H13C-0925
The Experimental Hydrology Wiki
Why do we need an Experimental Hydrology Wiki?
It allows us to learn about, recommend, question and discuss new and established, basic and advanced
methods of experimental hydrology.
It helps us to avoid reinventing the wheel each time we start out measuring something we haven't measured
before.
It also helps us not to make the same mistakes others have made before us.
It makes it easier to share new ideas and concepts.
It supports us in finding the methodology and the equipment suitable for our investigation.
But: a wiki needs users! Without users and contributors it cannot exist.
The general idea and layout of the Experimental Hydrology Wiki is presented here along with an invitation to
all experimental hydrologists to contribute with their knowledge and experience!
http://www.experimental-
hydrology.net/
H13C-0926
Long-term Water Quality and Stream Nutrient Responses to Forest Harvest and Disturbance at US Forest Service Experimental Forests and Ranges
Changes in land use and management as well as natural disturbances can dramatically affect water quality in some but not all headwater streams and downstream rivers. For decades, researchers have studied stream hydrology and solute chemistry in disturbed and undisturbed watersheds at Experimental Forests – a continental-scale research platform that spans gradients of precipitation, atmospheric deposition, nutrient limitation, biomes, topography, and soil types. To increase our understanding of water quality responses to disturbance, we examine: 1) How the short and long-term responses of stream nutrients to forest harvest, biomass removal, and disturbances vary across North America; 2) Which biotic and abiotic factors explain why the magnitude and timing of stream solutes responses vary within and among sites; 3) What factors control the magnitude and duration of disturbance responses; and 4) How atmospheric deposition, geologic, climatic, topographic, and hydrologic parameters affect responses. We find that the magnitude and longevity of nutrient responses varies within and among ecoregions. Following harvest or disturbance, concentrations and fluxes of stream nutrients generally increased but with varying response times across gradients of nutrient limitation. Changes in nutrient fluxes reflect gradients of nutrient limitation as well as the role of vegetation as a primary influence on nutrient cycling. These findings have implications about the role of watershed disturbances on instream biotic communities and downstream users.
H13C-0927
A Century of Experimental Forest and Range Research at the US Forest Service: Recent Synthesis Efforts and Opportunities for Cross-Site Collaboration
The US Forest Service Experimental Forest and Range (EF&R) network was established a century ago at Fort Valley, Arizona, now includes 81 sites, and is the backbone of a continental-scale research platform for studying ecosystem functions. These sites span the nation and encompass large gradients of climate, forest types, environmental factors, and disturbance regimes. Long-term data collected from network sites are 1) publicly-available, important records of climate, hydrology, and ecosystem productivity, and 2) instrumental in quantifying how diverse ecosystems respond to disturbances, atmospheric deposition, and climate change. We present examples of success stories that outline the significant contributions of these sites to science, land management, and society, and outline emerging synthesis efforts that are realizing the legacy of long- term data collected at these networked sites to advance fundamental and emerging topics in ecosystem sciences.
H13C-0928
Hydrological Response to Mountain Pine Beetle Infestation in Western Subalpine Watersheds
Water supply in western North America is controlled primarily by snow accumulation and melt in forested headwater basins. Trees impact runoff through wintertime canopy interception losses of snowfall and summertime transpiration losses. The mountain pine beetle (Dendroctonus ponderosae) epidemic attacking western pine forests will produce an estimated 90% mortality in lodgepole (Pinus contorta) stands, and will likely impact other tree species at significant levels over large areas of the US and Canada. Management studies suggest that changes in water quantity and quality will occur in response to beetle induced tree mortality. Hydrological responses in beetle killed forests are dependent on local climatology, forest age and species composition, understory response, and severity of infestation. Changes in discharge measured at the watershed level are typically quantified using statistical methods applied to time series data. Critical analysis elements are stationarity, and a sufficient data record for statistically significant detection of change. Short-term studies comparing statistical properties of flow often lack these critical elements and should be examined with caution. We show why short-term studies related to pine beetle impact on hydrology are unreliable. During the past five years, significant forest mortality has resulted from the current bark beetle epidemic, yet double mass plots using control basins and analysis of covariance (ANCOVA) show no significant response to date.
H13C-0929
Streamwater Chemistry and Nutrient Export During Five Years of Bark Beetle Infestation of Subalpine Watersheds at the Fraser Experimental Forest
Forested watersheds of western North America are currently undergoing rapid and extensive canopy mortality caused by a variety of insect species. The mountain pine bark beetle (Dendroctonus ponderosae) began to attack lodgepole pine (Pinus contorta) at the USFS Fraser Experimental Forest in central Colorado in 2002. By 2007, bark beetles had killed 78% of the overstory pine in Fraser research watersheds on average. The hydrologic, climatic, biogeochemical and vegetation records at the Fraser Experimental Forest provide a unique opportunity to quantify the impacts of this widespread, but poorly understood forest disturbance relative to a multi-decade pre-disturbance period. Here we compare seasonal streamwater chemistry and annual nutrient export for the five years since the bark beetle outbreak began with the pre- attack record. Patterns in post-outbreak streamwater biogeochemistry are compared to changes is species composition and proportional loss of overstory basal area for four basins. The influence of the outbreak will depend upon an aggregate of short (i.e. halted overstory water and nutrient use) and longer-term (i.e. altered canopy interception, windthrow, and understory growth) processes, so the hydrologic and biogeochemical implications of current beetle activity will not be fully realized for decades.
