COUPLED HYDROLOGIC AND BIOGEOCHEMICAL RESPONSE TO INSECT- INDUCED FOREST DISTURBANCE by

COUPLED HYDROLOGIC AND BIOGEOCHEMICAL RESPONSE TO INSECT- INDUCED FOREST DISTURBANCE by
COUPLED HYDROLOGIC AND BIOGEOCHEMICAL RESPONSE TO INSECTINDUCED FOREST DISTURBANCE
by
Joel A. Biederman
A Dissertation Submitted to the Faculty of the
DEPARTMENT OF HYDROLOGY AND WATER RESOURCES
In Partial Fulfillment of the Requirements
For the Degree of
DOCTOR OF PHILOSOPHY
WITH A MAJOR IN HYDROLOGY
In the Graduate College
THE UNIVERSITY OF ARIZONA
2013
2
THE UNIVERSITY OF ARIZONA
GRADUATE COLLEGE
As members of the Dissertation Committee, we certify that we have read the dissertation
prepared by Joel A. Biederman entitled “Coupled hydrologic and biogeochemical
response to insect-induced forest disturbance” and recommend that it be accepted as
fulfilling the dissertation requirement for the Degree of Doctor of Philosophy.
_______________________________________________________________________
Date: September 24, 2013
Dr. Paul Brooks
_______________________________________________________________________
Date: September 24, 2013
Dr. Shirley Papuga
_______________________________________________________________________
Date: September 24, 2013
Dr. Thomas Meixner
_______________________________________________________________________
Date: September 24, 2013
Dr. David Gochis
Final approval and acceptance of this dissertation is contingent upon the candidate’s
submission of the final copies of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under my direction and
recommend that it be accepted as fulfilling the dissertation requirement.
________________________________________________ Date: September 24, 2013
Dissertation Director: Dr. Paul Brooks
3
STATEMENT BY AUTHOR
This dissertation has been submitted in partial fulfillment of the requirements for
an advanced degree at the University of Arizona and is deposited in the University
Library to be made available to borrowers under rules of the Library.
Brief quotations from this dissertation are allowable without special permission,
provided that an accurate acknowledgement of the source is made. Requests for
permission for extended quotation from or reproduction of this manuscript in whole or in
part may be granted by the head of the major department or the Dean of the Graduate
College when in his or her judgment the proposed use of the material is in the interests of
scholarship. In all other instances, however, permission must be obtained from the
author.
SIGNED: Joel A. Biederman
4
ACKNOWLEDGMENTS
I would like to thank the students, collaborators and mentors that contributed
many hours of effort to this project. In the field Traeger Meyer and Nick Ludolph of
Colorado University braved both winter and summer at 10,000 feet to keep the field sites
running. Meanwhile Allison Peterson, Gary Gold, Chris Ferlin and Shane Clark of the
University of Arizona braved the Tucson summer to analyze many hundreds of water
samples. David Reed, Gary Bolton, Scott Peckham and Lisa Philander of the University
of Wyoming made it possible to conduct research at a remote mountain site. Thanks to
David Gochis and Ethan Gutmann of the National Center for Atmospheric Research for
hospitality and training during my summer visits to NCAR. Thank you to Drs. Adrian
Harpold, Brent Ewers, Elise Pendall, Peter Troch, James Shuttleworth, Hoshin Gupta and
Ty Ferre for academic and research guidance.
The following funding sources made it possible for me to complete this work by
either providing scholarship and fellowship funds, or funds for field equipment and
analytical services: National Science Foundation, Water Sustainability Program Graduate
Fellowship, Science Foundation Arizona, Institute of the Environment, the Department of
Hydrology and Water Resources, and the Graduate Student and Professional Council.
I would like to thank my advisory and technical committee, Dr. Thomas Meixner,
Dr. Shirley Papuga, Dr. David Gochis, and of course, my major advisor, Dr. Paul D.
Brooks. Thank you for your guidance and friendship. This has been a transformative
experience.
5
DEDICATIONS
Thank you:
Joel Cahoon, my undergraduate advisor, for training me rigorously and supporting me
endlessly in all aspects of life.
Warren Jones, undergraduate professor, for teaching me about engineering, life, and
especially teaching.
My parents Chuck and Loraine: Your endless faith motivated me to earn this Ph.D. at
last.
My wife, Linda: This research would not have been possible without your support, faith,
patience and encouragement.
My children Axel and Lucie: You gave up snow and sledding to move to Arizona and
help me accomplish this dream.
6
TABLE OF CONTENTS
ABSTRACT...................................................................................................................... 11
1.
2.
INTRODUCTION ................................................................................................... 13
1.1.
PROBLEM STATEMENT ........................................................................... 13
1.2.
BACKGROUND .......................................................................................... 14
1.3.
DISSERTATION FORMAT ........................................................................ 20
PRESENT STUDY ................................................................................................. 22
2.1.
APPENDIX A. MULTISCALE OBSERVATIONS OF SNOW
ACCUMULATION AND PEAK SNOWPACK FOLLOWING WIDESPREAD,
INSECT-INDUCED LODGEPOLE PINE MORTALITY..................................... 22
2.2.
2.1.1.
Study Objective .......................................................................... 22
2.1.2.
Major Study Findings and Implications ..................................... 22
APPENDIX B. COMPENSATORY VAPOR FLUX REDUCES WATER
FOR STREAMFLOW FOLLOWING SEVERE BARK BEETLE-INDUCED
FOREST MORTALITY. ........................................................................................ 24
2.3.
2.2.1.
Study Objective .......................................................................... 24
2.2.2.
Major Study Findings and Implications ..................................... 24
APPENDIX C. CATCHMENT BIOGEOCHEMICAL IMPACTS OF BARK
BEETLE INFESTATION ATTENUATED IN THE RIPARIAN ZONE AND
HEADWATER STREAMS .................................................................................... 26
2.3.1.
Study Objective .......................................................................... 26
2.3.2.
Major Study Findings and Implications ..................................... 26
7
TABLE OF CONTENTS - Continued
3. REFERENCES ............................................................................................................. 28
APPENDIX A: MULTISCALE OBSERVATIONS OF SNOW ACCUMULATION
AND PEAK SNOWPACK FOLLOWING WIDESPREAD, INSECT-INDUCED
LODGEPOLE PINE MORTALITY................................................................................. 33
Abstract ................................................................................................................... 34
1.
Introduction .................................................................................................. 36
2.
Methods ......................................................................................................... 39
3.
2.1.
Study Sites ................................................................................... 40
2.2.
Winter Climate............................................................................ 42
2.3.
Intensive Study Plots................................................................... 43
2.4.
Snow Surveys at peak accumulation........................................... 44
2.5.
Isotopic tracers of sublimation:.................................................. 47
Results ........................................................................................................... 48
3.1.
Weather....................................................................................... 48
3.2.
Snowfall Interception.................................................................. 49
3.3.
Spatial distribution of peak snowpack ........................................ 50
3.4.
Comparison of peak snowpack between MPB and Unimpacted 51
3.5.
Subcanopy environment.............................................................. 52
3.6.
Stable isotope tracers of sublimation ......................................... 53
4.
Discussion ..................................................................................................... 53
5.
Conclusions ................................................................................................... 61
8
TABLE OF CONTENTS - Continued
6.
Acknowledgments ......................................................................................... 62
7.
References ..................................................................................................... 63
8.
Tables ............................................................................................................ 69
9.
Figures .......................................................................................................... 72
APPENDIX B: COMPENSATORY VAPOR FLUX REDUCES WATER FOR
STREAMFLOW FOLLOWING SEVERE BARK BEETLE-INDUCED FOREST
MORTALITY. .................................................................................................................. 79
Abstract ................................................................................................................... 80
1.
Introduction .................................................................................................. 81
2.
Study sites ...................................................................................................... 84
3.
Methods ......................................................................................................... 85
4.
3.1.
Quantification of MPB-driven Forest Mortality ........................ 85
3.2.
Local Climate and Soil Moisture Measurements ....................... 86
3.3.
Eddy Covariance Vapor Flux Measurements............................. 87
3.4.
Streamflow and Groundwater Measurements (MPB site only) .. 88
3.5.
Water Balance Quantification .................................................... 89
3.6.
Stable Isotope Observations ....................................................... 90
3.7.
Computational methods and software ........................................ 91
Results ........................................................................................................... 91
4.1.
MPB Infestation .......................................................................... 91
4.2.
Local Climate and Soil Moisture ................................................ 92
9
TABLE OF CONTENTS - Continued
4.3.
Tower Footprint Water Balance ................................................. 93
4.4.
Stable Isotope Indicators of Evaporation ................................... 94
4.5.
Catchment Water Balance (MPB Site Only) .............................. 95
4.6.
Comparison of Sites and Water Balance Approaches ................ 97
5.
Discussion ..................................................................................................... 97
6.
Conclusions ................................................................................................. 102
7.
Acknowledgments ....................................................................................... 103
8.
References ................................................................................................... 104
9.
Tables .......................................................................................................... 110
10.
Figures ........................................................................................................ 115
APPENDIX C: CATCHMENT BIOGEOCHEMICAL IMPACTS OF BARK BEETLE
INFESTATION ATTENUATED IN THE RIPARIAN ZONE AND HEADWATER
STREAMS. ..................................................................................................................... 126
Abstract: ................................................................................................................ 127
1.
Introduction ................................................................................................ 128
2.
Study Sites ................................................................................................... 131
3.
Methods ....................................................................................................... 132
3.1.
Phsyical Hydrology and Weather ............................................. 132
3.2.
Biogeochemical Sampling and Analysis ................................... 133
3.3.
Catchment Spatial Analysis and Statistics ............................... 135
10
TABLE OF CONTENTS - Continued
4.
Results ......................................................................................................... 135
4.1.
Weather and Hydrological Response ....................................... 135
4.2.
Unimpacted Biogeochemistry From Uplands to RSS ............... 136
4.3.
MPB Biogeochemistry From Uplands to RSS .......................... 138
4.4.
Spatial Patterns of Chemistry in Nested MPB Streams............ 140
5.
Discussion ................................................................................................... 141
6.
Implications and Conclusions .................................................................... 147
7.
Acknowledgments ....................................................................................... 148
8.
References ................................................................................................... 149
9.
Tables .......................................................................................................... 157
10.
Figures ........................................................................................................ 160
11
ABSTRACT
Forest disturbance is expanding in rate and extent and is affecting many montane
catchments critical to water resources.
Western North America is experiencing an
epidemic of mountain pine beetle (MPB) that has affected 20 million hectares of forest in
Canada and the United states. This epidemic may have long-lasting consequences for
coupled cycles of water, energy, and biogeochemicals.
While impacts of forest
disturbance by fire and harvest have been studied for more than a half-century, insectdriven mortality differs from these events in the timing and accompanying biophysical
impacts. In this work, we quantified catchment hydrologic and hydrochemical response
to severe MPB infestation in a lodgepole pine ecosystem. Observations were organized
laterally in a nested fashion from soil observations to nested headwater catchments.
Vertical observations encompassed what is often termed the critical zone, from
atmospheric interactions at the top of the forest through the ground surface and the
rooting zone to the interface with groundwater. We quantified responses manifest in
snowpack, the primary hydrologic input to this montane ecosystem, in water partitioning
between vapor flux and streamflow, and in biogeochemical patterns across the landscape.
Key findings of this study include 1) Loss of shelter from the atmosphere caused
compensatory sublimation of snowpack to offset decreased interception losses after
MPB-driven canopy loss; 2) Vaporization at multiple scales increased over time and in
comparison to control forest, reducing water available for streamflow; 3) Nitrogen (N)
concentrations were elevated in hillslope groundwater, but attenuation in the riparian
zone protected streams from major N influx; and 4) headwater streams rapidly attenuated
12
dissolved carbon (C) and N inputs. Collectively these results demonstrate compensatory
negative feedbacks which help explain the lack of strong response to streamflow and
stream chemistry observed in the recent MPB epidemic.
13
1. INTRODUCTION
1.1. PROBLEM STATEMENT
A rapid, extensive, and ongoing epidemic of mountain pine beetle (MPB) has
damaged forests in more than 20 million hectares of western North America since the
mid-1990s (Safranyik et al., 2010; USFS, 2012).
Although an extensive body of
literature documents hydrological and biogeochemical effects of forest disturbance in
headwater catchments, there is a paucity of observational evidence related to mortality
that is gradual and initially leaves the forest and soils physically intact. Specifically, it is
unclear how:
1) tradeoffs between interception losses and sheltering of snowpack will
combine to alter net snowpack accumulation,
2) tradeoffs among reduced interception, reduced overstory transpiration, and
increased evaporation will influence partitioning of precipitation between
vaporization and water available for streamflow,
3) increased litter inputs, changes to microclimates, and physical hydrologic
impacts of items 1) and 2) will alter hydrochemical concentrations in soil
water and groundwater, and
4) how the above effects will be propagated downstream from the headwaters
toward streams of relevance to water resources.
These points are addressed in order to improve current conceptual models
regarding forest ecosystem function under insect-driven disturbance and to provide
14
relevant information that can aid in the development of forest and water management
strategies in affected regions.
1.2. BACKGROUND
Forested montane catchments are critical to downstream water resources and the
regulation of biogeochemical fluxes. Research from the last half-century in experimental
watersheds has demonstrated that forested watersheds are reliable sources of clean water
(Brown et al., 2005).
In western North America, much of this water derives from
snowmelt, and more than 60 million water users rely on snowmelt as a primary supply
(Bales et al., 2006).
In recent decades, forest die-off has increased due to increased temperature
(Williams et al., 2010) drought (Breshears et al., 2009; Adams et al., 2012; Allen et al.,
2010 ), fire, and pathogen infestation (Williams et al., 2013; Huber, 2005; Tokuchi et al.,
2004). Western North American forests have suffered extensive tree die-off due to
mountain pine beetle (MPB; Dendroctonus ponderoseae) and beetle-associated fungal
pathogens (Raffa et al., 2008; Kurz et al., 2008; USFS, 2012). Since forests regulate
water partitioning and biogeochemical fluxes, the unprecedented MPB epidemic
challenges our ability to predict availability of water and limiting nutrients for
ecosystems (Edburg et al., 2012) and the amount and quality of downstream resources
(Ellison et al., 2012).
Understanding response in headwater catchments and their streams is critical for
several reasons. First, headwater streams comprise 86% of total stream length and the
majority of shoreline in most drainage networks (Leopold, 1964). Second, headwater
15
streams are dominated by inputs from the hillslopes, and therefore they may be more
sensitive to hillslope disturbance. Third, headwater streams are relative hot spots for
biogeochemical transformation (McLain et al., 2003) due to their low water velocities,
turbulent flow, and high ratio of surface area to volume, all of which promote hyporheic
exchange (Peterson et al., 2001). Finally, larger-order streams integrate the response of
the headwaters. For example, in the snowmelt-dominated Rocky Mountains, Troendle et
al. (2001) found similar fractional flow increases in response to harvest in a basin of 17
km2 and in small headwater catchments.
Insect-driven forest mortality may affect hydrology differently from other forest
disturbances including non-lethal insect infestation (e.g. Swank, 1981; Bormann and
Likens, 1979), harvest (e.g. Bosch and Hewlett, 1982; Jones and Post, 2004; Troendle et
al., 2001), ice storms (Houlton et al., 2003; Bernhardt et al., 2003) and fire (e.g. DeBano,
2000).
MPB may kill a large fraction of available host trees without immediately
changing forest structure and soil hydraulic conductivity (Brown et al., 2010).
While
subsequent needle drop results in decreased snowfall interception loss (Boon, 2012; Pugh
and Small, 2012), compensatory increases to winter snowpack sublimation may negate
any change to the volume of snowmelt entering the catchment (Boon, 2012; Biederman et
al., 2012). Increased energy inputs to spring snowpack increase melt rates (Boon, 2009;
Biederman et al., 2012; Pugh and Small, 2012) which could reduce snowmelt infiltration,
advance flow timing, and increase the magnitude of peak streamflow (Pugh and Gordon,
2012). While decreased transpiration could be expected to increase soil moisture and
streamflow (Pugh and Gordon, 2012), observations indicate that growing season vapor
16
fluxes remain near to or exceed those of control forests (Brown et al., 2013; Reed et al.,
2010; Biederman et al., in review) and that streamflow may decline Somor et al. 2010)
Direct observations of response to MPB have been grouped into small-scale
response (e.g. scale of plots, trees, and stands of trees) and large-scale response (e.g.
records at long-term gauged basins of many square km). Reduced snow interception and
transpiration are expected to increase streamflow (Pugh and Gordon, 2012), while
increased organic matter decay is expected to increase biogeochemical stream fluxes
(Morehouse et al., 2008; Griffin et al., 2011). Plot-scale observations have documented
expected biophysical and biogeochemical changes, but there has been little significant
change to streamflow or water quality at the larger scales relevant to water resources
(Clow et al., 2011; Rhoades et al., 2013). A critical gap exists in our understanding of
process changes that have prevented small-scale impacts observed in the uplands from
propagating to the watershed scales of water resource relevance.
Streamflow response to forest disturbance has been the subject of research for
more than a half century (Jones and Post, 2004). However, mountain pine beetle-driven
mortality differs in its biophysical impacts from the more well-studied cases of fire and
harvest. It kills the majority of host trees but does not remove any biomass or disturb the
soils as do both harvest and fire (DeBano, 2000).
Further, the majority of disturbance
studies have occurred in more humid environments as opposed to the boreal and
subalpine zones where the MPB epidemic is concentrated (Raffa et al., 2008). A review
of disturbance experiments in 14 experimental watersheds showed a median value of
mean annual precipitation of 1450 mm, which is nearly twice as large as the mean annual
17
precipitation at the sites used in the present study (Jones and Post, 2004; Biederman et al.,
2012).
Streamflow in all seasons is dominated by snowmelt in many of the montane
catchments of western North America affected by MPB, and various studies have aimed
to quantify snowpack response to forest disturbance (e.g. Connaughton, 1935; Troendle
and King, 1983; Golding and Swanson, 1986). Plot-based studies of MPB and control
stands have shown either increased snowpack beneath killed trees (Pugh and Small,
2012; Boon, 2012) or no response (Boon, 2012). The prevailing expectation has been
that decreased interception would lead to larger snowpack accumulation, eventually
driving greater streamflow (Pugh and Gordon, 2012). This interpretation is codified in
the structures of some land surface models, where simulated vaporization declines with
leaf area index under simulated disturbance, increasing streamflow (Bewley et al., 2010,
Pomeroy et al., 2012; Mikkelson et al., 2013).
Empirical observations of streamflow following MPB are less clear. Bethlahmy
(1974) found a small streamflow increase peaking 15 years after infestation, Potts (1984)
observed a small increase immediately, and Somor (2010) reported significantly less
streamflow in one catchment and no significant change in seven others. Our inability to
explain the variable response in streamflow of gauged basins highlights the need for
understanding of underlying process responses in the headwaters. While recent work has
sought to quantify individual water fluxes after MPB (Boon, 2012; Pugh and Small,
2012; Brown et al., 2013), we are not aware of any previous studies combining
18
observations of vaporization and streamflow to quantify water balance terms in
catchments affected by MPB.
Changes in the physical hydrology may combine with a rapid increase in dead
organic matter to produce biogeochemical response to MPB disturbance. Recent work on
biogeochemical response to MPB has provided conceptual frameworks and initial results
to guide ongoing research. Kurz et al. (2008) used a land surface model to predict
changes to net biome production of carbon (NBP) over a 14-year period in the region of
western Canada most affected by the present MPB outbreak. They reported that during
the peak of the epidemic and for several years afterward, the impacted forest was
converted from a small carbon sink to a large carbon source. Edburg et al., (2012)
suggested that NPB would be depressed for 5-10 years or more following mortality,
though they discussed variability related to coupling of biogeochemical and biophysical
responses (e.g. snowmelt dynamics, understory growth). Edburg et al. (2012) proposed a
variety of possible ecosystem responses to additional nitrogen inputs from decaying plant
matter. These include short-term retention by microbial biomass and eventual transfer to
new plant growth (Brooks et al., 1998) and the possibility that denitrification (Brooks et
al, 1997) or export in solution (Brooks and Williams, 1999) might limit plant growth and
NBP in the nitrogen-poor ecosystems typically infested by MPB. Mikkelson et al. (2012)
reported increased total organic carbon and production of disinfection byproducts at
Colorado water treatment plants with source water from MPB-infested catchments. Clow
et al. (2011) found increased soil C and N beneath MPB-killed trees but no changes in C
or nitrate at the outlets of catchments with areas of 101-102 km2. Rhoades et al. (2013)
19
attribute the surprising lack of stream nitrate response to a combination of low regional
atmospheric N deposition in the Rocky Mountains, patchy and time-varying tree
mortality, and compensatory uptake by understory and surviving trees. There is therefore
a lack of consensus regarding the fate of increased inputs of C and N.
The dichotomy of strong response in uplands and weak response in streams
highlights our need to address MPB impacts on catchment biogeochemical cycling across
spatial scales from uplands to streams. Evaluating response across the landscape will
improve our understanding of how MPB impacts propagate through catchment fluxes and
pools.
Given the variability of responses reported in the literature, a number of questions
arise on the response of headwater catchments to MPB-driven insect mortality:
1) How does canopy loss affect net snowpack accumulation in headwater
catchments?
2) How does tree mortality affect partitioning of precipitation into vaporization and
water available for streamflow?
3) How do hydrological and other biophysical changes propagate across landscape
biogeochemistry from the uplands to the riparian-stream system?
4) How do changes in snowpack, vaporization and hydrochemistry propagate
downstream from headwater catchments?
To quantify hydrological and biogeochemical response to MPB, we observed a
site with heavy MPB infestation and compared it to a similar unimpacted site (i.e. control
site). Continuous snow depth observations quantified changes to snow inputs, while
20
distributed snow surveys of peak accumulation determined net impacts.
Water
partitioning response was quantified using streamflow observations in a catchment water
balance as well as eddy covariance vaporization observations in a vertical “1dimensional” water balance framework. Stable water isotopes were used to evaluate
abiotic vaporization (i.e. the sum of sublimation and evaporation) at multiple scales.
Biogeochemical response was quantified using daily to biweekly sampling in a nested
design of soil water, groundwater, riparian groundwater and stream water in catchments
of four stream orders.
1.3. DISSERTATION FORMAT
This dissertation introduction is followed by a section titled “Present Study”,
which summarizes the findings of 3 manuscripts drafted for journal submission. The
publications drafted for publication follow the “Present Study” as 3 separate appendices,
Appendix A, B and C. The titles and my personal contributions to Appendices A, B and
C are:
Appendix A: “Multiscale observations of snow accumulation and peak snowpack
following widespread, insect-induced lodgepole pine mortality.”
My personal
contribution was the collection and analysis of the data and publication of a scientific
article. This work is accepted by the journal Ecohydrology and although not yet included
in an issue, it has been made available online by the publisher at DOI: 10.1002/eco.1342
This work is included in this dissertation in accepted manuscript format, for consistency
of style with the other appendices. Wiley Science and Ecohydrology do not require any
copyright permissions for inclusion of a published article in the dissertation by the same
21
author as described at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)19360592/homepage/Permissions.
Appendix B: “Compensatory vapor flux reduces water for streamflow following
severe bark beetle-induced forest mortality.” My personal contribution was the collection
of the streamflow and water isotopic data, analysis of these and eddy covariance data, and
generation of a manuscript for publication which is under review at the journal Water
Resources Research.
Appendix C: “Catchment biogeochemical impacts of bark beetle infestation
attenuated in the riparian zone and headwater streams.” My personal contribution was
the installation of instrumentation, collection of water samples, data analysis and
generation of a manuscript for publication.
22
2.
2.1.
APPENDIX
A.
PRESENT STUDY
MULTISCALE
OBSERVATIONS
OF
SNOW
ACCUMULATION AND PEAK SNOWPACK FOLLOWING WIDESPREAD,
INSECT-INDUCED LODGEPOLE PINE MORTALITY.
2.1.1.
Study Objective
This study addresses the overarching question: How does insect-induced forest
die-off affect peak snowpack and the potential snow water inputs to affected catchments?
2.1.2.
Major Study Findings and Implications
The major findings of this study are:
1)
Snowfall interception declined from 20% of annual snowfall in healthy
lodgepole pine forest to a number indistinguishable from zero in grey-phase trees 3-4
years following mortality.
2)
There was no significant difference in peak snowpack between healthy
forest and MPB-infested forest in either red or grey phases of mortality, indicating that
reduced interception losses were compensated by winter sublimation of snowpack from
the ground.
3)
Stable water isotopes confirmed greater kinetic fractionation of snowpack
under MPB forest, confirming increased snowpack sublimation.
4)
Solar radiation reaching the snowpack was on average 70% greater in the
MPB forest as compared to the Unimpacted forest, while subcanopy wind speeds were
relatively similar.
23
5)
Peak snowpack in MPB forest showed reduction in the spatial variability
associated with interception and shading by trees.
This work was unique in surveying snowpack across small headwater catchments
as compared to prior work that focused mainly on adjacent plots or stands. These results
show that reduced snowfall interception losses may be counteracted by increased
sublimation of snowpack on the ground. This contrasts with a common perception that
interception is the dominant control on peak snowpack accumulation and that snow
reaching the snowpack is relatively immune to ablation. Snowmelt water is the dominant
annual input to many montane catchments affected by forest disturbance, and our results
help explain why expected streamflow increases have not been reported after nearly 15
years of the current MPB epidemic in the central Rocky Mountains.
24
2.2.
APPENDIX B. COMPENSATORY VAPOR FLUX REDUCES WATER FOR
STREAMFLOW FOLLOWING SEVERE BARK BEETLE-INDUCED FOREST
MORTALITY.
2.2.1.
Study Objective
This study addresses the overarching question: How does insect-induced forest
mortality affect the partitioning of precipitation between vapor losses and water available
for streamflow?
2.2.2.
Major Study Findings and Implications
The major findings of this study are:
1)
Net vapor losses increased following MPB-induced forest mortality
2)
Streamflow declined in headwater catchments with heavy insect-induced
forest mortality
3)
A large fraction of the vapor losses occurs during snowmelt, suggesting
that abiotic evaporation from wet surfaces plays an important role.
4)
Stable water isotopes show greater kinetic fractionation in soil and stream
water in MPB-infested catchments, indicating increased abiotic evaporation.
5)
Solar radiation reaching the forest floor under grey-phase MPB forest was
300-350% larger during the snow-free period as compared to unimpacted forest.
In addition to showing that vapor losses may increase and streamflow may
decrease following severe forest mortality, these results demonstrate that abiotic
evaporation is more important in hydrological response to forest disturbance than
previously reported. This study shows that evaporation of snowmelt water, soil water
25
and stream water may increase sufficiently to collectively offset the reduced transpiration
resulting from forest die-off.
This result is consistent with our previous analysis
showing increased snowpack sublimation during winter (Appendix A), as both
sublimation and evaporation represent abiotic fluxes from the subcanopy. Model-based
studies of MPB impacts on water balance have generally predicted that decreased
transpiration would lead to lower vapor losses and greater streamflow until forest
regeneration could make a significant difference. Our results suggest that the land surface
models commonly used to evaluate and predict water balance following forest
disturbance should include improved representation of abiotic vapor fluxes.
26
2.3.
APPENDIX C. CATCHMENT BIOGEOCHEMICAL IMPACTS OF BARK
BEETLE INFESTATION ATTENUATED IN THE RIPARIAN ZONE AND
HEADWATER STREAMS.
2.3.1.
Study Objective
This study addresses the overarching question: Where on the landscape does
attenuation of dissolved carbon and nitrogen occur in MPB-infested catchments?
2.3.2.
Major Study Findings and Implications
The major findings of this study are:
1)
Soil water concentrations of dissolved organic carbon (DOC) and nitrogen
(DON) are relatively similar in MPB-infested and unimpacted uplands 3 to 5 years
following infestation.
2)
Nitrate concentrations are 1 to 2 orders of magnitude greater in MPB-
infested uplands.
3)
Both DON and nitrate are elevated in MPB upland groundwater,
suggesting significant flushing of N from surface soils has already occurred 3 to 5 years
following MPB infestation.
4)
Most of the elevated N concentrations in MPB groundwater are attenuated
in the riparian-stream system (RSS), while DOC concentrations increase upon entering
the RSS.
5)
DOC and DON are rapidly attenuated downstream in headwater streams,
with 55% and 69% lower concentrations, respectively, within 5 km of flow distance in
the channel network.
27
These results fill a critical knowledge gap in our understanding of the locations
and spatial scales of C and N attenuation following MPB forest mortality. Existing work
has documented increased C and N in the top 20 cm of soils as well as very low response
in the chemistry of rivers draining MPB-infested mountains, but here we show that 1)
within 3-5 years following infestation, the primary impact is observed to be elevated N in
groundwater, 2) both organic and inorganic N are attenuated between hillslope
groundwater and the RSS and 3) attenuation of DON and DOC occurs rapidly over just a
few km in the headwaters, consistent with the lack of response previously reported for
large watersheds.
28
3. REFERENCES
Adams, H. D., Luce, C. H., Breshears, D. D., Allen, C. D., Weiler, M., Hale, V. C.,
Huxman, T. E. (2012). Ecohydrological consequences of drought- and infestationtriggered tree die-off: insights and hypotheses. Ecohydrology, 5(2), 145-159. doi:
10.1002/eco.233
Allen, C. D., Macalady, A. K., Chenchouni, H., Bachelet, D., McDowell, N., Vennetier,
M., Cobb, N. (2010). A global overview of drought and heat-induced tree
mortality reveals emerging climate change risks for forests. Forest Ecology and
Management, 259(4), 660-684. doi: 10.1016/j.foreco.2009.09.001
Bales, R. C., Molotch, N. P., Painter, T. H., Dettinger, M. D., Rice, R., & Dozier, J.
(2006). Mountain hydrology of the western United States. Water Resources
Research, 42(8). doi: 10.1029/2005wr004387
Bernhardt, E. S., Likens, G. E., Buso, D. C., & Driscoll, C. T. (2003). In-stream uptake
dampens effects of major forest disturbance on watershed nitrogen export.
Proceedings of the National Academy of Sciences of the United States of America,
100(18), 10304-10308. doi: 10.1073/pnas.1233676100
Bethlahmy, N. (1974). More streamflow after a bark-beetle epidemic. Journal of
Hydrology, 23(3-4), 185-189.
Bewley, D., Alila, Y., & Varhola, A. (2010). Variability of snow water equivalent and
snow energetics across a large catchment subject to Mountain Pine Beetle
infestation and rapid salvage logging. Journal of Hydrology, 388(3-4), 464-479.
doi: 10.1016/j.jhydrol.2010.05.031
Biederman, J. A., Brooks, P. D., Harpold, A. A., Gutmann, E., Gochis, D. J., Reed, D. E.,
& Pendall, E. (2012). Multi-scale Observations of Snow Accumulation and Peak
Snowpack Following Widespread, Insect-induced Lodgepole Pine Mortality.
Ecohydrology.
Boon, S. (2012). Snow accumulation following forest disturbance. Ecohydrology. doi:
doi 10.1002/eco.212.
Bormann, F. H., & Likens, G. E. (1979). Catastrophic disturbance and the steady-state in
northern hardwood forests. American Scientist, 67(6), 660-669.
