmen
nt of the stu
udy catchmeent.
Table 1: Flow gagging sites in
n Krycklan catchment
SIITENO 1 2 3 4 5 6 7 8 9 10 12 13 14 15 16 SITE NAME RB SVV LMB SVE STO STB SVW FB NÄB SMB NMB LB ÅB ÖKL KL X_COORD
DINATE 11694040
11692550
11692840
11692310
11691640
11692330
11692560
11692140
11693320
11692940
11694430
11693140
11692410
11695500
11697360
Y_
_COORDINA
ATE CATCH
HMENT AREA
A (m2) 71330
030
438617
4
71332
270
126070
1
71338
830
32
2814.5
71341
150
172171
1
71342
260
643354
6
71332
200
10
084540
71332
260
465478
4
71327
790
21
193080
71317
770
28
879970
71339
910
33
315120
71322
200
53
396080
71312
220
69
900410
71304
430
137
759600
71330
080
172
285700
71276
650
649
914300
3.2.De
erivationofterrain
nindices andflowdirection
ns
D8 algoorithm by O’Callaghan
O
n & Mark (11984) was used
u
to calcu
ulate flow ddirections in
n PCRaster..
To com
mpute the floow direction
ns in PCRasster, Deurseen (1995) ussed the folloowing proceedure:
Creatioon of local drain
d
direcction (LDD
D) and pits removing
r
The conncept of loccal drain dirrection is baased on an analysis
a
of each cell inn a DEM. The
T analysiss
of eachh cell in thee DEM is performed
p
inn a 3x3 cellls window.. The slopee of a center cell to itss
eight neeighboring grid
g cells iss first determ
mined. The material (ffor examplee, surface water) wouldd
flow froom that cennter cell to th
he steepest downward cell as show
wn in figuree 3. This flo
ow directionn
is then defined as the local drrain directioon. If the lo
ocal drain direction
d
off that cell iss found, forr
instancee, the 3x3 cells window
w is moved to the next cell. This reefers to a sim
mple case.
Fig
gure 3: LDD
D code for direction
d
of runoff
d, this mean
ns the centeer cell can have eitherr
In contrrast, if the local drain direction iis not found
equal ellevation (thhe cell is parrt of a flat) or lower eleevation (thee cell is partt of a pit) th
han its eightt
neighbooring cells. In this casse, special LDD codee will be asssigned andd processed
d further too
computte the locall drain direection. The solution fo
or finding the local ddrain directiion, in thiss
6
regard, is to start with identifying two types of flat areas: flat area of type 1 and flat area of type
2.
Flat areas of type 1 are defined as a set of neighboring cells of the same elevation, and at least
one of those cells has lower elevation than the core cell. The local drain direction for this case
can be resolved by assuming that is like a simple case, which means the water will flow from the
cell under consideration to the cell of lower elevation.
Flat areas of type 2 are defined as a set of neighboring cells of the same elevations, and all of
these cells have exactly the same elevation as the cell under consideration (the core cell). To
resolve this, each neighboring cell of the core cell is assigned a drain direction.
As mentioned earlier, pits are those cells whose neighbors point towards them. Pit removing
process is to assign artificial local drain direction to those depressed cells. This process is
performed because there might be problems with digitizing or discretization of digital elevation
data, which is not natural phenomenon.
Calculation of upslope contribution area based on D8 algorithm
Upslope contributing area was computed based on local drain direction derived from D8
algorithm. The calculation assumed that, for each cell, accumulated amount of material would
flow out from itself to neighboring downstream cell following the local drain direction. The
accumulated material is the amount of material in a cell under consideration plus the amount of
material flowing from the upstream cells of the cell under consideration. Based on local drain
direction network, the catchment for an outflowing cell is determined, which contains the cell
itself and the upstream cells that drain into it. To get the value of upslope contributing area for
each outflowing cell, the total cell area was considered.
Calculation of terrain slope based on DEM
Terrain slope is the ratio of vertical distance to horizontal distance (dz/dx), whose value ranges
between 0 and 1, and can be turned into percentage by multiplying that ratio with 100.
For each cell, computation of slope is performed in a 3x3 cell window, based on an analysis of
DEM. There might be cases that the elevation of neighboring cells of the core cell is unknown. If
this is the case, for each missing value cell, the elevation of the core cell is assigned with average
elevation of non-missing value cells in the 3x3 cell window.
3.3.Developmentofadynamicstreamnetworkmodel
Flow-Groundwater table relationship
The dynamic stream network model examined the expansion and contraction of stream network
during different streamflow conditions. The basic principle for model development was based on
the combination of landscape analysis and field measurements of streamflow and riparian
groundwater tables. The model assumes that groundwater flow can be scaled with specific
contributing area
and that hydraulic gradient can be estimated by terrain slope tan . Thus,
total groundwater discharge
at a riparian location of homogeneous soil can be related to
groundwater table position based on the Darcy’s law as following (Grabs, 2010).
. tan .
[1]
Where,
: Total groundwater discharge [l/s]
7
: Width [m]
: Terrain slope
: Soil depth [m]
: Groundwater table position [m]
K(z) : Hydraulic conductivity at depth Z [l.s-1m-1]
tan
If the riparian soil and hydraulic conductivity vary exponentially with depth, then the equation
becomes
. tan . ′ . .
[2]
Furthermore, the total groundwater discharge
can be related to streamflow measured at the
catchment outlet as following:
.
.
[3]
Where
: Specific hillslope contributing area derived from topography [m2/m]
A
: Catchment area [m2]
By solving the last two equations, [2] and [3], we get
ln
.
.
,
.
[4]
Where,
o b
o
,
o
: Transmissivity shape parameter [m-1]
: Streamflow measured at the catchment outlet [l/s]
: Transmissivity parameter [l.s-1.m-1]
The dynamic model for estimating groundwater table distribution in the whole catchment was
then based on equation [4].
For the whole study period, based on the available discharge series from April 17, 2007 to
December 31, 2008, the frequency of surface flow occurrence for each cell in the stream
networks can be calculated by taking up the number of days that the streams may generate flow
divided by the total number of observations (total time steps in the model).
To increase the chance of finding the actual locations where the stream starts, a map with at least
90% chance of flow surface occurrence was created and overlaid with geology class map, and
used as a referencing source for locating the stream heads.
Hence, it is practical to create a frequency map of surface flow occurrence in the stream to
facilitate the finding of stream head locations in the field. The frequency was calculated by
assuming that if the groundwater table was at least 0.1m above the ground surface, there would
be surface flow in the stream channels. Days for which the groundwater table was above 0.1m
was summed up for each grid cell and divided by the total number of days
The model generated the stream networks based on local drain directions, and so the stream
network that went across the lake and wetland in the extracted map did not reflect reality. Also
the model did not perform very well at the zone of gentle-slope terrain because groundwater
level distribution in the catchment was interpolated in the sense that its distribution is the same
way as the terrain topography does. Thus, the stream networks that come across lakes and
wetlands in the catchment were not considered.
8
Determination of model parameters
There are three main parameters in the model: a transmissivity shape parameter ( , a
transmissivity parameter ( ′ ), and threshold groundwater table (Zmin) parameter, i.e., the
minimum groundwater table above which surface flow occurring in the stream channels.
The transmissivity parameters value ( ′ ), of the model were primarily set based on the
study location geology. Since the catchment geology is sediment deposit and moraine (glacial
till), then the values for these two parameters can be approximately estimated to be
0.034 (Grabs, 2010, p.27)
7.9
′
The value of groundwater level, based on the equation [4], is negative if the considered location
has a groundwater table below the surface, and is zero or higher if the groundwater table position
is at or above the ground surface. Within the stream, the estimated groundwater level will take a
positive value. Thus, it is important to define a threshold level above which the presence of
water in the stream might generate flow. In this study, the threshold value for the stream to
generate flow was primarily assumed to be Zmin=0.1 m.
Terrain slope
From equation [4], the groundwater table position is related to terrain slope, which is the
denominator of fraction, and this value should not be zero. In the model, terrain slope was set to
take the value of minimum slope (assumed to be 1/10000) for all cases where DEM-derived
slopes were less than the minimum slope value.
Time series of flow
In the dynamic stream network model, groundwater table position for each cell in the catchment
was estimated based on corresponding flow for each time step of the model. The average daily
flow series from April 17, 2007 to December 31, 2008, measured at site 7 of the catchment was
used in the model. Figure 4 below illustrates the flow series for each corresponding time step of
the model.
80
70
Flow(l/s)
60
50
40
30
20
10
0
0
50
100
150
200
250
300
350
400
450
Timesteps
Figure 4: Average daily flow series at site 7
9
500
550
600
Table 2 shows the statistical characteristic of the flow series used in the model. The maximum
discharge (Qmax) occurs during the spring flood period after most of the snow melts, while the
minimum discharge (Qmin) happens at the beginning of the summer. Q75% and Q90% are discharge
of 75th percentile and 95th percentile respectively. From Table 2, it is observed that up to 90% of
the discharge record falls below 8.59 l/s, and 75% falls below 4.91l/s, which is close to the
average discharge of the total time series.