H13C-0930
Summer lowflow deficits after two decades of forest regeneration in the western Cascades, Oregon
This study explored the long-term response of summer water yields to past forest management practices, specifically the conversion of mature and old growth conifer forests to young forest plantations, in seasonally drought-stressed conifer forests of western Oregon. Results were based on long-term (40 to 50-year) paired watershed experiments in the HJ Andrews Forest in the Willamette National Forest and Coyote Creek in the South Umpqua National Forest. In the third decade after 100 percent clearcutting, streamflows were 30 to 80 percent lower in the young forest than the reference (mature and old forest) watershed during August to November. In the third decade after patch-clearcutting, summer streamflows were 20 to 40 percent lower in the cut watershed compared to the control. In the third decade after a 50 percent overstory thin, almost all summer low flows were within 25 percent of the flows at the control watershed. A 12 percent understory thin in a clearcut watershed during the third decade led to a temporary, minor abatement in summer low flow declines, but within five years, summer low flows from the thinned watershed were similar to those from an adjacent, unthinned forest plantation of similar age. Streamflow deficits emerged as early as March or April and persisted into October and November in the warmer, drier site in southern Oregon (Coyote Creek), whereas summer streamflow deficits emerged later and persisted for fewer weeks in the cooler, wetter Andrews Forest. These findings are consistent with previous studies demonstrating (1) increases in water use in certain conifer species relative to others (e.g. Douglas-fir versus pine); (2) higher water use in young (i.e., 10 to 50- yr-old) compared to old (100 to 250-yr-old) stands of many tree species; and (3) decreased interception capacity of young relative to old forest stands associated with loss of canopy epiphytes. Results appear to be robust, despite gaps in data availability, uncertainties associated with changes in stream gauging, streamflow trends over time in control watersheds, and multi-decadal fluctuations in regional climate over the study period. These findings support the notion that variable-intensity logging prescriptions over small areas to approximate natural forest structure may have the least long-term effect on summer low streamflows. However, more research, preferably new paired-watershed experiments, is needed to quantify the magnitude and duration of summer streamflow effects from various levels of overstory and understory thinning treatments.
H13C-0931
Hydrological control over stream nitrate loss in an aggrading New Hampshire forest
Stream chemistry of a small watershed in the Hubbard Brook Experimental Forest (Watershed 4) displays higher nitrate export than expected for an early successional forest in this region. Within Watershed 4, a small tributary (300 m in length) has a chemical signature far different from the main channel (1000 m in length). Previous monitoring has shown that the pH of the side tributary was significantly higher than that of the main channel and contains detectable levels of nitrate whereas the main channel had no detectable nitrate, and it is suspected that this side tributary significantly contributes to watershed export under base flow conditions. We expected that watershed's dominant water and chemical sources would vary with flow conditions, especially during summer thunderstorms. We hypothesized that the side tributary is the dominant source area under the normal base flow conditions of the summer (usually under 1 L/s) and that the main stem exerts dominance under the high flow conditions brought on by events. Daily water samples were taken throughout summer 2008 with three ISCO automated samplers: One ISCO was placed at the main stem of the stream, a second at the small internal tributary, and a third was placed at the weir. The samples were analyzed for pH, specific conductivity, and ANC, DOC, DON, and major anions and cations. The chemical data was compared to precipitation and rate of watershed flow calculated at each sampling hour to detect associations between chemical dominance and hydrological conditions. Under the base flow conditions of the summer, the chemistry of watershed outflow was dominated by that of the short side tributary, with lower acidity and higher nitrate levels than the longer main channel, but with notable contributions from the main channel. During each of the three high flow events that occurred over the summer, flow in the main channel increased dramatically and flow at the weir corresponded to temporarily increased acidity and decreased ANC. Preliminary nitrate data shows that a dilution response was associated with these events since there were significant declines in the nitrate levels of both the side tributary and watershed export during high flow conditions.