Bosch, J. M., & Hewlett, J. D. (1982). A review of catchment experiments to determine
the effect of vegetation changes on water yield and evapo-transpiration. Journal
of Hydrology, 55(1-4), 3-23. doi: 10.1016/0022-1694(82)90117-2
Breshears, D. D., Myers, O. B., Meyer, C. W., Barnes, F. J., Zou, C. B., Allen, C. D., . . .
Pockman, W. T. (2009). Tree die-off in response to global change-type drought:
mortality insights from a decade of plant water potential measurements. Frontiers
in Ecology and the Environment, 7(4), 185-189. doi: 10.1890/080016
29
Brooks, P. D., Schmidt, S. K., & Williams, M. W. (1997). Winter production of CO2 and
N2O from Alpine tundra: Environmental controls and relationship to inter-system
C and N fluxes. Oecologia, 110(3), 403-413. doi: 10.1007/pl00008814
Brooks, P. D., & Williams, M. W. (1999). Snowpack controls on nitrogen cycling and
export in seasonally snow-covered catchments. Hydrological Processes, 13(1415), 2177-2190. doi: 10.1002/(sici)1099-1085(199910)13:14/15<2177::aidhyp850>3.0.co;2-v
Brooks, P. D., Williams, M. W., & Schmidt, S. K. (1998). Inorganic nitrogen and
microbial biomass dynamics before and during spring snowmelt.
Biogeochemistry, 43(1), 1-15. doi: 10.1023/a:1005947511910
Brown, A. E., Zhang, L., McMahon, T. A., Western, A. W., & Vertessy, R. A. (2005). A
review of paired catchment studies for determining changes in water yield
resulting from alterations in vegetation. Journal of Hydrology, 310(1-4), 28-61.
doi: 10.1016/j.jhydrol.2004.12.010
Brown, M., Black, T. A., Nesic, Z., Foord, V. N., Spittlehouse, D. L., Fredeen, A. L.,
Trofymow, J. A. (2010). Impact of mountain pine beetle on the net ecosystem
production of lodgepole pine stands in British Columbia. Agricultural and Forest
Meteorology, 150(2), 254-264. doi: 10.1016/j.agrformet.2009.11.008
Brown, M. G., Black, T. A., Nesic, Z., Foord, V. N., Spittlehouse, D. L., Fredeen, A. L.,
Meyer, G. (2013). - Evapotranspiration and canopy characteristics of two
lodgepole pine stands following mountain pine beetle attack. doi: 10.1002/hyp.9870
Clow, D. W., Rhoades, C., Briggs, J., Caldwell, M., & Lewis, W. M., Jr. (2011).
Responses of soil and water chemistry to mountain pine beetle induced tree
mortality in Grand County, Colorado, USA. Applied Geochemistry, 26, S174S178. doi: 10.1016/j.apgeochem.2011.03.096
Connaughton, C. A. (1935). The accumulation and rate of melting snow influenced by
vegetation. Journal of Forestry, 33, 564-569.
DeBano, L. F. (2000). The role of fire and soil heating on water repellency in wildland
environments: a review. Journal of Hydrology, 231, 195-206. doi: 10.1016/s00221694(00)00194-3
Edburg, S. L., Hicke, J. A., Brooks, P. D., Pendall, E. G., Ewers, B. E., Norton, U., . . .
Meddens, A. J. H. (2012). Cascading impacts of bark beetle-caused tree mortality
on coupled biogeophysical and biogeochemical processes. Frontiers in Ecology
and the Environment, 10(8), 416-424. doi: 10.1890/110173
Ellison, D., Futter, M. N., & Bishop, K. (2012). On the forest cover-water yield debate:
from demand- to supply-side thinking. Global Change Biology, 18(3), 806-820.
doi: 10.1111/j.1365-2486.2011.02589
30
Golding, D. L., & Swanson, R. H. (1986). Snow distribution patterns in clearings and
adjacent forest. Water Resources Research, 22(13), 1931-1940.
Griffin, J. M., Turner, M. G., & Simard, M. (2011). Nitrogen cycling following mountain
pine beetle disturbance in lodgepole pine forests of Greater Yellowstone. Forest
Ecology and Management, 261(6), 1077-1089. doi: 10.1016/j.foreco.2010.12.031
Houlton, B. Z., Driscoll, C. T., Fahey, T. J., Likens, G. E., Groffman, P. M., Bernhardt,
E. S., & Buso, D. C. (2003). Nitrogen dynamics in ice storm-damaged forest
ecosystems: Implications for nitrogen limitation theory. Ecosystems, 6(5), 431443. doi: 10.1007/s10021-002-0198-1
Huber, C. (2005). Long lasting nitrate leaching after bark beetle attack in the highlands of
the Bavarian Forest National Park. Journal of Environmental Quality, 34(5),
1772-1779. doi: 10.2134/jeq2004.0210
Jones, J. A., & Post, D. A. (2004). Seasonal and successional streamflow response to
forest cutting and regrowth in the northwest and eastern United States. Water
Resources Research, 40(5), 19. doi: 10.1029/2003wr002952
Kurz, W. A., Dymond, C. C., Stinson, G., Rampley, G. J., Neilson, E. T., Carroll, A. L.,
Safranyik, L. (2008). Mountain pine beetle and forest carbon feedback to climate
change. Nature, 452(7190), 987-990. doi: 10.1038/nature06777
Leopold, L. B., Wolman, M. G., and Miller, T. P. (1964). Fluvial Processes in
Geomorphology,W. H. Freeman and Company, San Francisco. Washington, DC:
U.S. Department of Agriculture, Forest Service, Forest Health Monitoring
Program. 123 p.
McClain, M. E., Boyer, E. W., Dent, C. L., Gergel, S. E., Grimm, N. B., Groffman, P.
M., Pinay, G. (2003). Biogeochemical hot spots and hot moments at the interface
of terrestrial and aquatic ecosystems. Ecosystems, 6(4), 301-312. doi:
10.1007/s10021-003-0161-9
Mikkelson, K., Dickenson, E., Maxwell, R., McCray, J., & Sharp, J. (2012). Waterquality impacts from climate-induced forest die-off Nature- Climate change. doi:
DOI: 10.1038/NCLIMATE1724
Mikkelson, K. M., Maxwell, R. M., Ferguson, I., Stednick, J. D., McCray, J. E., & Sharp,
J. O. (2013). Mountain pine beetle infestation impacts: modeling water and
energy budgets at the hill-slope scale. Ecohydrology, 6(1), 64-72. doi:
10.1002/eco.278
Morehouse, K., Johns, T., Kaye, J., & Kaye, A. (2008). Carbon and nitrogen cycling
immediately following bark beetle outbreaks in southwestern ponderosa pine
forests. Forest Ecology and Management, 255(7), 2698-2708. doi:
10.1016/j.foreco.2008.01.050
31
Peterson, B. J., Wollheim, W. M., Mulholland, P. J., Webster, J. R., Meyer, J. L., Tank, J.
L., Morrall, D. D. (2001). Control of nitrogen export from watersheds by
headwater streams. Science, 292(5514), 86-90. doi: 10.1126/science.1056874
Pomeroy, J., Fang, X., & Ellis, C. (2012). Sensitivity of snowmelt hydrology in Marmot
Creek, Alberta, to forest cover disturbance. Hydrological Processes, 26(12),
1892-1905. doi: 10.1002/hyp.9248
Potts, D. F. (1984). Hydrologic impacts of a large-scale mountain pine beetle
(Dendroctonus Ponderosae Hopkins) epidemic. Water Resources Bulletin, 20(3),
373-377.
Pugh, E., & Gordon, E. (2013). A conceptual model of water yield effects from beetleinduced tree death in snow-dominated lodgepole pine forests. Hydrological
Processes, 27, 2048-2060. doi: 10.1002/hyp.9312
Pugh, E., & Small, E. (2012). The impact of pine beetle infestation on snow accumulation
and melt in the headwaters of the Colorado River. Ecohydrology, 5(4), 467-477.
doi: 10.1002/eco.239
Raffa, K. F., Aukema, B. H., Bentz, B. J., Carroll, A. L., Hicke, J. A., Turner, M. G., &
Romme, W. H. (2008). Cross-scale drivers of natural disturbances prone to
anthropogenic amplification: The dynamics of bark beetle eruptions. Bioscience,
58(6), 501-517. doi: 10.1641/b580607
Rhoades, C. C., McCutchan, J. H., Jr., Cooper, L. A., Clow, D., Detmer, T. M., Briggs, J.
S., Lewis, W. M., Jr. (2013). Biogeochemistry of beetle-killed forests: Explaining
a weak nitrate response. Proceedings of the National Academy of Sciences of the
United States of America, 110(5), 1756-1760. doi: 10.1073/pnas.1221029110
Somor, A. J. (2010). Quantifying streamflow change following bark beetle outbreak in
multiple central Colorado catchments. (MS), University of Arizona, Tucson, AZ.
Swank, W. T., Waide, J. B., Crossley, D. A., & Todd, R. L. (1981). Insect defoliation
enhances nitrate export from forest ecosystems. Oecologia, 51(3), 297-299. doi:
10.1007/bf00540897
Tokuchi, N., Ohte, N., Hobara, S., Kim, S. J., & Masanori, K. (2004). Changes in
biogeochemical cycling following forest defoliation by pine wilt disease in Kiryu
experimental catchment in Japan. Hydrological Processes, 18(14), 2727-2736.
doi: 10.1002/hyp.5578
Troendle, C. A. (1983). The potential for water yield augmentation from forest
management in the Rocky Mountain Region. Water Resources Bulletin, 19(3),
359-373.
Troendle, C. A., Wilcox, M. S., Bevenger, G. S., & Porth, L. S. (2001). The Coon Creek
Water Yield Augmentation Project: implementation of timber harvesting
technology to increase streamflow. Forest Ecology and Management, 143(1-3),
179-187. doi: 10.1016/s0378-1127(00)00516-8
32
United States Forest Service. Potter, K.M.; Conkling, B.L., eds. 2012. Draft Forest
Health Monitoring 2011 National Technical Report, Version 1/20/2012.
Williams, A. P., Allen, C. D., Macalady, A. K., Griffin, D., Woodhouse, C. A., Meko, D.
M., . . . McDowell, N. G. (2013). Temperature as a potent driver of regional forest
drought stress and tree mortality. Nature Climate Change, 3(3), 292-297. doi:
10.1038/nclimate1693
Williams, A. P., Allen, C. D., Millar, C. I., Swetnam, T. W., Michaelsen, J., Still, C. J., &
Leavitt, S. W. (2010). Forest responses to increasing aridity and warmth in the
southwestern United States. Proceedings of the National Academy of Sciences of
the United States of America, 107(50), 21289-21294. doi:
10.1073/pnas.0914211107
33
APPENDIX A: MULTISCALE OBSERVATIONS OF SNOW ACCUMULATION
AND PEAK SNOWPACK FOLLOWING WIDESPREAD, INSECT-INDUCED
LODGEPOLE PINE MORTALITY.
Manuscript accepted for publication and published online (2012) but not yet
included in an issue of the journal Ecohydrology. The publisher (Wiley) does not require
copyright permission for inclusion in a dissertation by the same author, as described at
http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1936-0592/homepage/Permission.
Title:
Multiscale observations of snow accumulation and peak snowpack following
widespread, insect-induced lodgepole pine mortality.
Running title:
Snowpack following insect-induced pine mortality
Authors:
Joel A. Biederman1*, P. D. Brooks1, A. A. Harpold1, D. J. Gochis3, E. Gutmann3, D. E.
Reed4, E. Pendall5 and B. E. Ewers5
1
SAHRA and Department of Hydrology and Water Resources
University of Arizona, Tucson AZ 85721
3
4
NCAR, Boulder, CO 80301
Department of Atmospheric Science and Program in Ecology
University of Wyoming, Laramie, WY 82071
5
Department of Botany and Program in Ecology
University of Wyoming, Laramie, WY 82071
34
Abstract
Seasonal snowpack in forested lands is the primary source of fresh water in
western North America, where mountain pine beetle (MPB) infestation has resulted in
rapid and extensive tree die-off. Forests significantly influence the amount and spatial
distribution of peak seasonal snowpack, but the impacts of large-scale tree mortality on
the processes controlling peak snowpack are not well understood. We evaluate the effects
of widespread tree mortality on winter snow accumulation and peak seasonal snowpack
across multiple spatial scales and several levels of MPB impact in the Central Rocky
Mountains. Observations for winters 2010 and 2011 include continuous snow depths in
20 plots, distributed snow surveys at peak accumulation, and climate observations above
and below canopy including precipitation, temperature, humidity, wind, and shortwave
radiation. Stable water isotopes were observed for fresh snowfall and for snowpack.
Plot-scale snowfall observations showed 20 % lower interception (p < 0.05) in greyphase stands (needles lost) than Unimpacted stands. However, distributed snow surveys
found no differences in peak seasonal snow water equivalent (SWE) between
Unimpacted and grey-phase stands. Water isotopes of snowpack from MPB-killed stands
indicated kinetic fractionation; enriched values demonstrated higher winter snowpack
sublimation than in Unimpacted forest. Following MPB infestation, reduced canopy
sublimation of intercepted snow appeared to be compensated by increased snowpack
sublimation, consistent with observations of higher snowpack insolation. Consequently,
the effects of widespread tree mortality on peak seasonal snowpack, which is crucial for
35
downstream water resources, will be influenced by compensation for lower interception
by higher snowpack sublimation.
36
1. Introduction
Snow dominates the water cycle inputs to the semi-arid western portions of
Canada and the United States (Connaughton, 1935; Troendle, 1983; Bales, 2006) and is
the primary control on the fluxes of energy and biogeochemicals. More than 60 million
people in the western U.S. rely on mountain snowmelt as a primary source of water
(Bales, 2006). Most of this water derives from forested areas, where accumulation and
ablation of snow are strongly controlled by vegetation (Troendle and King, 1987; Jost et
al., 2007; Veatch et al., 2009; Rinehart et al., 2008; Musselman et al., 2008; Molotch et
al., 2009; Varhola et al., 2010). Vegetation influences how montane catchments capture,
store, partition and release snow water, yet the impacts of insect-induced tree mortality
on snow-vegetation interactions are not well understood, especially against the backdrop
of changing or extreme climate (Varhola et al., 2010). Many seasonally snow-covered
areas in western North American have recently suffered rapid and extensive tree die-off
caused by mountain pine beetle (MPB; Dendroctonus ponderoseae; Raffa et al., 2008;
Kurz et al., 2008). Though MPB is endemic to the region, its activity has reached
unprecedented levels during the last 15 years. In 2010, MPB infested 2.77 million
hectares of forests, accounting for 75% of all U.S. tree mortality related to insects and
other pathogens (USFS, 2012). Causes of the present epidemic are thought to include
historically high stocks of mature pine trees following a century of fire suppression,
increasing beetle survival and reproduction rates due to longer summers and warmer
winters (Mitton and Ferrenberg, 2012), and greater tree susceptibility due to water stress
from a warmer and dryer climate (Hicke et al., 2006; Ayres and Lombardero 2000; Logan
37
et al. 2003). Understanding the effects of insect-induced changes to canopy structure on
snow accumulation and peak seasonal snowpack is critical to effective management of
forest and water resources (Troendle, 1983; Troendle and King, 1987) and the prediction
of ecosystem response and recovery following MPB (Edburg et al., 2012).
Forest canopy controls snowpack through interception, sublimation of snow from
the canopy, attenuation of wind speeds, creation of wind eddies and alteration of the
radiation balance.
A variable fraction of snowfall is retained in the forest canopy
(Schmidt and Troendle, 1992; Hedstrom and Pomeroy, 1998; Link and Marks, 1999), and
sublimation of this canopy snow may reduce peak seasonal snowpack by 25 – 45 % in
mature pine forest (Pomeroy et al., 1998). Studies by Boon (2012) and Pugh and Small
(2011) attributed greater peak seasonal snowpack under grey-phase than living lodgepole
forest to reduced sublimation of intercepted snow. Modeling studies have predicted
increased peak water yield following MPB infestation using model equations that
increased snowpack as interception (modeled as a function of canopy density) was
decreased (Bewley et al., 2010; Mikkelson et al., 2011, Pomeroy et al., 2012), reinforcing
the expectation of reduced interception as a dominant response. Vegetation interacts with
wind to affect spatial distribution of snow through development of eddies across the top
of forest canopy that may preferentially deposit snow in canopy gaps (Golding and
Swanson, 1978) and by altering redistribution of snow on the ground (Troendle, 1983).
Canopy also attenuates wind (Bergen et al., 1971), thus offering protection against wind
scour (e.g. Winstral et al., 2002) and sublimation (Bernier, 1990). A forest canopy
protects snowpack from ablation by attenuating solar radiation and reducing near-surface
38
wind speeds (Link and Marks, 1999; Musselman et al., 2008; Molotch et al., 2009;
Veatch et al., 2009; Lopez-Moreno and Stahli, 2008). A significant body of literature
indicates the importance of shading from solar radiation in determining snowpack
ablation both during winter and spring melt (e.g. Anderson, 1956; Molotch et al., 2009;
Veatch et al., 2009; Lopez-Moreno & Stahli, 2008; Musselman et al., 2008, Rinehart et
al., 2008).
Because canopy can both prevent snowfall from reaching the snowpack and
protect the snowpack from ablation (Musselman et al., 2008; Rinehart et al, 2008; Veatch
et al., 2009; Molotch et al., 2009, Pugh and Gordon, 2012), reduced canopy density
following MPB tree mortality is likely to lead to compensatory process changes (Somor
et al., 2010). Prior studies of canopy removal by harvest (e.g. Hoover and Leaf, 1967;
Gary, 1974; Troendle, 1983; Golding and Swanson, 1978, 1986; Troendle and King,
1987; Woods et al., 2006) found more snow water equivalent (SWE) in harvested
clearings than nearby forest plots for clearing sizes ranging up to about 5-8 times the
mean surrounding tree height. However, these studies reported a mix of increased and
unchanged SWE at the stand or hillslope scales, leading to the ideas that 1: increases in
clearings were supplied by decreases in undisturbed forest (Gary, 1974; Golding and
Swanson, 1986) and 2: reduced interception could be compensated by increased
sublimation (Troendle, 1983; Golding & Swanson, 1986; Troendle & King, 1987).
Differences were observed at plots scales but sometimes not at larger scales,
demonstrating the importance of multi-scale observation. Understanding snow process
responses (e.g interception reduction) may require observation at the scale of individual
39
trees to plots, but understanding larger-scale snowpack response to canopy reduction calls
for observations across multiple stands. Insect-induced mortality may have some similar
effects to harvest, but it differs because much of the tree remains intact and may continue
to intercept snow, attenuate wind and alter radiation fluxes (Edburg et al., 2012).
Prediction of hydrologic response to forest mortality should cover a range of
scales allowing process-based understanding to inform the larger-scale impacts on
ecosystem function and water resource availability. Studies comparing adjacent stands
within a given site are often hampered by lack of stands with similar structure but
differing MPB impact (Pugh and Small, 2011), but comparison of multiple sites requires
consideration of topography, forest structure, and climate forcing. To address the effect
of MPB tree mortality on peak seasonal snowpack, we used nested multi-scale
observations of snow depth and SWE, onsite climate data, and isotopic tracers of
hydrologic partitioning to answer two questions. 1: How does MPB tree mortality affect
the balance of canopy-mediated snow processes influencing peak SWE?, and 2: What is
the net effect of MPB tree mortality on peak SWE at the scale of small headwater
catchments?
2. Methods
Two sites with varying levels and timing of forest mortality from MPB infestation
were studied during winters 2010 and 2011, with observations organized in a nested
fashion at multiple spatial scales (Figure 1). Each study site comprised a small headwater
catchment of approximately 1 square kilometer containing one or more precipitation
gauges and a tower for above-canopy climate observation. Within each site, forest stands
40
were categorized by their level of MPB mortality and year of initial major infestation
(e.g. Unimpacted, MPB-2007). Stand-scale observations included canopy characteristics,
snow depth, SWE, snow chemistry and sub-canopy wind speed. Within three stands
chosen for intensive study from the Unimpacted, MPB-2008 and MPB-2007 types, 20
plots were instrumented for continuous observation of snow depth and sub-canopy
temperature, humidity, and downward shortwave radiation.
2.1. Study sites
Two study sites were identified in late 2009 along the Front Range of the Central
Rocky Mountains in northern Colorado and southern Wyoming (Figure 1), a region of the
U.S. severely affected by MPB during the present epidemic (since circa 1996). Sites with
adjacent uninfested and infested stands of otherwise similar characteristics were not
found.
The sites at Chimney Park, WY and Niwot, CO offered observations of
meteorology and snowpack from prior and concurrent studies. To isolate the effects of
vegetation mortality from those of topography, we selected headwater catchments with
similar elevation (2750 - 3,000 m) and gentle terrain (Table 1). Both sites have annual
mean precipitation of about 800 mm and mean annual air temperatures of 1-3 °C. Forest
characteristics and MPB infestation status were obtained from the Niwot Ameriflux
project (http://ameriflux.ornl.gov), from H. Barnard (personal communication) and from
ground-based vegetation surveys conducted annually since 2008 at Chimney Park.
Although of similar species and age, the Chimney Park stands had larger mean tree
height and diameter at breast height (DBH) as well as lower stem density than the Niwot
stands (Table 1).
41
The Niwot site (Figure 1) is within the University of Colorado’s Long-Term
Ecological Research site, 8 km east of the continental divide and 50 km northwest of
Denver. It includes an above-canopy Ameriflux tower (Blanken and Monson, 2012) and
a Natural Resources Conservation Service (NRCS) SNOTEL station (NRCS, 2012) all
within approx. 400 m distance and 30 m elevation of the forest stands studied.
Forest
cover is natural re-growth following logging a century ago, dominated by 97 % lodgepole
pine (Pinus contorta) with a few individuals of Engelmann spruce (Picea engelmannii)
and subalpine fir (Abies lasiocarpa). MPB activity was not observed within the forest
stands studied, and these were labeled Unimpacted.
The Chimney Park site (Figure 1) is in the Medicine Bow National Forest 110 km
north of the Niwot site and 50 km southwest of Laramie, WY. It includes a tower with
above-canopy observations of wind velocity, temperature, humidity and shortwave
radiation, which began collecting data in January 2009. The site is primarily an evenaged forest of lodgepole pine (82 %) with small amounts of aspen (Populus tremuloides,
11 %), Douglas fir (Pseudotsuga menziesii, 5 %) and Engelmann spruce (1 %), (Reed et
al., in preparation), with the last stand-replacing fire 135 years ago (Knight et al, 1985).
Chimney Park experienced extensive MPB infestation beginning in 2007, with
approximately 75 % of mature lodgepole pine killed by 2011. MPB travel to new host
trees during summer, where they introduce a blue-stain fungus that inhibits sap flow and
usually kills the tree within several weeks to months (Yamaoka et al., 1990). Dead trees
retain their needles for 1-3 years (Wulder et al., 2006), termed the red phase of mortality.
Once the needles are lost, trees are said to be in the grey phase of mortality, which lasts
42
5-20 years or more, until the stems fall. Study stands at Chimney Park were categorized
by their year of initial MPB infestation including MPB-2007, MPB-2008, and MPB2009. The MPB-2007 and 2008 stands experienced approximately 75 % mortality,
nearly all in the first year of infestation. Most dead trees (75 % of total) in the MPB2007 stand were in the grey phase during the entire study. Most dead trees in the MPB2008 stand (75 % of total) were in the red phase during the first winter of the study
(2010) but lost their needles and progressed to the grey phase in the second winter
(2011). The MPB-2009 stand experienced 10 % mortality before winter 2010 and an
additional 15 % by winter 2011.
2.2. Winter Climate
We defined the winter snow accumulation season as lasting from the first day of
persistent snow cover on the Niwot SNOTEL station 663 (NRCS, 2012) in mid-October
to our April 7-9 peak accumulation surveys, conducted approximately one week prior to
the snowpack becoming isothermal and commencing to melt. We will refer to this snow
accumulation period as “winter” hereafter. We characterized winter climate at each site
with above-canopy observations of temperature, vapor pressure deficit, wind, and
incident shortwave radiation (Table 2). Wind speed at the Niwot site was corrected to the
same height above canopy as the Chimney Park observations (3.7 m above canopy) using
a neutral atmosphere logarithmic profile with zero-plane displacement height of twothirds the mean tree height and a roughness length of 1.1 m. Open-area winter
precipitation observations were made at the Niwot SNOTEL supplemented with data
from a nearby (~200 m) NOAA Climate Reference Network site when necessary.
43
Chimney Park precipitation was observed by a cluster of weighing-type gauges located
250 m west of the site meteorological tower in a small forest clearing similar in
dimension to the clearing at the Niwot SNOTEL. These different instruments within the
cluster were bias-corrected to a Geonor T-200B reference weighing precipitation gauge at
the site.
2.3. Intensive study plots
During February and March 2010, digital Judd ultrasonic snow depth sensors
(Judd Communications, Salt Lake City, Utah) were installed at 20 plots. Eight sensors
were located in an Unimpacted stand, six in an MPB-2008 stand and six in an MPB-2007
stand. In each stand type, half of the depth sensors were located under tree canopy, while
the others were located in canopy gaps. Gaps were defined by open areas with widths
approximately 1/4 to 1/2 of the mean surrounding tree height. Although the timing of
precipitation at the two sites was correlated at better than 99 % (p < 0.0001), Chimney
Park received 10-14 % more precipitation. Depth records were therefore normalized to
the mean snow depth in gaps from each stand at peak accumulation to account for this
difference; this approach treats gap sensors as experimental controls for analysis of new
snowfall. Sensor power was lost for portions of the winter in the MPB-2007 and MPB2008 plots (Figure 2), and manual snow depth observations were made to supplement the
record at the MPB-2007 plots on the days bracketing the 2011 snow survey. New
snowfall inputs during each storm (mean value in gaps) were quantified for each stand,
with storms defined as periods of precipitation rate ≥ 1.0 mm/hr (water equivalent) and
total precipitation ≥ 5 mm (water equivalent).
44
Sub-canopy climate observations included temperature, humidity, incoming
shortwave radiation and wind speeds. Silicon cell pyranometers (model SP, Apogee
Instruments, Logan, Utah) measured downward shortwave radiation below the canopy at
a height of 2 m in all 20 plots (half under canopy and half in gaps) beginning in April,
2010. These sensors directly sample wavelengths from 300 – 1100 nm, the spectral range
containing 80-90% of solar radiation reaching the land surface, and estimate the
remaining 10-20% of radiation in wavelengths of 1100 - 2800 nm. Vegetation or high
humidity can bias incoming shortwave radiation towards shorter wavelengths, reducing
the accuracy.
Power loss issues disrupted observation in the MPB-2007 plots, but
comparisons were made between the Unimpacted and MPB-2008 plots for winter 2011.
Temperature and relative humidity sensors (model EM50, Decagon Devices, Pullman,
Washington), were installed in all 20 plots at 2 m height in April 2010. These were gillshielded and aspirated with solar-powered fans.
Two sub-canopy (3 m) sonic
anemometers (model CSAT3, Campbell Scientific, North Logan, UT) were deployed
from April 16 to July 29, 2011 in the MPB-2009 stand near the Chimney Park met tower
and another two were deployed in Unimpacted study plots to investigate the effects of
MPB mortality on wind attenuation. Above-canopy winds were on average 13 % lower
at each site during this spring/summer season than during the two winters of the study.
2.4. Snow surveys at peak accumulation
Snow surveys were conducted April 7-9 2010 and 2011 (Figure 1) over small
headwater catchments including 5-6 stands at each site (somewhat less area was surveyed
at Niwot during the 2nd year due to overall lower variability in canopy cover and snow
45
distribution). In recognition of the potentially important roles of slope and aspect on
snow accumulation and ablation, we focused on areas of low slope at both sites.
Chimney Park snow surveys had slopes ranging from 1.9 – 17 % with mean (standard
deviation) of 4.3 % (1.8 %) and a relatively uniform mixture of all aspects. The Niwot
surveys were made on gentle opposing hillslopes with northeast (15 % of survey) and
southeast (85 % of survey) aspect and slopes ranging from 3.7 – 18 % with mean
(standard deviation) of 9.0 % (2.3 %).
Surveys were made along orthogonal transects
(e.g. Anderton et al, 2004) aligned with cardinal magnetic directions.
Depth
measurements were made at 5-m centers as well as one meter to the front, back, left and
right for a total of 100 depth observations along each 100 linear meters of survey. An
observation of canopy density was made at each 5-m center point using the categories
none, sparse, medium and dense. Transects were selected to characterize Unimpacted,
MPB-2009, MPB-2008 and MPB-2007 stand types but extended beyond these to broadly
characterize small headwater catchments of ~1 km2 at each site (Figure 1) with over
8,000 total observations.
The present analysis focused on MPB mortality in mature
forests and used the 4,634 observations in mature lodgepole pine stands, excluding those
closer to a stand edge than twice the mean tree height to minimize edge effects (Woods et
al., 2006). Determination of differences among mean SWE of stand types and canopy
densities was performed by one-way ANOVA followed by a t-test-based comparison of
multiple means (Matlab R2012a).
46
Patterns of spatial correlation were studied within each stand type using
experimental and modeled semivariograms (e.g. Schabenberger & Gotway, 2004). The
experimental semivariogram :
1 N h 

 zxi   zxi  h 
2 N h  i 1
2



is half the average squared difference in depth observation pairs z at locations x,
separated by a lag distance h. N(h) is the number of such depth observation pairs in the
given dataset. Snow depth variability may be anisotropic, so semivariograms were
calculated separately for 200-m South-North and West-East transect portions using the
120 depth observations falling along the centerline of each survey at positions of 0, 1, 2,
5, 6, 7, . . . 199, 200, 201 m. Experimental semivariograms were fitted with exponential
models (Matlab R2012a) to allow calculation of the sill, or variance approached at large
lag distances, and the range, calculated as the lag distance at which the model
semivariogram reaches 95 % of the sill variance. Observations closer to one another than
the range are considered autocorrelated, while those farther apart than the range are
considered independent.
With each peak seasonal snow survey, densities were determined from 2-4 snow
pits at each site using the method described by Cline et al. (2001) in a variety of forest
stands and both under canopy and in gaps. Pits were excavated to the ground surface,
taking snow density, grain type and size, and temperature measurements at 10-cm vertical
increments. Fewer snow density measurements were required than snow depths because
density varies in space much less than depth (Elder et al., 1998; Anderton et al., 2004).
Densities did not vary significantly in space within either site, so mean density values
47
(range: 261 to 342 kg/m3) for each site and date were applied in calculation of SWE
(Anderton et al., 2002; Jost et al., 2007).
2.5. Isotopic tracers of sublimation
Stable water isotopes of snowpack were compared to the isotopic content of fresh
snowfall.
Snowfall samples were collected within 24-36 hours of storms in forest
clearings, which minimized potential contamination by previously intercepted snow
sloughed from the canopy, where it would be more likely to have partially sublimated.
Fresh snowfall determined the local meteoric water line (LMWL) at each site, which was
considered the starting point for any snowpack isotopic enrichment. Snowpack isotope
samples were collected from snow pits dug approximately monthly at paired undercanopy and gap locations. Samples were collected from each 10-cm vertical layer, with
composite values volume-weighted using the snow density measurements. The samples
were analyzed for stable water isotope ratios (δ2H and δ18O) at the University of Arizona
using a DLT-100 liquid water isotope analyzer (Los Gatos Research (LGR), Inc., model
908-000). Further analytical details are described in Lyon et al. (2009).