Table 2: Statistical characteristics of the flow series
Qmax (l/s) Qmin (l/s) Qmean (l/s) Qmedian (l/s) Q75% (l/s) Q90% (l/s) 67.3 0.4 4.6
2.6
4.9
8.5 3.4.Fieldinvestigationformappingstreamheads
Field investigation was done to validate the model results, i.e. to see how the locations of stream
heads identified in the map extracted from the model differ from the real head locations in the
catchment. Furthermore, from field observation, it was possible to see how simple understanding
represented by the model works, and to determine if this simple process needs to be revised.
It is inevitable that the results of the model are to some extent not consistent with the real
systems due to certain model uncertainties and assumptions in the processes, i.e. the locations of
the stream heads identified in the map extracted from the model might be different from those in
real catchment.
In the study catchment of Krycklan, there are 16 flow gaging sites including site 16, the one at
the main outlet of the catchment. In this study, three subcatchments upstream of the gaging sites
7, 9, and 1 were selected for detailed field investigations. Additionally, stream heads were also
searched in other subcatchments where depending on the accessibility of the terrain.
Accessibility was most limited at the onset of spring flood period when the snow layer was
relatively deep at some places.
The three subcatchments were selected for field observations because they have more available
data and are underlain by the majority of till deposit whose transmissivity parameters of the
model were based on. Another interesting aspect is that those subcatchments are also partially
underlain by sediment deposits, which make it possible not only to discuss how stream network
extractions from DEM differs topographically, but also how differently the sources of the stream
heads are formed according to geological differences. Another advantage for choosing these
subcatchments is that they are relatively easily accessible from nearby roads (Figure 5).
10
Figure 5:: Investigateed subcatch
hments of sitte 9, 7, and 1
To facilitate the fiield investig
gation, a haand held Gaarmin Oreg
gon 450 GPPS with a base
b
map off
stream nnetwork exttracted from
m DEM of 55-meter reso
olution was used. The bbase map composes off
contourr lines of 5 meter
m
equid
distance andd >90% freq
quency of co
onnecting suurface flow
w (Figure 5)..
During field obserrvation, the mainstream
ms of each selected
s
sub
bcatchmentts were follo
owed up soo
all headdwater streams or thee branches of the maiinstream co
ould be couunted. Whether or nott
observeed headwateer streams appear
a
in thee map extraacted from th
he DEM waas verified.
11
3.5.Testingofmodelparameters’value
GPS coordinates of observed stream heads were imported into ArcGis and spatially compared
with the stream networks extracted from the model. Then, the model parameter values were
changed as shown in Table 3 and Table 4 so the observed stream heads and the simulated stream
heads were spatially matched.
There are three main parameters in the model (b, K0, and Zmin), but only two parameters were
used in the testing: b and Zmin. The reason is that, from equation [2], the total groundwater
discharge is to the exponent of b. A small change in b value would increase the total
groundwater discharge value a lot, and so increase the surface flow in the stream network since
the total groundwater discharge is proportional to the stream flow (equation [1]).
Case 1:
Table 3: Model calibration for b value
-1
-1
Ko[l.s .m ] b (m) Numbers of matched heads 0.033466
7.9024
24
0.033466
8.2975
24
0.033466 8.6926 24 0.033466 9.0878 24 Case 2:
Table 4: Model calibration for Zmin value
-1
-1
Ko [l.s .m ] Zmin (m) Numbers of matched heads 0.033466
‐0.1
25
0.033466
0.0
26
0.033466 0.1 24 0.033466 0.2 25 Table 3 and Table 4 show the model parameter testing procedure for b and Zmin values. Ko was
kept constant for all cases because it is less sensitive to model response. In case 1, the value of b
was increase by 5% for each time, and its initial value is bolded in Table 3. For case 2, the value
of threshold groundwater table, Zmin, was increased by 0.1 m and later decreased by 0.1 m
instead. The bold value (Table 4) corresponds to its initial value.
For each case, the observed and simulated stream heads were visually examined in the map and
spatially compared. Since there might be spatial error in the GPS, it was assumed that the
observed and simulated stream heads are matched if the relative error between them is less than
15 m. From each step of the model parameter testing, it can be seen that there was almost no
improvement on the results, which means the number of matched stream heads did not increase
even after changing the parameter values. Therefore, afterwards the parameter values from the
initial assumption were used for the whole study, i.e. Zmin=0.1 m, Ko= 0.033466 [l.s-1.m-1], and
b=7.9024 [m-1].
4.Results
Since modeled stream network may be inaccurately mapped, field observation was carried out to
verify the model results. The results as presented in the following sections were divided into two
main parts: model results and field observation. Section 4.1, 4.2, and 4.3 show the results from
the model.
12
4.1.Sttreamnettworkand
dfrequen cyofsurffaceflowg
generatin
ng
Figure 6: >90%
>
frequuency map of
o surface flow generatiing
The maap of at leasst 90% frequ
uency (Figuure 6) was extracted frrom the freqquency map
p of surfacee
flow geeneration. The
T grid cellls of the loocal drain direction
d
map, which hhad a frequ
uency valuee
rangingg from 0.900 to 1.00, were
w
extractted to exam
mine the arreas in the catchment that wouldd
frequenntly have suurface flow. From this map, it can
n be seen th
hat most off the times many
m
smalll
branchees of the strream netwo
ork had disccontinuous surface flow
w, especiallly those loccated in flatt
areas. T
These smalll branches can be connsidered as first or zerro-order stre
reams. Som
me areas nott
belongiing to the stream
s
netw
work, in add
ddition, had also the presence of discontinuo
ous surfacee
flow. Siince the moodel assumeed that grouundwater tab
ble value from 0.1 m w
would generrate surfacee
flow wiithin the strream, thus, it
i can be infferred that the
t initial groundwaterr table thresshold’ valuee
in the m
model assum
mption has strong influuence on strream netwo
ork density. If the threshold valuee
were inncreased, thhere would be fewer sm
mall stream
m branches formed beccause of a decrease inn
numberr of saturateed areas at th
he ground ssurface.
13
Figure 7: >90%
>
frequeency map of
o connectin
ng surface fllow
Figure 7 representts a frequen
ncy map of at least 90%
% chance su
urface flow
w would con
nnect to thee
c
nuous surfaace flow (Fig
gure 6), nott
main ouutlet of the catchment.
Despite maany areas wiith discontin
all of thhose areas had
h connectiing surface flow to the main outlett (Figure 7)).
14
Figure 8: Stream nettwork map at
a maximum
m streamflow
w
ng the stud
dy period frrom April 17,
1 2007 too
The maaximum vallue of time series disccharge durin
Decembber 31, 20008 was 67.3 l/s. The m
map in Figurre 8 illustraates the streeam network
k extensionn
corresponding to maximum
m
sttreamflow. It can be seeen that durring the higgh flow periiod most off
the areaas in the cattchment waas saturated and had su
urface flow. It seems thhat the streaam networkk
was so dense and the
t stream started
s
from
m almost ev
verywhere in
n the catchm
ment. Is it realistic? Itt
would nnot be realistic becausse this resuult was soleely based on the modeel assumptiion that thee
stream would havee surface flo
ow if the grroundwater table is equ
ual or higheer than 0.1m
m above thee
uld be much less densse than this in spite off
ground surface. Inn reality, thee stream neetwork shou
maximuum streamfllow becausee the saturatted areas with the preseence of surfface water would
w
drainn
into streeam channeel and surfacce flow usuually occurs only in those existing stream chan
nnels.
15
Figure 9:: Stream nettwork map at minimum
m streamflow
w
w
studyy
Unlike the period of highest flow in thee catchmentt, during thee lowest floow in the whole
period, the presencce of surfacce flow dram
matically deecreases in most
m areas (Figure 9). The modell
also preedicted a seeveral disco
ontinuous ppatches of surface wateer that weree not conneected to thee
main strream netwoork.
16
4.2.Sttreamnettworkand
dgeologyofthestu
udycatchm
ment
Figure 10: >90% freequency maap and geolo
ogy of site 9
Figure 10 shows thhe frequenccy map of sttream netw
work in subccatchment oof gaging sitte 9 with att
least 900% chance of surface flow generaating within
n the stream
ms. In overaall, it can be
b seen thatt
most off the time during
d
the sttudy periodd from Apriil 17, 2007 to Decembber 31, 2008
8, not manyy
small sstream brannches exist in the catcchment. On
n the otherr hand, thee catchmentt is mostlyy
underlaain by till deeposit, on which
w
is thee model parrameters’ values
v
were based. In the
t areas off
sedimennt deposits, there were a few shorrt stream branches form
med along th
the mainstreeams, whilee
in till ddeposit areaas, the stream networkks tended to
t be longeer and havee more bran
nches. Onee
reason may be because
b
thiis sedimentt area hass less genttle slope aand located
d closer too
subcatcchment bordder.