H13C-0932
Effects of Clearcut Logging on Channel Form in a Coastal Redwood Forest
The morphological characteristics of channels were mapped in recently logged areas and in a mature second-growth redwood forest at Caspar Creek, California, USA, to evaluate the extent to which logging has influenced channel conditions. Channels are pervasively incised in the area. Gullying appears to have originated with first-cycle logging of the late 1800s, and sequences of headcuts have continued to migrate up channels since then. Second-cycle logging of the early 1990s in the North Fork Caspar Creek is associated with increased spacing between headcuts in channels within clearcut tributary watersheds, suggesting that some discontinuous incisions have coalesced. Upstream channel limits in logged watersheds are now located significantly farther upslope than in control watersheds, and the magnitude of observed increases in peakflows and storm runoff is consistent with the observed change in drainage density. Relations between channel morphological variables and indices of stream power show higher variance in logged watersheds than in controls, suggesting that previously established channel forms have been disrupted after logging. Correlations between suspended sediment yield and indices of gully and surface erosion suggest that in- channel erosion associated with hydrologic change is the dominant source of post-logging sediment during years lacking major landslides. Common sediment-control measures, such as use of riparian buffer strips and control of road surface erosion, would not be effective for reducing sediment input from this in-channel source.
H13C-0933
Water Yield and Runoff Timing Across the Rain-snow transition at the Kings River Experimental Watersheds in California's Southern Sierra Nevada
The hydrologic response of eight headwater catchments located at and above the rain-snow transition, 1,500-2,500 m elevation, was investigated over five years using hourly streamflow, precipitation, snowpack and weather-station data. The Kings River Experimental Watersheds (KREW) is a watershed-level, integrated ecosystem project for long-term research on nested headwater streams in the southern Sierra Nevada. The four lower elevation watersheds also comprise the Southern Sierra Critical Zone Observatory. Water yield increased and retention decreased with elevation consistently for all years, but the form of the relationship was distinctly different for the wet years (2005 and 2006) when compared to the dry years (2004 and 2007). Stream base flows range from less than 1 to 10 liters per second (l/s), but during spring snowmelt flows of 400 l/s occur for a month or more. Maximum peak flows of more than 1,000 l/s were measured for a single rain event. While the amount of precipitation is similar for all watersheds, about 40 percent leaves the lower elevation watersheds as stream discharge compared to 60 percent for the higher elevation watersheds. The lower elevation watersheds are representative of how the higher watersheds may function with climate change. The rain-snow transition zone in these forested, mountain catchments undergoes dramatic seasonal changes, going from a snow-covered, water-saturated state with modest evapotranspiration (ET) to a system dominated by ET to a relatively dry state over a period of only a few weeks. The spring runoff period typically lasts from March to May at the lower elevation site. Runoff consistently lags about one month across this elevation range, with interannual variability also accounting for a 1-month difference in timing. Snowmelt occurs over a period of about one month in higher-elevation catchments, followed by a 2-week-long transition to evapotranspiration dominance of streamflow. Lower-elevation catchments are rainfall-dominated in spring, with the transition to evapotranspiration dominance being less distinct. Lags in the system between peak temperature and peak snowmelt were consistent at about 10 hours; the summer lag between solar maximum and streamflow minimum was similar, but offset.
H13C-0934
Research design for hydrologic response to watershed treatments in the mixed conifer zone of California's Sierra Nevada
Water quantity response to forest management is of great interest in California's Sierra Nevada, owing to
shifts in the rain-snow transition elevation associated with climate change, increasing value of hydropower
from high-elevation dams, and the re-examination of adaptive management strategies for wildfire mitigation.
In 2006 we initiated a multi-disciplinary research program to inform adaptive management for Forest Service
lands in the Sierra Nevada. The forest treatment approach is based on disconnected, overlapping fuel
treatment patches (forest thinning) to reduce the rate and intensity of fire. As little as 30% of the area in a
given catchment will be treated. Controlling for confounding influences is particularly challenging when the
experimental unit is a whole landscape and the inferential reference is an entire region. To isolate water and
ecosystem impacts related to forest thinning, we are using a Before After Control Impact (BACI) design in
conjunction with mechanistic modeling. BACI compensates for the sparse replication (2 sites) and the non-
random assignment of the treatments by providing robust longitudinal controls. BACI design defines two
treatments, a control and an impact. For modeling fire and wildlife response we chose subdivided the region
into two 40-km2 sub-firesheds; within each is a 1-km2 hydrologic study catchment. The control site in
this a measure of natural variation rather than a true control. Meta-replication using parallel studies in the
Sierra Nevada with different approaches is also an important component and involves a creative combination
of data from multiple sources. Rather than statistical comparisons or traditional hypothesis testing, we will
measure the support in the data for our a priori expectations using mechanistic models. We are currently
evaluating how to extend this research design to private forest lands with a wider range of management
options.