The isotopic enrichment of snow is similar to that of evaporating water, with mass
exchange between snow and vapor (Moser and Stichler, 1975). The sublimated fraction
of snowpack, f, was estimated via the method given by Clark and Fritz (1997) and
demonstrated in a recent field study by Gustafson et al. (2010) using both δ18O and δ2H
(deuterium) values. Equation 2 dictates the equilibrium fractionation enrichment factor
expressed as per mil (‰) due to the transition of water from the solid to vapor phase:
 18Ol s  (1   ) 103
(2)
48
where  Ol  s is the per mil (‰) fractionation and is the fractionation factor ( =
18
1.015) (O’Neil, 1968, cited in Clark and Fritz, 1997) Kinetic fractionation is governed by
equation 3, where h is the humidity at the snow-atmosphere boundary.
 18 Ol  s  14.2  (1  h)
(3)
Deuterium excess of snow pit isotopic values, relative to the LMWL, showed a humidity
of approximately 80-90% at the snow-atmosphere boundary. The total fractionation of
18
O
 O  due to sublimation is the sum of equilibrium and kinetic fractionation,
18
T
expressed as:
 18 OT   18 Ol s   18 Ol s
(4)
The fractional water loss f from the snowpack due to sublimation enrichment then was
calculated according to Rayleigh distillation:
18OObs.  18OSnowfall  18OT  ln( f )
(5)
18
where  OObs was the maximum snowpack value and  OSnowfall was the intercept of the
18
least-squares linear regression of snowpack samples with the LMWL (Matlab R2012a).

The sublimation estimates were not sensitive to temperature and relative humidity, which
were assumed based on conservative conditions to be -2 °C and 90 %.
3. Results
3.1. Weather
Cold temperatures and frequent storms (mean interstorm interval of 4.2 days)
prevailed throughout the winter snow accumulation seasons (winters) of 2010 and 2011,
resulting in the development of continuous snow cover across the study sites. Snowpack
temperatures and crystal stratigraphy excluded melt during the winters. 2010 winter
49
precipitation was climatologically average (379 mm at Niwot and 422 mm at Chimney
Park) while 2011 winter precipitation was 113 % (calculated from 30-year Niwot
SNOTEL history) of average (422 m at Niwot and 466 mm at Chimney Park).
The
Niwot and Chimney Park sites received comparable climate forcing as characterized by
temperature, vapor pressure deficit, wind speed, and downward shortwave radiation
(Table 2) and largely experienced the same regional-scale winter storms, evidenced by
similar timing and magnitude of snow depth increases in plots (Figure 2). Cumulative
daily winter precipitation plots (Chimney Park vs. Niwot, plots not shown) had slopes of
1.10 and 1.14 for winters 2010 and 2011, reflecting the 10-14% greater winter
precipitation at Chimney Park (see methods).
However, these plots showed linear
correlation greater than 99 % (p < 0.0001), meaning the precipitation timing was quite
similar.
3.2. Snowfall interception
To address this study’s first question about the balance of canopy-mediated snow
process responses (see Introduction), continuous snow depth records were used to
compare new snowfall under canopy and gaps in the Unimpacted, MPB-2008 and MPB2007 plots. In the Unimpacted plots, locations under gaps accumulated deeper snow than
under adjacent canopy throughout the winter (Figure 2). Differences between canopy
and gaps were smaller in the MPB-2008 plots and slightly reversed in the MPB-2007
plots, where the under-canopy mean depth was slightly larger than in gaps. At the end of
the winter, differences in mean snow depth under canopy as compared to gaps were 11%,
4% and -7% for the Unimpacted, MPB-2008 and MPB-2007 plots.
Taking one minus
50
the ratio of new snowfall during storms under canopy to that in gaps (Figure 3) as an
estimate of interception, we calculated interception of 20 %, 14 % and 0 % in the
Unimpacted, MPB-2008 and MPB-2007 plots, respectively.
Assuming 60 % of
intercepted snow sublimates and the remaining 40 % is unloaded to the snowpack
(Pomeroy et al., 1998), these correspond to reductions in peak seasonal snowpack of 12.0
%, 8.4% and 0.0%.
3.3. Spatial distribution of peak snowpack
Marked contrasts in spatial distribution of peak snow depth across the different
stages of MPB infestation were used to make further inferences about the balance of
canopy-mediated snow process responses.
Representative South-to-North transects
(Figure 4) show least depth under denser canopy and greatest depth under canopy
observed to be “sparse” or “none”, but the overall variability decreased with increasing
time since MPB infestation. Standard deviations of depth were 21.9, 10.6 and 6.4 cm for
Unimpacted, MPB-2008 and MPB-2007 transects (Figure 4), respectively. Numerous
field observations indicated that gaps larger than half the surrounding tree height had
deepest snow at their southern edges and least snow at their northern edges (e.g. Figure
4a Unimpacted transect near 0 m and +70 m); the phenomenon was weaker in grey stands
that had lost their needles than in healthy or red-phase stands as illustrated in Figure 4b
(MPB-2008) near 0 m and +75 m and in Figure 4c (MPB-2007) just south of 0 m.
Semivariograms (Figure 5) had clear sills, indicating that snow surveys were of sufficient
extent to characterize the mean and variability (Figure 6) of each stand type (Cressie,
1993; Legendre & Legendre, 1998; Jost et al., 2007). Semivariograms of normalized
51
2010 peak snow depth were anisotropic (Isaaks & Srivastava, 1989), showing an order of
magnitude lower sill variance along South-North transects in an MPB-2007 stand (grey
phase, sill = 0.015) than an Unimpacted stand (sill = 0.120) with an MPB-2008 stand (red
phase, sill = 0.053) of intermediate variability (Figure 5). In the Unimpacted forest
(Figure 5a), the sill variance along a South-North transect (0.120) was double the sill
variance along a West-East transect. The semivariogram ranges, which represent the
distance beyond which depths were independent, decreased from 12.2 m in an
Unimpacted stand to 7.4 in an MPB-2008 stand (red phase) to 3.3 m in an MPB-2007
stand (grey phase) along South-North transects.
3.4. Comparison of peak snowpack between MPB and Unimpacted
Amounts and spatial distribution of SWE were used to compare mean effects of
MPB infestation on peak seasonal SWE as well as the role of canopy within each stand
type. SWE at the snow surveys ranged from 23.0 to 29.3 cm over the two years (Table
3). SWE values were normalized by onsite observations of winter precipitation for each
site and year, with the ratio SWE:P indicating the fraction of snowfall remaining in the
snowpack at peak seasonal accumulation. Consistent with expectations based on low
slope, there was no significant relationship between peak SWE:P and slope at either site.
At Niwot the small portion (15 %) of the dataset with Northeast aspect had higher SWE:P
than the points with Southeast aspect (SWE:P of 0.658 and 0.625, respectively). Mean
normalized peak seasonal snowpack (SWE:P) was used to address this study’s second
question regarding the net impact of forest mortality on SWE. Unimpacted, MPB-2008
and MPB-2007 stand types were statistically indistinguishable with a mean SWE:P of
52
0.62 (Figure 5), while the MPB-2009 stand had 2 % greater SWE:P (p < 0.05). Total
winter season ablation was therefore 36 - 38 % of snowfall. When SWE:P values were
categorized by canopy density (i.e. none, sparse, medium and dense), the number of
unique (p < 0.05) canopy groups was highest in the Unimpacted stand type (4 groups)
and lowest in the MPB-2007 stand type (2 groups).
With increased time since
infestation, areas of medium to dense canopy tended to higher SWE:P, while areas of
sparse or no canopy showed monotonically decreasing SWE:P (i.e. gap areas retained a
smaller fraction of winter snowfall at peak seasonal accumulation).
3.5. Subcanopy environment
The balance of canopy-mediated snow process responses (study question 1) was
further informed by sub-canopy climate observations during the second winter (2011,
Table 2). Sub-canopy mean temperatures and vapor pressure deficits for the winter
season were similar among all plots, with temperature ranging from -1.9 °C (Unimpacted)
to -0.9 °C (MPB-2008) and vapor pressure deficit ranging from a winter mean of 0.20
kPa (MPB-2007) to 0.23 kPa (Unimpacted). Incoming shortwave radiation averaged 14
W/m2 in the Unimpacted plots as compared to 25 W/m2 in the MPB-2008 plots (in grey
phase in 2011), equating to 70 % greater insolation of the MPB-2008 snowpack than
Unimpacted snowpack during the winter. During the April-July 2011 assessment of subcanopy (3-m height) wind speeds, above- and below-canopy wind speeds were 2.73 m/s
and 0.36 m/s in the Unimpacted stand and 3.00 m/s and 0.43 m/s in the MPB-2009 stand.
53
3.6. Stable isotope indicators of sublimation
The enrichment of stable water isotopes in snowpack differed between
Unimpacted and MPB-impacted stands. Unimpacted stand snowpack plotted along the
LMWL, indicating no kinetic fractionation.
This does not exclude equilibrium
fractionation (saturated conditions at the snow-air interface) which is not detectable with
this method. MPB-impacted stand snowpack plotted on a line below (Figure 7) and
statistically different from the LMWL (p < 0.05) with a slope of 6.2, indicating kinetic
sublimation at an average relative humidity of 90 %. We calculated δ 18Oreact - δ 18O0-react
as 3.6 and δ 2Hreact - δ 2H0-react as 22.2. The fraction of SWE lost from the combined
MPB-impacted stand snowpack via kinetic sublimation is estimated at 24 % using
18
O
and 18 % using 2H. These estimates of snowpack kinetic sublimation loss would include
neither sublimation of intercepted snow completely sublimated from the MPB canopy,
nor sublimation of snowpack under equilibrium (saturated) conditions. The estimate may
be further reduced by condensation back into the snowpack, which may bring the isotopic
signature of snowpack back towards the LMWL and obscure the sublimation signal.
These 18-24 % snowpack loss estimates are therefore lower bounds of total winter
snowpack sublimation in the MPB stands, consistent with estimates of an additional 0-8.4
% lost by canopy sublimation and snow surveys showing a total of 36-38 % winter
season ablation of snowfall.
4. Discussion
In this study we observed shifts in two canopy-mediated snow processes
following MPB forest mortality: decreased interception (Figure 3) and increased
54
snowpack sublimation (Figure 7).
Ultimately these compensated for one another,
resulting in no difference to peak normalized SWE (Figure 6), in contrast with the few
previous direct observations of snow accumulation following insect-induced mortality
(Boon 2007, 2012; Pugh and Small, 2011). Sublimation of intercepted snow from the
canopy is often considered the dominant flux representing differences between snowfall
and peak SWE (Schmidt and Troendle, 1992; Hedstrom and Pomeroy, 1998), leading to
an expectation of increased peak SWE following forest mortality. Sublimation requires
available energy for the change of phase between ice and water vapor and the turbulence
to transport vapor away from the ice surface, both of which are high in the forest canopy
and much lower at the level of the snowpack on the forest floor. Given sufficient energy
and turbulence however, snowpack sublimation may represent a large fraction of winter
snowfall (Hood et al, 1999; Molotch et al., 2009). Molotch et al. (2009) reported winter
snowpack sublimation rates averaging 0.41 mm d-1 and a mean contribution by snowpack
to total sublimation of 45 % for healthy forest at the Niwot site used in this study. Our
results indicate that the sum of winter sublimation losses from the canopy and snowpack
remained constant (Figure 6). Snowpack sublimation increased (Figure 7), compensating
for lower sublimation of snow from the canopy. These results were similar to studies in
which canopy removal treatments altered spatial distribution of snow with little to no
effect on peak SWE at larger scales (e.g. Woods, 2006; Troendle and King, 1987; Gary,
1974). A recently developed method (Gustafson et al., 2010) of comparing stable water
isotopes in fresh snowfall to those in snowpack confirmed greater sublimation of
snowpack in insect-killed forest (Figure 7), and season-long observations of temperature,
55
vapor pressure deficit, solar radiation and wind speed (Table 2) provide inferences to
explain why snowpack sublimation increased, compensating for reduced canopy snow
sublimation.
Continuous snow depth sensors identified the expected pattern of lower
interception in the high mortality forests (Figure 3), presumably in response to reduced
canopy density (Hedstrom and Pomeroy, 1998; Pomeroy et al., 2002). Interception is a
function of canopy area, temperature, humidity, and snowfall characteristics (e.g.
Stroebel, 1978; Hedstrom and Pomeroy, 1998; McNay et al., 1988; Pomeroy et al., 2002),
and has been implicitly accepted as a driver of increased snowpack in modeled response
to MPB (Bewley et al., 2010; Mikkelson et al., 2011, Pomeroy et al., 2012). Interception
rates observed in healthy forest (Figure 3) were similar to recent regional observations
(Musselman et al., 2008; Molotch et al., 2009), suggesting that differences in new
snowfall under canopy and in gaps are due primarily to interception and not influenced
by redistribution of canopy snowfall preferentially deposited in gaps (Golding and
Swanson, 1978). Although unloading of intercepted snow from the canopy may also
influence estimates of interception (Clark et al., 2011) low new snow densities result in
an immediate depression of snow depth following unloading (Mussleman et al., 2008,
Molotch et al., 2009) which allowed us to exclude these events from the analysis
summarized in Figure 3. It is possible that higher subcanopy wind speeds (Table 2)
during storm events reduced the observed interception below impacted stands, and that
both lower stem density and taller trees (Table 1) could enhance this effect. However,
the 20 % difference in interception between the Unimpacted and MPB-2007 plots (Figure
56
3) is of similar magnitude to findings by Pugh and Small (2011) in lodgepole pine forests
in Colorado and by Boon (2012) in British Columbia.
In spite of reduced interception, peak SWE was unchanged by widespread (~75
%) MPB tree mortality (Figure 6), indicating that lower interception (Figure 3) was
compensated by higher winter snowpack sublimation. Differences in topography between
the Niwot and Chimney Park sites could have influenced the SWE comparison. The
Unimpacted stands at Niwot showed slightly higher (~ 5 %) peak normalized SWE in an
area with 9 % slope and Northeast aspect as compared to an area with 9 % slope and
Southeast aspect. Because 85 % of the dataset was collected on the Southeast aspect, this
may have led us to underestimate SWE:P in the Unimpacted stands, meaning that net
snow accumulation at the larger scale might actually be lower in MPB impacted stands
than the Unimpacted stands. Melt was excluded during the winter because profiles of
cold content and crystal structure indicated a continuously cold snowpack. Wind scour
was excluded by more than 8,000 snow depth observations collected over areas of
approximately one square kilometer at the two sites in two years, which showed no major
areas of net deposition. Minimal wind redistribution is consistent with our observations
of low subcanopy wind speeds (Table 2).
With melt and wind excluded as important process of winter snowpack ablation at
these sites, we infer that reduced canopy sublimation in MPB-impacted stands was
compensated by increased snowpack sublimation. Snowpack sublimation is difficult to
observe, especially when concurrent with canopy sublimation, which confounds standscale vapor flux observations (Molotch et al., 2009). Increases in concentrations of
57
snowpack solute ions have been used successfully in open areas (Gustafson et al., 2010),
but ion contamination from throughfall and needle litter in MPB-impacted stands made
stable isotopes preferable in this study (Gustafson et al., 2010; Sommerfeld et al., 1987;
Sommerfeld et al., 1991; Stichler, 1986). Comparison of stable water isotopes in
snowpack with those in fresh snowfall offered evidence that snowpack sublimation was
indeed higher in MPB-impacted stands (Figure 7). Although this method (Eq. 2-5)
allows quantification of just the kinetic portion of sublimation, the significant difference
in kinetic fractionation (i.e. snowpack sublimation loss under subsaturated conditions)
between Unimpacted (no loss) and MPB-impacted (20-25 % loss) stand types confirms a
major increase in sublimation from the snowpack below the canopy following MPB
mortality. Such a difference could have been caused by enhancement of vapor transport
away from the surface (i.e. increased vapor pressure deficits and/or turbulent transport) or
by increased available energy from solar radiation (Veatch et al., 2009, Gustafson et al.,
2010, Rinehart et al., 2008). Isotopic contents of fresh snowfall inputs to the snowpack
could be influenced by partial sublimation of snow initially retained in the canopy and
then sloughed to the ground. The collection of samples in open areas immediately after
storms was intended to minimize such effects. Given the lower wind speeds and higher
density of canopy available for interception, we would expect any such effect to be larger
in Unimpacted stands.
Since canopy sublimation would likely exhibit kinetic
fractionation, the consistency with which snowfall and snowpack samples remained near
the Global Meteoric Water Line in the Unimpacted stands (Figure 7a) suggests that
partial canopy snow sublimation was of minimal influence on the isotope results.
58
Larger snowpack sublimation beneath MPB-impacted forest canopy could have
been controlled by higher wind speeds (Stegman, 1996) and turbulence (Molotch et al.,
2007) or by greater transmission of solar radiation (Boon, 2009; Pugh and Small, 2011),
consistent with the major role that radiation plays in the energy balance of forested
snowpacks (Link and Marks, 1999). Our observation of similar wind speeds in the
Unimpacted and MPB-2009 stand types (Table 2) makes it unlikely that wind controlled
differences in snowpack sublimation. 71 % more solar radiation reached the snowpack in
the MPB-2008 plots than in the Unimpacted plots (Table 2). Although the type of
pyranometer used might lead to over-estimation of below-canopy shortwave by 10-20 %
(see Methods), any bias should be larger for Unimpacted stands, suggesting that the true
difference could be larger. The 71 % difference we observed is much larger than the 1113 % shortwave difference calculated by Pugh and Gordon (2012) for grey-phase stands
using a modified Beer-Lambert model (Hellstrom, 2000).
More detailed canopy
radiation transmission models could be more informative (e.g. Hellstrom, 2000; Hardy et
al., 2004); caution is warranted in using spatially lumped models, becuase winter season
sublimation can vary among forest stands with similar average canopy density but
different geometry (i.e. spatial arrangement of trees and gaps, Woods et al., 2006; Veatch
et al., 2009) and/or topography (slope, aspect, elevation e.g. Rinehart et al., 2009).
Although vegetation is the most important control of SWE in forested terrain, even in
complex topography (Varhola et al., 2010), solar interactions with slope and aspect may
be of increased importance following canopy density reduction (Troendle and King,
1987; Somor et al., 2010).
Net shortwave inputs to the snowpack may be further
59
increased by litter-fall beneath killed trees and associated albedo reductions (Winkler et
al., 2010; Pugh and Small, 2011). If the observed 11 W/m2 (Table 2) were the only
difference in available energy from December 1 to April 1, it would be sufficient to
sublimate an additional 4.1 cm of SWE, or 14 % of the mean peak SWE observed in this
study, more than compensating for the estimated 12 % reduction in canopy sublimation
between the Unimpacted and grey phase MPB-2007 (Figure 3).
Spatial observations from our distributed snow surveys supported the importance
of solar radiation in controlling snowpack sublimation and net annual snow water input.
First, the observation of greater snow depth at the shaded southern edges of forest gaps
and less depth near the less-shaded northern edges (Figure 4) indicated the importance of
solar shading (Veatch et al., 2009; Molotch et al., 2009, Musselman et al., 2008; Golding
and Swanson, 1986). Second, snow surveys showed decreased snowpack (SWE:P) in
gaps (Figure 6) of MPB-killed forest, which may be attributed to reduced shading by
adjacent canopy. Third, the importance of solar shading was demonstrated by higher
semivariogram sill variance along the South-North transect than the West-East transect of
Unimpacted forest (Figure 5). A South-North transect in healthy forest crosses the deep,
shaded southern portion and shallower, unshaded northern portion of snow in each gap
traversed, whereas a West-East transect crosses each gap parallel to its snow depth
contours, resulting in lower variability. This directional difference was not observed in
MPB-killed forest, consistent with a reduction in canopy shading.
Compensatory process changes affect peak SWE following insect-induced tree
mortality, with a potentially variable balance between lower interception and increased
60
higher snowpack sublimation. Reduced sublimation of intercepted snow with reduced
canopy density is intuitive, and our estimates of 8.4 – 12.0 % more snowfall reaching the
snowpack in grey-phase stands (Figure 3) are consistent with both empirical observations
(Boon, 2012; Pugh and Small, 2011) and modeled results (Bewley et al., 2010; Pugh and
Gordon, 2012) ranging from 7 - 19 %. Much less is known about the effects of insectinduced tree mortality on winter snowpack sublimation. The present study’s conclusion
of no change in peak seasonal SWE (Figure 6) contrasts with increased SWE observed
following MPB in Colorado by Pugh and Small (2011) and for one of two winters in
British Columbia by Boon (2012). We suggest that such contrast may arise because the
observations in this study are nested across a larger range of spatial scales (tree, plot,
stand, hillslope) than previous studies. Earlier studies of harvest effects on snowpack
reported forest-to-clearing snowpack differences but no net change averaged over larger
scales (Hoover and Leaf, 1967; Gary 1974; Troendle, 1983; Golding and Swanson, 1986;
Troendle and King, 1987), highlighting the importance of large-scale snowpack
characterization.
Peak SWE is a primary indicator of long-term water availability in forested
mountain catchments, and this study has implications for ecosystem function and
downstream water resources. The conventional wisdom is that widespread tree mortality
results in increased water yield based on harvest studies (see review by Pugh and Gordon,
2012), but the few empirical studies of water yield response to insect-induced mortality
are inconclusive in this regard. Bethalmy (1974) reported a small water yield increase
that did not peak until 15 years following major infestation, while Potts (1984) found a
61
small increase immediately, and Somor et al. (2010) found less runoff in one catchment
and no change in seven others following widespread MPB forest mortality. These mixed
responses in runoff likely reflect variable SWE response (i.e. changes to interception and
winter sublimation), time-varying changes to transpiration (Edburg et al., 2012), and
mediation of impacts due to complex hydrologic flow paths and the patchy, transient
nature of insect infestation. Our results showed that increased winter sublimation
compensated for reduced interception and demonstrated that insect-induced forest
mortality will not necessarily result in increased water availability.
5. Conclusions
This study provided extensive empirical evidence from two winters that snow
process responses to insect-induced forest mortality can compensate for one another,
resulting in no change to peak SWE.
Observations of reduced interception were
consistent with expectations from prior studies, but surprisingly, there was no difference
in peak SWE between healthy and grey-phase forest.
Higher rates of snowpack
sublimation were confirmed by the recently developed method of comparing stable water
isotopes in fresh snowfall to those in snowpack. Several lines of evidence suggested that
vegetation shading of snowpack from solar radiation was a primary control on winter
sublimation of snow at the study sites and that shading was reduced in insect-killed
forest. The plot-scale observations of new snowfall and accumulated depth showed a
decreased association between canopy density and snow depth, consistent with the
decreased spatial variability of depth, the decreased spatial scale of variability, and the
lower importance of canopy density in controlling SWE: P at peak accumulation. The
62
contrast between our results and those of two prior studies emphasizes the importance of
winter snowpack sublimation as well as the need for large-scale observation when
characterizing peak SWE. The significance of peak SWE as an indicator of hydrologic
response to forest disturbance highlights the need for improved understanding of the
balance between greater net snow inputs and increased snowpack sublimation; factors
needing further investigation include forest geometry, topographic exposure to radiation
and wind, and the dynamics of a given disturbance.
6.
Acknowledgements
The National Science Foundation EAR-0910831, United States Geological
Service, Wyoming Agricultural Research Station through McIntire-Stennis and Wyoming
Water Development Commission supported this research. Science Foundation Arizona
and the Arizona Water Sustainability Program provided additional support. The National
Science Foundation Niwot Ridge Long-Term Ecological Research project and the
University of Colorado Mountain Research Station provided logistical support and data.
Snowfall isotope data for Niwot were provided by Rory Cowie and Dr. Mark Williams of
INSTAAR.
The National Center for Atmospheric Research is supported through a
cooperative agreement from the National Science Foundation.
We gratefully
acknowledge the work of two anonymous reviewers whose comments helped clarify and
strengthen the key contributions of this work.
63
7. References
Anderson, H.W. 1956. Forest-cover effects on snowpack accumulation and melt. Central
Sierra Snow Laboratory. Transactions AGU. 37(3): 307–312.
Anderton, S. P., White, S. M., & Alvera, B. 2002. Micro-scale spatial variability and the
timing of snow melt runoff in a high mountain catchment. Journal of Hydrology,
268(1-4), 158-176. DOI: 10.1016/s0022-1694(02)00179-8
Anderton, S. P., White, S. M., & Alvera, B. 2004. Evaluation of spatial variability in
snow water equivalent for a high mountain catchment. Hydrological Processes,
18(3), 435-453. DOI: 10.1002/hyp.1319
Ayres, M. and Lombardero, M. (2000). Assessing the consequences of global change for
forest disturbance from herbivores and pathogens. Science of the Total
Environment, 262(3), 263–286.
Bales, R. C., Molotch, N. P., Painter, T. H., Dettinger, M. D., Rice, R., & Dozier, J. 2006.
Mountain hydrology of the western United States. [Article]. Water Resources
Research, 42(8), 13. DOI: W0843210.1029/2005wr004387
Bergen JD. 1971. Vertical Profiles of Windspeed in a Pine Stand. Forest Science 17(3):
314–321.
Bernier, P. Y. 1990. Wind-speed and snow evaporation in a stand of juvenile lodgepole
pine in Alberta. Canadian Journal of Forest Research-Revue Canadienne De
Recherche Forestiere, 20(3), 309-314. DOI: 10.1139/x90-045
Bethlahmy, N. 1974. More streamflow after a bark-beetle epidemic. Journal of
Hydrology, 23(3-4), 185-189.
Bewley, D., Alila, Y., & Varhola, A. 2010. Variability of snow water equivalent and
snow energetics across a large catchment subject to Mountain Pine Beetle
infestation and rapid salvage logging. [Article]. Journal of Hydrology, 388(3-4),
464-479. DOI:10.1016/j.jhydrol.2010.05.031
Blanken, P. and Monson, R. 2012. Accessed May 1, 2012. Niwot Ridge AmeriFlux site
information available at:
http://public.ornl.gov/ameriflux/Site_Info/siteInfo.cfm?KEYID=us.niwot_ridge.0
1
Boon, S. 2007. Snow accumulation and ablation in a beetle-killed pine stand in Northern
Interior BC. BC Journal of Ecosystems and Management, 8(3), 1-13.
Boon, S. 2009. Snow ablation energy balance in a dead forest stand. Hydrological
Processes, 23(18), 2600-2610. DOI: 10.1002/hyp.7246
Boon, S. 2012. Snow accumulation following forest disturbance. Ecohydrology. DOI
10.1002/eco.212.
64
Clark, I., and Fritz, P. 1997. Environmental Isotopes in Hydrology, CRC Press, Boca
Raton, FL.
Clark, M. P., Hendrikx, J., Slater, A. G., Kavetski, D., Anderson, B., Cullen, N. J., . . .
Woods, R. A. 2011. Representing spatial variability of snow water equivalent in
hydrologic and land-surface models: A review. [Article]. Water Resources
Research, 47, 23. DOI: W0753910.1029/2011wr010745
Cline D., Armstrong R., Davis R., Elder K., Liston G. 2001. NASA Cold Land Processes
Field Experiment
Plan 2001 – 2004. NASA Earth Science Enterprise,
Land Surface Hydrology Program.
Connaughton, C.A. 1935. The accumulation and rate of melting snow influenced by
vegetation. Journal of Forestry, 33: 564-569.
Cressie, N.A.C. 1993. Statistics for spatial data. Revised Edition. John Wiley and Sons,
New York, 900 p.
Edburg, S. L., Hicke, J. A., Brooks, P.D., Pendall, E.G., Ewers, B.E., Norton, U., Gochis,
D., Gutmann, E.D., Meddens, A.J.H. 2012. Cascading impacts of bark beetlecaused tree mortality to coupled biogeophysical and biogeochemical processes.
Frontiers in Ecology. (e-view, DOI: TBD).
Elder, K., Rosenthal, W., & Davis, R. E. 1998. Estimating the spatial distribution of snow
water equivalence in a montane watershed. Hydrological Processes, 12(10-11),
1793-1808. DOI: 10.1002/(sici)1099-1085(199808/09)12:10/11<1793::aidhyp695>3.3.co;2-b
Gary, H. L. 1974. Snow accumulation and snowmelt as influenced by a small clearing in
a lodgepole pine forest. Water Resources Research, 10(2), 348-353. DOI:
10.1029/WR010i002p00348
Golding, D. L., & Swanson, R. H. (978. Snow accumulation and melt in small forest
openings in Alberta. Canadian Journal of Forest Research-Revue Canadienne De
Recherche Forestiere, 8(4), 380-388. DOI: 10.1139/x78-057
Golding, D. L., & Swanson, R. H. 1986. Snow distribution patterns in clearings and
adjacent forest. Water Resources Research, 22(13), 1931-1940.
Gustafson, J. R., Brooks, P. D., Molotch, N. P., & Veatch, W. C. 2010. Estimating snow
sublimation using natural chemical and isotopic tracers across a gradient of solar
radiation. Water Resources Research, 46. DOI: W1251110.1029/2009wr009060
Hardy, J. P., Melloh, R., Koenig, G., Marks, D., Winstral, A., Pomeroy, J. W., & Link, T.
2004. Solar radiation transmission through conifer canopies. Agricultural and
Forest Meteorology, 126(3-4), 257-270. DOI: 10.1016/j.agrformet.2004.06.012
Hedstrom, N. R., & Pomeroy, J. W. 1998. Measurements and modelling of snow
interception in the boreal forest. Hydrological Processes, 12(10-11), 1611-1625.
65
Hellstrom, R. A. 2000. Forest cover algorithms for estimating meteorological forcing in a
numerical snow model. Hydrological Processes, 14(18), 3239-3256. DOI:
10.1002/1099-1085(20001230)14:18<3239::aid-hyp201>3.0.co;2-o
Hicke, J. A., Logan, J. A., Powell, J., & Ojima, D. S. 2006. Changing temperatures
influence suitability for modeled mountain pine beetle (Dendroctonus
ponderosae) outbreaks in the western United States. Journal of Geophysical
Research-Biogeosciences, 111(G2). DOI: G0201910.1029/2005jg000101
Hood, E., Williams, M., & Cline, D. 1999. Sublimation from a seasonal snowpack at a
continental, mid-latitude alpine site. Hydrological Processes, 13(12-13), 17811797. DOI: 10.1002/(sici)1099-1085(199909)13:12/13<1781::aidhyp860>3.0.co;2-c
Hoover, M.D. and Leaf, C.F. 1967. Process and Significance of Interception in
Colorado Subalpine Forest in: W.E. Sopper and H.W. Lull (Editors), Forest
Hydrology, Pergamon, New York, NY.
Isaaks, E.H. and R.M. Srivastava. 1989. An introduction to applied geostatistics. Oxford
University Press, New York, 561 p.
Jost, G., Weiler, M., Gluns, D. R., & Alila, Y. 2007. The influence of forest and
topography on snow accumulation and melt at the watershed-scale. Journal of
Hydrology, 347(1-2), 101-115. DOI: 10.1016/j.jhydrot.2007.09.006
Knight, D. H., Fahey, T. J., & Running, S. W. 1985. Water and nutrient outflow from
contrasting lodgepole pine forests in Wyoming. Ecological Monographs, 55(1),
29-48.