17
Figure 11: >90% freequency maap and geolo
ogy of site 1
Additioonally, the stream
s
netw
work with att least 90% frequency at subcatchhment 1 was examinedd
(Figure 11). Underrlain by alm
most till deeposit area, most frequ
uently, the sstream netw
work in thiss
subcatcchment seem
med to havee very few small bran
nches. The stream
s
netw
work appeaaring in thiss
map shhould be thee perennial streams, w
which were usually fou
und in the rreal terrain during thee
field obbservation.
18
Figure 12
2: > 90% frrequency maap and geollogy of site 7
Subcatcchment of site
s 7 (Figu
ure 12) was not differeent from thee previous ccase (subcaatchment off
site 1). It was undeerlain by alm
most till depposits. Almost none off the small sstream brancches appearr
in this map, and the
t stream network prresent in th
his map sho
ould be perrennial streeams of thee
subcatcchment becaause it is thee stream neetwork extraacted with >90%
>
frequuency, accorrding to thee
model. One of the mainstream
ms in the suubcatchmentt ends up in
n the wetlannd area. In the
t wetlandd
area, m
many small short brach
hes of headdwater streaam were formed and tthese stream
m networkss
usually are not reealistic sincce this areaa has a verry gentle sllope, wheree the modeel does nott
simulate well.
4.3Mo
odeledstrreamleng
gthandsttreamord
dersdurin
ngdiffereentflowco
onditions
19
80
70
5000
60
4000
50
3000
40
30
2000
Flow (l/s)
Total stream length (Km)
6000
20
1000
10
0
0
18‐Jan‐07 28‐Apr‐07 6‐Aug‐07 14‐Nov‐07 22‐Feb‐08 1‐Jun‐08 9‐Sep‐08 18‐Dec‐08 28‐Mar‐09
Total stream length (Km)
Flow (l/s)
Figure 13: Relationship between total stream length and flow
The total length of the stream network was calculated by summing up frictional length from
upstream grid cell to downstream grid cell along local drain direction (flow directions extracted
from the DEM) in the stream network for each corresponding value of streamflow. The total
stream length varied with the same trend as the streamflow (Figure 13). The total stream length
at maximum discharge value was around 10 times longer than that at the minimum discharge
value (Table 5). The total length of the stream network would be overestimated because the flow
directions extracted from the DEM, which was used in the calculation, also took into account the
discontinuous patches in the catchment.
Table 5: Modeled total stream length during highest and lowest flow conditions
Date 4‐May‐08 25‐Jun‐07 Discharge(l/s) 67.3
0.4
20
Total Stream Length (Km) 5620 509 Total stream length (Km)
6000
y = 763,29x0,4477
R² = 0,9988
5000
4000
3000
2000
1000
0
0
10
20
30
40
50
60
70
80
Flow (l/s)
Figure 14: Relation between total stream length and corresponding streamflow
After fitting different functions, the streamflow could be best approximated by a power relation
to the total stream length (Figure 14). The power trend line seemed to fit the points very well
when the flow is less than 30 l/s and somehow deviates from those points at the higher
streamflow values. The line is steeper at low discharge value. This depicts that during the high
flow episodes the total stream length dramatically increase, which means during this episode a
small change in discharge values would increase the total stream length a lot.
3000
2750
Total stream length (Km)
2500
2250
2000
1750
1500
1250
1000
750
500
250
0
18‐Jan‐07 28‐Apr‐07 6‐Aug‐07 14‐Nov‐07 22‐Feb‐08 1‐Jun‐08
Order 1
Order 2
Order 3
9‐Sep‐08 18‐Dec‐08 28‐Mar‐09
Order 4
Figure 15: Total stream length versus different stream orders
Figure 15 above shows that the stream length of the first, second and third-order stream changed
very much at different flow conditions, especially during the high flow period. The stream of
these low orders can be regarded as ephemeral streams, the streams that may not be channelized
or that form only during the storm event and snowmelt.
21
140
Total stream length (Km)
120
100
80
60
40
20
0
18‐Jan‐07 28‐Apr‐07 6‐Aug‐07 14‐Nov‐07 22‐Feb‐08 1‐Jun‐08
Order 5
Order 6
9‐Sep‐08 18‐Dec‐08 28‐Mar‐09
Order 7
Figure 16: Total stream length versus different stream orders
In contrast, the length of the fourth (Figure 15) and fifth-order stream (Figure 16) did not change
very much at various flow conditions. The stream of these orders, for instance, can be classified
as intermittent streams, the stream that are formed seasonally, after spring flood for example,
and that could be channelized. On the other hand, the streams of higher orders, i.e. from order 6
seemed to have little change in length. At certain flow conditions, the stream length does not
change even when the flow increases, usually in the low flow episodes. These high order streams
may be considered as perennial streams.
22
4.4.Fieldobservations
Descriptions of observed stream heads
The majority of the observed stream heads were man-made ditches (Table 6). Generally, these
ditches were dug to drain water away from agricultural lands, forested area, or wetland and
connected to downstream channels. Furthermore, the ditches had a straight regular channel.
During the spring flood period in 2012, most of the ditches had active and continuous surface
flow from the heads till the first confluence with the water depth ranging from 0.02m to 0.27 m.
Only few ditches were found to have discontinuous flow at the upstream channels. Of all the 49
observed stream heads, only two points were road ditches. These road ditches were generally
connected to other downslope man-made ditches before evacuating the water to the mainstream.
Besides this, 10 stream heads were found in the form of a natural stream. These natural streams,
overally, were unchannelized and had irregular channel shape governed by the terrain
topography.
Table 6: Observed stream heads
Types of headwater streams Natural stream Man‐made ditch Road ditch Total Numbers of observed stream heads 10 37 2 49 Water depth (m) 0.02‐0.27 Sources of headwater stream Seepage erosion/ Subsurface flow/ Saturated overland flow Stream head formations
Of all the 49 investigated stream heads, it was found that generally those stream heads were
formed by either subsurface flow or saturated overland flow (Table 7). Subsurface flow was seen
in the form of seepage erosion and seepage from saturation at the upslope of the stream heads,
while saturated overland flow was formed because certain area around the stream heads were
saturated with presence of surface water. Table 7 illustrates that the majority of the stream heads
were formed by subsurface flow in the form of seepage from the saturated soil at the upslope of
the stream heads. Because of convergent topography or the hollow at the stream heads, the water
drained downwards to the stream heads.
Table 7: Formation of stream heads
Formations Natural stream Seepage erosion 3 Seepage from saturation 3 Overland flow 4 Man‐made ditch 2 35 0 Road ditch 0 2 0 Total Total 5 40 4 49 Figure 17 below shows an example of the head of a ditch which was formed by seepage erosion
in a hollow. Because the surrounding soil of the ditch head was saturated, it formed subsurface
flow and recharged water to the hollow (usually the ditch head) before generating surface flow
within the ditch channel. From observations in the field this kind of stream head had presence of
debris, dead understory vegetation, and tree roots.
23
Figu
ure 17: Streeam head wiith seepage erosion
t
picture (Figure 188) was form
med by saturrated area w
which had presence
p
off
Overlannd flow in this
surface water. Usuually in thiss kind of sttream head,, there was the presennce of dead understoryy
vegetatiion and debbris with black color bbecause the water surfaace at this ssaturated so
ource lastedd
long ennough, and the
t water flo
ows accordiing to topog
graphical ex
xpression.
Figure 18:
1 Stream hhead with saturated
s
ov
verland flow
w
24
Figure 19 below is another example of stream head that was formed by subsurface flow. Thiss
stream hhead was loocated at thee base of hiillslope and the headwaater stream was a man--made ditchh
ending with a holllow. The upslope
u
of this stream
m head had
d water graass and wass saturated..
Becausee of conveergent topog
graphy of tthe head, the
t water accumulated
a
d into the hollow
h
andd
formed surface flow
w downwarrd in the dittch channel.
Figure 19:
1 Stream hhead with seepage from
m saturationn
4.5.Co
omparisonofobserrvedandpredicted
dstreamheads
In the ccomparison of observeed and preddicted stream
m heads, th
he observedd stream heaads and thee
predicteed stream network werre mapped aand visually
y examined. The observved stream heads weree
consideered matcheed with the predicted oones if they fell on the simulated sstream netw
work withinn
15 m distance. In
I the dyn
namic streeam networrk model, for each streamflow
w value, a
ork for thee study caatchment was
w generatted. During
g the fieldd
corresponding streeam netwo
observaation, the daaily average streamfloow was 48 l/s. The strream netwoork in this section
s
wass
extracteed from thhe model with correesponding streamflow
s
value durring which
h the fieldd
observaation was doone. A streaam head in tthis study was
w defined as a point ffrom which continuouss
surface flow of at least 5 m in
i the stream
m channel generates. A headwateer stream, on
o the otherr
hand, w
was regardeed as a streaam connectting the streeam head to
o its first coonfluence. The stream
m
heads w
were primarrily observeed within suubcatchmen
nt of site 1, 9, and 7. T
The stream heads weree
also obbserved in other
o
subcatchments raather than these
t
three subcatchm
ments and th
hose stream
m
heads rrepresented almost half of the colllected sam
mple (49). During
D
the ffield investigation, thee
coordinnates for each observ
ved stream heads weere recorded
d with a hhand held GPS. Thee
coordinnates of som
me first con
nfluence pooints were also record
ded, while some otherrs were nott
accessibble due to thhick snow accumulatio
a
on.
25
Subcatchment of site 7
Figure 20:: Observed stream head
ds within su
ubcatchmennt 7
During the field observation
o
, the two m
mainstreams of subcattchment off site 7 werre followedd
upwardds from its outlet so as to counnt all possib
ble headwaater streamss connectin
ng to thosee
mainstrreams. The results
r
from
m the field oobservation showed thaat only two active ditch
hes that hadd
surface flow conneecting to thee mainstreaam were fou
und, while there
t
were a lot of stream brachess
connectting to the mainstream
m
according tto the modeel prediction
ns. One obseerved stream
m head wass
located near the border
b
of the
t subcatcchment, and
d the otherr one appeaared in weetland area,,
m (Figuree 20), none of observed
d and simullated stream
m heads werre matched..
According to the map
This illuustrates the inconsisten
ncy betweenn the model simulation and the fielld observatiion.
Table 8: Number
N
of surveyed strream heads falling with
hin subcatchhment of site 7
Stream heead No. 26
Descriptio
ons 34
41
Natural strea
N
am Man‐made d
M
ditch Subcatchment of site 9
Figure 21: Observedd stream heaads in catchm
ment of sitee 9
Of all 15 stream heeads observ
ved along thhe mainstreaams of the catchment
c
ass seen in Fiigure 21, 100
w were foun
nd on the m
map extracteed from thee
headwaater streams (bold text rows in Tabble 9) below
model. One headw
water stream
m was a rroad ditch that conneccted to othher ditches. The otherr
headwaater streamss did not ap
ppear in thee map becaause some ditches werre dug receently, whilee
others m
might not bee covered in
n the DEM rresolutions..
27
Table 9: Number of stream heads falling within catchment of site 9
Stream head No. Descriptions 4 7 1 2 3 11 37 36 39 31 32 30 41 29 28 40 10 Man‐made ditch Natural stream Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Road ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch Man‐made ditch 28
Subcatchment of site 1
Figure 22
2: Observedd stream heaad in catchm
ment of site 1
Like suubcatchmentt of site 7, the mainsttreams of th
his subcatch
hment weree followed upwards too
count aall possible active head
dwater streaams or bracches that co
onnect to thhose mainsttreams. Thee
results from the field investig
gation (Figuure 22) sho
ow that all headwater
h
sstreams are man-madee
ditches and only five
f
of them
m had activee flow into the mainstreams, whiile the otherrs were dryy
and paartly covereed by snow
w. The streeam head No.19 (Taable 10) feell within one
o
of thee
mainstrreams, and were
w found to be dry dduring the day
d of field observationns with the presence
p
off
little suurface waterr further up of the stream
m head, wh
here water sttarted flowiing.
29
Tablle 10: Numb
ber of stream
m heads fallling within catchment of site 1
Stream heead No. Descriptio
ons 22
21
19
18
20
Man‐made d
M
ditch Man‐made d
M
ditch Man‐made d
M
ditch Man‐made d
M
ditch Man‐made d
M
ditch Other ssubcatchments
ds in other subcatchmen
s
ents
Figure 23:: Observed sstream head
30
Table 11: Number of stream heads falling within other subcatchments
Stream head No.
Descriptions
42
43
44
45
13
47
48
49
14
17
16
5
6
8
9
36
12
24
27
26
25
15
23
46
33
Man-made ditch
Man-made ditch
Man-made ditch
Natural stream
Man-made ditch
Natural stream
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Natural stream
Natural stream
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Natural stream
Natural stream
Natural stream
Natural stream
Man-made ditch
Man-made ditch
Man-made ditch
Road ditch
From Table 11, the stream heads observed in other subcatchments other than subcatchment of
site 1, 7, and 9 represented half of the collected sample. More importantly, there were up to 12
simulated stream heads that matched with observed stream heads whose headwater streams were
either ditches or natural stream (Figure 23).
31
Headw
water stream
ms
Figure 24: O
Observed heeadwater streams
32
Table 12: Characteristics of headwater streams
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Streamlength
13
152
8
94
26
18
49
24
20
43
27
5
284
68
154
160
162
48
62
102
397
32
86
16
14
20
18
95
5
220
Descriptions
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Naturalstream
Man‐madeditch
Naturalstream
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Man‐madeditch
Naturalstream
Man‐madeditch
Man‐madeditch
Naturalstream
Naturalstream
Naturalstream
Man‐madeditch
Naturalstream
Man‐madeditch
A headwater streams was defined as the stream that connect its head to its first confluence.
During the field observation, some first confluence points of the observed headwater streams
were covered by snow, and so were recorded. The length of the headwater streams presented in
Table 12 was calculated in ArcGis, based on the recorded coordinates of the observed stream
heads. The observed headwater stream’s length, for instance, ranged from 5m to 397m and those
streams could be man-made or natural. Generally, the man-made streams tend to be longer,
while the natural ones are shorter.
33
4.6.Slope‐arearelationship
Slope-area relationship was defined as the relationship between the local slope and the source
area (drainage area) of the stream heads. The value for local slope of a stream head was
extracted from the DEM-derived slope map, while the value for the source area of the stream
head was extracted from the map of upslope accumulated area, whose calculation procedure was
previously described. This relationship can describe how stream channels form naturally in
different landscapes.
Figure 25 below shows the relationship between local slope and source area of the stream head
for man-made ditches. It is no doubt that the relationship was not significant because the value
of R2 was too low.
Local slope of stream heads (%)
45
40
35
30
25
20
15
10
5
0
y = ‐0,0009x + 8,2731
R² = 0,0024
0
200
400
600
800
1000
1200
1400
1600
1800
Source area of stream heads (m^2)
Figure 25: Relationship of local slope/source area of stream head for man-made ditches
Local slope of stream heads (%)
Not different from the case of man-made ditches, any relationship between the local slope and
source area of the stream heads for the natural streams was not found (Figure 26).
20
15
10
y = 0,0016x + 11,339
R² = 0,0033
5
0
0
50
100
150
200
250
300
350
400
Source area of stream heads (m^2)
Figure 26: Relationship of local slope/source area of stream head for natural stream
34
450
5.Discussions
5.1.Comparisonofobservedandsimulatedstreamheads
Model parameter testing was performed to increase the number of matched stream heads
between the model simulation and the field observation. However, there was almost no
improvement on the results. There might be two reasons for this none-response to changing
model parameters’ value. First, it may be because the most of the streams in the Krycklan study
catchment are man-made ditches. People dug these ditches on purpose, for example, to drain the
water away to and from specific areas, which alters from the natural processes. In contrast, the
stream network extraction from the DEM is primarily based on the topographical expressions in
the digital elevation data, which tends to be more natural processes. Second reason is that the
calibrated parameters were probably less sensitive to the expansion and contraction of the stream
network, but the streamflow was much more sensitive.
From the results, only 24 out of 49 observed stream heads fell onto branches of the modeled
stream network, which accounted for 49% of the collected sample. This result, to some extent,
illustrates the inconsistency between the stream network extracted from the model and the
observed one. This finding supports other studies which suggest that this inconsistency could be
either due to the low resolution of DEM (McMaster, 2002) or unsound performance of DEM in
gentle slope areas (Mackay & Band, 1998). For instance, from the field observation, it was
observed that the real stream network of the catchments was generally formed by the network of
man-made ditches that drain the water away from the agricultural land, forested area, or wetland.
These ditches might not be incorporated in the gridded digital elevation data of 5 meter
resolution, which was used to extract the stream network.
On the other hand, all of the observed stream heads were identified to have been formed by
either subsurface flow or overland flow, which is consistent with the finding from Jaeger et al.,
(2007) that the stream heads can be formed by subsurface flow, or overland flow, or landslide
failure. In this study, the form of landslide failure was not encountered. Seepage from saturation
at the upslope of the stream heads was the common form of subsurface flow formed at the heads
of the stream, most of which were the man-made ditches. Other form of subsurface flow such as
seepage erosion was also found.
These findings illustrate human impacts on natural systems. The flow in the stream network of
the study catchment was, to large extent, governed by the artificial ditch network.
5.2.Variationofstreamlengthandstreamordersatvariousflowconditions
From the results (Figure 15), the modeled total stream length of the stream network appears to
have a power relation to the streamflow. This suggests that the change in modeled stream length
of the stream network was very sensitive to the changes in discharge. However, the power
relation might only be an artifact of the model structure which used an exponential relation
between groundwater tables and streamflow. Moreover, the model appeared to largely
overpredict the occurrence of surface flow and therefore, the modeled total length of the stream
network might not present reality.
If the flow increases, more streams will be formed and the stream network would expand in the
catchment. In contrast, if the flow decreases, the stream network will contract and result in more
inactive streams. Thus, during the increasing trend of the flow, the stream network, especially
the headwater streams that connect the surrounding stream head locations to the downstream
35
channel, potentially receives more chemical substances from those connected areas such as
DIC/TOC or metals.
In addition, during different flow conditions (low, medium, and high episode) the stream length
for each stream order also changed. This means that if the flow increases, the number of stream
orders also increases, especially the lower-order streams. This low-order stream such as first,
second and third order are considered ephemeral streams, and usually the transitional streams to
the higher-order stream. Thus, in the study of DIC/TOC evasion from the stream, one can
assume that stream order and stream length are constant during the same flow condition. The
assumption that the stream order and stream length does not change during different flow
condition, for example in dry and wet condition, would more or less overestimate the amount of
DIC/TOC evasion during dry period, but underestimate the amount during the wet period,
especially during the high flow episode. This result is potentially significant for models such as
the TRIM model (Grabs, 2010), which assumed a static stream network regardless of the
discharge conditions.
5.3.Modeluncertainties
The results of the study shows considerable differences between the stream network extracted
from DEM and the real stream network observed in the field. This inconsistency may be caused
by uncertainties and assumptions in the model structure. The raster based dynamic stream
network model used in this study had assumptions and uncertainties as following:
o The soil types in the study catchment were assumed to be homogeneous and the same
transmissivity parameters’ value were used in the whole catchment areas; however,
Grabs (2010) mentions that the threshold groundwater above which the surface flow
may generate varies with different riparian zones.
o The groundwater table position in the whole catchment was estimated directly from the
streamflow and some terrain indices, and only the streamflow measured at site 7 of the
catchment was used. This streamflow was then multiplied by the percentage of each
contributing area to the total catchment area to scale up the streamflow for the other
subcatchments.
o Groundwater table was not monitored in the study area, but assumed that the
groundwater table position would follow the topographical expression of the catchment.
o The surface flow in the stream channels was assumed to occur if the groundwater table
was 0.1 m above the ground surface. This assumption might not be true because the
elevation of the surface flow in the stream channel may be lower than the ground surface.
o The stream network in the study catchment was formed mostly by man-made ditches,
some of which were recently dug. There might be chance that these ditches were not
incorporated in the DEM.
6.Concludingremarks
From the results of the study, it was clearly seen that the observed stream heads and the stream
heads appearing in the stream network map extracted from the model did not agree very well.
49% of the observed stream heads (49) fell onto the branches/headwater streams of the modeled
stream network and only few of them exactly matched the predicted stream heads. One reason
for this inconsistency is because the majority of the stream network in the Krycklan study
catchment is man-made ditch, most of which would not be incorporated in the DEM. This result
illustrates the human intervention on the natural landscape. People dug the ditches to increase
the forest productivity, but they also altered the natural formation of the stream network.
36
Furthermore, it can be seen that the developed raster based stream network model seemed to
overestimate the stream network, which cannot represent the dynamic expansion of the stream
network very well. For instance, the modeled stream network was so dense, forming many zero
and first-order stream almost everywhere, especially during the high flow episode. The modeled
stream network was somehow unrealistic because much fewer streams were observed during the
field observation carried out during the spring flood of 2012. This inconsistency encourages
further field surveys to be conducted in the study of where the streams begin to verify the
observed and predicted stream network extracted from the DEM because the stream network
extracted from the DEM can be inaccurately mapped.
From model results, the number of streams of lower order and their lengths were very sensitive
to change in streamflow, especially during the high flow episode. Even though the model
seemed to overestimate the stream network, large difference in number of stream with active
flow was observed between the peak discharge period and the decreasing discharge period
during the field observation. Therefore, studies on the variation of stream-water chemistry
should take into account dynamic flow and stream network conditions.
The headwater streams are important transitional zones that connect the stream head and its
surrounding upslope areas to the downslope stream networks. During the dry conditions such as
in summer, many headwater streams may be inactive, while in the wet conditions, for example
during the spring flood of 2012, many headwater streams turned to be very active and
substantially influence the flow conditions in the catchment.
7.Suggestionsforfuturework
Since the modeled stream network was usually not accurately mapped, field observations are
always necessary for the study of where the stream starts and the field survey should be done at
different flow and wetness conditions in the catchment. For example, a specific stream network
belonging to certain subcatchments may be examined during and after spring flood and in late
fall to see where the stream begins at different times.
Specifically for the study of where the stream begins in the Krycklan study catchment, one
would have a clearer picture on the stream network expansion and contraction if one tries to
investigate the real stream networks, most of which are man-made ditches, by tracing all the
networks that connect to the main streams and include those networks in the map. One would
just record the coordinates of each confluence of the stream networks and trace the networks
based on the available gridded digital elevation data, and/or the contour lines. This field
observation could be done after the spring flood because during the spring flood period, there is
still very thick snow accumulation in many areas of the catchment and that is a big challenge
that hides the ditch networks from visual observation.
37
8.Acknowledgements
I remember when I got a list of topics for master thesis in hydrology/hydrogeology from Roger
Herbert. I scanned through the list many times, but still I couldn’t keep my eyes away from a
line that wrote “Where does the stream begin?”, and then it became the topic for my master
thesis.
I would like to express special thanks to Thomas Grabs, my helpful supervisor. He gave me not
only the new idea and comments regarding the research, but also a warm encouragement from
the start to the end of the writing. Without his immense support, I wouldn’t have completed this
thesis.
I can’t forget Reinert Huseby Karlsen, who always accompanies me since the start of my master
thesis research. He also gave me a lot of helpful comments on my thesis writing and assisted me
in technical issues. During the field investigation, he drove me to and from the field.
I’d like to say thank to Prof: Kevin Bishop, my subject reviewer, who put me into the research
topic at the beginning and for his intervention in making it possible for me to conduct the field
observation in Vindeln. His comments on my thesis were very helpful to improve the quality of
my research.
Pianpian and Anna, thank both of you for accompanying me during the field investigation. You
walked with me in the quiet forest, with thick snow depth, sometimes went up to the high
hillslope for searching the stream head locations with me.
Last but not least, I remember Prof: Allan Rodhe, who gave me the colorful tracer that could be
used to identify the movement of the water flowing in the stream channels.
38
References
Deursen, V.WPA (1995). Geographical Information Systems and Dynamic Models;
development and application of a prototype spatial modelling language. Doctoral
dissertation, University of Utrecht, The Netherlands.
Grabs, T., Seibert, J., Bishop, K., & Laudon, H. (2009). Modeling spatial patterns of saturated
areas: a comparision of the topographic wetness index and a dynamic distributed model.
Journal of Hydrology, 373, 15-23.
Grabs, T., Bishop, K.H., Laudon, H., Lyon, S.W., Seibert J. (2010). Riprian zone processes and
soil-water total organic carbon (TOC): Implication for spatial varability, upscaling and
carbon exports. Manuscript.
Hancock, G. R., & Evans, K. G. (2006). Channel head location and characteristics using digital
elevation models. Earth Surface Processes and Landforms, 31, 809-824.
Jaeger, K. L., Montgomery, D. R., & Bolton, S. M. (2007). Channel and Perennial Flow
Initiation in Headwater Streams: Management Implications of Variability in SourceArea Size. Environ Manage, 40, 775-786.
Köhler, S., Buffam, I., Seibert, J., Bishop, K., & Laudon, H. (2009). Dynamics of stream water
TOC concentrations in a boreal headwater catchment: Controlling factors and
implications for climate scenarios. Journal of Hydrology, 373, 44-56.
Lyon, S. W., Walter, M. T., Gerard-Marchant, P., & Steenhuis, T. S. (2004). Using a
topographic index to distribute variable source area runoff predicted with the SCS
curve-number equation. Hydrological Processes, 18, 2757-2771.
Mackay, D. S., & Band, L. E. (1998). Extraction and representation of nested catchment areas
from digital elevation models in lake-dominated topography. Water Resources
Research, 34(4), 897-901.
Montgomery, D. R., & Dietrich, W. E. (1988). Where do channels begin?. Nature, 336, 232-234.
Montgomery, D. R., & Dietrich, W. E. (1989). Source areas, drainage density, and channel
initiation. Water Resources Research, 25, 1907-1918.
McMaster, K. J. (2002). Effects of digital elevation model resolution on derived stream network
positions. Water Resources Research, 38(4), 13-1.
O'Callaghan, J. F., & Mark, D. M. (1984). The extraction of drainage networks from digital
elevation data. Computer Vision, Graphics, and Image Processing, 28(3), 323-344.
PCRaster documentation — PCRaster v3.0.1 documentation. (n.d.). PCRaster | Software for
environmental modelling. Retrieved June 13, 2012, from
http://pcraster.geo.uu.nl/documentation/PCRaster/html/index.html
Tarboton, D. G., Bras, R. L., & Rodriguez-Iturbe, I. (1991). On the extraction of channel
networks from digital elevation data. Hydrological Processes, 5, 81-100.
Wallin, M., I. Buffam, M. Öquist, H. Laudon, and K. Bishop (2010), Temporal and spatial
variability of dissolved inorganic carbon in a boreal stream network: Concentrations and
downstream fluxes, J. Geophys. Res., 115, G02014, doi:10.1029/2009JG001100
39
Appendix1:Fieldprotocol
Physical characteristics of stream heads
o Begin with convergent topography;
o Located at the base of a valley down slope of a hollow
o Formed by landslide failure or seepage erosion or subsurface flow
Physical characteristics of headwater streams
o Can be continuous or discontinuous from the downstream channel to the first
confluence
o Can be first order or zero order stream
o Having flowing water continuous for at least 5 meters
o Can be channelized (man-made ditches) or unchannelized (natural streams)
Headwater channels have four distinguished topographic units:
o Hillslope: it has divergent or straight contour lines with unchannelized flow or
dispersion flow.
o Zero-order basins: there is the presence of unchannelized hollow with convergent
contour lines. It can be saturated with overland flow and return flow into flood plain.
o Ephemeral/transitional channels: it emerges from zero-order basins, with definable
banks if the channels exist at the outlet of the basins. In addition, it has ephemeral
flow and can be considered temporary storage of organic carbon, which may also
contain discontinuous segments prior to entering first-order channels.
o First order stream channels: it may be directly originated from zero-order basin.
Field materials
o
o
o
o
o
o
Tracer for indentifying if the water is moving
Global Positioning System (GPS) device to mark the stream heads
Field phones to communicate between field observers
A plastic container for maintain the tracer
A tape/meter to measure the water depth and stream length
Waterproof field notes with back-ups
Field procedures
a. Identify the formation of the headwater stream
- Overland flow: presence of debris, brown /killed vegetation like pastures/grass
seed, or
- Landslide failure: presence of recent mass wasting scars, or
- Convergent subsurface flow: if the two cases above are not applied or there is
the presence of observed seeping from the channel head.
b. Record the coordinates of the stream heads with GPS
c. Determine the stream order and identify if the stream is channelized or not.
d. Determine if the water is moving in the stream
In a steep slope stream, the movement of water flow can be easily visualized, but in a
gentle slope stream, it was really difficult to indentify if the water is moving. In the
later case, the colorful tracer was used:
- Mix little amount of tracer with a cup of stream water and stir it to a solution
a
-
Choose any point at the upper part of the headwater stream, record the time,
and inject the solution to the selected point
- After several minutes, measure the distance of which the tracer solution moves
along with the water in the stream channels.
e. Approximately find the deepest point of the stream heads and measure the depth of
the water.
b
Appendix2:Dynamicstreamnetworkmodel’sscripts
binding
# =========================================
# INPUT
# =========================================
# Maps
#-----------------------------------------# Network DEM
STREAM_DEM = dem.map;
# Stream slope (Downslope Index)
STREAM_SLOPE = stream_slope.map;
# Stream catchment area
STREAM_CATCHM_AREA = stream_ca.map;
# Gaging sites
GAGING_SITE = gaging_site.map;
# Setting based map
STREAM_LDD=stream_ldd.map;
#Geology map classes
GEO_CLASS=geo_class.map;
# Mainstream of more than 4 ordered
MAIN_STREAM=main_stream.map;
#-------------------------------------------# Time series
TS_FLOW_SITE = flow.tss; # [l/s]
#
#-----------------------------------------# CONSTANTS
#-----------------------------------------# Start time (make sure it is smaller or equal END_TIME)
START_TIME = 1;
# End time (make sure it is smaller or equal to N_TIMESTEPS)
END_TIME = 625;
# Total number of time steps
N_TIMESTEPS = 1;
# Since the catchment geology is kind of sediment deposit and moraine
b=7.9024; # [1/m]
Ko=0.033466; # [l/s.m], Flow per m flow width
# Site upslope area
UAA_SITE=465478 ; #[m2]
# Minimum slope
SLOPE_MIN =0.0001;
# Minimum GWLevel above which the stream generate flow
Z_MIN=0.1; #[m]
# Total number of time series data
TOTAL_TS=625;
# =========================================
c
# OUTPUT
# =========================================
# Maps
#-----------------------------------------# Groundwater table
#-----------------------------------------areamap
STREAM_LDD;
timer
START_TIME END_TIME 1;
initial
ONES = STREAM_DEM/STREAM_DEM;
ZEROS = STREAM_DEM*0;
# Smooth dem
DEM_ = lddcreatedem(STREAM_DEM,1E35,1E35,1E35,1E35);
# Create LDD
STREAM_LDD = lddcreate(DEM_,1E35,1E35,1E35,1E35);
report stream_ldd.map=STREAM_LDD;
# Calculate stream slope
STREAM_SLOPE = slope(STREAM_DEM);
STREAM_SLOPE=if(STREAM_SLOPE<SLOPE_MIN then SLOPE_MIN else
STREAM_SLOPE);
report stream_slope.map = STREAM_SLOPE;
# Stream order
STREAM_ORDER=streamorder(STREAM_LDD);
report stream_order.map=STREAM_ORDER;
#MAIN_STREAM=STREAM_ORDER>5;
MAIN_STREAM = if(STREAM_ORDER>5,STREAM_ORDER);
report main_stream.map = MAIN_STREAM;
# Calculate stream cathcment area
STREAM_CATCHMENT_AREA=accuflux(STREAM_LDD,25);
report stream_ca.map=STREAM_CATCHMENT_AREA;
# Catchment for the gaging site whose discharge series used in the model
GAGED_CATCHMENT=catchment(STREAM_LDD,GAGING_SITE);
report site_catchment.map=GAGED_CATCHMENT;
# Determine upslope accumulated area at the outlet
UAA_OUTLET=
STREAM_CATCHMENT_AREA==mapmaximum(STREAM_CATCHMENT_AREA);
# Determine the main outlet of the catchment
MAIN_OUTLET = nominal(scalar(UAA_OUTLET)/scalar(UAA_OUTLET)); #Remove
missing values
SURFACE_FLOW_CONNECTED_SUM = ZEROS;
SURFACE_FLOW_SUM = ZEROS;
d
dynamic
# Flow at gauging site
FLOW_SITE = timeinputscalar(TS_FLOW_SITE,2);
# Estimate the discharge at the main outlet
Q = FLOW_SITE*(STREAM_CATCHMENT_AREA/ UAA_SITE);
# Estimate the groundwater table distribution within study period
Zgwt=(1/b)*ln(Q/(celllength()*STREAM_SLOPE*Ko));
# Determine GWLevel above which the sream generates flow
SURFACE_FLOW = if(Zgwt>Z_MIN then Zgwt);
report SF=SURFACE_FLOW;
# Count the number of days for which the stream may generate the flow
SURFACE_FLOW_ = if(Zgwt>Z_MIN then ONES else ZEROS);
SURFACE_FLOW_ = cover(SURFACE_FLOW_, ZEROS);
SURFACE_FLOW_SUM = SURFACE_FLOW_ + SURFACE_FLOW_SUM;
SURFACE_FLOW_SUM_CLEAN = if(SURFACE_FLOW_SUM> 0 then
SURFACE_FLOW_SUM);
report N_DAY_SF.map = SURFACE_FLOW_SUM_CLEAN;
# Calculate the probability for which the stream might generate surface flow
PROB_SF = SURFACE_FLOW_SUM_CLEAN/TOTAL_TS;
PROB_SF_CLEAN = if(PROB_SF>0.1,PROB_SF);
report P_SF.map = PROB_SF_CLEAN;
P95_SF = if(PROB_SF_CLEAN>=0.95,PROB_SF_CLEAN);
report P95_SF.map=P95_SF;
P90_SF = if(PROB_SF_CLEAN>=0.90,PROB_SF_CLEAN);
report P90_SF.map = P90_SF;
# LDD map corresponding to the days with surface flow
LDD_SF=if(Zgwt>Z_MIN then STREAM_LDD);
report ldd_sf.map=LDD_SF;
STREAM_ORDER_=streamorder(LDD_SF);
STREAM_ORDER_SF =if(STREAM_ORDER_>=1,STREAM_ORDER_);
report stream_order_sf.map =STREAM_ORDER_SF;
# Catchment whose flow at the outlet connecting with surface flow
CATCHMENT_FLOW=boolean(catchment(LDD_SF,MAIN_OUTLET));
report SFC=CATCHMENT_FLOW;
# Compute the number of days the surface flow connecting to the outlet
SURFACE_FLOW_CONNECTED = if(CATCHMENT_FLOW then ONES else ZEROS);
SURFACE_FLOW_CONNECTED = cover(SURFACE_FLOW_CONNECTED, ZEROS);
SURFACE_FLOW_CONNECTED_SUM = SURFACE_FLOW_CONNECTED +
SURFACE_FLOW_CONNECTED_SUM;
report N_DAY_SFC.map = SURFACE_FLOW_CONNECTED_SUM;
# Calculate the probability for which the stream flow connect to the outlet
PROB_SFC = SURFACE_FLOW_CONNECTED_SUM/TOTAL_TS;
PROB_SFC_CLEAN = if(PROB_SFC>0.1,PROB_SFC);
report P_SFC.map = PROB_SFC_CLEAN;
P95_SFC= if(PROB_SFC_CLEAN>=0.95,PROB_SFC_CLEAN);
report P95_SFC.map = P95_SFC;
P90_SFC= if(PROB_SFC_CLEAN>=0.90,PROB_SFC_CLEAN);
e
report P90_SFC.map = P90_SFC;
# Count the number of cell with surface flow connecting to the stream orders
CELL_SF = areatotal(SURFACE_FLOW,MAIN_STREAM);
CELL_SFC = areatotal(scalar(CATCHMENT_FLOW),MAIN_STREAM);
report cell_sf=CELL_SF;
report cell_sfc=CELL_SFC;
# Assign stream length for all streams
STREAM_LENGTH_OUTLET = downstreamdist(LDD_SF);
# Calculate total total legnth of all streams in the catchment
STREAM_LENGTH_SUM = maptotal(STREAM_LENGTH_OUTLET);
TS_STREAM_LENGTH_OUTLET_TOTAL = timeoutput(1,STREAM_LENGTH_SUM);
# Assign the longest stream of all streams
STREAM_LENGTH_ = slopelength(LDD_SF,1);
# Calculate the stream length for the longest stream
STREAM_LENGTH_LONGEST =mapmaximum(STREAM_LENGTH_);
TS_STREAM_LENGTH_LONGEST = timeoutput(1,STREAM_LENGTH_LONGEST);
# Assing stream length by stream orders
STREAM_LENGTH_ORDER1 = if(STREAM_ORDER_SF==1,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER1_SUM = maptotal(STREAM_LENGTH_ORDER1);
TS_STREAM_LENGTH_TOTAL_ORDER1 =
timeoutput(1,STREAM_LENGTH_ORDER1_SUM);
STREAM_LENGTH_ORDER2 = if(STREAM_ORDER_SF==2,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER2_SUM = maptotal(STREAM_LENGTH_ORDER2);
TS_STREAM_LENGTH_TOTAL_ORDER2 =
timeoutput(1,STREAM_LENGTH_ORDER2_SUM);
STREAM_LENGTH_ORDER3 = if(STREAM_ORDER_SF==3,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER3_SUM = maptotal(STREAM_LENGTH_ORDER3);
TS_STREAM_LENGTH_TOTAL_ORDER3 =
timeoutput(1,STREAM_LENGTH_ORDER3_SUM);
STREAM_LENGTH_ORDER4 = if(STREAM_ORDER_SF==4,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER4_SUM = maptotal(STREAM_LENGTH_ORDER4);
TS_STREAM_LENGTH_TOTAL_ORDER4 =
timeoutput(1,STREAM_LENGTH_ORDER4_SUM);
STREAM_LENGTH_ORDER5 = if(STREAM_ORDER_SF==5,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER5_SUM = maptotal(STREAM_LENGTH_ORDER5);
TS_STREAM_LENGTH_ORDER5 = timeoutput(1,STREAM_LENGTH_ORDER5_SUM);
STREAM_LENGTH_ORDER6 = if(STREAM_ORDER_SF==6,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER6_SUM = maptotal(STREAM_LENGTH_ORDER6);
TS_STREAM_LENGTH_TOTAL_ORDER6 =
timeoutput(1,STREAM_LENGTH_ORDER6_SUM);
STREAM_LENGTH_ORDER7 = if(STREAM_ORDER_SF==7,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER7_SUM = maptotal(STREAM_LENGTH_ORDER7);
f
TS_STREAM_LENGTH_TOTAL_ORDER7 =
timeoutput(1,STREAM_LENGTH_ORDER7_SUM);
STREAM_LENGTH_ORDER8 = if(STREAM_ORDER_SF==8,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER8_SUM = maptotal(STREAM_LENGTH_ORDER8);
TS_STREAM_LENGTH_TOTAL_ORDER8 =
timeoutput(1,STREAM_LENGTH_ORDER8_SUM);
STREAM_LENGTH_ORDER9 = if(STREAM_ORDER_SF==9,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER9_SUM = maptotal(STREAM_LENGTH_ORDER9);
TS_STREAM_LENGTH_TOTAL_ORDER9 =
timeoutput(1,STREAM_LENGTH_ORDER9_SUM);
STREAM_LENGTH_ORDER10 = if(STREAM_ORDER_SF==10,downstreamdist(LDD_SF));
STREAM_LENGTH_ORDER10_SUM = maptotal(STREAM_LENGTH_ORDER10);
TS_STREAM_LENGTH_TOTAL_ORDER10 =
timeoutput(1,STREAM_LENGTH_ORDER10_SUM);
g
Appendix3:Coordinatesanddescriptionsofobservedstreamheads
Note: Lat/Lon: X/Y: Name: Site No: Site No
1
2
3
4
5
6
7
8
9
10
11
49
35
30
36
31
32
33
26
27
28
29
34
45
46
47
48
18
19
20
21
22
23
24
25
Imported from GPS with UTM_WGS84 coordinate system Calculated in ArcGIS from imported Lat/Lon, projected to RT90_25_gon_V coordinate system Modified name of stream heads saved in GPS (012=stream head of first point, 112=stream head of 11th point) Observed stream heads No corresponding to original field data Name
Lat
Lon
012
022
032
042
052
062
072
082
092
102
112
122
132
142
152
162
172
182
192
202
212
222
232
242
252
262
272
282
292
302
312
322
332
342
352
64.24398
64.24662
64.24638
64.24274
64.23888
64.23891
64.24225
64.24527
64.24525
64.24774
64.24774
64.20282
64.25129
64.25738
64.24999
64.25845
64.2587
64.25658
64.25599
64.25618
64.25506
64.25202
64.25128
64.21952
64.21796
64.21774
64.2182
64.25861
64.25672
64.25553
64.25297
64.25325
64.25974
64.26015
64.26195
19.79073
19.79445
19.79125
19.79079
19.79097
19.79076
19.78672
19.77278
19.77288
19.782
19.782
19.83581
19.80894
19.80015
19.80805
19.79947
19.80013
19.80723
19.80697
19.80547
19.80724
19.80919
19.8098
19.75886
19.75526
19.75572
19.75464
19.771
19.77107
19.77454
19.77494
19.77632
19.78383
19.78057
19.77469
X
Y
Descriptions
Geology Classes
1693046.49
1693208.16
1693055.23
1693057.97
1693093.98
1693083.51
1692864.33
1692168.23
1692173.54
1692597.46
1692597.46
1695520.23
1693877.31
1693409.08
1693843.30
1693368.60
1693398.73
1693757.12
1693748.94
1693675.04
1693768.29
1693884.32
1693918.94
1691672.59
1691508.82
1691532.80
1691477.27
1691989.73
1692006.17
1692182.07
1692219.28
1692284.51
1692602.74
1692441.88
1692144.76
7132345.48
7132650.39
7132613.98
7132208.21
7131778.94
7131781.52
7132141.32
7132434.80
7132433.00
7132737.38
7132737.38
7127902.10
7133214.51
7133866.10
7133066.89
7133982.11
7134012.16
7133797.93
7133731.05
7133747.85
7133629.02
7133296.08
7133216.46
7129526.20
7129342.06
7129318.74
7129366.69
7133914.31
7133703.71
7133581.61
7133297.89
7133333.59
7134078.88
7134114.30
7134297.30
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Natural stream
Natural stream
Natural stream
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Natural stream
Natural stream
Natural stream
Natural stream
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Road ditch
Road ditch
Natural stream
Man-made ditch
Sediment deposits
Till deposits
Till deposits
Sediment deposits
Sediment deposits
Sediment deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Sediment deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Wetlands
h
12
13
14
15
16
17
37
38
39
40
41
42
43
44
362
372
382
392
402
412
422
432
442
452
462
472
482
492
64.24993
64.25163
64.2511
64.2511
64.25536
64.25534
64.24908
64.24833
64.2498
64.2499
64.25238
64.25372
64.25476
64.25647
19.77275
19.7749
19.77654
19.77654
19.77714
19.7772
19.80714
19.80533
19.8035
19.80352
19.80306
19.8046
19.80506
19.80509
1692134.46
1692226.67
1692310.13
1692310.13
1692309.18
1692312.24
1693805.28
1693723.17
1693624.19
1693624.68
1693585.08
1693649.89
1693664.86
1693654.28
i
7132953.15
7133149.30
7133094.51
7133094.51
7133571.01
7133568.63
7132963.59
7132874.17
7133031.98
7133043.62
7133318.36
7133471.83
7133589.20
7133778.83
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Man-made ditch
Natural stream
Man-made ditch
Natural stream
Man-made ditch
Man-made ditch
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Till deposits
Appendix4:Fielddatarecords
Date Site No 16/5/2012 1 2 3 4 5 6 7 8 9 10 Head WD (m) 0.06 0.08 0.1 0.09 0.1 0.13 0.12 0.13 0.11 0.11 CD WW (m) (m) 0.3 0.3 0.3 0.4 0.4 0.4 0.35 0.4 0.6 0.6 0.7 0.7 CW (m) 0.6 0.8 0.6 0.8 0.8 1.2 1.2 j
Descriptions Ditch, no flow further upward Ditch discontinuous further upward with water grass Ditch continuous flow slow flow Not channelized falling tree over Natural flow Not channelized Natural flow with many small hollows Flood plain or semi‐wetland Not channelized Ditch with continuous flow Ditch with continuous Formation Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow Seepage erosion from under the tree roots Seepage erosion from under the tree roots Saturated overland flow Seepage from saturated flow Seepage from saturated WD (m) 0.05 0.09 0.1 flow Sharing head with site 17 at the highest dividing elevation Sharing head with site 16 at the highest dividing elevation flow Seepage from saturated flow Seepage from saturated flow 0.09 0.1 WW (m) 0.02 0.3 0.06 0.6 Confluence CW (m) Descriptions Connecting the other ditch Confluence with that of site 9 Spreading fast flow Natural fast disperse flow 0.7 0.8 Natural setream connecting to ditch at the upstream 1 1.1 0.35 0.4 Natural stream not channelized connecting to ditch at the upstream CD (m) 0.35 0.3 17/5/2012 11 12 13 14 15 16 17 18 19 20 0.05 0.08 0.08 0.02 0.04 0.03 0.06 0.05 0.6 0.5 0.35 0.45 0.3 0.3 0.4 0.3 0.6 0.45 0.3 0.4 0.25 0.4 0.85 0.7 1 1.2 0.8 1 0.5 0.6 1.7 Continuous flow ditch Continuous flow ditch with snow covered Ditch covered by snow the whole stream length sharing head with site 21 Ditch covered by snow the whole stream length sharing head with site 20 Covered by snow ditch Covered by snow ditch continuous flow snow melting contribute ditch continuous flow ditch continuous flow Ditch continuous flow snowmelting contribute Ditch no flow upward Seepage from saturated flow with the presence of water grass Seepage from saturated flow and snowmelting contribution Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow Seepage from saturated flow k
0.08 Natural stream not channelized connecting to ditch at the upstream Natural stream not channelized covered by snow Covered by snow covered by snow Covered by snow Connecting to road ditch 18/5/12 21 22 23 24 25 26 27 28 29 30 31 0.05 0.06 0.1 0.12 0.06 0.05 0.05 0.05 0.06 0.04 0.08 0.8 0.3 0.25 0.3 0.4 0.45 0.04 0.4 0.4 0.7 0.2 0.35 0.4 0.57 0.8 0.4 0.4 1.2 0.5 0.4 0.6 1.05 0.3 1 Not channelized widen road titch Widen road ditch Covered by the snow at the end Not channelized killed grass, natural flow continuous flow Ending with the tree snow covered nearby hillslope Discontinuous upwards of the stream head presence of channel erosion Ditch with discontinuous flow upwards with water grass presence of surface water Ditch with discontinuous flow upwards Ditch with continuous flow till the end Debris, killed grass, organic matters, at the base of hillslope ditch Presence of surface water upwards of the head Ditch Dividing at the highest elevation between 39 & 40 covered by snow Saturate surface flow Saturated flow Seepage from saturated flow Seepage from saturated head Seepage from saturated head Seepage erosion Seepage from saturated head Seepage from saturated head Seepage erosion from under the tree roots l
0.5 0.05 0.05 0.05 0.3 0.45 0.4 0.8 0.7 272 19/5/12 32 33 34 35 36 37 38 39 40 41 42 0.08 0.05 0.1 0.06 0.04 0.06 0.02 0.27 0.05 0.05 0.4 0.02 0.4 0.4 0.25 0.5 0.3 0.6 0.25 0.4 0.45 0.8 1.2 0.09 0.75 0.5 0.35 1 0.7 1 1.5 0.7 0.9 0.8 0.5 Dividing at the highest elevation between 39 & 40 covered by snow Debris and organic matter no flow upwards but little surface water no flow upwards of the head with water grass No flow upwards of the head covered by snow drain to road ditch Presence of surface water upwards Organic matter, killed grass rocks, water grass in the channel Not channelized, small hollows under the rocks and roots of tree, channelized downwards Ditch with little sufrace flow saturated upwards, under falling die tree with killed grass Not channelized, old shallow Ditch with little sufrace flow and water grass Ditch covered by snow head Subsurface flow covered bys snow Seepage from saturated head Seepage from saturated head Seepage from water grass upwards Seepage from saturated Seepage from saturated head Seepage from saturated head Seepage from saturated head Seepage from saturated m
43 44 45 46 47 Note: WW WD CD CW 0.07 0.3 0.25 0.6 no flow upwards No flow upwards, but saturated with water grass Natural stream, not channelized black color, presence of killed grass and organic matters Natural flow Natural flow Natural flow CW
CD
WW
head Seepage from saturated head Seepage erosion from under the died falling tree Overland flow Overland flow Overland flow WD
Width of water surface within the channel Depth of water measured at the deepest point Channel depth Channel width n
Tidigare utgivna publikationer i serien ISSN 1650-6553
Nr 1 Geomorphological mapping and hazard assessment of alpine areas in
Vorarlberg, Austria, Marcus Gustavsson
Nr 2 Verification of the Turbulence Index used at SMHI, Stefan Bergman
Nr 3 Forecasting the next day’s maximum and minimum temperature in Vancouver,
Canada by using artificial neural network models, Magnus Nilsson
Nr 4 The tectonic history of the Skyttorp-Vattholma fault zone, south-central Sweden,
Anna Victoria Engström
Nr 5 Investigation on Surface energy fluxes and their relationship to synoptic weather
patterns on Storglaciären, northern Sweden, Yvonne Kramer
Nr 237 Structural Model of the Lambarfjärden Area from Surface and Subsurface Data
in Connection with the E4 Stockholm Bypass Anna Vass, June 2012
Nr 238 Mechanisms Controlling Valley Asymmetry Development at Abisko, Northern
Sweden and Sani Pass, Southern Africa, Carl-Johan Borg, August 2012
Nr 239 Effect of Orientation on Propagation of Pre-existing fractures, Hajab Zahra,
August 2012
Nr 240 Mobility of multi-walled carbon nanotubes in unsaturated porous media,
Abenezer Mekonen, August 2012
Nr 241 Re-processing of Shallow and Deep Crustal Reflection Seismic Data along
BABEL Line7, Central Sweden, Hanieh Shahrokhi, August 2012
Nr 242 Usability of Standard Monitored Rainfall-Runoff Data in Panama,
Juan Diaz River Basin, José Eduardo Reynolds Puga, August 2012
Nr 243 Numerical Model of a Fossil Hydrothermal System in the Southern East Pacific
Rise Exposed at Pito Deep, Páll Halldór Björgúlfsson, September 2012
Nr 244 Regional Precipitation Study in Central America, Using the WRF Model
Tito Maldonado, September 2012
Nr 245 Reduktion av järn, mangan och CODMn i dricksvatten – Ett pilotförsök vid
Högåsens vattenverk, Tommy Olausson, September 2012
Nr 246Short-term Variations in Ice Dynamics During the Spring and Summer Period on
Storglaciären, Kebnekaise, Sweden, Helena Psaros, September 2012
Nr 247 Reassessment of the ‘last’ Goniopholidid Denazinosuchus kirtlandicus Wiman,
1932 from the Late Cretaceous of New Mexico, Oskar Bremer, October2012
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

Related manuals

Download PDF

advertisement