http://snamp.cnr.berkeley.edu/
H13C-0935
Design And Implementation Of A Water-Quality Monitoring Program In Support Of Establishing User Capacities In Yosemite National Park
Under the 1968 Wild and Scenic Rivers Act, the managing agency of the designated river is required to
develop a comprehensive management plan that protects the characteristics of the river segment meriting
special protection and that addresses user capacities. We analyzed water quality in the upper Merced and
Tuolumne Rivers of Yosemite National Park, designated as Wild and Scenic, to provide baseline data in
support of establishing park-specific water-quality standards as part of an adaptive management framework
that addresses user capacities. Samples collected from 2004 to 2007 were analyzed for nitrate, total
dissolved nitrogen, total dissolved phosphorus, and total phosphorus at all sites, and for E.coli, total
petroleum hydrocarbons, and waste water compounds at a subset of sites. Water quality in the upper Merced
and Tuolumne River basins was dilute, with concentrations in most samples falling near or below minimum
detection limits for the constituents of concern. Results indicate that spatial and seasonal variability in
nutrient concentrations is a function of local basin characteristics, particularly elevation. Generally, nitrate-
plus-nitrite increased with elevation, while total dissolved nitrogen decreased with elevation. Anomalous high
nitrate-plus-nitrite concentrations occurred during low flow at several sites, such as below Yosemite National
Park's primary waste water treatment facility, making those sites good candidates for further investigation.
Bootstrapping was used to characterize the distribution of concentrations at each site and to make inferences
regarding background reference conditions, which could be used by the National Park Service to select anti-
degradation water-quality standards and establish water-quality based user capacities. Results suggest that
future monitoring efforts should emphasize storm-event sampling, which is necessary to capture infrequent
elevated nutrient concentrations, and sampling in the head-waters of each basin where further data is
needed to characterize possible impacts from atmospheric deposition of pollutants. Additional water-quality
monitoring in Merced River tributaries throughout Yosemite Valley is needed to identify sources of nutrient
inputs and to determine whether visitor use is contributing to the overall water quality.
http://www.nps.gov/yose/parkmgmt/ucmp.htm
H13C-0936
Long-Term Response of High-Elevation Lakes to Changes in Atmospheric Deposition and Air Temperature
High-elevation lakes are often sensitive indicators of environmental and climate change. Here we evaluate over two decades of water-quality data from a long-term monitoring network of remote high-mountain lakes in three Colorado wilderness areas. In two of the areas, lake sulfate concentrations have decreased by a factor of 2-3 times since the mid-1980s, which is consistent with declines in precipitation sulfate measured at 11 high-elevation NADP stations in Colorado. The downward trends in lake and precipitation chemistry are likely the result of regional declines in sulfate emissions from power plants and nonferrous metal smelters. By contrast, lake sulfate concentrations increased dramatically in the third wilderness area despite declines in atmospherically deposited sulfate. In lakes in these areas, sulfate is derived primarily from pyrite oxidation in contrast to the other wilderness areas where sulfate is dominated by atmospheric sources. It is hypothesized that the upward trend in sulfate reflects enhanced pyrite weathering caused by an increase in annual air temperature of about 2-degrees Celsius since the mid-1980's. These results indicate the water-quality response of remote lakes to improvements in air quality may be complicated by long-term changes in climate.
H13C-0937
Stream Nitrate Concentrations in a Small Catchment in South West England over a Period of 35 Years (1970-2005)
A 35-year record of nitrate concentration for the Slapton Wood stream, a small agricultural catchment in south west England, is presented. The study reconsiders earlier work in order to assess whether upward trends have been maintained and how controls on catchment nitrate processes have altered. The study has shown that: (i) the catchment has reached a new position of equilibrium and increases in nitrate concentration have levelled off; (ii) the occurrence of severe droughts means that records of less than a decade are misleading and only longer records can illustrate changes of system state; (iii) the change of state observed in the catchment is illustrated in the switching of long-term memory effects from a negative to a positive annual memory; (iv) a significant long-term impulsivity relationship with rainfall becomes insignificant over the course of the study period. The study shows the importance of long records in exposing changes in state in catchment systems and understanding the time constants of a range of driving processes. The study by its very nature also demonstrates the importance of maintaining long-term monitoring programmes.