Kurz, W. A., Dymond, C. C., Stinson, G., Rampley, G. J., Neilson, E. T., Carroll, A. L.,
Safranyik, L. 2008. Mountain pine beetle and forest carbon feedback to climate
change. Nature, 452(7190), 987-990. DOI: 10.1038/nature06777
Link, T., & Marks, D. 1999. Distributed simulation of snowcover mass- and energybalance in the boreal forest. Hydrological Processes, 13(14-15), 2439-+. DOI:
10.1002/(sici)1099-1085(199910)13:14/15<2439::aid-hyp866>3.0.co;2-1
Logan, J. A., Regniere, J., & Powell, J. A. 2003. Assessing the impacts of global
warming on forest pest dynamics. Frontiers in Ecology and the Environment,
1(3), 130-137. DOI: 10.2307/3867985
Lopez-Moreno, J. I., & Latron, J. 2008. Influence of canopy density on snow distribution
in a temperate mountain range. Hydrological Processes, 22(1), 117-126. DOI:
10.1002/hyp.6572
Lyon, S. W., Desilets, S. L. E., & Troch, P. A. 2009. A tale of two isotopes: differences
in hydrograph separation for a runoff event when using delta D versus delta O-18.
Hydrological Processes, 23(14), 2095-2101. DOI: 10.1002/hyp.7326
66
McNay, R. S., Peterson, L. D., & Nyberg, J. B. 1988. The influence of forest stand
characteristics on snow interception in the coastal forests of British Columbia.
Canadian Journal of Forest Research-Revue Canadienne De Recherche
Forestiere, 18(5), 566-573. DOI: 10.1139/x88-082
Mikkelson, K. M., Maxwell, R. M., Ferguson, I., Stednick, J. D., McCray, J. E. and
Sharp, J. O. 2011. Mountain pine beetle infestation impacts: modeling water and
energy budgets at the hill-slope scale. Ecohydrology. DOI: 10.1002/eco.278
Ministry of Forests and Lands. 2011. Accessed April 1, 2012. Available at
http://www.for.gov.bc.ca/hfp/mountain_pine_beetle/facts.htm
Mitton, J. B., & Ferrenberg, S. M. 2012. Mountain Pine Beetle Develops an
Unprecedented Summer Generation in Response to Climate Warming. American
Naturalist, 179(5), E163-E171. DOI: 10.1086/665007
Molotch, N. P., Blanken, P. D., Williams, M. W., Turnipseed, A. A., Monson, R. K., &
Margulis, S. A. 2007. Estimating sublimation of intercepted and sub-canopy snow
using eddy covariance systems. Hydrological Processes, 21(12), 1567-1575. DOI:
10.1002/hyp.6719
Molotch, N. P., Brooks, P. D., Burns, S. P., Litvak, M., Monson, R. K., McConnell, J. R.,
& Musselman, K. 2009. Ecohydrological controls on snowmelt partitioning in
mixed-conifer sub-alpine forests. Ecohydrology, 2(2), 129-142. DOI:
10.1002/eco.48
Moser, H., and W. Stichler. 1975. Deuterium and Oxygen-18 contents as an index of the
properties of snow covers, paper presented at Snow Mechanics; Proceedings of
the Grindelwald Symposium, IAHS Publication No. 114.
Musselman, K. N., Molotch, N. P., & Brooks, P. D. 2008. Effects of vegetation on snow
accumulation and ablation in a mid-latitude sub-alpine forest. Hydrological
Processes, 22(15), 2767-2776. DOI: 10.1002/hyp.7050
Natural Resources Conservation Service (NRCS) National Water and Climate Center.
SNOTEL data network. 2012. Accessed May 1, 2012. Available at:
http://www.wcc.nrcs.usda.gov/nwcc/site?sitenum=663&state=co
Pomeroy, J. W., Gray, D. M., Hedstrom, N. R., & Janowicz, J. R. 2002. Prediction of
seasonal snow accumulation in cold climate forests. Hydrological Processes,
16(18), 3543-3558. DOI: 10.1002/hyp.1228
Pomeroy, J. W., Parviainen, J., Hedstrom, N., & Gray, D. M. 1998. Coupled modelling of
forest snow interception and sublimation. Hydrological Processes, 12(15), 23172337.
Pomeroy, J.W., Fang, X., & Ellis, C. 2012. Sensitivity of snowmelt hydrology in
Marmot Creek, Alberta to forest cover disturbance. Hydrological Processes,
26(12, 1891-1904. DOI: 10.1002/hyp.9248
67
Potts, D. F. 1984. Hydrologic impacts of a large-scale mountain pine beetle
(Dendroctonus Ponderosae Hopkins) epidemic. Water Resources Bulletin, 20(3),
373-377.
Pugh E.T. and Gordon E.S. 2012. A conceptual model of water yield effects from beetleinduced tree death in snow-dominated lodgepole pine forests. Hydrological
Processes. DOI: 10.1002/hyp.9312
Pugh E.T. and Small E.E. 2011. The impact of pine beetle infestation on snow
accumulation and melt in the headwaters of the Colorado River. Ecohydrology.
DOI: 10.1002/eco.239
Raffa, K. F., Aukema, B. H., Bentz, B. J., Carroll, A. L., Hicke, J. A., Turner, M. G., &
Romme, W. H. 2008. Cross-scale drivers of natural disturbances prone to
anthropogenic amplification: The dynamics of bark beetle eruptions. Bioscience,
58(6), 501-517. DOI: 10.1641/b580607
Rinehart, A. J., Vivoni, E. R., & Brooks, P. D. 2008. Effects of vegetation, albedo, and
solar radiation sheltering on the distribution of snow in the Valles Caldera, New
Mexico. Ecohydrology, 1(3), 253-270. DOI: 10.1002/eco.26
Schabenberger, O. and C.A. Gotway. 2004. Statistical methods for spatial data analysis,
CRC Press, Boca Raton, FL, 488 p.
Schmidt R.A., and Troendle, C.A. 1992. Sublimation of intercepted snow as a global
source of water vapor. In Proceedings of the 60th Western Snow Conference,
Jackson, WY, 60, 1–9.
Sommerfeld, R. A., I. Friedman, and M. Nilles 1981. The fractionation of natural
isotopes during temperature gradient metamorphism of snow. Presented at
Seasonal Snowcovers: Physics, Chemistry and Hydrology, NATO-ASI series C,
211.
Sommerfeld, R. A., J. Clark, and I. Friedman 1991. Isotopic changes during the
formation of depth hoar in experimental snowpacks,. Presented at Stable Isotope
Geochemistry: A Tribute to Samuel Epstein, Special Pub 3:205, The Geochemical
Society, San Antonio.
Somor, A. J. (2010). Quantifying streamflow change following bark beetle outbreak in
multiple central Colorado catchments. (MS), University of Arizona, Tucson, AZ.
Stegman, S. V. 1996. Snowpack changes resulting from timber harvest: Interception,
redistribution, and evaporation. Water Resources Bulletin, 32(6), 1353-1360.
Stichler, W. (1986), Snowcover and snowmelt processes studied by means of
environmental isotopes. Presented at Seasonal Snowcovers: Physics, Chemistry,
and Hydrology, NATO-ASI Series C, 211.
Strobel, T., 1978. Schneeinterzeption in Fichten-Bestaenden in den Voralapen des
Kantons Schwyz. In: Proceedings, IUFRO Seminar on Mountain, Forests and
Avalanches, Davos, Switzerland, pp. 63–79.
68
Troendle, C. A. 1983. The potential for water yield augmentation from forest
management in the Rocky Mountain Region. Water Resources Bulletin, 19(3),
359-373.
Troendle, C. A., & King, R. M. 1987. The effect of partial and clearcutting on streamflow
at Deadhorse Creek, Colorado. Journal of Hydrology, 90(1-2), 145-157. DOI:
10.1016/0022-1694(87)90177-6
United States Forest Service, Potter, K.M.; Conkling, B.L., eds. 2012. Draft Forest
Health Monitoring 2011 National Technical Report, Version 1/20/2012.
Washington, DC: U.S. Department of Agriculture, Forest Service, Forest Health
Monitoring Program. 123 p
Varhola, A., Coops, N. C., Bater, C. W., Teti, P., Boon, S., & Weiler, M. 2010. The
influence of ground- and lidar-derived forest structure metrics on snow
accumulation and ablation in disturbed forests. Canadian Journal of Forest
Research-Revue Canadienne De Recherche Forestiere, 40(4), 812-821. DOI:
10.1139/x10-008
Veatch, W., Brooks, P. D., Gustafson, J. R., & Molotch, N. P. 2009. 'Quantifying the
effects of forest canopy cover on net snow accumulation at a continental, midlatitude site'. Ecohydrology, 2(2), 115-128. DOI: 10.1002/eco.45
Winkler, R., Boon, S., Zimonick, B., & Baleshta, K. 2010. Assessing the effects of postpine beetle forest litter on snow albedo. Hydrological Processes, 24(6), 803-812.
DOI: 10.1002/hyp.7648
Winstral, A., & Marks, D. 2002. Simulating wind fields and snow redistribution using
terrain-based parameters to model snow accumulation and melt over a semi-arid
mountain catchment. Hydrological Processes, 16(18), 3585-3603. DOI:
10.1002/hyp.1238
Woods, S. W., Ahl, R., Sappington, J., & McCaughey, W. 2006. Snow accumulation in
thinned lodgepole pine stands, Montana, USA. Forest Ecology and Management,
235(1-3), 202-211. DOI: 10.1016/j.foreco.2006.08.013
Wulder, M. A., Dymond, C. C., White, J. C., Leckie, D. G., & Carroll, A. L. 2006.
Surveying mountain pine beetle damage of forests: A review of remote sensing
opportunities. Forest Ecology and Management, 221(1-3), 27-41. DOI:
10.1016/j.foreco.2005.09.021
Yamaoka, Y., Swanson, R. H., & Hiratsuka, Y. (1990). Inoculation of lodgepole pine
with 4 blue-stain fungi associated with mountain pine beetle, monitored by a heat
pulse velocity (HPV) instrument. Canadian Journal of Forest Research-Revue
Canadienne De Recherche Forestiere, 20(1), 31-36. doi: 10.1139/x90-005
69
8. Tables
Table 1. Mean characteristics of the two sites and four stand types
Site
Stand
Elevation
(m)
Latitude
Longitude
Age
(yr)
Stem
Density
(per ha)
Tree Ht.
(m)
DBH
(cm)
Niwot, CO
Unimpacted
3000
40° 02′
-105° 33′
100
3900
11.4
12.1
2750
41° 04′
-106° 07′
90-110
2250
1160
1310
14.0
17.6
17.0
14.2
20.1
19.9
Chimney Park, WY
MPB-2009
MPB-2008
MPB-2007
70
Table 2. Total precipitation and above- and below-canopy climate by site for winters
2010 and 2011. NWT = Niwot, CO and CP = Chimney Park, WY. Standard deviations
are shown to reflect the variability in daily climate variables. *All data are for the winter
snow accumulation season except sub-canopy wind speeds, which were observed April
16 – July 29, 2011, during which the mean of daily above-canopy wind speeds were:
NWT = 2.7 (1.3) m/s and CP = 3.0 (0.9) m/s.
Year
Site
2010
NWT Above Canopy
Winter
Precip.
(mm)
Mean Daily (Standard Deviation)
Wind
Temp.
Speed
VPD
(°C)
(m/s)
(kPa)
Rsw
(W/m2)
379
-3.0 (4.8)
3.3
(1.7)
0.26 (0.19)
158 (80)
CP Above Canopy
433
-3.0 (4.5)
3.2
(1.1)
0.24 (0.17)
165 (80)
2011
NWT Above Canopy
Unimpacted Sub-canopy
422
-2.1 (6.8)
-1.9 (5.1)
4.1 (2.4)
*0.38 (0.15)
0.28 (0.22)
0.23 (0.17)
131 (65)
14 (9)
466
-2.0 (7.1)
-1.8 (5.3)
-0.9 (2.4)
-
3.6 (1.3)
*0.43 (0.30)
0.27 (0.25)
0.20 (0.21)
0.21 (0.14)
-
126 (65)
25 (15)
-
CP Above Canopy
MPB-2007 Sub-canopy
MPB-2008 Sub-canopy
MPB-2009 Sub-canopy
71
Table 3. Distributed snow depth and SWE values from surveys conducted in headwater
catchments (~1km2) near peak accumulation for winters 2010 and 2011. MPB-2007 and
MPB 2008 stand types experienced 75 % mortality in their initial infestation year; The
MPB-2009 stand type experienced 15 % mortality by winter 2010 and a total of 25 % by
winter 2011. The sample mean and standard error of the mean are shown to reflect the
uncertainty associated with estimating the population mean. Sample sizes of depth
observations are Unimpacted = 1766, MPB-2009 = 746, MPB-2008 = 923 and MPB2007 = 1199.
Year
Stand
2010
Unimpacted
MPB-2009
MPB-2008
MPB-2007
Depth
(cm)
SWE
(cm)
SWE:P
88.3 (0.6)
92.9 (0.3)
96.4 (0.8)
94.0 (0.3)
23.0 (0.2)
28.2 (0.1)
29.3 (0.2)
28.6 (0.1)
0.61
0.65
0.68
0.66
2011
Unimpacted
MPB-2009
MPB-2008
MPB-2007
87.8 (0.7)
82.6 (0.5)
83.2 (1.5)
79.0 (1.6)
27.0 (0.2)
28.2 (0.2)
28.5 (0.5)
27.0 (0.5)
0.64
0.61
0.61
0.58
72
9. Figures
Figure 1. Site maps of a. Chimney Park, Wyoming and b. Niwot, Colorado sites showing
canopy taller than 2 m (shaded area). Chimney Park includes lodgepole pine stands with
ongoing low-level infestation (MPB-2009) as well as stands with greater than 75% mortality
from infestations in 2007 (MPB-2007) and 2008 (MPB-2008). The Niwot site was Unimpacted
by MPB through 2011
73
Figure 2. Mean normalized depths under canopy (solid) and gaps (dashed) for a. Unimpacted b.
MPB-2008 and c. MPB-2007 plots (n = 4 each for Unimpacted gap and Unimpacted canopy and
n = 3 for all other categories). Depths were normalized to the mean gap value at the snow survey
date of April 9 (day of water year 191, indicated with vertical dotted line) to account for the
small differences in seasonal precipitation between sites (Table 2). Mean depth values on Apr. 9
for gap and canopy plots in each stand type are also shown.
74
Figure 3. Ratios of snowfall under canopy to snowfall in gaps for all snowstorms of more than 5
cm between March 10, 2010 and April 9, 2011. Sample sizes were n = 56, 36 and 38 for the
Unimpacted, MPB-2008 and MPB-2007 plots, where each sample is one storm observed at one
plot. Bars show mean values, and error bars show standard error of the mean. The MPB-2007
plots mean ratio was higher than those of the Unimpacted and MPB-2008 plots (p < 0.05).
75
Figure 4. Snow survey depths (line and markers) and canopy densities (shaded bars) April 7-8,
2010 for sample transects in the a. Unimpacted (mean, sd: 85.3, 21.9 cm), b. MPB-2008 (95.6,
10.6 cm), and c. MPB-2007 (95.9, 6.4 cm), stands. The MPB-2008 and MPB-2007 stands were
21 and 33 months post-infestation, respectively. The MPB stands (panels b and c) had received
14 % more snowfall than the Unimpacted stands at this date (Table 2). Each point represents the
mean and standard deviation (bars) of 5 observations made within a 1-m radius on 5-m centers.
76
Figure 5. Experimental semivariograms (markers) of peak seasonal snow depth normalized to
the mean depth of sample transects in a. Unimpacted, b. MPB-2008 and c. MPB-2007 stands
(same as in Fig 4) April 8-9, 2010, with best-fit exponential model semivariograms (curves). All
semivariograms reached clear sills, indicating surveys were of sufficient spatial extent to
characterize mean and variability of snow depth. Larger sill variance along the South-North than
East-West Unimpacted transect illustrated solar shading. Sill variance and the range of spatial
autocorrelation were lower in stands with ~75 % MPB-induced tree mortality.
77
Figure 6. Peak SWE normalized to winter precipitation for both the 2010 and 2011 surveys in a.
Unimpacted, b. MPB-2009, c. MPB-2008 and d. MPB-2007 stand types. All points were in
mature lodgepole stands, and canopy density was recorded using the categories None, Sparse,
Medium and Dense. Dashed lines are means of each forest cover type, bars are means of each
canopy type, and error bars are standard error of the mean.
The means were equal for
Unimpacted, MPB-2008 and MPB-2007 stand types but slightly higher (p < 0.05) for the MPB2009 stand type. Letters indicate unique canopy density means within each stand type (p <
0.05).
78
Figure 7. Water isotopic contents of fresh snowfall and evolved snowpack for a. Unimpacted
stands and b. combined MPB stands. “+” symbols denote fresh snowfall and “Δ” denote
snowpack. The MPB stands snowpack slope was significantly less than that of the Local
Meteoric Water Line or LMWL (p < 0.05).
79
APPENDIX B: COMPENSATORY VAPOR FLUX REDUCES WATER FOR
STREAMFLOW FOLLOWING SEVERE BARK BEETLE-INDUCED FOREST
MORTALITY
Manuscript in review by the journal Water Resources Research.
Title
Compensatory vapor flux reduces water for streamflow following severe bark beetleinduced forest mortality
Running title:
Reduced streamflow following bark beetle disturbance
Joel A. Biederman1*, P. D. Brooks1, A. A. Harpold2, D. Gochis, B.E. Ewers4, D. E. Reed4,5
1
Department of Hydrology and Water Resources
University of Arizona, Tucson AZ 85721
2
INSTAAR, Boulder, CO 80309
3
NCAR, Boulder, CO 80301
4
Department of Botany and Program in Ecology
University of Wyoming, Laramie, WY 82071
5
Department of Atmospheric Science and Program in Ecology
University of Wyoming, Laramie, WY 82071
80
Abstract
A North American epidemic of mountain pine beetle (MPB) has disturbed over 5 million
hectares of forest already under stress from shifts in climate and related disturbances and
containing headwater catchments crucial to water resources. There is limited empirical evidence
of how MPB infestation affects partitioning of precipitation between vapor flux and streamflow,
and to our knowledge these fluxes have not been observed simultaneously following forest
disturbance. We combined eddy covariance tower footprint water balance, catchment water
balance, and stable isotope indicators of evaporation to quantify hydrologic partitioning over
three years in MPB-impacted and Unimpacted forests. Annual vapor flux V at the Unimpacted
site was conservative, varying only from 573 to 623 mm, while MPB site V varied over a larger
range from 569 to 700 mm. Stable isotope fractionation of stream and soil water showed no
kinetic evaporation at the Unimpacted site. Annual evaporation line slopes at the MPB site
declined monotonically from 5.9 to 4.9 from 2010 – 2012, indicating abiotic evaporation
contributed significantly to V. Increased evaporation was likely driven by greater subcanopy
shortwave radiation, which averaged 21 Wm-2 during summer in the Unimpacted forest as
compared to 66 Wm-2 in MPB-impacted forest. Runoff coefficients at the Unimpacted site varied
from 14% to 40%, following the pattern of annual precipitation. MPB site runoff coefficients
varied from 0% to 37% but declined monotonically each year. Collectively, these results show
that compensatory vapor flux, including evaporation, may offset reduced interception and
transpiration, thereby decreasing water available for streamflow.
Key Words: Forest disturbance, evaporation, evapotranspiration, streamflow, insect, beetle
81
1.
Introduction
Forested watersheds are among the most reliable sources of clean water [Brown et al.,
2005]. Montane forests world-wide are undergoing unprecedented die-off due to combined
effects of increased temperature, drought, fire, and pathogen infestation [Williams et al., 2010,
2013; Breshears et al., 2009; Adams et al., 2012; Westerling et al., 2006; Huber, 2005; Tokuchi
et al., 2004]. Of particular interest in Western North American forests is tree mortality due to
endemic mountain pine beetle (MPB; Dendroctonus ponderoseae) [Raffa et al., 2008; Hicke et
al., 2012]. Recent MPB activity is unprecedented, affecting more than 5 million hectares in the
western US and British Columbia [Meddens et al., 2012]. MPB introduce blue-stain fungi that
inhibit sap flow and usually kill host trees within several weeks to months [Hubbard et al.,
2013]. Dead trees may retain their red-brown needles for one to three years, termed a red phase
[Wulder et al., 2006]. Once needles have fallen, the killed MPB hosts are said to be in a grey
phase, which may last years to decades. This progressive change in canopy structure over several
years without immediate soil disturbance is different than harvest or fire, challenging our ability
to predict MPB effects on water resources [Clow et al., 2011; Mikkelson et al., 2012].
A large body of literature provides initial context for hydrologic response to MPB
disturbance. Studies of forest disturbance impacts on streamflow have established that
streamflow usually increases in proportion to forest area removed, with lesser response in dry
regions [see Stednick, 1996; Brown et al., 2005]. However, only a small portion of studies has
been conducted within interior mountain conifer forests with annual precipitation of 600 to 1200
mm [Bosch and Hewlett, 1982], which describes much of the North American forest impacted by
MPB. Disturbance effects in interior mountains are the weakest and most variable (r2 = 0.01) of
all eco-regions in the United States [Stednick, 1996]. Even streamflow response to 100% forest
82
removal can be highly variable in interior mountains, with increases ranging from 0 to 350 mm
[Stednick, 1996]. Variable streamflow response reflects variability in vapor flux due to snow
sublimation [Troendle, 1983] and summer evapotranspiration [Brown et al., 2013; Hubbart et
al., 2007], which remain poorly understood.
Because the change in forest canopy following MPB is gradual, recent work has
examined how progressive hydrologic partitioning response may affect water resources and
forest succession [Edburg et al., 2012].
Forests regulate partitioning through interception,
transpiration, and sheltering the land surface [Troendle and King, 1987; Molotch et al., 2009;
Varhola et al., 2010].
Disturbance-driven changes in hydrologic partitioning may affect
downstream water resources, including the amount and timing of streamflow [Pugh and Gordon,
2013; Brown et al., 2005; Ellison et al., 2012] and water availability for forest successional
vegetation [Edburg et al., 2012]. Succession is likely to depend on both partitioning between
streamflow and vapor flux as well as how vapor flux comprises sublimation, evaporation, and
transpiration for plant growth [Edburg et al., 2012; Brown et al., 2013; Romme et al., 1986]. The
expectations for response to MPB include an initial, large decline in transpiration followed by
more gradual increases in sublimation, compensatory transpiration, and evaporation [Biederman
et al., 2012; Edburg et al., 2012, Romme et al., 1986]. The timing and relative magnitudes of
these changes are unknown, but the prevailing expectation is for greater streamflow [Pugh and
Gordon, 2013].
There is limited and conflicting empirical evidence of streamflow response to MPB in
support of these expectations. Increased streamflow was reported by Potts [1984] and Bethlahmy
[1974], though an increase reported by Bethlahmy [1974] peaking 15 years post-MPB contrasts
with the view that disturbance impacts are usually immediate and decline over years to decades
83
[Bosch and Hewlett, 1982; Stednick, 1996; Brown et al., 2005]. A recent study of eight MPBimpacted basins found both increases and decreases, attributed in part to variable interactions
among topography and canopy loss in controlling vapor flux [Brooks et al., 2012; Somor, 2010].
Snowpack observations have shown either that winter vapor flux declined or that increased
snowpack sublimation compensated for reduced canopy sublimation, resulting in no net change
[Biederman et al., 2012; Boon, 2012; Pugh and Small, 2012]. In summer, total vapor flux has
been observed to remain relatively constant several years following MPB infestation [Brown et
al., 2013]. Conversely, land surface models predict reduced vapor flux with declining leaf area
index (LAI) under simulated disturbance, leading to increased water availability for streamflow
[Bewley et al., 2010, Pomeroy et al., 2012; Mikkelson et al., 2013b]. Similarly, remote-sensing
products such as MODIS that use plant-mediated algorithms [Mu et al., 2011] predict reduced
vapor flux [Bright et al., 2013; Maness et al., 2013].
The variable streamflow response to MPB [Bethlahmy, 1974; Potts, 1984; Brooks et al.,
2012] and lack of consistency between observations and models demonstrates our limited ability
to predict water resources following insect-driven disturbance. There is therefore a need to
quantify hydrologic partitioning in insect-disturbed headwater catchments by observation of the
dominant fluxes.
Snow surveys quantify winter vapor flux over a range of spatial scales
[Anderton et al., 2004; Musselman et al., 2008; Veatch et al., 2009]. Eddy covariance vapor flux
observations above canopy [Baldocchi et al., 1988] allow computation of a tower footprint water
balance for a footprint similar in magnitude to a headwater catchment (~101 ha) [Wilson et al.,
2001].
Alternatively, using a traditional catchment water balance approach, the difference
between precipitation and observed streamflow can be used to estimate catchment-scale vapor
flux. Both water balance approaches usually assume that storage changes are small compared to
84
hydrologic fluxes and are therefore suited to seasonal to annual time scales [Brutsaert, 2005].
Finally, stable water isotope fractionation can be used to quantify both average humidity and the
cumulative kinetic evaporation at any desired scale [Jasechko et al., 2013; Clark and Fritz,
1997]. Kinetic evaporation is a lower estimate of vapor flux because it does not include plant
transpiration or the equilibrium component of evaporation.
We present three years of observations following MPB infestation at a site with 52-77%
MPB-driven overstory mortality and at a paired unimpacted site to answer the question: How
does MPB tree mortality affect partitioning of precipitation between vapor flux and streamflow?
To our knowledge, the present work is the first effort combining snow surveys, eddy covariance
water balance, catchment water balance, and stable isotope indicators of evaporation to quantify
hydrologic partitioning following forest disturbance.
2. Study sites
An MPB-impacted site “MPB” and a control site “Unimpacted” were identified in 2009
along the eastern edge of the Central Rocky Mountains, a region severely impacted by MPB over
the last 15 years (Figure 1). Due to extensive tree mortality, adjacent MPB-impacted and
unimpacted stands having otherwise similar climate and biophysical characteristics could not be
located. However, biophysical and meteorological observations at these sites over several prior
decades and including our study period showed them to be well-paired in both pre-MPB
condition and climate forcing.
The MPB site (Figure 1) at Chimney Park WY is 100 km west of Cheyenne in the
Medicine Bow National Forest and the headwaters of the Laramie River. The site experienced
overstory mortality of 52-77% between 2007 and 2010.
85
The Unimpacted site (Figure 1) is located 125 km SSE of MPB and 50 km NW of
Denver, CO within the Niwot Ridge Long-Term Ecological Research (LTER) observatory in the
headwaters of Boulder Creek. The study area, known as “C1,” contains an above-canopy eddy
covariance and meteorological tower and the Niwot SNOTEL station. MPB activity above
background levels was not observed at this site through 2012 (< 5% of trees).
To isolate MPB effects, we selected sites with gentle slopes (5 to 8%) and similar
elevation of 2750 to 3000 m (Table 1). The sites are characterized by long, cold winters with
continuous snow cover from October until May or June, and mean annual air temperatures of 1-3
°C. The Unimpacted site receives an average of 800 mm annual precipitation, with 400 mm as
snow (Niwot SNOTEL 1981-2012). The MPB site similarly receives an average of 400 mm
snowfall, but drier summers result in lower annual precipitation of 600 - 650 mm [Fahey et al.,
1985]. Soils at both sites are of sandy loam texture. Overstory trees at the two sites are
dominated by lodgepole pine (Pinus contorta) with most stands aged 110-140 years since the
prior stand-replacing event [Knight et al., 1985; http://ameriflux.lbl.gov]. The eddy covariance
tower footprints had trees of similar height (11 m). Mean diameters at breast height were 12.1
cm and 14.0 cm at the Unimpacted and MPB tower footprints, respectively, while stem densities
were 3900 and 2500 stems ha-1. Further site details are given in Biederman et al., [2012] and by
University of Colorado Mountain Research Station (http://culter.colorado.edu/NWT/index.html).
3. Methods
3.1. Quantification of MPB-driven Forest Mortality
Forest characteristics and die-off status were obtained for the Unimpacted tower footprint
using the AmeriFlux database [http://ameriflux.lbl.gov] and for the MPB tower footprint using
annual ground-based surveys. Tree mortality across the MPB site was mapped by classifying
86
Quickbird imagery acquired on August 25, 2011. Live, red-phase, and grey-phase tree pixels,
along with pixels of other land cover types, were identified on true-color imagery, and used to
train and evaluate a maximum likelihood classifier that separated classes using green band,
Normalized Difference Vegetation Index, and Red-Green Index variables [Coops et al., 2006].
Classification accuracy was 96.8% with a kappa of 0.962.
3.2. Local Climate and Soil Moisture Measurements
The study period comprised hydrologic years 2010 to 2012 divided into two seasons
termed winter and summer (all further reference to years indicates the hydrologic year, from
October 1 to September 30). Winter was October 1 until the onset of snowmelt runoff, marked
by five consecutive days of normalized streamflow increase of at least 0.05 mm d-2 in at least one
catchment, with summer lasting from this onset of snowmelt until September 30.
Observations were collected every 30 minutes of above-canopy temperature, vapor
pressure deficit VPD, wind speed, and incident solar (shortwave) radiation Rsw.
For the
Unimpacted site, these data were obtained as an AmeriFlux Level 1 product (quality checked,
calibrations applied and gaps filled) [http://ameriflux.lbl.gov], while gaps at the MPB site (~5%
annually after the solar power system upgrade in 2011) were filled using linear regressions
against the hourly NLDAS forcing data product (http://ldas.gsfc.nasa.gov/nldas/). Daily means
were calculated for days with at least 20 hours of data, with remaining days rejected. Wind
speeds were corrected to the same height above canopy at each site using a logarithmic profile
[Shuttleworth, 2012].
Precipitation at the MPB site was observed by a cluster of weighing-type gauges located
250 m west of the eddy covariance tower (Figure 1) and bias-corrected to an onsite Geonor T200B reference weighing precipitation gauge. A wind correction for snowfall was applied to raw
87
measurements. Precipitation observations for the Unimpacted site were obtained from an onsite
NOAA Climate Reference Network (CRN) site, gap filled with data from the SNOTEL station in
winter (<5%) and a tipping bucket (<5%) in summer.
In 20 plots (12 MPB and 8 Unimpacted), snow depths were observed using ultrasonic
sensors (Judd Communications), while subcanopy solar radiation was measured using silicon cell
pyranometers (model SP, Apogee Instruments, Logan, UT). Peak snow water equivalent SWE
for 2010 and 2011 was determined by distributed depth surveys (n ≈ 2,000 site-1 year-1) and
snow-pit based densities [Biederman et al., 2012]. In 2012, snowmelt commenced one month in
advance of normal (March 9, 2012) before peak snowpack surveys could be performed.
Volumetric soil moisture was observed in 2011 and 2012 using Decagon EC-5 and 5-TE
sensors (Decagon Corp. Pullman, WA) at depths of 10 cm (+2 cm), 30 cm (+ 5 cm) and 60 cm (+
10 cm) in each of six plots per site (n = 18 per site).
3.3. Eddy Covariance Vapor Flux Measurements
An above-canopy eddy covariance flux tower was instrumented in 2009 at the MPB site
with an open path infrared gas analyzer (IRGA) (LI-7500; Li-Cor Inc., USA) and a sonic
anemometer (CSAT3; Campbell Scientific Inc., USA). 10 hz data were processed to 30-minute
water vapor fluxes following Lee et al. [2004]. An unbiased statistical friction velocity filter
[Gu, 2005] was used to remove periods with insufficient turbulence.