H13C-0938
Model-based Analysis to Identify the Impacts of Climate Variability and Land Use Changes on Streamflow
With climate change and water use competition being ever more important considerations in catchment management, improved understanding of the sensitivity of catchment response to climate and land use is paramount. This understanding includes the susceptibility of a catchment's response characteristics to shifts in the magnitude and seasonality of rainfall, temperature increases, as well as the impact of large scale land use changes. The lumped conceptual rainfall-runoff model IHACRES (Catchment Moisture Deficit version) has been applied to catchments in the Canberra region, Australia to evaluate the separate impact of land use on catchment response. The evaluation involved comparison of model performance and identified responses for catchments affected by considerable surface water (farm dams) and groundwater development (Tinderry and Burbong catchments) with that for less developed catchments (Gingera, Brindabella and Orroral Crossing). Prior to application of the model, data analysis techniques (e.g. trend analysis, deconvolution and baseflow filtering) have been used to develop appreciation of the temporal variation in hydrologic response characteristics for each site. Techniques for non-parametric estimation of the total unit hydrograph have been applied to examine seasonal and long-term variation in the unit hydrograph response. Tinderry and Burbong catchments have lower baseflow proportions and greater sensitivity to climate variability than the less developed catchments. The Gingera catchment shows significant seasonal variation in its response function, probably due to increased summer evaporative loss from bogs within the catchment. The IHACRES model was calibrated on 4 to 6 year periods from 1968 to 2002 using daily rainfall, temperature and streamflow data. Various statistics were used to measure the performance of the models, and 'best' model fits in each calibration period were selected after considering trade-offs among these statistics. There is larger temporal variability in optimized parameter values at Tinderry and Burbong compared with the less developed catchments. The stronger temporal stability of the calibrated parameter values for the latter catchments suggests that the model is adequately (though not perfectly) representing the impact of climate variability on the response of these catchments.
H13C-0939
Evaluation of Methods for Estimating the Effects of Vegetation Change and Climate Variability on Streamflow
Changes in vegetation cover can significantly affect streamflow and our understanding of the vegetation effect is primarily based on paired catchment studies. Two commonly used methods for estimating vegetation effects on streamflow are paired catchment method and time-trend analysis. In this study, the applicability of these methods is evaluated using data from paired catchments in Australia, New Zealand and South Africa. Results showed that these methods generally yield consistent estimates of the vegetation effect and most of the observed streamflow changes are attributed to vegetation alteration. These estimates are appropriate and supported by the vegetation history. The accuracy of the estimates, however, is dependent on the length of calibration periods or pre-treatment periods. For catchments with short or no pre-treatment periods, statistically identified pre-change point periods can be used as calibration periods. In addition to the vegetation effect, streamflow also responds to climate variability and it is necessary to consider the climate effect in assessing streamflow changes. The climate effect on streamflow was estimated using sensitivity- based method by considering changes in rainfall and potential evaporation. A conceptual framework based on the assumption that climate and vegetation are the only drivers for streamflow changes, enables comparison of these three methods. It is shown that these methods provided consistent estimates of vegetation and climate effects on streamflow for the catchments considered. One of the advantages of the time-trend analysis and sensitivity-based method is their applicability to non-paired catchments, making them potentially applicable to large catchments for quantifying effects of plantation development on streamflow.
H13C-0940
Relative importance of evapotranspiration variability in a semi-arid urban environment
In semi-arid ecosystems, evapotranspiration (ET) is a significant portion of the water balance, sometimes accounting for over 70 percent of the annual water balance. In these water limited systems, spatial and temporal patterns of ET have large impacts on streamflow variability and storm response. In the western U.S., increasing human populations are resulting in an expansion of urban land uses in semi-arid areas. Urbanization may affect the local water balance in several ways. Replacing vegetation with impervious surface may decrease ET, as well as increasing storm runoff and annual streamflow. At the same time, importation of water from outside a drainage basin may increase opportunities for ET within the drainage area if this water is used to water lawns or other outdoor vegetation. Given the highly heterogeneous and fragmented nature of urban environments, these changes in ET are expected to show a high level of both spatial and temporal variability. We use the Regional Hydrologic Simulation System (RHESSys) to simulate ET and streamflow in the Mission Creek catchment in Santa Barbara, CA. This modeling relied heavily on input data collected through the Santa Barbara Coastal (SBC) LTER site. We consider a range of different urban development and climate scenarios to estimate how urbanization may alter ET and its impacts on streamflow. Results show that ET is highly variable on both an interannual and seasonal basis and its influence on storm flow response varies on both these scales. Results also show that urbanization is likely to significantly alter ET, with consequences for streamflow. The effect of urbanization is spatially variable and emphasizes the relative importance of different regions of the catchment, such as the riparian zone.
H13C-0941
Validation and Sensitivity Assessment of a Distributed Hydrologic Model for an Inland Pacific Northwest Forested Watershed
Physically-based, spatially distributed hydrologic models typically require a large numbers of parameters which must be measured or estimated for catchments of interest. Meteorological data used to drive models are critical to accurate output, but are frequently difficult to measure accurately and continuously at remote locations. To answer the question of how the error from each of the input parameters and variables affect model outputs, a sensitivity analysis of the Distributed Hydrology Soil-Vegetation Model (DHSVM) was completed. A rigorous five-year calibration of the model was completed and validated with internal catchment variables including soil water content, snow water equivalent, and sapflow at the Mica Creek Experimental Watershed (MCEW) in northern Idaho. Following the model validation, a series of systematic model runs were completed to explore the sensitivity of the catchment outflow to state parameters and climate variables. A ten percent increase in precipitation was found to increase water yield by 24.67% and peakflow by 19.67% in the five-year study period. An air temperature increase of two degrees C increased ET by 35.4% and reduced water yield by 15.9%. The 50th percentile of annual flow was advanced by 45 days for the 2 degree C air temperature increase, and by 16 days for a 1 degree C increase. The 5% peakflow was reduced by 24% and enhanced by 20% with the air temperature and precipitation increment, respectively. The model was surprisingly sensitive to relative humidity, where a 10% increase produced a 37.6% in water yield and increased the 5% peakflow by 20%. The soil properties greatly affect the water yield and annual streamflow timing in a by changing the soil water storage. For example, switching from silt loam to silty clay increases water yield by 24% and to loamy sand advances 50th percentile flow by 6 days. Although these results are site specific, the modeling results from air temperature and precipitation sensitivity tests indicate the importance of accurate hydrometeorological data collection. These results also suggest the degree of sensitivity to future climate scenarios that indicate warmer and potentially wetter conditions for this region.
H13C-0942
Application of Physics Based Distributed Hydrologic Models to Assess Anthropologic Land Disturbance in Watersheds
The Department of Defense (DoD) manages approximately 200,000 km2 of land within the United States on military installations and flood control and river improvement projects. The Watershed Systems Group (WSG) within the Coastal and Hydraulics Laboratory of the Engineer Research and Development Center (ERDC) supports the US Army and the US Army Corps of Engineers in both military and civil operations through the development, modification and application of surface and sub-surface hydrologic models. The US Army has a long history of land management and the development of analytical tools to assist with the management of US Army lands. The US Army has invested heavily in the distributed hydrologic model GSSHA and its predecessor CASC2D. These tools have been applied at numerous military and civil sites to analyze the effects of landscape alteration on hydrologic response and related consequences, changes in erosion and sediment transport, along with associated contaminants. Examples include: impacts of military training and land management activities, impact of changing land use (urbanization or environmental restoration), as well as impacts of management practices employed to abate problems, i.e. Best Management Practices (BMPs). Traditional models such as HSPF and SWAT, are largely conceptual in nature. GSSHA attempts to simulate the physical processes actually occurring in the watershed allowing the user to explicitly simulate changing parameter values in response to changes in land use, land cover, elevation, etc. Issues of scale raise questions: How do we best include fine-scale land use or management features in models of large watersheds? Do these features have to be represented explicitly through physical processes in the watershed domain? Can a point model, physical or empirical, suffice? Can these features be lumped into coarsely resolved numerical grids or sub-watersheds? In this presentation we will discuss the US Army's distributed hydrologic models in terms of how they simulate the relevant processes and present multiple applications of the models used for analyzing land management and land use change. Using these applications as a basis we will discuss issues related to the analysis of anthropogenic alterations in the landscape.
H13C-0943
Application of Existing Streamflow Gauge Networks and Remote Sensing for use in Paired Watershed Experiments
Drawing general conclusions about the effect of land-cover change on river flows in a region using paired watershed experiments tends to be limited by the lack of replicates in the experimental design. Paired watershed studies are expensive, require monitoring for many years and are usually based on small experimental watersheds. The possibility of exploiting existing gauged watersheds (e.g., from the U.S. Geological Survey network) for land-cover change studies has been proposed for a project that is being conducted in large semi-arid shrubland watersheds in Southern California and the Western Cape Region of South Africa. Satellite data are used to assess the stability of land-cover during the calibration phase of experiments and to quantify the magnitude of change following disturbance (e.g., fire, deforestation). The primary limitation of this approach was found to be the availability of control watersheds. Even with a large number of candidate watersheds, undisturbed watersheds were uncommon. Despite substantial distance between some watershed pairs, along with considerable fluctuation in the LAI's of the control and test watersheds, the streamflow calibration relationships were found to be good. However, it is uncertain whether this is sufficient to ensure meaningful analyses of land-cover change effects on river flow volumes. The analyses included two approaches, the use of prediction bounds and the widely used dummy variable method. No conclusive change in streamflow was evident from the paired watershed analyses in both regions, despite substantial land-cover change in some of the watersheds (e.g., 100 percent burnt). Large watersheds in Mediterranean-type Ecosystems appear to be resilient to land-cover change in terms of streamflow response.