Daily means were
calculated for days with at least 20 hours of data. MPB site gaps in daily vapor flux (25%)
resulted from solar power failure, lightning strikes and animal activity. To estimate cumulative
seasonal vapor flux, we compared using the mean daily flux within the same year and season
(e.g. summer 2011) against using linear interpolation of a 5-day running mean.
These
approaches produced cumulative seasonal fluxes that differed by < 5%, and the first approach
88
was selected. 30-minute vapor flux values for the Unimpacted site were provided as a Level 1
product by P. Blanken and S.P. Burns from [ftp://cdiac.ornl.gov//pub/ameriflux/data/Level1/
Sites_ByName//Niwot_Ridge/Fluxes/] and processed to daily fluxes. Energy balance closure at
each site averaged about 80%, which is typical for tall-canopy eddy covariance systems, and
corrections to force energy closure were not applied. Observations and data processing at both
sites were consistent with AmeriFlux guidelines, reducing potential bias, and cumulative error at
each site should be at or below 20% [Wilson 2002]. For winters 2010 and 2011, prior to MPB
site winter flux measurements, cumulative winter vapor flux at both sites was determined by the
difference of winter precipitation and peak snow water equivalent [Biederman et al., 2012]. The
MPB site solar power system was upgraded in 2011, enabling winter eddy covariance
observations thereafter.
3.4. Streamflow and Groundwater Measurements (MPB site only)
Streamflow was gaged in nested catchments of first, second and third order at the MPB
site (Figure 1) with respective drainage areas of 15, 136 and 725 ha using Hobo pressure sensors
(Onset Corporation). Stream stage stations were added progressively such that the first-order
catchment was observed in 2010 to 2012, the second-order catchment in 2011 and 2012, and the
third-order catchment only in 2012, and stage records were compared for consistency across
scales. The first-order catchment was gauged at a culvert that flowed partially full (tranquil flow
throughout), and a rating curve was developed using outlet stage (Type 4 culvert flow) [Chow,
1959]. In the other catchments, empirical power function rating curves were developed using
standard US Geological Survey streamflow methods [http://ga.water.usgs.gov/edu/streamflow2],
and these remained consistent from year to year. The second-order catchment stage exceeded
the observational rating curve for one week during 2011 snowmelt, during which Manning’s
89
equation [Chow, 1959] was applied using a cross-section survey. Daily streamflow for the 1100
km2 Laramie River above Woods Landing was obtained from USGS gauge 6659500
(http://waterdata.usgs.gov). We divided annual streamflow into that which occurred during
snowmelt and the remainder, which was defined as base flow.
Groundwater level was observed at the MPB site in a piezometer 114 cm deep placed 55
m east of and 3 m above the main channel of the second-order catchment in a topographically
convergent “zero-order” channel observed to flow during snowmelt and for two to four weeks
thereafter.
3.5. Water Balance Quantification
At the Unimpacted and MPB eddy covariance towers, the tower footprint water balance
was quantified as
(1)
where P is precipitation, V above-canopy eddy covariance vapor flux, ΔS the change in
water storage, and Q* is residual water available for streamflow (Wilson et al., 2001). Annual
changes in catchment water storage ΔS were assumed to be negligible in comparison to the other
fluxes [Wilson et al., 2001].
While observed fluxes (P, V) were reported seasonally, the
calculated water balance component Q* as only determined annually due to large intra-annual
changes in water storage. Annual tower footprint partitioning coefficients were calculated for
runoff
and vapor flux
.
Precipitation and peak SWE were assumed to be uniform in space at each site based on
the small site dimensions (< 8 km2), gentle terrain (maximum elevation difference of 120 m in
the third-order catchment), and prior observations indicating no effect of MPB mortality on
stand-scale peak SWE [Biederman et al., 2012].
90
At the MPB site, catchment water balance was quantified for the first-, second- and thirdorder catchments as:
(2)
where Q is observed streamflow, V* the total catchment vapor flux estimated as the
residual of the other terms [Brutsaert, 2005], and P and ΔS are as in Eq. 1. Annual catchment
partitioning
coefficients
were
calculated
for
runoff
and
vapor
flux
.
3.6. Stable Isotope Observations
Rainfall was collected biweekly in amber glass bottles and protected from evaporation
with mineral oil. Soil water was collected from using porous tension cup lysimeters (Prenart
Corp.) at depths of 10 cm (+2 cm), 30 cm (+ 5 cm) and 60 cm (+ 10 cm) during and after
snowmelt until the soils became too dry to produce samples under tension of 70 kPa. Twelve
lysimeters were located in forested hillslopes at the MPB site with another eight at the
Unimpacted site. Stream water samples were collected from the outlets of the zero-, first-,
second- and third-order MPB catchments and an Unimpacted site zero-order catchment draining
the study hillslope (Figure 1). Stream samples were collected daily at 15:00 by autosampler
(ISCO Corp.) during snowmelt and by grab samples every one to two weeks during base flow.
Isotopic signatures were not altered by up to 7 days storage inside the autosampler. Stable
isotope ratios (δ2H and δ18O) were determined at the University of Arizona using a DLT-100
liquid water isotope analyzer (Los Gatos Research (LGR), Inc., model 908-000).
Further
analytical details are described in Lyon et al. [2009].
Isotopic enrichments of δ18O and δ2H (deuterium) values were used to evaluate kinetic,
abiotic evaporation following the method given by Clark and Fritz [1997) and demonstrated
91
recently in field studies and regional- to global-scale analyses of evaporation, [Gustafson et al.,
2010; Biederman et al., 2012; Jasechko et al., 2013]. Stream evaporation lines were used to
quantify the catchment average humidity at the evaporating surface using the relationship of
Gonfiantini [1986] as adapted by Clark and Fritz [1997]. The total fractionations of both δ18O
and δ2H in the most enriched sample per site and year were used to quantify the maximum
observed kinetic evaporation according to Rayleigh distillation [Clark and Fritz, 1997].
3.7. Computational Methods and Software
Catchment delineation was performed in ARC-GIS 10.1 with the TauDEM toolkit
[http://hydrology.uwrl.usu.edu/taudem/taudem5.0/ accessed Feb. 2013] using a 1-m airborne
laser swath map (ALSM). The percentage northerly aspect in each MPB catchment was defined
as cells with slope greater than 5% and terrain aspect with a north component. Quickbird
imagery was classified using Environment for Visualizing Images (ENVI) software.
Statistical analyses were performed in MATLAB R2012a. Least squares regression with
ANOVA was used to test whether mean daily shortwave radiation, wind speed and vapor
pressure deficit were significant linear predictors of daily summer vapor flux, with ANCOVA
used to evaluate MPB effects. The same approach was used to determine the local meteoric
water line (LMWL) for each site from δ18O and δ2H in precipitation and to identify evaporation
lines significantly different from the LWML.
4. Results
4.1. MPB Infestation
The MPB site infestation began in 2007, and by August 2011, overstory mortality was
77% in the eddy covariance tower footprint and 62%, 61% and 52% in the first-, second-, and
third-order catchments, respectively. Most of the killed trees in the tower footprint had lost their
92
needles and progressed to grey phase by winter 2011. Understory appeared to respond rapidly to
the release of resources, with understory cover reaching 28% to 37% by summer 2011 in greyphase stands as compared to 3% in uninfested stands at the MPB site [Borkhuu et al., submitted
to Soil Biology and Biochemistry, 2013].
4.2. Local Climate and Soil Moisture
In the three years of this study (2010-12), precipitation at the Unimpacted site was 100%,
128% and 90% of average (onsite SNOTEL 1981-2012), while the Cinnabar Park SNOTEL 20
km distant from the MPB site indicated precipitation for 2010-12 was 121%, 127% and 74% of
the 2004-12 average. Winter precipitation ranged from 295 to 459 mm and accounted for 40%
to 52% of annual precipitation at the Unimpacted site and 55% to 72% of annual precipitation at
the MPB site. Winter precipitation was similar at the two sites, varying between sites in any year
by 5 to 9% (Table 2). Summer precipitation ranged from 125 to 566 mm and was more variable
among years and between sites than winter precipitation. The Unimpacted site experienced more
summer precipitation, with the largest seasonal difference between sites in summer 2012 (125
mm at MPB and 428 mm at Unimpacted).
Patterns of snow accumulation and ablation were similar at the two sites (Figure 2). Peak
SWE ranged from 212 to 287 mm and differed between the two sites in any year by 3% to 24%
(Table 2). Spring snowmelt commenced at the same time at each site and ended about one week
earlier at the MPB site each year.
Above-canopy observations of T, VPD, wind speed and Rsw showed similar seasonal
means (Table 2) and temporal patterns (Figure 2) at the two sites. Mean T ranged from -4.0 to 5.3 ˚C in winter and 7.5 to 10.5 ˚C in summer. Mean VPD averaged 0.13 to 0.27 kPa in winter
and 0.58 to 0.90 kPa in summer. Mean winter VPD was lower by 0.05 to 0.11 kPa at the MPB
93
site, while in summer, MPB site mean VPD was greater by 0.17 to 0.21 kPa. Mean daily wind
speeds in all seasons ranged from 2.9 to 6.1 m s-1 and were greater at the Unimpacted site by 0.6
to 2.0 m s-1. Above-canopy Rsw was similar at the two sites, but mean subcanopy Rsw under
MPB forest was greater by three times or more, averaging 61 to 70 W m-2 in summer as
compared to 20 to 21 W m-2 under Unimpacted forest (Table 2).
Volumetric soil moisture averaged 28 % at the MPB site and 19 % at the Unimpacted
site. Peak soil moistures of 36% for MPB and 39% for Unimpacted were reached each year
during late snowmelt.
Snowmelt was the dominant input to soil moisture, which usually
decreased monotonically each summer, reaching minima of 20% and 15% on average at the
MPB and Unimpacted sites, respectively. A few small responses of soil moisture (3 to 5%
increase) to summer rains were observed only at the Unimpacted site and shallowest (10 cm)
depth.
4.3. Tower Footprint Water Balance
Vapor flux time series showed similar patterns at each site both seasonally and daily
(Figure 3), consistent with the similar seasonal means (Table 2) and daily variability (Figure 2)
of likely environmental drivers of vapor flux: wind speed, T, VPD, and Rsw. Annual V at the
two sites was equal to one another in 2010, when P was very similar. In 2011, the year with
greatest P (Figure 4), the MPB tower V (Figure 3) averaged 50% to 100% greater than
Unimpacted V during snowmelt, and V remained greater at the MPB site throughout most of the
relatively wet summer in spite of greater precipitation at the Unimpacted site (Figure 4).
Similarly, in 2012 a pattern of greater V at the MPB site was observed up until June. The MPB
site subsequently received only 29% as much summer P as the Unimpacted site, and by August
and September 2012, V was 30% to 50% lower at the MPB site. Continuous eddy covariance
94
tower operation in winter 2012 showed similar V to the Unimpacted tower, consistent with prior
snow surveys showing similar winter V at each site in 2010 and 2011 [Biederman et al., 2012].
Annual eddy covariance tower V ranged from 569 to 691 mm (Table 3) and varied
among years by only 9% at the Unimpacted tower as compared to 21% variability among years
at the MPB tower. In 2010, when many of the recently-killed MPB trees remained in the red
phase, annual V differed by <1% between the two towers, leading to similar residual streamflow
Q* of 220 mm for Unimpacted and 182 mm for MPB. In 2011, when most MPB trees had lost
their needles and progressed to the grey phase, annual V was 13% (79 mm) greater at the MPB
tower compared to Unimpacted in spite of 30% (239 mm) lower precipitation, leading to MPB
Q* of only 95 mm as compared to 413 mm for Unimpacted.
In the dry year 2012, the
Unimpacted tower footprint received 10% less precipitation than average but recorded its highest
annual V of the three years, resulting in its lowest annual Q* (99 mm). Meanwhile 2012 MPB V
exceeded P, leaving no water available for Q*.
Expected climate drivers VPD, Rsw and wind speed were tested for prediction of summer
V at each tower. VPD was a significant predictor of daily vapor flux (p < 0.001, Figure 5), and
the linear regression slope was nearly four times greater (p < 0.001) at the Unimpacted tower
(slope = 1.04 mm kPa-1) than the MPB tower (slope = 0.27 mm kPa-1). Interestingly, the
intercept of this regression was larger at the MPB tower (2.1 mm) than the Unimpacted tower
(1.5 mm), resulting in greater predicted V at the MPB tower when VPD < 0.8 kPa. Mean daily
Rsw and wind speed were not significantly predictors of summer daily V at either site.
4.4. Stable Isotope Indicators of Evaporation
Isotopes suggested that snowmelt was the primary water source to both soil moisture and
streams. At both sites, isotope values in stream and soil water were much closer to those of
95
snowpack (mean δ18O of -21‰) than of summer rain (mean δ18O of -10 ‰). However, at the
Unimpacted site, six samples of 10-cm soil water and two stream samples had δ18O > -12‰
(Figure 6), suggesting occasional small increases in shallow soil moisture from summer
precipitation.
Soil water and stream samples at the MPB site followed evaporation lines
beginning with the isotopic signature of snowpack. All stream water isotopes plotted along these
evaporation lines, indicating snowmelt as the primary water source and consistent with the
observed lack of deep soil moisture response to summer rainfall.
Isotopes showed more kinetic evaporation at the MPB site, with evaporation line slopes
falling below the LMWL (p< 0.05) in both soil water and stream water at four spatial scales
(Figure 6). The lack of evaporation line at the Unimpacted site indicates no significant kinetic
fractionation. Equilibrium evaporation (humidity of 100%) is not excluded, but net evaporation
into a saturated transition layer would be less due to simultaneous condensation [Clark and Fritz,
1997]. Evaporation lines were not different (p > 0.05) within a given year between MPB soil and
stream water, so these were combined to determine one MPB site evaporation line per year
(Figure 6). The resulting slopes decreased from 5.9 to 5.7 to 4.9 for 2010 to 2012, respectively.
These correspond to average humidity of 87%, 84%, and 66% (Table 4). At the MPB site in
2010, the most enriched stream water showed greater evaporation than the most enriched soil
water (32% evaporative fraction from stream water and 25% from soil water). In 2011 and 2012,
however, the most enriched MPB soil water showed greater evaporation than the most enriched
MPB stream water, with stream evaporative fraction reaching 42% in 2011 and 38% in 2012.
4.5. Catchment Water Balance (MPB Site Only)
The onset of snowmelt runoff at the MPB site occurred April 9, March 25 and March 9 in
the years 2010 to 2012, respectively (Figure 7). Groundwater at the MPB site responded to
96
snowmelt, with the water table reaching the ground surface during snowmelt in 2010 and 2011
and rising to within 20 cm of the ground surface in 2012 (not shown). Summer groundwater
recession did not have observable responses to rain, consistent with the observed lack of
response in soil moisture and water isotopes. Minimum groundwater elevation at the end of the
hydrologic year was 10 cm greater in 2011 than in 2012.
Other than snowmelt, the only hydrologic inputs producing observable response in
groundwater and streamflow were during the first few weeks of October 2010 and 2011 (Figure
7). Precipitation during this period of low evaporative demand and before the establishment of
permanent snowpack produced 1 to 4 cm rise in groundwater and a discernible streamflow
response (> 0.2 mm d-1) in the second- and third-order catchments.
In the first-order MPB catchment, which overlapped the MPB tower footprint, annual
streamflow Q (Table 5) declined monotonically from 2010 to 2012 despite the greatest
precipitation in 2011. Two-thirds of the first-order catchment Q reduction (-167 mm) between
2010 and 2011 occurred during snowmelt (Figure 7; Table 5). Only 11 mm of annual Q were
measured in the dry year 2012, consistent with tower V exceeding annual precipitation (Table 3).
First-order catchment V* was greatest in 2011 and lowest in the dry year 2012 (Table 5),
reflecting the same inter-annual pattern as the observed eddy covariance V (Table 3).
The MPB second-order catchment and Laramie River hydrographs were broader and
peaked later than the first-order catchment hydrograph in 2010 and 2011, while streamflow was
very small at all spatial scales in 2012 (Figure 7). The 1100 km2 Laramie River produced 200
mm of area-normalized streamflow Q in 2010 and 280 mm in 2011, consistent with greater P in
2011 than 2010 (SNOTEL) but in contrast to the reduced Q in the first-order catchment. The
second- and third-order catchments produced greater Q than the first-order catchment, resulting
97
in lower V* (Table 5). The second-order catchment RC was 36% and 19% in 2011 and 2012,
while the third-order catchment RC was 17% in 2012, the only year it was observed.
4.6. Comparison of Sites and Water Balance Approaches
Annual hydrologic partitioning was summarized by the vapor flux coefficient VC and
runoff coefficient RC (Figure 8). Similar VC and RC* at each tower footprint indicated similar
partitioning in 2010, when many of the MPB site trees remained in the red phase. Unimpacted
RC* followed the pattern of annual P, with the highest partitioning to streamflow (40%) in the
wet year 2011 and the lowest percentage (14%) in the dry year 2012, while the MPB RC*
declined each year of the study, from 24% to 12% to zero. Similarly in the first-order MPB
catchment, which overlapped the tower footprint, RC declined from 37% to 16% to 2% over the
period 2010 – 12. Both tower footprint and catchment water balance approaches consistently
indicated an increased percentage of precipitation consumed by vapor flux (increasing VC or
VC*) and reduced availability of water for streamflow (decreasing RC or RC*) at the MPB site.
5. Discussion
Both eddy covariance observations of vapor flux and catchment surface discharge
consistently showed increased vapor flux and reduced streamflow following severe MPB
infestation in the central Rocky Mountains. Strong evidence of kinetic fractionation of stream
and soil water isotopes indicated that part of this response was due to an increase in evaporation
due to the canopy opening up following tree mortality.
Results from this comprehensive
approach differ from prior studies based on streamflow [Potts, 1984; Bethlahmy, 1974], model
simulations [Bewley et al., 2010; Mikkelson et al., 2013b; Pomeroy et al., 2012] and remote
sensing [Bright et al., 2013; Maness et al., 2013], that suggest a hydrologic partitioning shift
towards reduced vapor flux and increased streamflow following beetle-induced disturbance.
98
Given that the sign of observed response is opposite of expectations, it is necessary to examine
the evidence from each water balance approach.
Next, we evaluate evidence regarding
mechanisms and drivers of increased vapor flux. Finally, we discuss the impact of vapor flux
increases on our understanding of hydrologic partitioning response to forest disturbance.
Catchment and tower footprint water balances were consistent in showing reduced
streamflow at the MPB site in comparison to the Unimpacted tower water balance. Consistent
year-to-year reductions in MPB site Q* and first-order catchment Q (Table 3; Table 5), which
had the longest observational record and overlapped with the MPB tower footprint during
prevailing west winds, increased our confidence in interpreting the observed hydrologic
response. The first-order catchment residual between Q and Q* ranged from 11 to 95 mm (2 –
13% of annual P), similar to a mean annual residual of 60 mm (5% of mean annual P) over five
years reported by Wilson et al. [2001]. Lower annual vapor loss from the MPB catchments than
the MPB tower (Table 3; Table 5) suggest that lower mortality in the first-, second- and thirdorder catchments (62%, 61% and 52%, respectively) could have resulted in lower vapor loss than
in the tower footprint (mortality 77%). Alternatively, lower vapor losses could be related to the
greater percentages of land surface with northerly aspect in the larger catchments, which we
would expect to reduce available energy for vapor flux.
Our assumptions of negligible annual changes in storage ΔS and lateral advection did not
appear to affect our conclusions [Wilson et al., 2001; Hubbart et al., 2007; Scott, 2010]. Base
flows reached similar low values each year (Figure 7), and stream water isotopes reflected only
the signature of snowmelt progressively enriched by kinetic fractionation, suggesting no shift
from near-surface to deeper streamflow sources (Figure 6). MPB V exceeded P in the dry year
2012 (Table 3). Since inter-annual variability in MPB water table elevations was small, on the
99
order of 3-5 cm, it is possible that there was lateral subsidy into the tower footprint [Mackay et
al., 2002; Thompson et al., 2011], although we observed no corresponding increase in
streamflow.
Stand scale evaluations of the effects of forest disturbance often focus on precipitation,
particularly snow, that is intercepted by vegetation and subsequently lost back to the atmosphere
[Troendle et al., 1983; Pugh and Small, 2012]. However, our results show that evaporation can
remove a significant fraction of precipitation reaching the ground. Process-based research has
shown that intermediate forest density can minimize winter vapor loss [Veatch et al., 2009;
Gustaffson et al., 2010] and enhance summer soil moisture [Gray et al., 2002] due to effective
shading from solar radiation. Accordingly, it is not surprising that severe canopy loss can drive
increases in snowpack sublimation canceling or outweighing reduced canopy sublimation
(Biederman et al., 2012; Harpold et al., 2013] and that winter vapor flux was similar at MPB and
Unimpacted sites (Figures 3 & 4, winter 2012). Still, the larger MPB site V as compared to
Unimpacted during snowmelt 2011 and 2012 (Figure 3) showed that evaporation or sublimation
of the melting snowpack was important (Table 5) [Schelker et al., 2013].
Stable isotope
fractionation, which is not affected by transpiration, indicated greater kinetic evaporation at the
MPB site (Figure 6), with stream and soil water samples showing maximum kinetic evaporation
as large as 25 to 42%. Increases in evaporation are in addition to the well documented increase
in transpiration of surviving vegetation stimulated by increased availability of water, light and
nutrients [Kozlowski et al., 2002; McMillin et al., 2003; Van Pelt et al., 1999; Rhoades et al.,
2013; Romme et al., 1986].
In light of the existing expectations for more streamflow, our observations of higher V
(Figure 3) due in part to greater evaporation (Figure 6) are somewhat surprising. Higher MPB V
100
as compared to Unimpacted V was observed after most trees lost their needles (2011 and 2012)
and during periods when water was plentiful at the ground surface, namely during snowmelt
2011 and 2012 and the relatively wet summer of 2011. Increased abiotic evaporation from a wet
land surface suggests alterations to the subcanopy environment, though any differences in abovecanopy climate should be considered first when comparing different sites. Larger summer VPD
at the MPB site could have contributed to larger V. However, the weaker response of summer V
to VPD at the MPB site as compared to Unimpacted (Figure 5) suggests VPD differences did not
exert a strong effect at the MPB site.
The MPB site daily V was on average greater than the
Unimpacted site when VPD was below average (0.8 kPa), suggesting that other drivers, such as
wind or available energy, at least partially controlled MPB site vapor flux. The wind speed was
lower at the MPB site, which would tend to reduce the evaporative demand of the atmosphere
and therefore does not explain increased V. Despite similar above-canopy Rsw at the two sites,
the more than 300% increase in subcanopy Rsw (Table 2) was the most likely driver of increased
evaporation. Increases in remotely-sensed winter albedo following MPB infestation suggest
higher shortwave radiation may drive increased evaporation across the landscape for years to a
decade or more post-disturbance [Bright et al. 2013; Vanderhoof et al., 2013, O’Halloran et al.,
2012].
Reduced partitioning to streamflow shown by declining RC and RC* (Figure 8) is in
direct contrast to prevailing expectations and modeled results for hydrologic response to MPB
[Pugh and Gordon, 2013; Bewley et al., 2010; Bright et al., 2013, Maness et al., 2013]. On
closer examination however, our results are not inconsistent with the empirical record in
response to forest disturbance. Bosch and Hewlett [1982] established that streamflow increase
following disturbance was lower in dry regions, such as the semi-arid interior mountains which
101
are host to much of the present MPB epidemic. Stednick [1996] found that in the Rocky
Mountains, streamflow response to disturbance was both the weakest and most variable of all
regions in North America and documented examples of zero response when up to 100% of the
forest was removed. Bethlahmy [1974] did not detect streamflow response in the first decade
following MPB infestation in the Rocky Mountains, while a more recent study found a mixture
of increases and decreases, with reduced streamflow in one catchment being the only significant
change [Brooks et al., 2012; Somor, 2010].
Recently Brown et al. [2013] showed that total vapor flux was conserved for several
years following MPB, but there remains uncertainty in the degree to which a tower footprint
represents spatially heterogeneous catchments [Wilson et al., 2001; Foken, 2008; Scott, 2010].
Our results clarify that following severe MPB disturbance 1) vapor flux may increase,
particularly during periods when water is abundant at the land surface, including snowmelt and
rainy periods, 2) abiotic evaporation contributes a significant fraction of the total vapor flux, and
3) increased penetration of solar radiation beneath MPB-killed canopy is a likely driver of this
compensatory evaporation. These results agree reasonably well with the hydrologic response
expected by Edburg et al. [2012] but demonstrate that compensatory vapor fluxes may be as
large as or larger than reductions in interception and transpiration, reducing water for
streamflow.
Our results highlight several priorities for future research. First, our study evaluated
MPB disturbance in the central Rocky Mountains, and it is necessary to determine the
applicability of our results across the range of latitude and climate within the present MPB
epidemic from Canada [e.g. Rex and Dube, 2006] to Arizona [e.g. Morehouse et al., 2008] as
well as to disturbance to other species of trees and by other beetle species [Raffa et al., 2008;
102
Hicke et al., 2012]. Second, complex topography has been shown to influence hydrologic
partitioning [Scott, 2010; Flerchinger and Cooley, 2000; Flerchinger et al., 2010], yet it is
unclear how forest disturbance might interact with topographic effects. We hypothesize that
forest hillslopes with greater exposure to incident solar radiation may produce greater
compensatory vapor flux following disturbance, as suggested by Rinehart et al. [2008]. Third,
our observations of compensatory vapor flux contrast with land surface models [Bewley et al.,
2010; Mikkelson et al., 2013b; Pomeroy et al., 2012] and remote sensing products [Bright et al.,
2013; Maness et al., 2013], demonstrating a need to improve vapor flux representation in
models. Fourth, resources managers remain uncertain about how interactions among climate
change, forest disturbance, and forest management will affect water for forests [Grant et al.,
2013] and downstream water resources [Clow, 2010; Harpold et al., 2012]. Our observations of
increased subcanopy solar radiation and greater vapor fluxes during periods of wet land surface
suggest that forest disturbance may increase the sensitivity of hydrological processes to climate.
Finally, there is the possibility that increased abiotic evaporation may limit water availability for
forests, with feedbacks to progression of mortality and forest succession [Kaiser et al., 2012;
Grant et al., 2013].
6. Conclusions
Numerous forest disturbance studies [see Brown et al., 2005] suggest that any streamflow
increase will be largest initially and decline back to baseline over a period of 10-15 years or
more as regrowth occurs. Surprisingly, we are unaware of any studies demonstrating increased
streamflow in response to the present MPB epidemic (post-1994) using long-term gauge records
or paired catchments. While it remains possible that ongoing mortality will continue to alter
hydrologic partitioning, eddy covariance tower footprint and catchment water balances
103
consistently yielded a similar inference of increased vapor flux, while stable isotope fractionation
suggested a portion of the increase was due to abiotic evaporation. The resulting availability of
water for streamflow was lower, in contrast to expectations. Future work should evaluate the
ongoing progression of hydrologic response and the transferability of these results to other areas,
but the observations presented here help clarify the lack of streamflow response observed at the
larger scales relevant to water resources.
7. Acknowledgments
This work was supported by funding from the National Science Foundation EAR0910831, United States Geological Service, Wyoming Agricultural Research Station through
McIntire-Stennis and the Wyoming Water Development Commission.
Science Foundation
Arizona and the Arizona Water Sustainability Program provided additional support. LiDAR
mapping of the MPB site was supported by an NCALM Seed grant. We thank B. Bright and J.
Hicke of the University of Idaho for sharing QuickBird land cover classification of the MPB site.
The National Science Foundation Niwot Ridge Long-Term Ecological Research project and the
University of Colorado Mountain Research Station provided logistical support and data. The
Niwot Ridge AmeriFlux tower data was provided by P. Blanken and S. Burns. The AmeriFlux
tower has been supported by the US Department of Energy (DOE), the National Institute for
Climate Change Research (NICCR) and Terrestrial Carbon Processes Program (TCP) and the
National Science Foundation (NSF) Long-Term Research in Environmental Biology (LTREB).
Supplementary snow pit data for Niwot were provided by J. Knowles. Additional regional
atmospheric reference data were provided by J. Frank and the Glacier Lakes Ecosystem
Experiments Site.
The National Center for Atmospheric Research is supported through a
104
cooperative agreement from the National Science Foundation. The authors thank T. Meixnier,
B. Bright and Z. Guido for useful comments.
8. References
Adams, H. D., Luce, C. H., Breshears, D. D., Allen, C. D., Weiler, M., Hale, V. C., Huxman, T.
E. (2012). Ecohydrological consequences of drought- and infestation- triggered tree dieoff: insights and hypotheses. Ecohydrology, 5(2), 145-159. doi: 10.1002/eco.233
Anderton, S. P., White, S. M., & Alvera, B. (2004). Evaluation of spatial variability in snow
water equivalent for a high mountain catchment. Hydrological Processes, 18(3), 435-453.
doi: 10.1002/hyp.1319
Baldocchi, D. D., Hicks, B. B., & Meyers, T. P. (1988). Measuring biosphere-atmosphere
exchanges of biologically related gases with micrometeorological methods. Ecology,
69(5), 1331-1340. doi: 10.2307/1941631
Bethlahmy, N. (1974). More streamflow after a bark-beetle epidemic. Journal of Hydrology,
23(3-4), 185-189.
Bewley, D., Alila, Y., & Varhola, A. (2010). Variability of snow water equivalent and snow
energetics across a large catchment subject to Mountain Pine Beetle infestation and rapid
salvage logging. Journal of Hydrology, 388(3-4), 464-479. doi:
10.1016/j.jhydrol.2010.05.031
Biederman, J. A., Brooks, P. D., Harpold, A. A., Gutmann, E., Gochis, D. J., Reed, D. E., &
Pendall, E. (2012). Multi-scale Observations of Snow Accumulation and Peak Snowpack
Following Widespread, Insect-induced Lodgepole Pine Mortality. Ecohydrology.
Boon, S. (2012). Snow accumulation following forest disturbance. Ecohydrology. doi: doi
10.1002/eco.212.
Bosch, J. M., & Hewlett, J. D. (1982). A review of catchment experiments to determine the
effect of vegetation changes on water yield and evapo-transpiration. Journal of
Hydrology, 55(1-4), 3-23. doi: 10.1016/0022-1694(82)90117-2
Breshears, D. D., Myers, O. B., Meyer, C. W., Barnes, F. J., Zou, C. B., Allen, C. D., Pockman,
W. T. (2009). Tree die-off in response to global change-type drought: mortality insights
from a decade of plant water potential measurements. Frontiers in Ecology and the
Environment, 7(4), 185-189. doi: 10.1890/080016
Bright, B. C., Hicke, J. A., & Meddens, A. J. H. (2013). Effects of bark beetle-caused tree
mortality on biogeochemical and biogeophysical MODIS products. Journal of
Geophysical Research Letters: Biogeosciences, 118, 1-9. doi: 10.1002/jgrg.20078
Brooks, P. D., Harpold, A. A., Biederman, J. A., Litvak, M. E., Broxton, P. D., Gochis, D. J.,
Ewers, B. E. (2012). Insect, Fires and Climate Change: Implications for Snow Cover,
Water Resources and Ecosystem Recovery in Western North America. Paper presented at
the AGU Fall Meeting, 3-7 Dec. 2012, San Francisco, CA.