H13C-0944
Using Long-Term Data Sets to Constrain and Regionalize Transit Time Estimates in Mountainous Catchments
Mean transit time (MTT) is being increasingly used as a metric of catchment hydrological function. Estimating MTT usually involves relating the temporally varying input concentration of a conservative tracer to the signal in the stream, using models of various transit time distributions (TTD). Most studies have been confined to data collection periods of 1-2 years at single sites, or within individual geomorphic provinces. This may limit the significance and transferability of the findings as such short periods usually only capture a narrow range of the climatic variability within a spatially restricted area. In this study, we use, for the first time, long-term (>10 year) data sets of hydrochemical tracers from 19 headwater catchments (ranging in size from 1 - 35km2), in six geomorphologically and climatically distinct parts of the Scottish Highlands. In each catchment, weekly samples of precipitation and stream water with measured chloride concentrations were used to estimate the MTT using various TTD models within an uncertainty framework. The MTTs obtained from a Gamma Distribution model were the best identified and ranged from 40 to 1500 days for individual catchments. Moving window analysis revealed that at least 4 years data were needed to gain consistent MTT estimates. Shorter observation periods could produce widely ranging estimates, reflecting inter-annual climatic variability. The constrained MTTs estimated from long-term data at a wide range of contrasting sites, in conjunction with GIS analysis of the landscape characteristics, allowed a robust multiple regression analysis to establish the relative importance of topographic, pedological and climatic controls. The best model combines the prediction variables percentage responsive (i.e. overland flow generating) soil cover, drainage density, precipitation intensity and median subcatchment area and yields R2adj = 0.88. Cross validation shows small absolute error suggesting that the model can be used to estimate MTTs in ungauged headwater catchments throughout the whole region of the Scottish Highlands, and potentially in similar mountainous regions.
H13C-0945
Channel changes at Carnation Creek: The results from 38 years of annual surveys
In 1971 the Canadian Department of Fisheries and Oceans and the British Columbia Ministry of Forests established a detailed fish forestry interactions study at Carnation Creek, on the west coast of Vancouver Island, to evaluate the impact of logging on the aquatic ecology of the watershed. As part of this study, a longitudinally distributed series of 8 study areas were established, each comprised of 18 to 36 cross-sections separated by about 3 meters. The cross-sections have been re-surveyed annually to determine bed and bank topography. The cross-sections were initially laid out with a plane table approach and at the time cross-sections were seen to be the most efficient means of characterizing the morphology of the study sites. Here we present the first major analyses of the temporal changes in both the channel geomorphology at the study areas and the most approprate monitoring techniques. The data will be presented through slideshows that illustrate the DEM's and the evolution of the bed over time. The data illustrate that large cycles of aggradation and degradation are common at some of the study areas while being substantially reduced or absent at others. The key factor controlling the behavior of the channel being the presence of wood and log jams. The basic channel shape, indexed by the width to depth ratio, was found to be very resilient and return to a near constant value after disturbances. In addition, the width to depth ratio is nearly identical for all study sites, despite differences in the size of the channel at the different sites. The data also reveal that as the channel changed over time the stream course arranged itself oblique to the cross-section suggesting that cross-sections may not be the best means of monitor channel change over extended periods of time.
H13C-0946
DYNAMIC PROCESSES OF LARGE WOOD AND THEIR EFFECTS ON FLUVIAL EXPORT AT THE WATERSHED SCALE
The presence of large wood (LW) has a pronounced impact on the geomorphic and ecological character of river corridors, yet relatively little is known about the patterns and processes at the watershed scale. To understand these patterns we monitored the volumetric input of LW into 131 reservoirs and a suite of watershed characteristics. Of all geomorphic and hydrologic variables tested, watershed area was most important in explaining LW export. LW export per unit watershed area was relatively high in small watersheds, peaked in intermediate-sized watersheds and decreased in large watersheds. To explain these variations, we surveyed the amount of LW with respect to channel morphology in 78 segments (26 segments in each size class) in the Nukabira River, northern Japan, and examined the differences in LW dynamics, including its recruitment, transport, storage, and fragmentation and decay along the spectrum of watershed sizes. We found in small watersheds a larger proportion of LW produced by forest dynamics and hillslope processes was retained due to narrower valley floors and lower stream power. The retained LW pieces may eventually be exported during debris flows. In intermediate-sized watersheds the volume of LW pieces derived from hillslopes decreased substantially with reductions of proportion of channel length bordered by hillslope margins, which potentially deliver large quantities of LW. Because these channels have lower wood piece length to channel width ratios and higher stream power, LW pieces can be transported downstream. During transport, LW pieces are further fragmented and can be more easily transported; and therefore, the fluvial export of LW is maximized in intermediate-sized watersheds. Rivers in large watersheds, where the recruitment of LW is limited by the decreasing hillslope margins, cannot transport LW pieces because of their low stream power and thus LW pieces accumulate at various storage sites. Although these stored LW pieces can be re-floated and transported by subsequent flood events, they may be also combed by obstacles such as log jams and standing trees on floodplains and in secondary channels. Redistribution on these surfaces can be up to decades with eventual decay into fine organic particles resulting in the reduction of fluvial export of LW in larger watersheds. Our findings provide important information for a role of LW pieces in regulating the dynamic character of geomorphic processes and aquatic habitats, and the transfer and residence time of energy for stream-dwelling organisms at the watershed scale.
H13C-0947
Scaling the Geomorphic and Ecological Consequences of Contemporary Climate Change Within the Salmon River Watershed, Central Idaho: A View From Taylor Ranch Field Station
Established in 1970 by the University of Idaho, Taylor Ranch Field Station is located in the Frank Church
Wilderness of No Return, along Big Creek, a 1445 km2 tributary to the Middle Fork of the Salmon River.
The field station has provided a stable center for terrestrial and aquatic ecological studies within the Salmon
River for almost 40 years. Dr. Wayne Minshall began monitoring aquatic ecology indices at numerous sites
in the Salmon River basin in the late 1970's. This rare continuum of roughly 30 years of field data can be
coupled with publically available hydrologic, geomorphic and meteorological data sets to reveal a rich record
of how recent demonstrable changes in climate have affected this wilderness watershed.
As a consequence of improved access and automated and telemetered sensors of water quality and
quantity, contemporary studies continue through out the watershed at an increasing temporal and spatial
resolution. The impetuous is upon current researchers to understand both the role of the basin as a major
water source to the Snake and Columbia River systems and also the function of the basin as ideal habitat for
threatened native fish. Beyond these applied questions that directly impact management decisions, the
pristine nature of much of the Salmon River basin also favors studies of fundamental feedbacks between the
physical and biological systems. These interdisciplinary studies are augmented by increasing collections of
high resolution spatial data sets such as Hyperspectral Imagery, Distributed Sensor Networks and LiDAR
topography.
We present a study that explicitly examines the feedbacks between wildfire, sediment production, basin
hydrology and aquatic ecosystem function. Because the tributaries to the Salmon River span discrete
ranges in elevation across the snow- to rainfall-dominated hydrologic regimes, these studies reveal how
sensitive different portions of the Salmon River system are to projected changes in temperature. Depending
on the elevation range within a give catchment, these changes will result in different responses in hillslope
stability, wildfire susceptibility, stream ecology and channel form.
http://www.uihome.uidaho.edu/default.aspx?pid=41858
H13C-0948
Headwater thermal response to partial-retention forest harvesting: a process-based paired-catchment experiment
Paired-catchment experiments are the most rigorous empirical research design for estimating the effects of land use on aquatic systems. However, they have recently come under increasing criticism, in part because past studies typically treated catchments as black boxes. As a result, investigators could only speculate about the factors responsible for any observed effects, limiting their ability to generalize the experimental results in space and time. This study used a paired-catchment approach to investigate the effects of partial- retention forest harvesting with no riparian buffer on the thermal regime of a headwater stream in coastal British Columbia. In addition to monitoring stream temperature at three locations within the treatment reach, we monitored above-stream microclimate, water surface evaporation, bed temperature profiles, groundwater temperature, and reach-scale surface-subsurface interaction. Daily maximum stream temperatures increased after harvesting by over 5 °C during summer, with little effect in winter. The major driver of post- harvest warming was an increase in solar radiation, which was partially moderated by the increased effects of hyporheic exchange, bed heat conduction and evaporation. Incorporating process-based measurements into paired-catchment experiments not only allows the causes of treatment response to be assessed, but they provide a valuable data set for testing predictive models.
H13C-0949
Do water quality BMPs work? Combined monitoring and modeling hold the answer
Although water quality problems associated with agricultural non-point source (NPS) pollution have prompted the rapid and widespread adoption of a variety of so called "best management practices" (BMPs), it has proven difficult to assess their cumulative impacts and individual effectiveness in reducing NPS pollution at the watershed scale. In this project we combined long-term monitoring, paired-watershed analyses, and process-based watershed modeling to assess changes in dissolved phosphorus (DP) for a 160 ha catchment in the New York City Catskill water supply watersheds. The land use was a combination of forests and dairy farmland. A suite of BMPs were implemented in the mid-1990s aimed at reducing P loads. Using a nearby 86 ha forested watershed as a control site for a paired-watershed study, we found that the DP loads were reduced by 43% (+/-6%) and particulate P loads dropped by 29%. To assess the roles of individual BMPs in this reduction we used the Variable Source Loading Function (VSLF) model, a distributed watershed model and empirical relationships for DP concentrations in runoff based on on-site rain simulator experiments. The model analysis predicted a total reduction that was within 5% of the paired-watershed analysis and showed that the most effective BMPs were those that disassociated manure spreading and other P sources from areas prone to generating runoff, i.e., hydrologically sensitive areas. Interestingly, barnyard BMPs, which were generally the most expensive, appeared to have little impact on stream water quality. Unfortunately, because we cannot mechanistically model the processes that control particulate P across a whole watershed, the model was unable to make similar assessments of BMP impacts on particulate P. This body of work emphasizes demonstrates that combining both long-term monitoring and process-based modeling allows us to evaluate BMP effectiveness in the "living landscape" without necessarily establishing special research watersheds.