Brown, A. E., Zhang, L., McMahon, T. A., Western, A. W., & Vertessy, R. A. (2005). A review
of paired catchment studies for determining changes in water yield resulting from
105
alterations in vegetation. Journal of Hydrology, 310(1-4), 28-61. doi:
10.1016/j.jhydrol.2004.12.010
Brown, M. G., Black, T. A., Nesic, Z., Foord, V. N., Spittlehouse, D. L., Fredeen, A. L., Meyer,
G. (2013). - Evapotranspiration and canopy characteristics of two lodgepole pine stands
following mountain pine beetle attack. doi: - 10.1002/hyp.9870
Brutsaert, W. (2005), Hydrology: An Introduction, Cambridge University Press, New York.
Chow, V. T. (1959). Open-Channel Hydraulics. New York, USA: McGraw-Hill, Inc.
Clark, I., & Fritz, P. (1997). Environmental Isotopes in Hydrogeology. Boca Raton, Florida.:
Lewis Publishers.
Clow, D. W. (2010). Changes in the Timing of Snowmelt and Streamflow in Colorado: A
Response to Recent Warming. Journal of Climate, 23(9), 2293-2306. doi:
10.1175/2009jcli2951.1
Clow, D. W., Rhoades, C., Briggs, J., Caldwell, M., & Lewis, W. M., Jr. (2011). Responses of
soil and water chemistry to mountain pine beetle induced tree mortality in Grand County,
Colorado, USA. Applied Geochemistry, 26, S174-S178. doi:
10.1016/j.apgeochem.2011.03.096
Coops, N. C., Johnson, M., Wulder, M. A., & White, J. C. (2006). Assessment of QuickBird high
spatial resolution imagery to detect red attack damage due to mountain pine beetle
infestation. Remote Sensing of Environment, 103(1), 67-80. doi:
10.1016/j.rse.2006.03.012
Edburg, S. L., Hicke, J. A., Brooks, P. D., Pendall, E. G., Ewers, B. E., Norton, U., Meddens, A.
J. H. (2012). Cascading impacts of bark beetle-caused tree mortality on coupled
biogeophysical and biogeochemical processes. Frontiers in Ecology and the
Environment, 10(8), 416-424. doi: 10.1890/110173
Ellison, D., Futter, M. N., & Bishop, K. (2012). On the forest cover-water yield debate: from
demand- to supply-side thinking. Global Change Biology, 18(3), 806-820. doi:
10.1111/j.1365-2486.2011.02589.x
Flerchinger, G. N., & Cooley, K. R. (2000). A ten-year water balance of a mountainous semiarid watershed. Journal of Hydrology, 237(1-2), 86-99. doi: 10.1016/s00221694(00)00299-7
Flerchinger, G. N., Marks, D., Reba, M. L., Yu, Q., & Seyfried, M. S. (2010). Surface fluxes and
water balance of spatially varying vegetation within a small mountainous headwater
catchment. Hydrology and Earth System Sciences, 14(6), 965-978. doi: 10.5194/hess-14965-2010
Foken, T. (2008). The energy balance closure problem: An overview. Ecological Applications,
18(6), 1351-1367. doi: 10.1890/06-0922.1
Gonfiantini, R. (1986). Environmental isotopes in lake studies. In P. Fritz & J.-C. Fontes (Eds.),
Handbook of Environmental Isotope Geochemistry: The Terrestrial Environment (pp.
113-168). Amsterdam: Elsevier.
106
Grant, G. E., Tague, C. L., & Allen, C. D. (2013). Watering the forest for the trees: an emerging
priority for managing water in forest landscapes. Frontiers in Ecology and the
Environment, 11(6), 314-321. doi: 10.1890/120209
Gray, A. N., Spies, T. A., & Easter, M. J. (2002). Microclimatic and soil moisture responses to
gap formation in coastal Douglas-fir forests. Canadian Journal of Forest Research-Revue
Canadienne De Recherche Forestiere, 32(2), 332-343. doi: 10.1139/x01-200
Gu, L. H., Falge, E. M., Boden, T., Baldocchi, D. D., Black, T. A., Saleska, S. R., Xu, L. K.
(2005). Objective threshold determination for nighttime eddy flux filtering. Agricultural
and Forest Meteorology, 128(3-4), 179-197. doi: 10.1016/j.agrformet.2004.11.006
Gustafson, J. R., Brooks, P. D., Molotch, N. P., & Veatch, W. C. (2010). Estimating snow
sublimation using natural chemical and isotopic tracers across a gradient of solar
radiation. Water Resources Research, 46. doi: 10.1029/2009wr009060
Harpold, A., Brooks, P., Rajagopal, S., Heidbuchel, I., Jardine, A., & Stielstra, C. (2012).
Changes in snowpack accumulation and ablation in the intermountain west. Water
Resources Research, 48, 11. doi: 10.1029/2012wr011949
Harpold, A. A., Biederman, J. A., Condon, K., Merino, M., Korgaonkar, Y., Nan, T., Brooks, P.
D. (2013). - Changes in snow accumulation and ablation following the Las Conchas
Forest Fire, New Mexico, USA. doi: - 10.1002/eco.1363
Hicke, J. A., Allen, C. D., Desai, A. R., Dietze, M. C., Hall, R. J., Hogg, E. H., Vogelmann, J.
(2012). Effects of biotic disturbances on forest carbon cycling in the United States and
Canada. Global Change Biology, 18(1), 7-34. doi: 10.1111/j.1365-2486.2011.02543.x
Hubbard, R. M., Rhoades, C. C., Elder, K., & Negron, J. (2013). Changes in transpiration and
foliage growth in lodgepole pine trees following mountain pine beetle attack and
mechanical girdling. Forest Ecology and Management, 289, 312-317. doi:
10.1016/j.foreco.2012.09.028
Huber, C. (2005). Long lasting nitrate leaching after bark beetle attack in the highlands of the
Bavarian Forest National Park. Journal of Environmental Quality, 34(5), 1772-1779. doi:
10.2134/jeq2004.0210
Jasechko, S., Sharp, Z. D., Gibson, J. J., Birks, S. J., Yi, Y., & Fawcett, P. J. (2013). Terrestrial
water fluxes dominated by transpiration. Nature, 496(7445), 347-+. doi:
10.1038/nature11983
Kaiser, K. E., McGlynn, B. L., & Emmanual, R. E. (2012). Ecohydrology of an outbreak:
mountain pine beetle impacts trees in drier landscape positions first. Ecohydrology, 6(3),
444-454. doi: DOI: 10.1002/eco.1286
Knight, D. H., Fahey, T. J., & Running, S. W. (1985). Water and nutrient outflow from
contrasting lodgepole pine forests in Wyoming. Ecological Monographs, 55(1), 29-48.
Kozlowski, T. T. (2002). Physiological ecology of natural regeneration of harvested and
disturbed forest stands: implications for forest management. Forest Ecology and
Management, 158(1-3), 195-221. doi: 10.1016/s0378-1127(00)00712-x
Lee, X., Massman, W. J., & Law, B. (2004). Handbook of Micrometeorology: A Guide for
Surface Flux Measurement and Analysis. Dordrecht, The Netherlands: Springer.
107
Lyon, S. W., Desilets, S. L. E., & Troch, P. A. (2009). A tale of two isotopes: differences in
hydrograph separation for a runoff event when using delta D versus delta O-18.
Hydrological Processes, 23(14), 2095-2101. doi: 10.1002/hyp.7326
MacKay, D. S., Ahl, D. E., Ewers, B. E., Gower, S. T., Burrows, S. N., Samanta, S., & Davis, K.
J. (2002). Effects of aggregated classifications of forest composition on estimates of
evapotranspiration in a northern Wisconsin forest. Global Change Biology, 8(12), 12531265. doi: 10.1046/j.1365-2486.2002.00554.x
Maness, H., Kushner, P. J., & Fung, I. (2013). Summertime climate response to mountain pine
beetle disturbance in British Columbia. Nature Geoscience, 6(1), 65-70. doi:
10.1038/ngeo1642
McMillin, J. D., & Allen, K. K. (2003). Effects of Douglas-fir beetle (Coleoptera : Scolytidae)
infestations on forest overstory and understory conditions in western Wyoming. Western
North American Naturalist, 63(4), 498-506.
Meddens, A. J. H., Hicke, J. A., & Ferguson, C. A. (2012). Spatiotemporal patterns of observed
bark beetle-caused tree mortality in British Columbia and the western United States.
Ecological Applications, 22(7), 1876-1891.
Mikkelson, K. M., Bearup, L. A., Maxwell, R. M., Stednick, J. D., McCray, J. E., & Sharp, J. O.
(2013). Bark beetle infestation impacts on nutrient cycling, water quality and
interdependent hydrological effects. Biogeochemistry, 115, 1-21. doi: DOI
10.1007/s10533-013-9875-8
Mikkelson, K. M., Maxwell, R. M., Ferguson, I., Stednick, J. D., McCray, J. E., & Sharp, J. O.
(2013). Mountain pine beetle infestation impacts: modeling water and energy budgets at
the hill-slope scale. Ecohydrology, 6(1), 64-72. doi: 10.1002/eco.278
Molotch, N. P., Brooks, P. D., Burns, S. P., Litvak, M., Monson, R. K., McConnell, J. R., &
Musselman, K. (2009). Ecohydrological controls on snowmelt partitioning in mixedconifer sub-alpine forests. Ecohydrology, 2(2), 129-142. doi: 10.1002/eco.48
Morehouse, K., Johns, T., Kaye, J., & Kaye, A. (2008). Carbon and nitrogen cycling
immediately following bark beetle outbreaks in southwestern ponderosa pine forests.
Forest Ecology and Management, 255(7), 2698-2708. doi: 10.1016/j.foreco.2008.01.050
Mu, Q. Z., Zhao, M. S., & Running, S. W. (2011). Improvements to a MODIS global terrestrial
evapotranspiration algorithm. Remote Sensing of Environment, 115(8), 1781-1800. doi:
10.1016/j.rse.2011.02.019
Musselman, K. N., Molotch, N. P., & Brooks, P. D. (2008). Effects of vegetation on snow
accumulation and ablation in a mid-latitude sub-alpine forest. Hydrological Processes,
22(15), 2767-2776. doi: 10.1002/hyp.7050
O'Halloran, T. L., Law, B. E., Goulden, M. L., Wang, Z. S., Barr, J. G., Schaaf, C., . . . Engel, V.
(2012). Radiative forcing of natural forest disturbances. Global Change Biology, 18(2),
555-565. doi: 10.1111/j.1365-2486.2011.02577.x
Pomeroy, J., Fang, X., & Ellis, C. (2012). Sensitivity of snowmelt hydrology in Marmot Creek,
Alberta, to forest cover disturbance. Hydrological Processes, 26(12), 1892-1905. doi:
10.1002/hyp.9248
108
Pugh, E., & Gordon, E. (2013). A conceptual model of water yield effects from beetle-induced
tree death in snow-dominated lodgepole pine forests. Hydrological Processes, 27, 20482060. doi: 10.1002/hyp.9312
Pugh, E., & Small, E. (2012). The impact of pine beetle infestation on snow accumulation and
melt in the headwaters of the Colorado River. Ecohydrology, 5(4), 467-477. doi:
10.1002/eco.239
Raffa, K. F., Aukema, B. H., Bentz, B. J., Carroll, A. L., Hicke, J. A., Turner, M. G., & Romme,
W. H. (2008). Cross-scale drivers of natural disturbances prone to anthropogenic
amplification: The dynamics of bark beetle eruptions. Bioscience, 58(6), 501-517. doi:
10.1641/b580607
Rex, J., & Dube, S. (2006 ). Predicting the Risk of wet ground areas in the Vanderhoof Forest
District: Project description and Progress Report. BC Journal of Ecosystems and
Management, 7(2), 57-71.
Rhoades, C. C., McCutchan, J. H., Jr., Cooper, L. A., Clow, D., Detmer, T. M., Briggs, J. S.,
Lewis, W. M., Jr. (2013). Biogeochemistry of beetle-killed forests: Explaining a weak
nitrate response. Proceedings of the National Academy of Sciences of the United States of
America, 110(5), 1756-1760. doi: 10.1073/pnas.1221029110
Rinehart, A. J., Vivoni, E. R., & Brooks, P. D. (2008). Effects of vegetation, albedo, and solar
radiation sheltering on the distribution of snow in the Valles Caldera, New Mexico.
Ecohydrology, 1(3), 253-270. doi: 10.1002/eco.26
Romme, W. H., Knight, D. H., & Yavitt, J. B. (1986). Mountain pine beetle outbreaks in the
Rocky Mountains - regulators of primary productivity. American Naturalist, 127(4), 484494.
Schelker, J., Kuglerova, L., Eklof, K., Bishop, K., & Laudon, H. (2013). Hydrological effects of
clear-cutting in a boreal forest - Snowpack dynamics, snowmelt and streamflow
responses. Journal of Hydrology, 484, 105-114. doi: 10.1016/j.jhydrol.2013.01.015
Scott, R. L. (2010). Using watershed water balance to evaluate the accuracy of eddy covariance
evaporation measurements for three semiarid ecosystems. Agricultural and Forest
Meteorology, 150(2), 219-225. doi: 10.1016/j.agrformet.2009.11.002
Scott-Denton, L. E., Sparks, K. L., & Monson, R. K. (2003). Spatial and temporal controls of soil
respiration rate in a high-elevation, subalpine forest. Soil Biology & Biochemistry, 35(4),
525-534. doi: 10.1016/s0038-0717(03)00007-5
Shuttleworth, W. J. (2012). Terrestrial Hydrometeorology. Oxford, UK: John Wiley & Sons.
Somor, A. J. (2010). Quantifying streamflow change following bark beetle outbreak in multiple
central Colorado catchments. (MS), University of Arizona, Tucson, AZ.
Stednick, J. D. (1996). Monitoring the effects of timber harvest on annual water yield. Journal of
Hydrology, 176(1-4), 79-95. doi: 10.1016/0022-1694(95)02780-7
Thompson, S. E., Harman, C. J., Troch, P. A., Brooks, P. D., & Sivapalan, M. (2011). Spatial
scale dependence of ecohydrologically mediated water balance partitioning: A synthesis
framework for catchment ecohydrology. Water Resources Research, 47. doi:
10.1029/2010wr009998
109
Tokuchi, N., Ohte, N., Hobara, S., Kim, S. J., & Masanori, K. (2004). Changes in
biogeochemical cycling following forest defoliation by pine wilt disease in Kiryu
experimental catchment in Japan. Hydrological Processes, 18(14), 2727-2736. doi:
10.1002/hyp.5578
Troendle, C. A. (1983). The potential for water yield augmentation from forest management in
the Rocky Mountain Region. Water Resources Bulletin, 19(3), 359-373.
Troendle, C. A., & King, R. M. (1987). The effect of partial and clearcutting on streamflow at
Deadhorse Creek, Colorado. Journal of Hydrology, 90(1-2), 145-157. doi: 10.1016/00221694(87)90177-6
Van Pelt, R., & Franklin, J. F. (1999). Response of understory trees to experimental gaps in oldgrowth Douglas-fir forests. Ecological Applications, 9(2), 504-512. doi: 10.1890/10510761(1999)009[0504:routte]2.0.co;2
Vanderhoof, M., Williams, C. A., Ghimire, B., & Rogan, J. (2013). - Impact of mountain pine
beetle outbreaks on forest albedo and radiative forcing, as derived from Moderate
Resolution Imaging Spectroradiometer, Rocky Mountains, USA. doi: 10.1002/jgrg.20120
Varhola, A., Coops, N. C., Bater, C. W., Teti, P., Boon, S., & Weiler, M. (2010). The influence
of ground- and lidar-derived forest structure metrics on snow accumulation and ablation
in disturbed forests. Canadian Journal of Forest Research-Revue Canadienne De
Recherche Forestiere, 40(4), 812-821. doi: 10.1139/x10-008
Veatch, W., Brooks, P. D., Gustafson, J. R., & Molotch, N. P. (2009). 'Quantifying the effects of
forest canopy cover on net snow accumulation at a continental, mid-latitude site'.
Ecohydrology, 2(2), 115-128. doi: 10.1002/eco.45
Westerling, A. L., Hidalgo, H. G., Cayan, D. R., & Swetnam, T. W. (2006). Warming and earlier
spring increase western US forest wildfire activity. Science, 313(5789), 940-943. doi:
10.1126/science.1128834
Williams, A. P., Allen, C. D., Macalady, A. K., Griffin, D., Woodhouse, C. A., Meko, D. M.,
McDowell, N. G. (2013). Temperature as a potent driver of regional forest drought stress
and tree mortality. Nature Climate Change, 3(3), 292-297. doi: 10.1038/nclimate1693
Williams, A. P., Allen, C. D., Millar, C. I., Swetnam, T. W., Michaelsen, J., Still, C. J., &
Leavitt, S. W. (2010). Forest responses to increasing aridity and warmth in the
southwestern United States. Proceedings of the National Academy of Sciences of the
United States of America, 107(50), 21289-21294. doi: 10.1073/pnas.0914211107
110
9. Tables
Table 1. Characteristics of the Unimpacted and MPB study sites.a
Soils
Site
b
Unimpacted
mortality < 5%
c
MPB
mortality 52-77%
a
EC Tower
Elevation Latitude
(m)
Longitude
Stand
Age
(yr)
Stem
Density
(per ha)
Tree
Ht.
(m)
DBH
(cm)
3000
40° 02′
-105° 33′
100
3900
11.4
12.1
Cryochrepts
Sandy Loam
2750
41° 04′
-106° 07′
90-110
2500
11.1
14.0
Cryochrepts
Sandy Loam
The sites had similar elevation, soil texture, stand age and average tree sizes in lodgepole pinedominated overstory. Location and elevation are for the eddy covariance (EC) towers. The
Unimpacted site had greater stem density in the eddy covariance tower footprint. bScott-Denton
et al., 2003; cKnight et al., 1985.
111
Table 2. Local climate for the Unimpacted and MPB sitesa
Year
SWE
(mm)
YR
P
(mm)
W
S
T
(˚C)
YR
W
VPD
(kPa)
S
W
Wind Speed
(m s-1)
S
W
S
Rsw Above Canopy
(W m-2)
W
S
Rsw Below Canopy
(W m-2)
W
S
---------- Unimpacted --------2010
2011
2012
230 (2)
270 (2)
222 (6)
420
459
295
381 801
566 1025
428 722
-4.8 (5.9)
-4.0 (7.1)
-4.0 (5.9)
9.0 (6.2)
7.5 (7.3)
9.0 (6.0)
0.24 (0.19)
0.24 (0.20)
0.27 (0.19)
0.64 (0.40)
0.58 (0.40)
0.69 (0.41)
5.2 (2.9)
6.1 (3.2)
6.0 (3.4)
3.5 (1.5) 137 (63) 251 (79)
4.1 (2.1) 129 (59) 250 (81) 14 (9)
3.6 (1.4) 131 (50) 240 (79)
-
21 (12)
20 (12)
-
3.2 (1.0)
3.8 (1.2)
4.1 (1.3)
2.9 (0.8) 131 (62) 275 (74)
3.1 (1.0) 123 (55) 261 (85) 25 (15)
2.9 (0.9) 124 (45) 270 (74)
-
61 (39)
70 (40)
---------- MPB --------2010
2011
2012
a
286 (1)
270 (3)
266 (10)
440
430
323
311 751
356 786
125 448
-5.1 (5.5) 10.2 (6.2)
-4.0 (7.0) 9.1 (7.6)
-5.3 (5.9) 10.5 (6.2)
0.13 (0.15)
0.19 (0.22)
0.20 (0.16)
0.82 (0.45)
0.75 (0.51)
0.90 (0.47)
YR, yearly; W, winter; S, Summer; SWE, peak snow water equivalent; P, precipitation; T, temperature; VPD, vapor pressure deficit;
Rsw shortwave (solar) radiation. For SWE, the mean and one standard error of spatially distributed observations are shown.
Precipitation values are period totals. All other values shown are mean (std. dev.) of daily values. The sites received similar winter
precipitation and peak SWE, but the Unimpacted site received more summer P. T was similar at both sites. WS was lower at the
MPB site. MPB site VPD and Rsw were slightly lower in winter and higher in summer than at the Unimpacted site. Below-canopy
Rsw averaged 70% greater in winter and 300-350% greater in summer at the MPB site as compared to Unimpacted. In 2012 the sites
received the lowest P and had the greatest mean VPD of the study period.
112
Table 3. Eddy covariance tower footprint water balance at the Unimpacted and MPB
sitesa
Year
W
P
(mm)
S
Year
W
V
(mm)
S
Year
Q*
(mm)
Year
RC* = Q* ÷
P
(%)
Year
220
413
99
27
40
14
182
86
0b
24
11
0b
---------- Unimpacted ---------2010
2011
2012
420
459
295
381
566
428
801
1025
722
190
189
194
383
423
429
573
612
623
-------------- MPB --------------2010
2011
2012
a
440
430
323
311
356
125
751
786
448
154
160
207
416
540
384
569
700
591
P, precipitation; V, vapor flux, Q*, residual streamflow; RC* eddy covariance water
balance runoff coefficient Q* P-1; W, winter, S, summer; MPB, mountain pine beetle site.
Annual vapor flux varied by 9% at the Unimpacted site and 21% at the MPB site. RC*
coefficient declined each year at the MPB site, while at the Unimpacted site, the runoff
coefficient showed a similar pattern to annual precipitation, with the highest value in
2011 and lowest in 2012. bTotal vapor flux for 2012 exceeded precipitation.
113
Table 4. MPB site annual stable isotope fractionationa
Year
Evaporation
Line Slope
(-)
Transition
Layer
Humidity
(%)
Soil Water
Δ18O-max
(‰)
Stream Water
Δ18O-max
(‰)
Soil Water
fkin-max
(%)
Stream
Water fkinmax
(%)
2010
5.9
87
4.3
5.6
25
32
2011
5.7
84
9.3
5.0
42
28
2012
4.9
66
9.8
6.3
38
26
Δ O-max, enrichment of the most isotopically enriched sample with respect to the local
meteoric water line; fkin-max, fraction of precipitation evaporated kinetically in the most
enriched sample with respect to the local meteoric water line.
a 18
114
Table 5. Annual catchment water balance fluxes at the MPB sitea
Water
Year
P
(mm)
V*
(mm)
Snowmelt Q
(mm)
Base Flow Q
(mm)
Total Q
(mm)
RC = Q ÷ P
(%)
------------ first-order catchment 15 ha; 11% northerly aspect, 62% mortality----------2010
2011
2012
751
786
448
474
657
437
250
83
5
27
52
6
277
129
11
37
16
2
---------- second-order catchment 136 ha; 20% northerly aspect, 61% mortality ---------2011
2012
786
448
506
362
150
78
130
8
280
86
36
19
--------- third-order catchment 725 ha; 39% northerly aspect, 52% mortality -------2012
a
448
374
45
31
76
17
P, annual precipitation; V*, annual catchment-based residual vapor flow; Q, annual
observed streamflow. Q in the first-order catchment declined by 53% from 2010 to 2011
in spite of 5% greater precipitation, indicating increased V*. The decline in Q was due
primarily to a 67% reduction in snowmelt Q, highlighting the importance of this brief
period for annual hydrologic partitioning. Forest mortality was quantified August, 2011,
when most dead trees were in grey phase.
115
10. Figures
Figure 1. Site maps (upper panels) and detail maps (lower panels). Both sites include
precipitation gauges, eddy covariance and meteorological towers. At each site, stable
water isotopes were sampled from soil lysimeters, and zero-order streams. Streamflow
and stable water isotopes were observed at the outlets of three nested catchments at the
MPB site.
116
117
Figure 2. Precipitation at the a. Unimpacted and b. MPB sites, c. snow depth, d.
temperature, e. vapor pressure deficit VPD and f. incoming shortwave radiation Rsw
divided by winter W and summer S. Patterns of winter precipitation were very similar at
the two sites, but the Unimpacted site received more summer precipitation. Snow depth
time series show similar patterns of accumulation and peak amounts, with the MPB site
snowmelt 1 to 2 weeks earlier than at the Unimapcted site. Temporal patterns of T, VPD
and Rsw were very similar between the two sites for most of the year. The Unimpacted
site tended to have lower T, VPD and Rsw in late summer, consistent with more summer
precipitation. Shown are 5-day running means of 30-minute observations.
indicates winter (W) and summer (S).
Shading
118
119
Figure 3. Above-canopy eddy covariance vapor flux V divided into winter W, and
summer S, which begins with snowmelt SM. Values shown are 5-day running means of
30-minute observations. The MPB site showed larger vapor flux than the Unimpacted
site during snowmelt 2011 and 2012.
MPB site vapor flux was lower than at the
Unimpacted site in late summer 2012, consistent with dier and warmer conditions (Figure
2). 2012 winter vapor flux patterns were similar for the two sites, consistent with snow
surveys showing equal amounts of total winter sublimation in 2010 and 2011.
-1
Vapor Flux (mm d )
W
S
W
SM
S
W
SM
S
SM
4
3
Unimpacted
MPB
2
1
0
1-Oct-09
1-Apr-10
1-Oct-10
1-Apr-11
1-Oct-11
1-Apr-12
1-Oct-12
120
Figure 4. Cumulative annual (a.) precipitation P and (b.) eddy covariance vapor flux V
divided into winter W, and summer S, which begins with snowmelt SM. When
precipitation was very similar at each site in 2010, annual V was equal at each site. In
2011, the year with greatest P, MPB tower V (slope of cumulative flux) was greater than
Unimpacted V during snowmelt and throughout the summer. In 2012, a similar pattern
of larger flux at the MPB site was observed during snowmelt, but the summer was much
drier at the MPB site (P = 125 mm) than the Unimpacted site (P = 428 mm), and MPB
Vapor Flux (mm)
Precipitation (mm)
site V slowed during late summer of this dry year.
1000 a.
W
S
W
SM
S
W
SM
S
SM
750
500
250
0
Unimpacted
MPB
700 b.
500
300
100
0
1-Oct-09
1-Apr-10
1-Oct-10
1-Apr-11
1-Oct-11
1-Apr-12
1-Oct-12
121
Figure 5. Summer daily vapor flux as a function of vapor pressure deficit (VPD) at the
Unimpacted and MPB eddy covariance towers. VPD was a significant predictor of
summer vaporization at both sites (p < 0.01), but the MPB site vapor flux was less
sensitive to VPD. The intercept was significantly greater at the MPB site, while the slope
was lower (p < 0.01). When daily VPD was less than 0.8 kpa, the vapor flux was greater
on average at the MPB site.
Unimpacted
Daily Vapor Flux (mm)
6
V = 1.04*VPD+1.5
MPB
p < 0.001
V = 0.27*VPD+2.1
2
p < 0.001
2
R = 0.18
R = 0.01
4
2
0
0.5
1
1.5
2 0
0.5
Daily Mean VPD (kPa)
1
1.5
2
122
Figure 6. Stable water isotopes in soil water and stream water and the local meteoric
water line (LMWL) for the (a.) Unimpacted and (b.) MPB sites. Stream water was not
distinguishable from the LMWL at the Unimpacted site (p > 0.05). At the MPB site,
evaporation line slopes less than the LMWL (p< 0.05) demonstrate kinetic fractionation,
suggesting compensatory vapor flux due to abiotic evaporation.
123
Figure 7. MPB site area-normalized streamflow for nested first, second and third-order
study catchments in the headwaters of the Laramie River, divided into winter W, and
summer S, which begins with snowmelt SM. Greater vapor flux (Figure 3), especially
during snowmelt, contributed to decreasing annual streamflow in the first-order study
catchment. Time series shown are five-day running means.
W
-1
Streamflow (mm d )
12
S
W
Peak = 23 SM
S
W
SM
S
SM
10
1st-order
2nd-order
3rd-order
Laramie Riv.
8
6
4
2
0
1-Oct-09
1-Apr-10
1-Oct-10
1-Apr-11
1-Oct-11
1-Apr-12
1-Oct-12
124
Figure 8. Annual vapor flux coefficients VC = V÷P and runoff coefficients RC = Q÷P
for the (a.) Unimpacted tower footprint, (b.) MPB tower footprint and (c.) MPB firstorder catchment. Water balance calculated ratios are denoted by “ * ”. The MPB site
tower footprint partitioning was similar to Unimpacted in 2010, but in 2011 the MPB
tower V increased in spite of greater P, reducing RC*. In the dry year 2012, MPB tower
footprint V (panel b.) exceeded annual precipitation, leaving no water available for
streamflow and RC* of zero. MPB tower footprint VC were greater than catchment
VC*, possibly because higher mortality in the tower footprint (77%) than across the firstorder catchment (62%) led to greater compensatory vapor flux.
125
126
APPENDIX C: CATCHMENT BIOGEOCHEMICAL IMPACTS OF BARK BEETLE
INFESTATION ATTENUATED IN THE RIPARIAN ZONE AND HEADWATER
STREAMS
Manuscript for submission to the Journal of Geophysical Research-Biogeosciences
Title:
Catchment biogeochemical impacts of bark beetle infestation attenuated in the
riparian zone and headwater streams
Running title:
Attenuation of bark beetle biogeochemical impacts
Authors:
Joel A. Biederman1*, P. D. Brooks1, T. Meixner1, A. A. Harpold2, , D. E. Reed4,5,
E. Gutmann3
1
Department of Hydrology and Water Resources
University of Arizona, Tucson AZ 85721
2
INSTAAR, Boulder, CO 80309
3
NCAR, Boulder, CO 80301
4
Department of Atmospheric Science and Program in Ecology
University of Wyoming, Laramie, WY 82071
5
Department of Botany and Program in Ecology
University of Wyoming, Laramie, WY 82071
Keywords: carbon, nitrogen, forest disturbance, groundwater, soil, insect, beetle
127
Abstract
Forest disturbance is expanding in rate and extent and is affecting many montane
catchments critical to water resources. Western North America is experiencing a bark
beetle epidemic that has affected an estimated 20 million hectares of forest in Canada and
the United states. Consequences for biogeochemical cycling may include loss of limiting
nutrients critical to forest succession and contamination of ground and surface waters.
We followed biogeochemical response to severe MPB forest mortality along a putative
hydrological flowpath from the uplands to the riparian-stream system including soil
water, upland and riparian groundwater, and nested streams of zero- to third-order,
encompassing spatial scales of 10-1 to 103 m. Soil water and upland groundwater showed
NO3 elevated by two orders of magnitude as compared to an unimpacted site, while
upland groundwater also showed about five times greater DON.
Nearly 100% of the
NO3 impacts and most of the DON were attenuated in the riparian zone en route to the
stream. DOC and DON in the impacted site channel network were attenuated by 46%
and 65%, respectively, within 4.5 km downstream of channel initiation. These results
indicate that groundwater was contaminated with organic and inorganic nitrogen, riparian
zones protected streams from elevated groundwater N, and rapid attenuation of DOC and
DON in headwater streams is partly responsible for the lack of biogeochemical impacts
observed at the watershed scale.
128
1. Introduction
Montane forests are among the world’s cleanest and most reliable water sources
[Brown et al., 2005], but forests are changing rapidly due to combined effects of
increased temperature, drought, fire, and pathogen infestation [Williams et al., 2010,
2013; Breshears et al., 2009; Adams et al., 2012; Westerling et al., 2006; Huber, 2005;
Tokuchi et al., 2004]. Western North America has experienced rapid and extensive forest
mortality due to endemic mountain pine beetle (Dendroctonus ponderoseae), and climate
change is expected to enhance infestation dynamics [Ayres and Lombardero, 2000;
Mitton and Ferrenberg, 2012].
Over the last two decades, bark beetles including
mountain pine beetle (MPB) have affected approximately 20 million ha in western North
America [Meddens et al., 2012], an area similar in extent to that impacted by wildfire.
Mountain pine beetles introduce fungi, inhibiting sap flow and killing host trees within
weeks to months [Hubbard et al., 2013]. Dead trees retaining red needles are said to be
in a red phase for one to two years as needles are lost [Wulder et al., 2006]. During the
subsequent grey phase, snags, branches and boles fall to the ground over years to
decades. This progressive change in canopy structure and gradual input of organic matter
to soils differentiates beetle mortality from more well-studied disturbances by harvest,
extreme weather, or nonlethal pathogens, challenging our ability to predict
biogeochemical impacts for ecosystem succession and water resources [Edburg et al.,
2012; Adams et al., 2012; Mikkelson et al., 2013].
A significant body of literature provides context for expected biogeochemical
response to forest disturbance by diverse agents including nonlethal insects [Swank,
129
1981; Bormann and Likens, 1979), harvest [Bormann et al., 1968; Schelker et al., 2012;
Jones and Post, 2004] extreme weather [Houlton et al., 2003; Bernhardt et al.., 2005]
and fire [Mast and Clow, 2008]. These disturbance studies have led to expected increases
in aqueous loss of inorganic N, particularly NO3, which is of concern for both human
health and ecosystem impacts [Galloway et al., 2003].
Higher NO3 loss has been
reported for regions with high atmospheric N deposition, such as the northeastern United
States [Bormann et al., 1968] or Europe [Huber, 2005] as compared to regions with
lower N deposition, such as the western United states [Sollins et al., 1981; Rhoades et al.,
2013]. While forest disturbance impacts for aqueous C fluxes are less well-studied than
for N, streamwater DOC may increase following harvest [Schelker et al., 2012]. Unlike
fire or harvest, mountain pine beetle disturbance does not initially disturb soils or remove
biomass [Adams et al., 2012], so biogeochemical responses may differ.
There is emerging evidence that strong upland biogeochemical response to
mountain pine beetle translates poorly to the watershed scales relevant to water resources
[Rhoades et al., 2013; Mikkelson et al., 2013]. Upland response includes increased
needle litter with higher N concentrations [Griffin and Turner, 2012; Morehouse et al.,
2008] and increased soil concentrations of NO3 and NH4 [Griffin and Turner, 2012;
Morehouse et al., 2008; Clow et al., 2011; Huber, 2005]. Similar patterns were reported
by Tokuchi et al. [2004] following pine wilt disease, which has comparable biophysical
impacts to mountain pine beetle. Two studies of mountain pine beetle response in soil
DOC have shown one increase [Kana et al., 2013] and one decrease [Xiong et al., 2011].
While increased streamwater NO3 has been reported in response to pine wilt in Japan
130
[Tokuchi et al., 2004] and bark beetle in Bavaria [Zimmermann et al., 2000], stream
response to mountain pine beetle in North America has been weak for DOC [Mikkelson et
al., 2012] and NO3 [Clow et al., 2011; Rhoades et al., 2013]. Hypotheses for the lack of
apparent connection between upland and watershed-scale stream responses include
spatial and temporal variability of infestation [Clow et al., 2011; Edburg et al., 2011],
delayed response due to shifts in water fluxes and flowpaths [Mikkelson et al., 2012,
2013], dilution [Zimmermann et al., 2000] and N uptake by surviving vegetation
[Rhoades et al., 2013; Zimmermann et al., 2000]. However, there is a critical lack of
biogeochemical observation along hydrological flowpaths connecting the uplands and
riparian-stream system against which to evaluate these hypotheses.
Here we evaluate biogeochemical solution concentrations at an unimpacted site
and a site with > 50% forest mortality along a putative flowpath [Lohse et al., 2009]
including soil water, hillslope groundwater, riparian groundwater, and nested streams up
to third-order.
Our goals are to: (1) evaluate how biogeochemical impacts vary as
putative flowpath length increases from upland soils to the stream network; and (2)
identify flowpath elements critical to regulation of biogeochemical response (i.e. “hot
spots” [Mclain et al., 2003]) to bark beetle-driven forest mortality. The present work
forms part of a larger project quantifying coupled hydrological and biogeochemical
response to mountain pine beetle, and we evaluate results of the present work in the
context of observed physical hydrology [Biederman et al., 2012; Biederman et al., in
review]. To our knowledge, this work represents the most comprehensive effort to
observe biogeochemical impacts of the North American bark beetle epidemic on aqueous
131
biogeochemistry over such a spatial continuum from the uplands to the riparian-stream
system over scales of 10-1 to >103 m.
2. Study sites
An MPB-impacted site “MPB” and an unimpacted control site “UN” were
identified in 2009 along the eastern edge of the Central Rocky Mountains, a region
severely impacted by mountain pine beetle (Figure 1). Due to extensive tree mortality,
adjacent impacted and unimpacted stands having otherwise similar climate forcing and
biophysical traits were not found. However, biophysical and meteorological observations
from prior and present work showed the MPB and UN sites to be well paired. To isolate
MPB effects, we selected sites with gentle slopes (5 to 8%) and similar elevations of
2750 to 3000 m. Both sites had overstory dominated by lodgepole pine (Pinus contorta)
aged 110-140 years since the previous stand-replacing disturbance [Knight et al., 1985;
http://ameriflux.lbl.gov].
Soils at both sites are sandy loam cryochrepts and show
depositional layering, although soils at MPB are of schist parent material, while UN soils
are granitic (Table 1) [Knight et al., 1985; Scott-Denton et al., 2003]. Both sites contain
onsite precipitation gauges and above-canopy meteorological towers.
The sites are
characterized by long, cold winters with continuous snow cover from October until May
or June and mean annual air temperatures of 1-4 °C (Table 2). These sites were
previously compared in studies evaluating changes to snowpack [Biederman et al., 2012]
and water partitioning [Biederman et al., in review] in response to mountain pine beetle.
UN (Figure 1) is located 50 km NW of Denver, CO within the Niwot Ridge LongTerm Ecological Research (LTER) observatory. MPB activity above background levels
132
was not observed at this site through 2012 (< 5% of trees). The site contains the Niwot
SNOTEL station [www.wcc.nrcs.usda.gov] and Niwot Ridge AmeriFlux tower
[http://ameriflux.lbl.gov]. Numerous prior studies have been conducted at this site [e.g.
Brooks et al., 1998; Monson et al., 2002; Hood et al., 2005].
The MPB site (Figure 1) at Chimney Park, WY is located 125 km NNW of UN in
the Medicine Bow National Forest. From 2007 to 2010, the site experienced overstory
mortality of 50-80% [Biederman et al., in review]. The stream-riparian zone is dominated
by phreatophytes not directly affected by mountain pine beetles. Previous research at
MPB has addressed the effects of forest gaps on snow accumulation [Gary, 1974],
nitrogen cycling [Fahey et al., 1985; Knight et al., 1985], and impacts of simulated bark
beetle infestation on water and nutrient fluxes [Knight et al., 1991].
3. Methods
3.1. Physical Hydrology and Weather
At both sites, observations of soil water and groundwater were nested within
upland hillslopes drained by zero-order streams, defined as the convergent portions of
hillslopes where surface flow channels were observed to initiate (Figure 1). Additionally
at MPB, the upland hillslopes were nested within larger study catchments of first and
second order (Table 2). To investigate spatial patterns in MPB stream chemistry that
emerged from preliminary results, a third-order catchment was added to routine sampling
in 2012.
Volumetric soil water content was observed using Decagon EC-5 and 5-TE
sensors (Decagon Corp.) at depths of 10 cm (+2 cm), 30 cm (+ 5 cm) and 60 cm (+ 10
133
cm) in each of six plots per site (n = 18 per site). Water stages in all streams and one
groundwater piezometer at each site were recorded using Hobo pressure transducers
(Onset Corp.). Periods of missing data at the UN piezometer (~ 5% of all data) were gapfilled using linear regression (r2 = 0.98) to a 7-m deep well located 50 m away.
Observations of meteorology and snowpack were made within 300 m of the
hillslope and zero-order stream observations at each site (Figure 1) as described in
Biederman et al., [2012].
3.2. Biogeochemical Sampling and Analysis
At each site, routine sampling included soil water, hillslope groundwater, riparian
groundwater and streams (Figure 1) lasting from snowmelt onset until the source dried
out or the start of the subsequent winter. Twelve porous cup tension lysimeters (Prenart
Corp.) were located in forested MPB upland plots and another eight were located at UN
at depths of 10 (+2) cm, 30 (+ 5) cm) and 60 (+ 10) cm. Soil water samples were
collected every 1-2 weeks under tension of 70 kPa. Upland groundwater at depths of 110
(+ 20) cm, was sampled from 3-4 piezometers at each site every 1-2 weeks. Riparian
groundwater at depths of 90 cm (+ 20 cm) was sampled from 2-4 piezometers every 1-2
weeks. During snowmelt, stream samples were collected daily by automated sampler at
the outlets of the MPB and UN zero-order catchments of both sites as well as outlets of
the first- and second-order MPB catchments. Additional grab samples were collected
very 1-3 weeks at all catchment outlets and several intermediate stream locations.
Water samples were transported to the lab on ice and filtered within 24 hours
using 0.45-micron glass fiber filters. Filtered aliquots for analysis of dissolved organic
134
carbon (DOC) and total dissolved nitrogen (TDN) were stored at 4 ̊C and analyzed within
2-3 weeks on a TOC/TN analyzer (Shimadzu Corp.) with a method detection limit (DL)
of 0.05 mg l-1 for DOC and 0.05 mg l-1 for TDN. Filtered aliquots for analysis of
dissolved inorganic nitrogen (DIN), NO3 and NH4 were stored at 4 ̊C and analyzed within
1-2 weeks on a SmartChem Discrete Analyzer (Westco Scientific) with DL of 0.005 and
0.002 for NO3-N and NH4-N, respectively. Dissolved organic nitrogen (DON) was
determined by subtraction of DIN from TDN.
3.3. Catchment Spatial Analysis and Statistics
Contributing area, average hillslope slope, stream network, stream order
[Strahler, 1952], and length and average slope of the main channel were determined for
each
MPB
catchment
using
the
TauDEM
toolkit
[http://hydrology.uwrl.usu.edu/taudem/taudem5.0] and a 1-m airborne laser swath map
(ALSM) in ARC-GIS 10.1 software.
Statistical analyses were performed using
MATLAB 2012a (MathWorks Corp.) Standard errors of the means were used to quantify
how well sample means estimated population means, while standard deviations were used
to represent sample variability. To evaluate how concentrations of DON and DOC were
predicted by distance from the channel head, we evaluated exponential models of
concentration as a function of distance. Each solute was additionally evaluated using an
exponential model with a putative nonzero asymptote. DOC was observed to remain
relatively constant between the second-order and third-order streams, so the mean value
from these two streams of 16 mg l-1 DOC was used as the asymptote. DON declined
from the second-order to third-order stream, so the third-order mean value of 0.26 mg l-1
135
was used as the asymptote. Overall significance of channel distance as a predictor of
concentration was determined from the p value of the F statistic versus a constant model.
4. Results
The presentation of results begins with weather and physical hydrologic response
at the UN and MPB sites. Biogeochemical concentrations are then presented along a
putative flowpath beginning with UN, which provides context for the subsequent
evaluation of biogeochemistry at MPB. The putative flowpath begins with shallow soil
water, proceeds through deeper soil water into hillslope groundwater, continues across
riparian groundwater into zero-order streams, and follows the channel network to streams
of progressively higher order.
4.1. Weather and Hydrological Response
The two years of this study were characterized by precipitation that was above
average in 2011 and below average in 2012. For 2011 and 2012, precipitation at UN was
1025 mm and 722 mm, or 128% and 90% of average (onsite SNOTEL 1981-2012). At
MPB, the 2011 and 2012 precipitation totals were 786 mm and 448 mm. No long-term
record exists for MPB, but the Cinnabar Park SNOTEL 20 km distant indicated
precipitation for 2011 and 2012 was 127% and 74% of average (2004-2012). Winter
precipitation and snowpack accumulation were similar at the two sites (Figure 2), but
MPB received less summer precipitation than UN, particularly in 2012 (Figure 2). In
2011 the mean annual temperatures were 2.8 and 1.9 ̊C at MPB and UN, while 2012 was
warmer with mean temperatures of 3.4 and 3.2 C
̊ . In both years the snowmelt began
synchronously at the two sites but proceeded more rapidly at MPB and finished 5 days
136
earlier at MPB than at UN. Below-canopy solar radiation averaged 74% greater during
winter and 300-350% greater during summer at MPB as compared to UN.
Volumetric soil moisture averaged 29% at MPB and 20% at UN in 2011, while in
2012 the means were 27% and 18%. The sites reached similar peak soil moistures of
36% at MPB and 39% at UN each year at the end of snowmelt. Soil moisture decreased
throughout each summer, with the only response to summer rains observed being
increases of 3 – 5% at the 10-cm depth at UN (data not shown).
Stream and groundwater stage rose rapidly during snowmelt each year, with
groundwater reaching the surface for several weeks at each site in the wet year 2011 and
rising to within 10 to 20 cm of the surface in the dry year 2012 (Figure 3). Snowmelt
was the dominant driver of stage in groundwater and streams, consistent with prior
observations of stream water isotopes at both sites [Biederman et al., in review] and of
soil moisture and streamflow dynamics at MPB [Knight et al., 1985].
However,
groundwater and the zero-order stream stage at UN responded to a period of several
weeks with persistent rainfall in July and August 2012 (Figure 2, Figure 3). In the wetter
year 2011, annual runoff coefficients ranged from 11 to 36% at MPB as compared to
40% at UN, while in the drier year 2012, runoff coefficients ranged from 0-19% at MPB
as compared to 14% at UN [Biederman et al., in review].
4.2. Unimpacted Biogeochemistry From Uplands to RSS
At the unimpacted site, mean chemistry in shallow soil water (nominal depth 10
cm) was highly variable, with DOC mean (standard deviation) of 34.4 (25.7) mg l-1 and
DON of 1.32 (0.81) mg l-1, giving a ratio of mean DOC:DON of 26. (Table 3). Deeper
137
soil water (30-60 cm) showed 60-76 % lower concentrations and lower variability than
shallow soil water, with DOC of 13.7 (3.6) mg l-1 and DON of 0.32 (0.07). DIN (both
NO3 and NH4) remained near detection limits in soil water. While mean chemistry of
hillslope groundwater was similar to deeper soil water, riparian groundwater and zeroorder stream water showed progressive reductions in DOC and DON. Ratios of mean
DOC:DON for all locations except shallow soil water were relatively invariant, ranging
from 40 to 48.
Time series showed rising DOC in UN hillslope groundwater as the water table
rose during snowmelt (Figure 3), with peak values of 26 mg l-1 in the wet year 2011 and
20 mg l-1 in the dry year 2012. DOC in UN riparian groundwater and zero-order stream
water was comparatively constant through time. DON was consistently low in UN
groundwater, with only one sample greater than 0.5 mg l-1 (Figure 4). UN zero-order
stream water DON was likewise relatively constant in time, with 90% of samples in the
wet year 2011 falling between 0.2 and 0.4 mg l-1. Zero-order DON was lower in the dry
year 2012, with all samples below 0.1 mg l-1. NO3-N remained below 0.12 mg l-1 in the
UN groundwater and zero-order stream, with most stream samples below detection limit
(Figure 5).
Along the putative hydrological flowpath at UN, the largest change in mean
chemistry occurred between shallow and deep soil water (Figure 6). Unimpacted
groundwater and zero-order stream had similar DOC, DON and DOC:DON ratio to deep
soil water, and no large transitions in dissolved chemistry were observed between the
uplands and the riparian-stream system.
138
4.3. MPB Biogeochemistry From Uplands to RSS
At the MPB site, mean chemistry in shallow soil water was not significantly
different from the UN site in DOC or DON (Table 3), and the ratio of mean DOC:DON
was 42. NO3-N was much larger in shallow soil water at MPB (p< 0.001), with mean
(standard deviation) of 1.21 (2.22) mg l-1, although NH4 was near detection limits, similar
to soil water at UN. In deeper soil water, mean DOC and DON were reduced by 50%
and 60%, respectively, as compared to shallow soil water, similar to the pattern observed
at UN, while NO3 remained similarly high in deep and shallow soil water. Upland
groundwater at MPB showed a further reduction of 36% in mean DOC as compared to
deep soil water, but mean DON and NO3 in upland groundwater were the highest
observed for any landscape unit in the study at 1.65 (1.2) mg l-1 and 3.68 (3.47) mg l-1,
indicating a large accumulation of N in groundwater. MPB riparian groundwater and
zero-order stream water showed mean DOC increases of 62% and 79% as compared to
upland groundwater. Mean DON declined from the hillslope to riparian groundwater and
the zero-order stream, with reductions of 80% and 65%, respectively. Nearly all the NO3
was attenuated upon entering the riparian zone, and both riparian groundwater and
streams at all scales showed NO3 <= 0.05 mg l-1. Both DOC and DON were greater in
the MPB zero-order stream as compared to the UN zero-order stream (p < 0.01). DOC in
the higher-order streams (second- and third-order) was attenuated by 32% to 39% as
compared to the lower-order streams (zero- and first-order), while DON was attenuated
by 42% to 59%.
139
Time series of MPB upland groundwater showed similar DOC dynamics to the
UN site, with rising concentrations as the groundwater table increased and peak values of
24 mg l-1 and 25 mg l-1 in the wet year 2011 and dry year 2012, respectively (Figure 3).
DOC in MPB riparian groundwater was comparatively constant in time but with mean
concentration (21 mg l-1) approaching the upland groundwater peak DOC concentrations.
In the wet year 2011, MPB lower-order streams showed peak DOC near peak stage,
followed by falling concentration with falling stage and then a gradual return to
concentrations near those of riparian groundwater. Lower-order stream DOC was less
dynamic in the dry year 2012, remaining similar to DOC in riparian groundwater.
Higher-order stream DOC was comparatively constant in time through both years.
Time series of DON showed high variability in MPB upland groundwater, with
peak concentrations of 4.6 mg l-1 and 3.5 mg l-1 reached at peak stage in 2011 and 2012,
respectively (Figure 4). While upland groundwater DON fell below 0.5 mg l-1 in 2011 as
groundwater stage declined, 2012 concentrations remained above 0.5 mg l-1 during the
entire period when samples could be obtained.
Riparian groundwater DON was
relatively constant, similar to the UN site, remaining below 0.7 mg l-1. DON is greater
and more variable in lower-order than higher-order streams, similar to DOC, but the
inter-annual differences are larger for stream DON than stream DOC, with lower
concentrations in 2012.
Along the putative flowpath at MPB, mean biogeochemical concentrations
showed different patterns from those at UN (Figure 6). In the uplands, DOC and DON
patterns with soil depth were similar at the two sites, declining sharply from shallow soils
140
to deep soils. DOC increased across the transition from uplands to the riparian-stream
system at MPB, in contrast to UN, where no change was observed. Both NO3 and DON
were elevated in MPB upland groundwater as compared to UN, indicating substantial
groundwater N contamination, but most of this was attenuated across the transition from
uplands to riparian-stream system.
4.4. Spatial Patterns of Chemistry in Nested MPB Streams
Exponential decay regression models of DOC showed distance from the channel
head to be a significant (p< 0.05) predictor of stream DOC across nine sampling locations
at the MPB site ranging from 23 to 4580 meters (Figure 7). A simple exponential model
overestimated DOC over the channel distance 1-3 km and underestimated DOC at larger
distances, while an exponential with putative nonzero asymptote had greater curvature
and more accurately reflected the stable DOC concentration at flow distances larger than
1 km. Distance from channel head was also a significant (p < 0.05) predictor of stream
DON (Figure 8). A simple exponential model with lower curvature than the model with
nonzero asymptote better predicted continued attenuation of DON across the two largest
scales.
5. Discussion
Observations of biogeochemistry along a putative hydrological flowpath [Lohse et
al., 2009] showed expected patterns at an unimpacted site, while a site with severe
mountain pine beetle infestation showed elevated soil water NO3, large increases of DON
and NO3 in upland groundwater, and attenuation of DOC and DON in headwater streams.
Our observations of increased N storage in upland groundwater and attenuation of C and
141
N in headwater streams improve current understanding of the weak response to bark
beetle infestation observed at the watershed scale [Clow et al., 2011; Rhoades et al.,
2013; Mikkelson et al., 2012].
In the following discussion, we evaluate plausible
mechanisms for observed responses in the uplands and riparian-stream system (RSS) and
draw inferences about whether these mechanisms occur gradually along hydrological
flowpaths or locally at biogeochemical hot spots [McLain et al., 2003; Lohse et al.,
2009].
We conclude with discussion of how this work improves understanding of
biogeochemical response to bark beetle infestation and an outline of key priorities for
future study.
An unimpacted site (UN) provided context for expected hydrological and
biogeochemical patterns in a subalpine, lodgepole pine-dominated catchment. Snowmelt
was the dominant hydrological input to soil moisture and streams, which were supplied
primarily by groundwater (i.e. all water moving through the subsurface) [Biederman et
al., in review; Jensco et al., 2009; Frisbee et al., 2011]. Annual evapotranspiration (ET)
averaged 618 mm and varied by only 2% between years, so runoff available for
transporting biogeochemical solutes was determined by precipitation in excess of this
relatively constant ET [Biederman et al., in review]. Biogeochemical concentrations at
UN agreed reasonably with prior subalpine observations of DOC [Hood et al., 2003],
DON, and NO3 [Hood et al., 2003b]. Soil water DOC and DON indicated decreasing
organic matter (OM) with depth (Figure 6). Upland groundwater had similar mean DOC
and DON concentrations to deep soil water with peaks during snowmelt suggesting OM
flushing from shallower soils as groundwater approached the surface (Figure 3, Figure 4)
142
[Hornberger et al., 1994; Boyer et al., 1997]. Stream water DOC and DON reflected
similar concentrations to the relatively stable concentrations of riparian groundwater,
with small snowmelt peaks suggesting activation of OM-rich shallow flowpaths (Figure
3, Figure 4) [Bishop et al., 2004; Jencso et al., 2009; Lyon et al., 2011]. Low DIN at UN
is consistent with fierce competition for N among plants and microbes in N-limited
lodgepole pine ecosystems of western North America ([Fahey et al., 1985] as well as
limited NH4 mobility.
MPB received similar meteorological forcing to UN (Figure 2) and produced
similar physical hydrological response in contrast to expected increases in snowpack and
runoff outlined in recent reviews [Edburg et al., 2012; Pugh and Gordon, 2013;
Mikkelson et al., 2013]. Mountain pine beetle infestation did not affect peak snowpack
accumulation [Biederman et al., 2012], and ET was not reduced despite overstory
mortality [Biederman at al., in review], meaning that runoff water for transport of
biogeochemical solutes did not increase as expected [Mikkelson et al., 2013].
In the uplands, MPB soil water DOC and DON showed similar concentrations
and reduction with depth as at UN (Figure 6) suggesting that increased litter inputs
[Borhkhu et al., in review; Morehouse et al., 2008; Griffin et al., 2011] did not greatly
increase soluble OM in soils. While our observations of 48% greater mean DOC in deep
soil water at MPB than UN (Table 3, Figure 6) contribute to very limited observations of
soil water DOC following bark beetle infestation [Mikkelson et al., 2013], the soil water
DOC difference did not appear to have affected upland groundwater (Figure 6). The
greatest impacts observed at MPB were elevated NO3 in soil water, consistent with prior
143
studies [Huber, 2005; Morehouse et al., 2008; Griffin et al., 2011; Clow et al., 2011] and
elevated DON and NO3 in upland groundwater (Table 3; Figure 6). Research at MPB
pre-dating the infestation showed only trace N [Fahey et al., 1985], so it is likely that
bark beetle infestation caused groundwater N accumulation as has been observed in more
humid climates [Zimmermann et al., 2000; Hobara et al., 2001]. When viewed along a
putative vertical hydrological flowpath through surface soils, elevated N over a relatively
small scale (~ 1 m), classifying this as a hot spot impact. Increased production of mobile
NO3 was likely stimulated by higher litter N inputs [Morehouse et al., 2008; Griffin et al.,
2011] and more rapid mineralization and nitrification [Hobara et al., 2001;] possibly
related to higher soil moisture and incoming solar radiation favoring N mineralization
and nitrification [Griffin et al. 2011; Zimmermann et al., 2000]. Increased NO3 leaching
to groundwater was probably facilitated both by temporal asynchrony between periods of
high soil water NO3 during snowmelt and uptake by short (i.e. snow covered)
regenerating vegetation [Meixner et al., 2001] and by the mortality of more deeply rooted
overstory trees [Romme et al., 1986]. Brooks et al. [1998] showed that microbial uptake
of NO3 during snowmelt could store N for later use by plants, dampening NO3 flushing to
streams, but this mechanism did not appear to prevent significant accumulation in
groundwater. While upland piezometers were arrayed at variable distance from the
stream-riparian zone (Figure 1), no compelling spatial patterns in chemistry were
detected along the putative flowpath, suggesting upland groundwater behaved as a
relatively well-mixed biogeochemical reservoir.
144
At the transition from the MPB uplands to the RSS, removal of nearly 100% of
NO3 and the majority of DON (Table 3, Figure 6) classified this transition as a
biogeochemical hot spot, consistent with extensive examples in the literature [e.g. Vidon
et al., 2010; Dahm et al., 1998; Mulholland and Hill, 1997]. Such hot spots occur due to
the mixing of limiting reactants, adjacency of different redox zones, and prevalence of
plants and microorganisms [Lohse et al., 2009]. Possible mechanisms of NO3 removal
include denitrification [Oelsner et al., 2009; Brooks and Lemmon, 2007; Meixner et al.,
2007; Peters et al., 2011] and uptake by phreatophytes [Cirmo and McDonnell, 1997;
Dosskey et al., 2010]. Riparian groundwater DOC concentration in excess of 20 mg C l-1
(Figure 3) is consistent with adequate substrate for denitrification [Baker and Vervier,
2004]. In a parallel study at MPB, Borkhu et al. [in review] observed higher fluxes of
N2O from soils in wetter landscape positions, suggesting significant denitrification.
While we did not quantify N stores or fluxes in riparian vegetation, it is likely that
increased concentrations in water delivered from the uplands also resulted in greater plant
uptake. An increase in mean DOC between MPB upland groundwater and the RSS
(Figure 3, Figure 6) suggests the riparian zone as a source of DOC production [Bishop et
al., 2004] which could support both greater microbial immobilization and the
development of reduced redox conditions facilitating denitrification [Lohse et al., 2009]
The attenuation of DOC and DON with increasing downstream distance improves
our understanding of the responses to bark beetle infestation observed at the watershed
scale. While the zero-order stream departed from the overall pattern of higher
concentrations at lower flow lengths, steeper terrain (Table 2) may have reduced the
145
residence time of upland groundwater in the riparian zone and the resulting addition of
DOM to water entering the stream. While DOC was attenuated by an average of 35%
between the first-order (channel distance 600 m) and second-order streams (channel
distance 3200 m), and an additional 11% at the third-order stream (Figure 7), and the
residual concentration of 16 mg l-1 was well above stream DOC for the UN site (Table 3)
and similar unimpacted subalpine streams [Hood et al., 2003]. Since the observed DOC
response at larger watershed scales varies from weak [Mikkelson et al., 2012] to none
[Clow et al., 2011], it appears likely that some combination of further downstream
processing and dilution has muted the watershed-scale DOC response. In contrast, DON
declined with channel distance over the range of observations and was reduced by 65%
between the first-order and third-order streams, reaching the same concentration as the
UN stream after ~ 5 km (Table 3, Figure 8). It therefore appears that a lack of N response
to bark beetle infestation at the watershed scale [Clow et al., 2011; Rhoades et al., 2013]
may be partly due to processing in headwater streams.
While biogeochemical inputs related to bark beetle infestation could have varied
among catchments at MPB, the relatively uniform and synchronous mortality across the
study catchments (Table 2) suggests that mixing of different inputs does not likely
explain declining DOC and DON with channel distance.
Rapid attenuation of C and N
in headwater streams can occur by both physical and biological immobilization [Covino
et al., 2010; Bernal et al., 2012] and is facilitated by shallow, turbulent flow and high
ratio of channel bed width to cross-sectional area, which foster movement of water
between the stream and hyporheic sediments [Peterson et al., 2001; Wollheim et al.,
146
2001].
Dissolved C and N may be permanently removed by respiration and
denitrification or processed into other forms. These various removal mechanisms may be
viewed as occurring either continuously along flowpaths [Meixner et al., 2001; Berhnardt
et al., 2003] or in a sequence of biogeochemical hot spots in the stream-hyporheic system
created by spatial and temporal variability in physical hydrology [Brooks and Lemmon.,
2007].
Although the results of this study are consistent with the weak response to North
American bark beetles in watershed-scale biogeochemistry, they contrast with
expectations of greater impacts, particularly for N, from other disturbance studies. One
likely explanation is that bark beetle disturbance has mainly occurred in the western part
of the continent, where lower atmospheric deposition rates and greater N limitation often
result in lower export following forest disturbance [e.g. Sollins et al., 1981; Fahey et al.,
1985; Mast and Clow, 2008] as compared to disturbances from regions of higher N
deposition in eastern North America [Vidon et al., 2010; Bernhardt et al., 2003] or
Europe [Zimmermann et al., 2000; Huber, 2005]. An important corollary is that loss of N
from soils to upland groundwater could put this limiting nutrient beyond the reach of
regenerating vegetation, possibly impacting forest succession. A second factor limiting
N loss is the semi-arid nature of much of the interior mountains affected by North
American bark beetles. While N mineralization and nitrification have been observed in
this and other studies, low annual runoff ranging from zero to a few hundred mm in this
region [e.g. Flerchinger and Cooley, 2000] and low water flux may limit N loss [Vitousek
147
et al., 1979], consistent wtih lower stream DON in the dry year 2012 as compared to
2011 (Figure 4).
6. Implications and Conclusions
The complete removal of NO3 and significant addition of DOC by the MPB
riparian zone illustrates the importance of riparian hot spots. Efforts to evaluate or model
biogeochemical response to forest disturbance should occur in the context of a system
including both upland (catchment) and riparian-stream system (RSS) components (Figure
9) [Seidl et al., 2011], and efforts to manage disturbed forest should preserve riparian
zones. Site differences in riparian zone vegetation or organic matter production could
cause different biogeochemical responses, and future work should include similar
observations to this study at sites with contrasting riparian environments. Hydrological
differences would also be expected to alter outcomes if they affected residence times or
flowpaths through the riparian zone [Lohse et al., 2009; Peters et al., 2011] including
upland topography and soils and the amount, timing and partitioning of precipitation
[Stednick et al., 1996; Biederman et al., in review].
Since MPB upland groundwater NO3 concentrations approached USEPA drinking
water limits, this raises concern for how widespread bark beetle infestation may impact
regional groundwater quality. Recharge to deeper aquifers is largely unquantified, but it
is reasonable to additional N inputs will be attenuated within several years to decades as
litter is processed [Borhkuu et al., in review] and vegetation takes up N more rapidly and
from deeper rooting soil layers [Romme et al., 1986].
Future breakthrough of N from
148
upland groundwater to streams could occur [Hobara et al., 2001], raising concerns for
increased aquatic productivity.
Prior to this study, the weak stream response to bark beetle infestation has been
attributed mainly to variability of impacts in time and space (i.e. dilution) [Zimmermann
et al., 2000; Clow et al., 2011; Mikkelson et al., 2012] and compensatory uptake by
surviving upland vegetation [Rhoades et al., 2013].
Significantly, this study illustrates
that biogeochemical mechanisms at riparian hot spots and along flow paths in soils and
streams can collectively attenuate some or all of the dissolved fluxes within small
headwater catchments (Figure 9).
7. Acknowledgements
This work was supported by funding from the National Science Foundation EAR0910831, United States Geological Service, Wyoming Agricultural Research Station
through McIntire-Stennis and the Wyoming Water Development Commission. Science
Foundation Arizona and the Arizona Water Sustainability Program provided additional
support. LiDAR mapping of the MPB site was supported by an NCALM Seed grant. We
thank B. Bright and J. Hicke of the University of Idaho for sharing QuickBird land cover
classification of the MPB site. The National Science Foundation Niwot Ridge Long-Term
Ecological Research project and the University of Colorado Mountain Research Station
provided logistical support and data. M. Williams and M. Zeliff provided supplementary
water table data for the UN site. The National Center for Atmospheric Research is
supported through a cooperative agreement from the National Science Foundation. The
authors thank S. Papuga for useful comments.
149
8. References
Adams, H. D., Luce, C. H., Breshears, D. D., Allen, C. D., Weiler, M., Hale, V. C.,
Huxman, T. E. (2012). Ecohydrological consequences of drought- and infestationtriggered tree die-off: insights and hypotheses. Ecohydrology, 5(2), 145-159. doi:
10.1002/eco.233
Ayres, M. P., & Lombardero, M. J. (2000). Assessing the consequences of global change
for forest disturbance from herbivores and pathogens. Science of the Total
Environment, 262(3), 263-286. doi: 10.1016/s0048-9697(00)00528-3
Baker, M. A., & Vervier, P. (2004). Hydrological variability, organic matter supply and
denitrification in the Garonne River ecosystem. Freshwater Biology, 49(2), 181190. doi: 10.1046/j.1365-2426.2003.01175.x
Bernal, S., von Schiller, D., Marti, E., & Sabater, F. (2012). In-stream net uptake
regulates inorganic nitrogen export from catchments under base flow conditions.
Journal
of
Geophysical
Research-Biogeosciences,
117.
doi:
10.1029/2012jg001985
Bernhardt, E. S., Likens, G. E., Buso, D. C., & Driscoll, C. T. (2003). In-stream uptake
dampens effects of major forest disturbance on watershed nitrogen export.
Proceedings of the National Academy of Sciences of the United States of America,
100(18), 10304-10308. doi: 10.1073/pnas.1233676100
Bernhardt, E. S., Likens, G. E., Hall, R. O., Buso, D. C., Fisher, S. G., Burton, T. M.,
Lowe, W. H. (2005). Can't see the forest for the stream? - In-stream processing
and terrestrial nitrogen exports. Bioscience, 55(3), 219-230. doi: 10.1641/00063568(2005)055[0219:acstff]2.0.co;2
Biederman, J. A., Brooks, P. D., Harpold, A. A., Gutmann, E., Gochis, D. J., Reed, D. E.,
& Pendall, E. (2012). Multi-scale Observations of Snow Accumulation and Peak
Snowpack Following Widespread, Insect-induced Lodgepole Pine Mortality.
Ecohydrology.
Bishop, K., Seibert, J., Koher, S., & Laudon, H. (2004). Resolving the Double Paradox of
rapidly mobilized old water with highly variable responses in runoff chemistry.
Hydrological Processes, 18(1), 185-189. doi: 10.1002/hyp.5209
Bormann, F. H., & Likens, G. E. (1979). Catastrophic disturbance and the steady state in
northern hardwood forests. American Scientist, 67(6), 660-669.
Bormann, F. H., Likens, G. E., Fisher, D. W., & Pierce, R. S. (1968). Nutrient loss
accelerated by clear-cutting of a forest ecosystem. Science, 159(3817), 882-&.
doi: 10.1126/science.159.3817.882
Boyer, E. W., Hornberger, G. M., Bencala, K. E., & McKnight, D. M. (1997). Response
characteristics of DOC flushing in an alpine catchment. Hydrological Processes,
11(12), 1635-1647.
150
Breshears, D. D., Myers, O. B., Meyer, C. W., Barnes, F. J., Zou, C. B., Allen, C. D.,
Pockman, W. T. (2009). Tree die-off in response to global change-type drought:
mortality insights from a decade of plant water potential measurements. Frontiers
in Ecology and the Environment, 7(4), 185-189. doi: 10.1890/080016
Brooks, P. D., & Lemon, M. M. (2007). Spatial variability in dissolved organic matter
and inorganic nitrogen concentrations in a semiarid stream, San Pedro River,
Arizona. Journal of Geophysical Research-Biogeosciences, 112(G3). doi:
10.1029/2006jg000262
Brooks, P. D., Williams, M. W., & Schmidt, S. K. (1998). Inorganic nitrogen and
microbial biomass dynamics before and during spring snowmelt.
Biogeochemistry, 43(1), 1-15. doi: 10.1023/a:1005947511910
Brown, A. E., Zhang, L., McMahon, T. A., Western, A. W., & Vertessy, R. A. (2005). A
review of paired catchment studies for determining changes in water yield
resulting from alterations in vegetation. Journal of Hydrology, 310(1-4), 28-61.
doi: 10.1016/j.jhydrol.2004.12.010
Cirmo, C. P., & McDonnell, J. J. (1997). Linking the hydrologic and biogeochemical
controls of nitrogen transport in near-stream zones of temperate-forested
catchments: a review. Journal of Hydrology, 199(1-2), 88-120. doi:
10.1016/s0022-1694(96)03286-6
Clow, D. W., Rhoades, C., Briggs, J., Caldwell, M., & Lewis, W. M., Jr. (2011).
Responses of soil and water chemistry to mountain pine beetle induced tree
mortality in Grand County, Colorado, USA. Applied Geochemistry, 26, S174S178. doi: 10.1016/j.apgeochem.2011.03.096
Covino, T., McGlynn, B., & Baker, M. (2010). Separating physical and biological
nutrient retention and quantifying uptake kinetics from ambient to saturation in
successive mountain stream reaches. Journal of Geophysical ResearchBiogeosciences, 115. doi: 10.1029/2009jg001263
Dahm, C. N., Grimm, N. B., Marmonier, P., Valett, H. M., & Vervier, P. (1998). Nutrient
dynamics at the interface between surface waters and groundwaters. Freshwater
Biology, 40(3), 427-451. doi: 10.1046/j.1365-2427.1998.00367.x
Dosskey, M. G., Vidon, P., Gurwick, N. P., Allan, C. J., Duval, T. P., & Lowrance, R.
(2010). The Role of Riparian Vegetation in Protecting and Improving Chemical
Water Quality in Streams1. Journal of the American Water Resources
Association, 46(2), 261-277. doi: 10.1111/j.1752-1688.2010.00419.x
Edburg, S. L., Hicke, J. A., Brooks, P. D., Pendall, E. G., Ewers, B. E., Norton, U.,
Meddens, A. J. H. (2012). Cascading impacts of bark beetle-caused tree mortality
on coupled biogeophysical and biogeochemical processes. Frontiers in Ecology
and the Environment, 10(8), 416-424. doi: 10.1890/110173
151
Edburg, S. L., Hicke, J. A., Lawrence, D. M., & Thornton, P. E. (2011). Simulating
coupled carbon and nitrogen dynamics following mountain pine beetle outbreaks
in the western United States. Journal of Geophysical Research-Biogeosciences,
116. doi: 10.1029/2011jg001786
Fahey, T. J., Yavitt, J. B., Pearson, J. A., & Knight, D. H. (1985). The nitrogen cycle in
lodgepole pine forests, southeastern Wyoming. Biogeochemistry, 1(3), 257-275.
Flerchinger, G. N., & Cooley, K. R. (2000). A ten-year water balance of a mountainous
semi-arid watershed. Journal of Hydrology, 237(1-2), 86-99. doi: 10.1016/s00221694(00)00299-7
Frisbee, M. D., Phillips, F. M., Campbell, A. R., Liu, F. J., & Sanchez, S. A. (2011).
Streamflow generation in a large, alpine watershed in the southern Rocky
Mountains of Colorado: Is streamflow generation simply the aggregation of
hillslope runoff responses? Water Resources Research, 47. doi:
10.1029/2010wr009391
Galloway, J. N., Aber, J. D., Erisman, J. W., Seitzinger, S. P., Howarth, R. W., Cowling,
E. B., & Cosby, B. J. (2003). The nitrogen cascade. Bioscience, 53(4), 341-356.
doi: 10.1641/0006-3568(2003)053[0341:tnc]2.0.co;2
Gary, H. L. (1974). Snow accumulation and snowmelt as influenced by a small clearing
in a lodgepole pine forest. Water Resources Research, 10(2), 348-353. doi:
10.1029/WR010i002p00348
Griffin, J. M., & Turner, M. G. (2012). Changes to the N cycle following bark beetle
outbreaks in two contrasting conifer forest types. Oecologia, 170(2), 551-565.
doi: 10.1007/s00442-012-2323-y
Griffin, J. M., Turner, M. G., & Simard, M. (2011). Nitrogen cycling following mountain
pine beetle disturbance in lodgepole pine forests of Greater Yellowstone. Forest
Ecology and Management, 261(6), 1077-1089. doi: 10.1016/j.foreco.2010.12.031
Hobara, S., Tokuchi, N., Ohte, N., Koba, K., Katsuyama, M., Kim, S. J., & Nakanishi, A.
(2001). Mechanism of nitrate loss from a forested catchment following a smallscale, natural disturbance. Canadian Journal of Forest Research-Revue
Canadienne De Recherche Forestiere, 31(8), 1326-1335. doi: 10.1139/cjfr-31-81326
Hood, E., McKnight, D. M., & Williams, M. W. (2003). Sources and chemical character
of dissolved organic carbon across an alpine/subalpine ecotone, Green Lakes
Valley, Colorado Front Range, United States. Water Resources Research, 39(7).
doi: 10.1029/2002wr001738
Hood, E., Williams, M. W., & Caine, N. (2003). Landscape controls on organic and
inorganic nitrogen leaching across an alpine/subalpine ecotone, Green Lakes
Valley, Colorado Front Range. Ecosystems, 6(1), 31-45. doi: 10.1007/s10021002-0175-8
152
Hood, E., Williams, M. W., & McKnight, D. M. (2005). Sources of dissolved organic
matter (DOM) in a Rocky Mountain stream using chemical fractionation and
stable isotopes. Biogeochemistry, 74(2), 231-255. doi: 10.1007/s10533-004-43225
Hornberger, G. M., Bencala, K. E., & McKnight, D. M. (1994). Hydrological controls on
dissolved organic carbon during snowmelt in the Snake River near Montezuma,
Colorado. Biogeochemistry, 25(3), 147-165. doi: 10.1007/bf00024390
Houlton, B. Z., Driscoll, C. T., Fahey, T. J., Likens, G. E., Groffman, P. M., Bernhardt,
E. S., & Buso, D. C. (2003). Nitrogen dynamics in ice storm-damaged forest
ecosystems: Implications for nitrogen limitation theory. Ecosystems, 6(5), 431443. doi: 10.1007/s10021-002-0198-1
Hubbard, R. M., Rhoades, C. C., Elder, K., & Negron, J. (2013). Changes in transpiration
and foliage growth in lodgepole pine trees following mountain pine beetle attack
and mechanical girdling. Forest Ecology and Management, 289, 312-317. doi:
10.1016/j.foreco.2012.09.028
Huber, C. (2005). Long lasting nitrate leaching after bark beetle attack in the highlands of
the Bavarian Forest National Park. Journal of Environmental Quality, 34(5),
1772-1779. doi: 10.2134/jeq2004.0210
Jencso, K. G., McGlynn, B. L., Gooseff, M. N., Wondzell, S. M., Bencala, K. E., &
Marshall, L. A. (2009). Hydrologic connectivity between landscapes and streams:
Transferring reach-and plot-scale understanding to the catchment scale. Water
Resources Research, 45. doi: 10.1029/2008wr007225
Jones, J. A., & Post, D. A. (2004). Seasonal and successional streamflow response to
forest cutting and regrowth in the northwest and eastern United States. Water
Resources Research, 40(5), 19. doi: 10.1029/2003wr002952
Kana, J., Tahovska, K., & Kopacek, J. (2013). Response of soil chemistry to forest
dieback after bark beetle infestation. Biogeochemistry, 113(1-3), 369-383. doi:
10.1007/s10533-012-9765-5
Knight, D. H., Fahey, T. J., & Running, S. W. (1985). Water and nutrient outflow from
contrasting lodgepole pine forests in Wyoming. Ecological Monographs, 55(1),
29-48.
Knight, D. H., Yavitt, J. B., & Joyce, G. D. (1991). Water and nitrogen outflow from
lodgepole pine forest after 2 levels of tree mortality. Forest Ecology and
Management, 46(3-4), 215-225. doi: 10.1016/0378-1127(91)90233-l
Lohse, K. A., Brooks, P. D., McIntosh, J. C., Meixner, T., & Huxman, T. E. (2009).
Interactions Between Biogeochemistry and Hydrologic Systems. Annual Review
of
Environment
and
Resources,
34,
65-96.
doi:
10.1146/annurev.environ.33.031207.111141
153
Lyon, S. W., Grabs, T., Laudon, H., Bishop, K. H., & Seibert, J. (2011). Variability of
groundwater levels and total organic carbon in the riparian zone of a boreal
catchment. Journal of Geophysical Research-Biogeosciences, 116. doi:
10.1029/2010jg001452
Mast, M. A., & Clow, D. W. (2008). Effects of 2003 wildfires on stream chemistry in
Glacier National Park, Montana. Hydrological Processes, 22(26), 5013-5023. doi:
10.1002/hyp.7121
McClain, M. E., Boyer, E. W., Dent, C. L., Gergel, S. E., Grimm, N. B., Groffman, P.
M., Pinay, G. (2003). Biogeochemical hot spots and hot moments at the interface
of terrestrial and aquatic ecosystems. Ecosystems, 6(4), 301-312. doi:
10.1007/s10021-003-0161-9
Meddens, A. J. H., Hicke, J. A., & Ferguson, C. A. (2012). Spatiotemporal patterns of
observed bark beetle-caused tree mortality in British Columbia and the western
United States. Ecological Applications, 22(7), 1876-1891.
Meixner, T., Fenn, M., & Poth, M. (2001). Nitrate in polluted mountainous catchments
with Mediterranean Climates. The Scientific World, 1 (S2), 564-571. doi:
10.1100/tsw.2001.324
Meixner, T., Huth, A. K., Brooks, P. D., Conklin, M. H., Grimm, N. B., Bales, R. C.,
Petti, J. R. (2007). Influence of shifting flow paths on nitrogen concentrations
during monsoon floods, San Pedro River, Arizona. Journal of Geophysical
Research-Biogeosciences, 112(G3). doi: 10.1029/2006jg000266
Mikkelson, K., Bearup, L. A., Maxwell, R. M., Stednick, J. D., McCray, J. E., & Sharp, J.
O. (2013). Bark beetle infestation impacts on nutrient cycling, water quality and
interdependent hydrological effects. Biogeochemistry, 115, 1-21. doi: DOI
10.1007/s10533-013-9875-8
Mikkelson, K., Dickenson, E., Maxwell, R., McCray, J., & Sharp, J. (2012). Waterquality impacts from climate-induced forest die-off Nature- Climate change. doi:
DOI: 10.1038/NCLIMATE1724
Mitton, J. B., & Ferrenberg, S. M. (2012). Mountain Pine Beetle Develops an
Unprecedented Summer Generation in Response to Climate Warming. American
Naturalist, 179(5), E163-E171. doi: 10.1086/665007
Monson, R. K., Turnipseed, A. A., Sparks, J. P., Harley, P. C., Scott-Denton, L. E.,
Sparks, K., & Huxman, T. E. (2002). Carbon sequestration in a high-elevation,
subalpine forest. Global Change Biology, 8(5), 459-478. doi: 10.1046/j.13652486.2002.00480.x
Morehouse, K., Johns, T., Kaye, J., & Kaye, A. (2008). Carbon and nitrogen cycling
immediately following bark beetle outbreaks in southwestern ponderosa pine
forests. Forest Ecology and Management, 255(7), 2698-2708. doi:
10.1016/j.foreco.2008.01.050
154
Mulholland, P. J., & Hill, W. R. (1997). Seasonal patterns in streamwater nutrient and
dissolved organic carbon concentrations: Separating catchment flow path and instream effects. Water Resources Research, 33(6), 1297-1306. doi:
10.1029/97wr00490
Oelsner, G. P., Brooks, P. D., & Hogan, J. F. (2007). Nitrogen sources and sinks within
the middle Rio Grande, New Mexico. Journal of the American Water Resources
Association, 43(4), 850-863. doi: 10.1111/j.1752-1688.2007.00071.x
Peters, N., Bohlke, J., Brooks, P., Burt, T., Gooseff, M., Hamilton, D., Turner, J. (2011).
Hydrology and Biogeochemistry Linkages. In P. Wilderer (Ed.), Treatise on
Water Science (Vol. 2, pp. 271-304): Elsevier.
Peterson, B. J., Wollheim, W. M., Mulholland, P. J., Webster, J. R., Meyer, J. L., Tank, J.
L., Morrall, D. D. (2001). Control of nitrogen export from watersheds by
headwater streams. Science, 292(5514), 86-90. doi: 10.1126/science.1056874
Pugh, E., & Gordon, E. (2013). A conceptual model of water yield effects from beetleinduced tree death in snow-dominated lodgepole pine forests. Hydrological
Processes, 27, 2048-2060. doi: 10.1002/hyp.9312
Rhoades, C. C., McCutchan, J. H., Jr., Cooper, L. A., Clow, D., Detmer, T. M., Briggs, J.
S., Lewis, W. M., Jr. (2013). Biogeochemistry of beetle-killed forests: Explaining
a weak nitrate response. Proceedings of the National Academy of Sciences of the
United States of America, 110(5), 1756-1760. doi: 10.1073/pnas.1221029110
Romme, W. H., Knight, D. H., & Yavitt, J. B. (1986). Mountain pine beetle outbreaks in
the Rocky Mountains - regulators of primary productivity. American Naturalist,
127(4), 484-494.
Schelker, J., Eklof, K., Bishop, K., & Laudon, H. (2012). Effects of forestry operations
on dissolved organic carbon concentrations and export in boreal first-order
streams. Journal of Geophysical Research-Biogeosciences, 117. doi:
10.1029/2011jg001827
Scott-Denton, L. E., Sparks, K. L., & Monson, R. K. (2003). Spatial and temporal
controls of soil respiration rate in a high-elevation, subalpine forest. Soil Biology
& Biochemistry, 35(4), 525-534. doi: 10.1016/s0038-0717(03)00007-5
Seidl, R., Fernandes, P. M., Fonseca, T. F., Gillet, F., Jonsson, A. M., Merganicova, K.,
Mohren, F. (2011). Modelling natural disturbances in forest ecosystems: a review.
Ecological Modelling, 222(4), 903-924. doi: 10.1016/j.ecolmodel.2010.09.040
Sollins, P., & McCorison, F. M. (1981). Nitrogen and carbon solution chemistry of an old
growth coniferous forest watershed before and after cutting. Water Resources
Research, 17(5), 1409-1418. doi: 10.1029/WR017i005p01409
Stednick, J. D. (1996). Monitoring the effects of timber harvest on annual water yield.
Journal of Hydrology, 176(1-4), 79-95. doi: 10.1016/0022-1694(95)02780-7
155
Strahler, A. N. (1952). Hypsometric (area-altitude) analysis of erosional topography.
Geological Society of America Bulletin, 63(11), 1117-&. doi: 10.1130/00167606(1952)63[1117:haaoet]2.0.co;2
Swank, W. T., Waide, J. B., Crossley, D. A., & Todd, R. L. (1981). Insect defoliation
enhances nitrate export from forest ecosystems. Oecologia, 51(3), 297-299. doi:
10.1007/bf00540897
Tokuchi, N., Ohte, N., Hobara, S., Kim, S. J., & Masanori, K. (2004). Changes in
biogeochemical cycling following forest defoliation by pine wilt disease in Kiryu
experimental catchment in Japan. Hydrological Processes, 18(14), 2727-2736.
doi: 10.1002/hyp.5578
Vidon, P., Allan, C., Burns, D., Duval, T. P., Gurwick, N., Inamdar, S., Sebestyen, S.
(2010). Hot Spots and Hot Moments in Riparian Zones: Potential for Improved
Water Quality Management. Journal of the American Water Resources
Association, 46(2), 278-298. doi: 10.1111/j.1752-1688.2010.00420.x
Vitousek, P. M., Gosz, J. R., Grier, C. C., Melillo, J. M., Reiners, W. A., & Todd, R. L.
(1979). NITRATE LOSSES FROM DISTURBED ECOSYSTEMS. Science,
204(4392), 469-474. doi: 10.1126/science.204.4392.469
Westerling, A. L., Hidalgo, H. G., Cayan, D. R., & Swetnam, T. W. (2006). Warming
and earlier spring increase western US forest wildfire activity. Science,
313(5789), 940-943. doi: 10.1126/science.1128834
Williams, A. P., Allen, C. D., Macalady, A. K., Griffin, D., Woodhouse, C. A., Meko, D.
M., McDowell, N. G. (2013). Temperature as a potent driver of regional forest
drought stress and tree mortality. Nature Climate Change, 3(3), 292-297. doi:
10.1038/nclimate1693
Williams, A. P., Allen, C. D., Millar, C. I., Swetnam, T. W., Michaelsen, J., Still, C. J., &
Leavitt, S. W. (2010). Forest responses to increasing aridity and warmth in the
southwestern United States. Proceedings of the National Academy of Sciences of
the
United
States
of
America,
107(50),
21289-21294.
doi:
10.1073/pnas.0914211107
Wollheim, W. M., Peterson, B. J., Deegan, L. A., Hobbie, J. E., Hooker, B., Bowden, W.
B., Hershey, A. E. (2001). Influence of stream size on ammonium and suspended
particulate nitrogen processing. Limnology and Oceanography, 46(1), 1-13.
Wulder, M. A., Dymond, C. C., White, J. C., Leckie, D. G., & Carroll, A. L. (2006).
Surveying mountain pine beetle damage of forests: A review of remote sensing
opportunities. Forest Ecology and Management, 221(1-3), 27-41. doi:
10.1016/j.foreco.2005.09.021
Xiong, Y. M., D'Atri, J. J., Fu, S. L., Xia, H. P., & Seastedt, T. R. (2011). Rapid soil
organic matter loss from forest dieback in a subalpine coniferous ecosystem. Soil
Biology & Biochemistry, 43(12), 2450-2456. doi: 10.1016/j.soilbio.2011.08.013
156
Zimmermann, L., Moritz, K., Kennel, M., & Bittersohl, J. (2000). Influence of bark
beetle infestation on water quantity and quality in the Grosse Ohe catchment
(Bavarian forest national park). Silva Gabreta, 4, 51-62.
157
9. Tables.
Table 1. Physical description and mean forest stand characteristics of the Unimpacted
(UN) and MPB study sites.a
Site
Stand
Age
(yr)
Stem
Density
(per ha)
Tree
Ht.
(m)
DBH
(cm)
Soils
UN
mortality < 5%
3000
40° 02′
-105° 33′
100
3900
11.4
12.1
Cryochrepts
Sandy Loam
c
MPB
mortality 52-77%
2750
41° 04′
-106° 07′
90-110
2500
11.1
14.0
Cryochrepts
Sandy Loam
b
a
Elevation Latitude
(m)
Longitude
The sites had similar elevation, soil texture, stand age and average tree sizes in lodgepole
pine-dominated overstory. The Unimpacted site had greater stem density and lower
mean diameter at breast height (DBH). bScott-Denton et al., 2003; cKnight et al., 1985.
158
Table 2. Characteristics of the MPB study catchments.
Main
Channel
Length
(m)
Main
Channel
Avg. Slope
(%)
Avg. Hillslope
Slope (stdev)
(%)
MPB
Catchment
Area
(ha)
Overstory
Mortality
(%)
Zero-Order
6
50
96
9.4
7.6 (5.5)
First-Order
15
62
725
1.3
5.3 (4.3)
Second-Order
136
61
3257
1.8
7.9 (6.5)
Third-Order
725
52
4581
2.5
8.3 (6.5)
159
Table 3. Solute concentrations by landscape position at the Unimpacted (UN) and MPB
sites.
Landscape Position
DOC
(mg C l-1)
DON
(mg N l-1)
NO3-N
(mg N l-1)
NH4-N
(mg N l-1)
---------- UN ---------Soil Water 10 cm
34.4 (25.7)
1.32 (0.81)
BDL
0.03 (0.01)
Soil Water 30-60 cm
13.7 (3.6)
0.32 (0.07)
0.02 (0.02)
ND
Upland Groundwater
15.6 (5.4)
0.37 (0.14)
0.03 (0.03)
0.05 (0.02)
Riparian Groundwater
12.6 (9.6)
0.26 (0.14)
0.01 (0.01)
0.10 (0.14)
Zero-Order Stream
9.4 (1.8)
0.23 (0.10)
0.01 (0.01)
0.15 (0.17)
---------- MPB ----------
a
Soil Water 10 cm
40.5 (15.3)
0.97 (0.34)
1.21 (2.22)
0.05 (0.02)
Soil Water 30-60 cm
20.4 (12.9)
0.39 (0.22)
0.94 (1.60)
ND
Upland Groundwater
13.1 (5.9)
1.65 (1.2)
3.68 (3.47)
0.09 (0.04)
Riparian Groundwater
21.2 (10.8)
0.33 (0.21)
0.05 (0.06)
0.09 (0.12)
Zero-Order Stream
23.5 (4.7)
0.57 (0.24)
0.01 (0.01)
0.15 (0.16)
First-Order Stream
26.4 (4.0)
0.63 (0.31)
0.02 (0.04)
0.09 (0.14)
Second-Order Stream
16.0 (2.9)
0.33 (0.10)
0.03 (0.10)
0.10 (0.14)
Third-Order Stream
16.9 (2.2)
0.26 (0.13)
0.02 (0.01)
ND
Shown are mean (standard deviation). Soil water showed much higher NO3 at the MPB
site. DON and NO3 were elevated in MPB upland groundwater but mostly attenuated in
riparian groundwater. Both DON and DOC decreased from the MPB zero-order stream
to larger-scale streams. BDL = Below NO3-N detection limit of 0.005 mg l-1. ND = no
data.
160
10. Figures
Figure 1. Site maps (upper panels) and detail maps (lower panels). Both sites include
precipitation gauges and meteorological towers. At each site, solution chemistry was
sampled from tension lysimeters, upland groundwater (GW), riparian groundwater, and
zero-order (0-ord.) streams. Additionally, stream chemistry was observed at the outlets
of three nested catchments (1st through 3rd-order) at the MPB site.
161
Figure 2. Precipitation at the a. Unimpacted and b. MPB sites, c. snow depth, d.
temperature. Precipitation patterns were similar at the two sites, but the Unimpacted site
received more summer precipitation. Snow depth time series show similar patterns of
accumulation and peak amounts, with the MPB site snowmelt finishing five to seven days
earlier. Temperature was similar between the two sites, although the Unimpacted site
tended to lower temp late summer, consistent with more summer precipitation. Shown are
5-day running means of 30-minute observations.
162
Figure 3. Time series of dissolved organic carbon (DOC) concentration and hydrologic
stage for the mountain pine beetle (MPB) and Unimpacted (UN) sites showing (a) upland
and riparian groundwater and (b) zero-order streams. The bottom panel (c) shows MPB
nested streams of first through third order. Groundwater DOC from each source was
relatively constant in time.
UN groundwater decreased in DOC from hillslope to
riparian, while MPB groundwater increased in DOC. The MPB zero-order stream had
greater DOC than the UN zero-order stream.
in the first-order stream than the larger streams.
In the MPB streams (c), DOC was greater
163
164
Figure 4. Time series of dissolved organic nitrogen (DON). All other descriptions are as
for Figure 3.
Note the different vertical scale for panel a, where MPB upland
groundwater was elevated in DON as compared to all other water sources. Riparian
groundwater DON was similar at MPB and UN. Zero-order steam DON mirrored the
riparian groundwater concentration at UN, while at MPB, zero-order stream DON was
higher than in the riparian groundwater. In the MPB streams (c), DON was greater in the
first-order stream than the larger streams.
165
Figure 5. Time series of dissolved nitrate (NO3) concentration. Note the logarithmic
scale for NO3. All other descriptions are as for Figure 3. MPB upland groundwater had
one to two orders of magnitude more NO3 than UN groundwater, but most of this
difference was attenuated in riparian groundwater.
Zero-order steams at both sites
consistently showed only trace NO3, and NO3 remained low in higher-order MPB
streams.
166
Figure 6. Solution concentrations of (a) DOC, (b) DON, and (c) NO3 across landscape
units from the uplands to riparian-stream system at the MPB and Unimpacted (UN) sites.
Shown are mean values and one standard error. Landscape positions are: SW 10 = soil
water at 10 cm depth; SW 30-60 = soil water at 30-60 cm depth; HS-GW = hillslope
groundwater; R-GW = riparian groundwater; Stream = zero-order streams. DOC was
similar in SW 10 at both MPB and UN, but greater at the MPB site in SW 30-60. HSGW at each site had similar DOC, but MPB R-GW and stream water showed increases in
DOC as compared to HS-GW. MPB soil water had one to two orders of magnitude more
NO3 than UN soil water, and MPB HS-GW had much greater DON and NO3 than UN
HS-GW. The MPB stream had higher DON than the UN stream, but only trace NO3 was
observed in either stream. Only one sample of SW 10 was analyzed for NO3 at UN. BDL
= Below NO3-N detection limit of 0.005 mg l-1.
167
168
Figure 7. Stream DOC concentration as a function of distance from the channel head at all MPB
site sampling locations. Distance was taken via the longest upstream channel. Boxes
shown represent the 25th to 75th percentile of all observations, while black markers show
the median and whiskers show the remaining quartiles.
The dashed curve is an
exponential decay function fit by nonlinear least squares. The solid curve is a similar
function decaying to an asymptote equal to the mean value of the 2nd-order and 3rd-order
streams.
Channel distance was a significant predictor of DOC using either model
(p<0.05).
Exponential Decay
Exponential With Asymptote
50
DOC (ppm)
40
DOC = 14e-0.9D+16
DOC = 28e-0.16D
30
20
10
0
0
1
2
3
4
Channel Distance (km)
5
169
Figure 8. As for Figure 5, for DON. Channel distance was a significant predictor of
DON using either model (p<0.05).
1.4
Exponential Decay
Exponential With Asymptote
1.2
DON (ppm)
1
DON = 0.73e-0.21D
0.8
DON = 0.47e-0.36D+0.26
0.6
0.4
0.2
0
0
1
2
3
4
Channel Distance (km)
5
170
Figure 9. Conceptual model of biogeochemical solution concentrations across landscape
positions from the hillslope to the stream-hyporheic zone and several orders of nested
streams.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement