H21A-0167
Nonlinearity of basin response with respect to watershed size and storm magnitude
The objective of this study is to investigate the nonlinear characteristics of the basin response with respect to watershed size. Watershed geomorphologic and hydrologic data from Komarovsky, Centralny, Sadovy, and Saharny Zavod watersheds in the Komarovka Basin of Russia were collected for analysis. A kinematic-wave- based geomorphologic instantaneous unit hydrograph (KW-GIUH) model was adopted as an auxiliary tool to demonstrate the nonlinear characteristics. The results show that the linearity of the basin response with respect to watershed size is shown in small watersheds especially for large storm events. On the contrary, nonlinearity prevails in large watersheds with small storms. Moreover, the transition break for the basin response transferring from linear to nonlinear in the collecting area vs. peak discharge graph was detailed investigated. A numerical experiment was performed, from the most upstream point of the watershed along the mainstream to the downstream outlet, to investigate the variation of the collecting areas. The results show that once a large tributary is confluence into the mainstream, the collecting area shows an abrupt change in the graph of the collecting area vs. distance to downstream outlet. The collecting area corresponding to the abrupt change is just the transition break from linear to nonlinear in the graph of the collecting area vs. peak discharge. This finding can be well simulated by the KW-GIUH model and has been verified by using the hydrological records.
H21A-0168
Topographic Controls on Hillslope–Riparian Water Table Continuity in a set of Nested Catchments, Northern Rocky Mountains, Montana
Understanding how hillslope and riparian water table dynamics influence catchment scale hydrologic response remains a challenge. In steep headwater catchments with shallow soils, topographic convergence and divergence (upslope accumulated area-UAA) is a hypothesized first-order control on the distribution of soil water and groundwater. To test the relationship between UAA and the longevity of hillslope-riparian-stream shallow groundwater connectivity, we quantified water table continuity based on 80+ recording wells distributed across 24 hillslope-riparian-stream cross-sections. Cross-section upstream catchment areas ranged in size from 0.41 to 17.2 km2, within the Tenderfoot Creek Experimental Forest (U.S. Forest Service), northern Rocky Mountains, Montana, USA. We quantified toe-slope UAA and the topographic index (TI = ln a/tanβ) with a Multiple-D- Infinity (area routing in multiple infinite downslope directions) flow accumulation algorithm analysis of 1, 3, 10, and 30m ALSM derived DEMs. Indices derived from the 10m DEM best characterized subsurface flow accumulation, highlighting the balance between the process of interest, topographic complexity, and optimal grid scale representation. Across the 24 transects, toe-slope UAA ranged from 600-40,000 m2, the TI ranged from 5-16, and riparian widths were between 0-60m. Patterns in shallow groundwater table fluctuations suggest hydrologic dynamics reflective of hillslope-riparian landscape setting. Specifically, correlations were observed between longevity of hillslope-riparian water table continuity and the size of the UAA (r2=0.84) and its topographic index (r2=.86). These observations highlight the temporal component of topographic-hydrologic relationships important for understanding threshold mediated hydrologic variables. We are working to quantify the characteristics and spatial distribution of hillslope-riparian sequences and their water table dynamics to temporally link runoff source areas to whole catchment hydrologic response.
H21A-0169
Physics-based simulation of surface and near-surface hydrologic response for the Tarrawarra Catchment: parameterization for a synthetic dataset
Comprehensive physics-based simulation is a useful tool for furthering our understanding of complex hydrologic- response processes. The first step to using hypothetical realities as a base-case for concept development and hypothesis testing, is to confirm that the model / boundary-value problem (BVP) is capable of realistically representing known hydrologic-response processes. This study presents the development of a synthetic dataset designed to emulate the Tarrawarra Catchment (Southeastern Australia) using the physics-based, Integrated Hydrology Model (InHM). The Tarrawarra Catchment is a 10.5 ha, unchanneled rangeland catchment, with gentle slopes; runoff generation is dominated by the Dunne overland flow mechanism. The Tarrawarra BVP was parameterized based on topography and observed soil characteristics, then calibrated for a six month period in 1996 during which there exist continuous records of rainfall, potential evapotranspiration, and surface runoff, as well as episodic sampling of soil-water content and piezometric head distributed through out the catchment. While in practice no simulation could ever incorporate all the intricacies present at the Tarrawarra Catchment, the simulations conducted for this study honor both the observations and our best understanding of the physics governing near-surface hydrologic response. The simulations show that runoff occurs only after the locally perched water table rises to the land surface, in agreement with observations at Tarrawarra. Quantitative comparison of observed and simulated hydrologic-response also shows that the model accurately captures both the integrated response (surface discharge) and the distributed response (pressure head and soil-water content), supporting the use of InHM to generate synthetic response data for a Tarrawarra-like catchment. The continuous simulations have the added advantage of greater spatial and temporal resolution than any comprehensive dataset, allowing further insight into the location and timing of runoff generation.
H21A-0170
Developing a Framework for Testing Distributed Hydrologic Models at the Hillslope Scale
Numerous hydrologic models solve Richards equation for the variably saturated subsurface domain. However, the scarcity of measured hydrologic states and variables and the scale discrepancies between observations and simulations pose a challenge in testing and evaluating such models. We develop a flexible framework for testing distributed hydrologic models at the hillslope scale. The proposed method consists of three major steps. First we generate "hypothetical realities" representing the hydrologic response of a synthetic watershed modeled after the 10.5 ha Tarrawarra catchment in Australia. The catchment was extensively monitored and has a relatively simple geometry with 0.5-1.5m deep soils overlaying bedrock and a fairly uniform grass cover. Eleven years of half-hourly time increment hydrological states and fluxes generically termed "hypothetical realities" have been generated using the complex Integrated Hydrology Model (InHM) representing fully coupled 3D variably saturated subsurface and 2D surface flow with high resolution. In the second step, simpler distributed hydrologic models can be evaluated against the hypothetical realities, which represent an error-free data set of hydrologic variables. The simpler distributed models are run first without calibration and then with calibration against different combinations of the observed data from the hypothetical realities. In the third step, further tests of distributed models incorporate event based and continuous simulations, variable spatial and temporal scales and increasing amounts and types of model input data and observed data.
H21A-0171
An Integration Of Subgrid Physics Into Coarse Grid Simulations Of River Models
In river flows, the large eddies scale on river width and depth, thus the uncertainties and inhomogeneities in river boundary structure have length scales only slightly smaller the large-eddy scale. Consequently, the turbulence generated by the river bottom inhomogeneity is hydrodynamically significant. Conventionally, uncertainty and inhomogeneity in large-scale river modeling are generally addressed together by calibrating a roughness coefficient. However, this kind of calibration is inherently grid dependent due to nonlinearity associated with local flow inhomogeneity. In other words, the work to calibrate a model at one grid scale is lost if the model grid is altered, and such grid-dependent calibration cannot provide insight into the unresolved processes or features. To address the grid dependence, a Coarse Grid Simulation (CGS) framework is proposed (Fu and Hodges, 2005). The CGS will provide a model structure for representing the relationship between empirically-known subgrid features and grid-resolved flow. Therefore, the CGS will lessen the calibration efforts associated with changing model grid scales. In the present work, we demonstrate that grid resolution must affect turbulence calibration for inhomogeneous subgrid-scale physics in standard RANS formulations. Examples are provided to show that through use of the proposed CGS formalism, fine-scale process observed in a fine-grid model can be represented at arbitrary grid scales in a coarse-grid model without ad hoc calibration. Reference Fu, S. and B.R. Hodges (2005), ¡°Grid-scale dependency of subgrid-scale structure effects in hydraulic models of rivers and streams," Mechanics and Materials Conference (McMat 2005), June 1-3, 2005, Louisiana State University, Baton Rouge, Electronic Proceedings (CD-ROM), 5 pgs.
H21A-0172
Development of a Distributed Watershed Model Coupling River Flow, Surface Runoff, and Subsurface Flow
Lumped watershed models are often suitable for long-term simulations but unable to provide adequate spatial and temporal resolutions. A distributed watershed model was developed in this study to couple river flow, surface runoff, and subsurface flow for resolving hydrological responses in different spatial and temporal scales. River flow, surface runoff, and subsurface flow are explicitly and interactively solved. The river flow is simulated by the 1-D diffusive wave approach for each river segment with the conservation of mass and the continuity of stage for river junctions. River cross-sectional characteristics, including cross-sectional area, wetted perimeter, top width, are parameterized as functions of water depths. The surface runoff is simulated by the 2-D diffusive wave approach and eight flow directions are allowed for runoff in/out of each surface grid. The interactions between river flow and surface runoff is determined by the continuity of stage and the conservation of mass when water of two regimes are connected and stage differences exist. When surface grid is ponding and the corresponding river stage is lower than its river bank, the surface discharge is estimated by the weir formula. Daily evapotranspiration is estimated by multiplying the potential evapotranspiration, either prescribed or computed, by the crop coefficient determined from land use. The subsurface flow is a quasi 3-D approach, including soil moisture movements in the vertical direction and groundwater fluxes in the horizontal direction, that water flows are described by the Darcy's law. Vertical soil moisture distributions are updated after groundwater table is determined. The interactions between subsurface water and surface water (river and surface runoff) is described by the direct connection approach that flux and head continuities are conserved. Each subsurface grid is designed as a variable-saturated soil column right below its corresponding surface grid. River grids are designed not to overlap surface grids for conserving surface area. Time step size will be automatically reduced by examining the Courant number to increase numerical stability. The newly developed model is then applied to simulate daily flow of the Shiehman Reservoir watershed from 1977 to 2005. Results demonstrate our approach of integrated watershed modeling is suitable for applications in continuous and long-term watershed study, such as climate change impact and land use change analyses.
H21A-0173
Quantitative evaluation of perched groundwater storage using electric resistivity method in mountainous headwater catchments at Shiranui, Kumamoto, Japan
The selected headwater catchments for this study have been done many hydrometric observations to understand the rainfall-runoff process during last five years and the distinctive seasonal perched groundwater aquifer system has existed only in the wet season to support the local natural spring discharge. According to the previous study using electric resistivity method along one survey line in the study catchments (Ono, 2006), the perched groundwater system exists in the strongly weathered auto-brecciated Andesite lava and a continuous aquifer belt is developed seasonally to support the local spring discharge. In this study, the same electric resistivity technique was applied along the selected representative three survey lines to understand the three dimensional structure of this perched aquifer system. Also the seasonal resistivity difference between the resistivity distributions along the particular survey line was calculated to evaluate the groundwater storage of the perched aquifer. The evaluated groundwater storage capacity along three survey lines were used to synthesize the catchments scale three dimensional aquifer storage volume of the studied perched aquifer. The estimated storage volume of the perched aquifer was compared with the observed cumulative spring discharge rate during the wet season to evaluate the method. Although there are many uncertainties for this estimation, the estimated three dimensional storage volume of the perched aquifer by the electric resistivity method and the observed spring discharge during the wet season show relatively close value each other. This result strongly supports that the estimated three dimensional structure of the perched aquifer system by electric resistivity method seems to be reasonable for the studied head water catchments.
H21A-0174
Runoff generation in SE Spain
We are working to improve a hydrological model for prediction of runoff in medium-scale semi-arid catchments in SE Spain. The aim is to develop and improve understandings of runoff generation in semi-arid areas and to improve modelling of runoff. Objectives are to investigate the influence of geology, landuse and seasonality on infiltration rates and use remonte sensing (RS) and GIS to classify an area into Hydrologically Similar Surfaces (HYSS) categories. The research includes investigating the impact of different landscape elements on runoff within two 150 km2 catchments, the Rambla Nogalte and the Rambla de Torrealvilla. Most storms within these catchments are of short duration. HYSS are defined as areas with similar1-D (vertical) partitioning of net rainfall between infiltration and overland flow. HYSS are identified from field measurements of soils, micro and macro- topography and infiltration rates; then combined with analysis of multi-spectral airborne RS images. HYSS are selected to minimise internal variability in the relationship between rainfall and local runoff generation and are scaled up to cover larger areas. The overall sampling strategy for measurements has been to undertake constant intensity rainfall simulator measurements within provisional HYSS categories, and to augment this with a large number of minidisk infiltrometer measurements. This strategy captures as much of the variability in the landscape as possible. The wide variability within even small areas has led to the final adoption of only a few large classes that can be effectively distinguished. The final part of the research is to link the spatial partitioning of the two catchments into HYSS with the detailed rainfall records for the areas, and combine these two sets of data into a grid-based model for runoff generation across the area. The applied Green-Ampt modelling approach gave 63 possible combinations of surface properties (9 HYSS) and areas in the Rambla Nogalte each represented by a raingauge (7). The model was initially run with uniform rainfall over the catchment, next the model was run for selected storms using collected rainfall data. Selected storms were either intense short duration rainfall, or records with several days of precipitation. Soil moisture was stored in the surface layer over days, since the evaporation needs time and temperature to be effective and dry the upper soil layer. The data from the model runs have been linked to the map of HYSS categories, and the resulting map of at a point runoff gave an indication of potential runoff in the catchments.
H21A-0175
Modeling Watershed Hydrologic Water Balance Using the Soil and Water Assessment Tool (SWAT), Lower Coastal Plain, South Carolina USA
Understanding the hydrology at the watershed scale can be complex due to the variability of land use, soil type, climate, and vegetative cover. Hydrological models are valuable tools to fill the gaps in not only understanding the processes, but also assessing the impact of anthropogenic and natural disturbances on water quantity and quality. The objective of this study is to test a spatially-distributed hydrologic model, SWAT, for a third-order, low- gradient forested watershed using spatial watershed characteristics and 2.5-years of continuous stream flow and climate data. SWAT is a public domain model, actively supported by the USDA-Agricultural Research Service; it couples a Geographic Information System (GIS) with a distributed parameter hydrological model in order to predict runoff, water quality and other hydrologic parameters. The model has been developed to evaluate the impact of land management practices on the hydrology and water quality in large, complex watersheds. The watershed investigated in this study, Turkey Creek, is located approximately 60 km northwest of Charleston, South Carolina on the lower Atlantic Coastal Plain of the Southeast U.S. Turkey Creek's approximately 7200- hectare drainage basin is comprised of mostly pine and pine-hardwood mixed forest of varying stand ages on predominantly poorly drained soils. The elevations in this low gradient watershed vary from 3 to 14 meters above mean sea level. This model is currently being tested to calculate the water budget for the watershed and to predict hydrological responses such as stream outflow including surface and subsurface flows and evapotranspiration to storm events. Calibrated and validated to the field measured data, the model can be further applied to predict the long- term hydrologic and water quality impacts of land use change, climate variability, and natural disturbances on this watershed. The results from this model will be used to evaluate the utility of the SWAT model for application in the low gradient watersheds that make up the Atlantic Coastal Plain region of the Southeast U.S.
H21A-0176
Characterizing Storm Event Dynamics of a Forested Watershed in the Lower Atlantic Coastal Plain, South Carolina USA
Hydrology research in the Southeast U.S. has primarily focused on upland mountainous areas; however, much less is known about hydrological processes in Lower Coastal Plain (LCP) watersheds. Such watersheds are difficult to characterize due to shallow water table conditions, low topographic gradient, complex surface- subsurface water interaction, and lack of detailed soil information. Although opportunities to conduct long term monitoring in relatively undeveloped watersheds are often limited, stream flow and rainfall in the Turkey Creek watershed (third-order watershed, about 7200 ha in the Francis Marion National Forest near Charleston, SC) have been monitored since 1964. In this study, event runoff-rainfall ratios have been determined for 51 storm events using historical data from 1964-1973. One of our objectives was to characterize relationships between seasonal event rainfall and storm outflow in this watershed. To this end, observed storm event data were compared with values predicted by established hydrological methods such as the Soil Conservation Service runoff curve number (SCS-CN) and the rational method integrated within a Geographical Information System (GIS), to estimate total event runoff and peak discharge, respectively. Available 1:15000 scale aerial images were digitized to obtain land uses, which were used with the SCS soil hydrologic groups to obtain the runoff coefficients (C) for the rational method and the CN values for the SCS-CN method. These methods are being tested with historical storm event responses in the Turkey Creek watershed scale, and then will be used to predict event runoff in Quinby Creek, an ungauged third-order watershed (8700 ha) adjacent to Turkey Creek. Successful testing with refinement of parameters in the rational method and SCS-CN method, both designed for small urban and agricultural dominated watersheds, may allow widespread application of these methods for studying the event rainfall-runoff dynamics for similar watersheds in the Lower Coastal Plain of the Southeast U.S.
H21A-0177
Groundwater Ages and Stable Isotope Fingerprints of Contaminated Water to Examine Potential Solute Sources at a Uranium Processing Mill
To evaluate sources of high solute concentrations in groundwater near a uranium processing facility, groundwater recharge dates are correlated to specific solute concentrations and depth in the water column. Stable isotopes are also used as potential fingerprints of water sourced from mill tailing cells. Passive diffusion samplers, to be analyzed for 3He/4He ratio, were deployed in 15 different wells with samplers at two depths in the saturated interval. Low-flow purging and sampling was then conducted to isolate sampling points at different depths in the wells, with sampling at multiple depths being completed in 4 of the 15 wells sampled. Laboratory analyses were conducted for CFC recharge age, as well as T/3He recharge age. Contract laboratories analyzed for: deuterium and oxygen-18 isotopes of water; sulfur-34 and oxygen-18 isotopes of sulfate; trace metals uranium, manganese, and selenium; and nitrate and sulfate. Analysis for 235U/238U isotope ratios will be conducted to further identify fingerprint signals of source water. Groundwater recharge ages determined using CFC analysis show some vertical stratification in ages across the water column. Upon initial data processing and analysis, measured CFC ages ranged from 30 to 40 years within the water column of one well to only several years difference in another well. Additional results for trace metal concentrations, stable isotope ratios, and T/3He recharge ages will be reported when results are received. Further post-processing of CFC laboratory analysis and noble gas analyses will provide greater clarity as to groundwater ages within the aquifer and, combined with field pumping data, will allow for a comprehensive groundwater model to be constructed. This study provides great insight to potential mine tailings leakage problems and using isotopes and groundwater age dating techniques as a means of tracing contaminated groundwater to the leakage source. Utilizing stable isotopes of water and sulfate, combined with isotope ratios of uranium, create a powerful tool for fingerprinting mine tailings waters and tracing contaminated waters back to a potential leakage point.
H21A-0178
Is Spatial Proximity of Watersheds a Sufficient Guide to Hydrologic Similarity?
A continuing challenge for hydrology is to advance our understanding of the relationship between watershed structure and watershed response characteristics. Here we are particularly interested in the following question: Under what conditions should we expect two watersheds to behave hydrologically similarly? One aspect that is unresolved in this context is the importance of spatial proximity with respect to hydrologic similarity of watersheds. Several studies have made conclusions about this issue, however, some of them concluded that spatial proximity is a guide to hydrologic similarity and others concluded that it is not. In our study we investigate the characteristics of landscape and climate, location and streamflow response behavior of over 400 US watersheds across spatial and temporal scales to investigate this question. We will present initial results about what controls the similarity of watershed behavior across scales, and the role of proximity in this context.
H21A-0179
Evaluation of Catchment Storage Volume for Moderating Flow Fluctuations
Any catchment has a storage volume to some extent as compared with a virtual catchment covered with impervious surface, e.g., glass. Such storage volume (or capacity) contributes to moderate fluctuations of discharge from the catchment. That is, any catchment has a lower peak discharge at flood and higher discharge at drought than the virtual catchment. With respect to comprehensive water management including flood prevention and effective use of water resources, it is important to evaluate the storage capacity in various catchments. In this study, the following two indices calculated by daily rainfall, discharge and evapotranspiration data: water use storage depth (WUSD) and total storage depth (TSD) are used for evaluating the catchment storage capacity. The WUSD is derived as the water storage that is performing for assuring the minimum daily flow. That is, it is the minimum storage capacity when the daily water demand rate is equal to the discharge that gives a stable water supply through the year. Meanwhile, the TSD can be obtained from the changes in storage volume in a catchment. The TDS means the maximum difference of deposited water within one year when the catchment is considered as a vessel. Taking up two forest catchments and one reclaimed farmland catchment in a snow area as case studies, those storage depths are compared. The WUSD and the TSD in the forest catchments within one year are 57-287mm and 166-444mm, respectively, while those in the reclaimed farmland catchment are about 47-159mm and 142-285mm. The results show that the forest catchments are on average more excellent for the function to ensure a stable water supply and regulate floods, although both the WUSD and the TSD in any catchment range widely depending on the climate conditions in each year, and that the amount of snow cover has a significant effect especially on the TSD.
H21A-0180
Travel Time Distribution Modeling in the Valles Caldera, New Mexico
Modeling the transit times of catchment waters is of paramount importance in hydrology. The distribution of the time it takes for individual water molecules to move through a hydrologic system (a.k.a., the travel time distribution) is a fundamental characterization of a catchment. Travel time distributions are affected by a variety of physical characteristics of catchments (e.g., vegetation type, degree of soil development) that depend on the amount of solar energy the catchment receives. These characteristics, therefore, can be considered a function of aspect. The goal of this research is to constrain travel time distributions on a series of eight radial mountain streams having different slope aspects on Redondo Peak, a resurgent dome in the center of the Valles Caldera, near Los Alamos, New Mexico. Redondo Peak is an excellent natural laboratory for this type of experiment because all aspects are represented on different sides of the mountain while the internal geology and climate are relatively consistent. To model the transit time distributions of each catchment, variations of chemical load of the snowpack, isotopic compositions of meltwater samples, and snowcover distribution data from closely related studies are coupled with periodic stream and precipitation samples that are analyzed for stable water isotopes content. Additional information comes from a network of temperature sensors to monitor the distribution of snowmelt and headwater stream discharge as well as a series of flumes to capture the flows from the streams. The travel time distributions determined in this project provide a bottom up approach to verify catchment-scale models. http://hwr.arizona.edu/~surface/fieldsites/valles.html
H21A-0181
Quantifying Groundwater Contribution to Surface Flow in an Agricultural Watershed
There is increasing evidence that groundwater withdrawals are contributing to reductions in stream flow in the agricultural area of the Fraser Valley of British Columbia, Canada. Results of a two-year data collection effort within the Bertand Creek Watershed are presented which focus on creating a data set for development of a coupled surface water-groundwater flow model to investigate the effect of groundwater withdrawals on the magnitude of surface water flows. The field program was carried out during the summers of 2006 and 2007, the former among the driest rainfall summers on record while the latter received higher than normal rainfall. A multi- scale approach was adopted for the field program which employed streamflow gauging, seepage meter measurements, thermal gradient measurements, and eighty-five mini-piezometers to measure channel bed groundwater levels, hydraulic conductivities and to investigate near stream chemistry patterns. To overcome the challenge of measuring the low magnitude gain in streamflow between successive stream gauging locations, weirs were installed in segments of the creek in 2007. Preliminary results of the study indicate that groundwater contribution to surface flow within the watershed is dependant on overall dry season rainfall and on irrigation demand. Hydraulic conductivity testing and vertical hydraulic gradient record of the mini-piezometers at multiple depths within the creek bed were found to be instrumental in characterizing the magnitude of groundwater - surface water exchange. Comparison of successive streamflow measurements required continuous stream discharge record at gauging locations due to variations imposed on stream flow by water users.
H21A-0182
A new method to separate rain and snow
Water balance calculations in cold climates require the separation of precipitation into snow and rain. This separation is critical to determine whether water is available for runoff and soil infiltration, or if it is stored as snow. There are currently several approaches available using air temperature to separate precipitation into rain and snow. Based on analyses of daily precipitation data from 113 climate stations in south-western Alberta and south-eastern British Columbia, a curvilinear method is proposed that can be calibrated based on snow observation at local stations, with the two key variables being the threshold mean daily temperature (TT), where 50% snowfall occurs, and the range of temperatures within which mixed precipitation occurs (TR). A sensitivity and verification analysis of the two input variables was carried out for 15 climate stations in southern Alberta, and evaluated using three objective functions. Results reveal that the optimum TT ranges between 1.1 and 4.5°C, with a TR between 10 and 17°C. Verification analyses, using almost 700 years of daily precipitation observation, reveals that the new method significantly improves the estimation of daily snowfall when compared with the static threshold method, and methods developed by Leavesley et al. (1983) and Pipes and Quick (1977). On average, mean annual snowfall (MAS) was estimated to within 0.2% when detailed climate records where available, and within 7.1% when mean estimates of TT (2.6°C) and TR (13.3°C) were used. Respective r2 values were 0.789 and 0.802. The investigation of monthly variations of TT and TR revealed that both have strong seasonal variations. These variations generally follow a sin curve, with TT following a sin curve with one annual cycle, while TR follows a sin curve with two annual cycles. Varying monthly TT and TR values based on observations, MAS was estimated on average to within 3.5%, and the r2 was 0.828. By varying the mean TT (2.6°C) and TR (13.3°C) using an appropriate sin cycle, MAS was estimated on average to within 2.2%, and the r2 was 0.821. While all other tested precipitation separation methods generally work quite well, the newly proposed method appears to offer a significant improvement in the estimation of daily and annual snowfall from precipitation records.
H21A-0183
Water Resources Estimation of the Biosphere Reserve "Sierra de las Minas" in Guatemala, by Using a Distributed Hydrological Model and Considering Lack of Data
The UNESCO Biosphere Reserve "Sierra de las Minas" is part of the Guatemalan System of Protected Areas and contains the largest cloud forests of Guatemala. Its southern slopes flow into the Motagua Valley, one of the more active zones in the country and paradoxically, the most arid and driest zone of Central America. The main objective of this work was to obtain a better estimation of the water resources coming from the southern slopes of the "Sierra de las Minas", and to have an analysis tool to better understand key hydrological processes involved on water sustainable management practices, as the environmental services initiatives on the zone. Unfortunately, the lack of data was the general framework. The selected model was the so called TETIS, which is a conceptual distributed model developed by our research group. A simple and efficient method was used for parameter maps estimation with a 100 m cell size. The model was calibrated at daily time step in the Teculután basin with 187 km2, by applying a new split-parameter structure coupled with the SCE-UA automatic optimization algorithm, in order to obtain the set of optimal correction factors of the model. For a period of medium daily precipitation within the validation period, we obtained a simulated discharge very close to the observed data (the monthly Nash and Sutcliffe model efficiency coefficient was 0.83). We present here a precipitation spatial and sensitivity analysis, and a simple approach for estimate the contribution of cloud forests to the hydrological balance. Finally, the calibration results were extrapolated to the Uyús ungaged basin to simulate its flow regime and to evaluate its water resources.
H21A-0184
Diagnostic evaluation of distributed physically based model at the REW scale (THREW) using rainfall-runoff event analysis
The importance of diagnostic analysis of hydrological models is increasingly recognized by the scientific community (M. Sivapalan, et al., 2003; H. V. Gupta, et al., 2007). Model diagnosis refers to model structures and parameters being identified not only by statistical comparison of system state variables and outputs but also by process understanding in a specific watershed. Process understanding can be gained by the analysis of observational data and model results at the specific watershed as well as through regionalization. Although remote sensing technology can provide valuable data about the inputs, state variables, and outputs of the hydrological system, observational rainfall-runoff data still constitute the most accurate, reliable, direct, and thus a basic component of hydrology related database. One critical question in model diagnostic analysis is, therefore, what signature characteristic can we extract from rainfall and runoff data. To this date only a few studies have focused on this question, such as Merz et al. (2006) and Lana-Renault et al. (2007), still none of these studies related event analysis with model diagnosis in an explicit, rigorous, and systematic manner. Our work focuses on the identification of the dominant runoff generation mechanisms from event analysis of rainfall-runoff data, including correlation analysis and analysis of timing pattern. The correlation analysis involves the identification of the complex relationship among rainfall depth, intensity, runoff coefficient, and antecedent conditions, and the timing pattern analysis aims to identify the clustering pattern of runoff events in relation to the patterns of rainfall events. Our diagnostic analysis illustrates the changing pattern of runoff generation mechanisms in the DMIP2 test watersheds located in Oklahoma region, which is also well recognized by numerical simulations based on TsingHua Representative Elementary Watershed (THREW) model. The result suggests the usefulness of rainfall-runoff event analysis for model development as well as model diagnostics.
H21A-0185
A Semi-Distributed Hydrologic Model for Stream Flow Simulation Using the Sacramento Soil Moisture Accounting Model (SAC-SMA)
Distributed hydrologic modeling is currently viewed as a potential pathway to improve streamflow simulations regarding the sensitivity of runoff predictions to spatial and temporal variability of precipitation, land use, and soil properties. The US National Weather Service (NWS) started Distributed Model Intercomparison Project (DMIP) to guide NWS's distributed modeling research. In our participation of DMIP-2 activity, a semi-distributed version of the Sacramento Soil Moisture Accounting Model (SAC-SMA) was used to conduct hourly streamflow simulations. This model uses sub-basins as the computational elements of rainfall-runoff modeling. Each sub-basin consists of a lumped SAC-SMA to generate the response components of the runoff. Fast response components are routed over the hillslopes using the unit hydrograph of the sub-basin outlet. Slow response components bypass the hillslopes and are added directly to the sub-basin outlet. Then, a sub-basin-to-sub-basin channel routing is done using kinematic wave routing method. The Illinois River basin at South of Siloam Spring, AR is chosen as the study test basin. Eleven years of grid-based multi-sensor (NEXRAD+gauge) precipitation dataset are used for the calibration and verification periods. Weekly adjustment factors to account for vegetation type and cover are estimated using Global Vegetation Fraction (GVF) product from NOAA/NESDIS/STAR to produce PET demand for the watershed modeling. Several calibration scenarios are tested using the Shuffled Complex Evolution Method (Duan et al, 1992) and Multi Step Automatic Calibration Scheme (Hogue et al, 2000). Results of different calibration scenarios, besides simulations of a priori parameters of the SAC-SMA model (Koren et al, 2000) are shown. The results show that the semi-distributed version of the SAC-SMA has the potential to improve the streamflow simulations.
H21A-0186
Relating Streamflow Recession to Soil and Watershed Properties
By exploring the empirical relationship between streamflow recession data from USGS streamgauges and soils data from the Soil Survey Geographic (SSURGO) Database, we test the general hypothesis that a relationship exists between soil properties derived from soil data integrated across a watershed and streamflow recession parameters. This empirical study provides a background for developing a priori parameters with SSURGO soil survey data for parameterizing watersheds for distributed hydrologic modeling and streamflow prediction. Using a random sample of 48 watersheds from across the continental United States, we found a significant correlation between streamflow recession parameters and soil sensitivity, which we define as a combination of hydraulic conductivity, porosity, and soil depth. Our work improves the understanding of streamflow generation processes by exploring how this correlation changes between different geographic regions and climate regimes.
H21A-0187
Analysis of the Two-Dimensional Effect in the H/V Spectral Ratio Method for Bedrock Depth Estimation
The horizontal-to-vertical spectral ratio (H/V) method for analysis of microtremor surveys is widely used as a practical and economical way to obtain bedrock depth information for hydrogeological studies and to assess potential site effects for earthquake damage. This method is effective because it relies on the resonance frequency of vibration of the sedimentary layers, even when there are certain uncertainties in the vibration's amplitude. Because of its ease of use, many scientists use the H/V method as an exploration tool since, in the one-dimensional case, the resonance frequency is linked to the shear-wave velocity and the sediment thickness. The accuracy of the H/V method suffers, however, when the sediment-bedrock interface has small-wavelength undulations because the method assumes a horizontally layered earth. To investigate the effect of unevenness of the bedrock interface on the peak H/V value and to quantify the applicability of this method for estimation of depth to bedrock at a watershed scale, we conducted a series of tests using a finite-difference time-domain numerical model. The general conclusion is that when the wavelength of the bedrock undulation is less than three times the depth to bedrock, there can be a significant effect on the estimate of the fundamental frequency as determined by the H/V method.
H21A-0188
An analytical model of the effects of catchment hypsography on the flood frequency distribution
The role of the temperature regime on the flood frequency distribution in alpine basins is examined through a minimalist analytical model of the flood formation mechanisms. We represent rainfall as a marked Poisson process of storm arrivals in time with rate λ in which the depth h of each storm is modelled as an exponentially distributed random variable with mean α. Each rain event occurs in a liquid form over a fraction of the basin area, the contributing area, while in the upper part of the basin the precipitations are in a solid form and do not contribute to the discharge at the outlet. The contributing area depends on the elevation of the snow line at the date when the event occurs. As a consequence, the watershed elevation characteristics, which are represented through a simple hypsographic curve, play a role in the derivation of the flood frequency curve. Under these assumptions the probability distribution of the annual maxima of discharge is analytically derived. Reasonable representations of the relations between moments of the distribution and mean basin elevation are obtained. They are in good agreement with the experimental trends and prove the ability of the model to mimic the dependency of flood formation on elevations. The relevance of this work derives from its simple structure, that allows the user to characterize each basin with a set of fundamental descriptors. Behaviour of basins in transition regions and mountainous basins sensitivity to climate change are possible interesting applications of the model.
H21A-0189
Applying C2VSIM, an integrated hydrologic model of California's Central Valley, to assess local and regional impacts of conjunctive use projects
Applications integrating land surface and root-zone processes with groundwater and surface water flow at the watershed scale (e.g. IWFM, SHE and the MODFLOW Farm Process) provide more robust and flexible models by internally calculating inter-process fluxes. The direct linkage of multiple processes within a single application facilitates model calibration because each model parameter is constrained by a greater number of observations and observation types, and the sensitivity of model results to the components of each process can be used to focus improvement efforts on the parameters and processes where they will provide the greatest impact on simulation results. Integrated models can also reduce the complexity of model input files by accepting generally available spatial and time-series data sets (e.g. precipitation, evapotranspiration and land use) and internally computing inter-process fluxes, eliminating the need to preprocess large amounts of data or to develop quasi- steady-state linkages between two or more process-level models. This increases model flexibility, facilitating for example analysis of watershed response to changes in precipitation or land use, while reducing the potential for introducing errors. Some of the advantages of integrated watershed-scale applications are demonstrated by applying the California Central Valley Groundwater-Surface Water Simulation Model (C2VSIM), which simulates land-surface, groundwater and surface water flow in the alluvial portion of California's Central Valley, to assess local and regional impacts of conjunctive use projects. http://baydeltaoffice.water.ca.gov/modeling/hydrology/IWFM/index.cfm
H21A-0190
Innovative Springshed Mapping for Trout Stream Management
Springs in the Paleozoic karst lands of southeastern Minnesota provide water to trout streams, whose existence depends on a steady supply of clear, cold water. 14 Minnesota springsheds or recharge areas of the source springs have been mapped by dye tracing over the past three decades. The boundaries of these springsheds bear little resemblance to surface watershed boundaries. However, dye traces are time and labor intensive. The 14 defined springsheds represent only 8% of the 173 designated trout streams in southeastern Minnesota. Water management associated with the increasing human impacts of intensive agriculture, new water demands, climate change, and landscape alteration requires more efficient means of defining the remaining springsheds. Methods are being investigated to improve the efficiency of dye tracing. We are also investigating alternative methods of defining springsheds. One promising method combines gamma logs of a few water wells with conventional drilling logs to construct detailed structural contour maps. Conduits present in southeastern Minnesota karst systems generally run parallel to the local, sub-horizontal bedding planes, and multiple joint systems enable water to move down-dip within and along local bedrock units. Another method uses temperature and conductance loggers to augment conventional spring hydrographs. Comparison of spring thermo- and chemo-graphs allows better hydrograph separation. This refined hydrograph analysis allows improved estimates of basin size and connectivity. Additional methods involve the use of major element chemistry and stable isotopes of hydrogen and oxygen in combination with pre-existing water quality data to look for seasonal influxes of recharge water. The results of these alternative methods can then be compared with new and pre-existing dye trace results to gauge their effectiveness in delineating springsheds. The integrated use of a variety of hydrogeologic tools allows a better understanding of these highly dynamic karst systems.
H21A-0191
Modeling Surface Runoff in the Mustang Creek Basin, California, Using the Soil and Water Assessment Tool (SWAT)
A study was made of the Mustang Creek Basin, a small agricultural basin within the Merced River Basin of the San Joaquin Valley, California, using the 2005 version of the Soil and Water Assessment Tool (SWAT). This research is part of a national study being conducted by the National Water-Quality Assessment program of the U.S. Geological Survey on the processes affecting agricultural chemical movement through hydrologic systems. The SWAT model was used to simulate streamflow in the Mustang Creek Basin using a set of model inputs derived and modified from various data sources. These data sources include soils, land use, precipitation, and elevation data. The model was calibrated using daily-flow data for 29 days in February 2004, and validated using daily-flow data for 58 days in January and February 2005. Measured streamflow from a U.S. Geological Survey gaging station on Mustang Creek was used for calibration and validation. Results of the simulated monthly streamflow had a Nash Sutcliffe efficiency (NSE) value during the calibration period of 0.72, indicating a reasonable fit (the best fit will have a NSE of 1). However, the model was unsuccessful in simulating streamflow during the validation period, with a NSE value of 0.33. This lack of a successful simulation is probably due to the limited quantity and quality of measured streamflow data available for model validation. This application of the SWAT model did not simulate daily discharge on Mustang Creek well; however, additional daily-flow data for a longer period of record may enhance the capability of adequately simulating streamflow in the watershed.
H21A-0192
Statistical Analysis of Agricultural BMP Effectiveness using Measured and Simulated Water Quality Data
Detecting changes in surface water quality due to the implementation of best management practices poses an interesting statistical problem. Several techniques are presented that range from simple, one-site, univariate methods, to more sophisticated multivariate techniques that incorporate both measured and simulated water quality. The Cannonsville, NY watershed, a basin that has implemented widespread BMP's over the last decade, is presented as a case study. Results from the statistical analyses demonstrate that BMP implementation has led to significant reductions in dissolved phosphorus loads near the watershed outlet over the course of a decade. The results also demonstrate the necessity of using multiple techniques to assess BMP effectiveness at the watershed level.
H21A-0193
Assessment of Uncertainty of Forest Road Hydrology Modeling with the Distributed Hydrology Soil Vegetation Model (DHSVM)
The effects of forest roads on catchment hydrology and sediment production continue to be the focus of concern for impacts to aquatic habitat. However, these processes are complex and difficult to measure at catchment scales. Consequently, the effects of forest roads on watershed hydrology are often studied using distributed hydrologic models. We examined a popular distributed hydrology model used in the Western USA, the Distributive Hydrology Soil Vegetation Model (DHSVM), to predict changes in peak flows, storm run-off volume, and interception of sub-surface flow from forest roads. We apply DHSVM to a 630 hectare watershed in the headwaters of Oak Creek in the McDonald/Dunn Research Forest, managed by the College of Forestry, Oregon State University. The Generalized Likelihood Uncertainty Estimation (GLUE) approach was used to determine the uncertainty of model output for 10,000 model structures. Estimates of road ditchflow and streamflow from the GLUE assessment of DHSVM were compared to observed road ditchflow and streamflow for 2003-2006. Generally DHSVM simulations provided a reasonable fit to the time series for Oak Creek streamflow. The fit of the time series data diminished with spatial scale and for road ditchflow locations. From the GLUE assessment the percentage of DHSVM model structures that exceeded a Nash/Sutcliffe Efficiency of 0.5 were 44% for Oak Creek streamflow but reduced to a range of 0-9% for road locations. The conceptual model for road interception used within DHSVM did not accurately predict road ditchflow throughout the catchment. Given the many uncertainties observed from DHSVM road ditchflow results we question whether cumulative effects analysis of road influences on peak flows and storm volumes with DHSVM is appropriate for catchments with highly variable hillslope responses, such as Oak Creek.
H21A-0194
Characterization of sediment sources in the Le Sueur River watershed, southern Minnesota
Low-relief, tectonically quiescent landscapes are fundamentally important for fulfilling economic and societal needs. In historic times humans have, no doubt, greatly enhanced sediment yield from these landscapes to an extent that is comparable to, if not greater than that from many tectonically active areas, yet they have received little attention from the geomorphic community. Currently, our understanding of integrated watershed sediment dynamics in low-relief landscapes is insufficient for proper development of widely-applicable, predictive models of sediment sources, transport and storage. This insufficiency limits our ability to deconvolve anthropogenic versus natural drivers of landscape change and limits our ability to develop reasonable management goals in these economically important landscapes. The Le Sueur River, southern Minnesota, is impaired by sediment under the Clean Water Act and is one of the largest contributors of sediment and colloidal phosphorus to the Minnesota River. The goal of this project is to develop an integrated sediment budget for the Le Sueur watershed that can be used to inform best management practices to achieve the greatest reduction of in-stream turbidity. We are combining field mapping and granulometry with geomorphic modeling, LiDAR technology, and strategic cosmogenic nuclide analysis to determine the proportion of the suspended load that is coming from erosion of uplands, streambanks, bluffs, and ravines, respectively. Preliminary analysis of the relationship between local channel slope and contributing drainage area indicates two distint knickzones migrating up the Le Sueur River system. A graded reach (river km 35-80) is bound on either side by the two knickzones. Our experimental design will test two competing hypotheses. If erosional hotspots are linked to knick migration, we hypothesize that most proximate to the lower extent of the knicks: a) ravines will be steeper and contain convex banks and b) erosion rates from ravines and tributaries will be highest. If, on the other hand, the erosional system is being driven by a distributed, watershed-scale process, such as land use or climate change, we expect that ravine morphology will be consistent and erosion rates will be elevated throughout the basin.
H21A-0195
An Integrated Sediment Budget for the Le Sueur River in Southern Minnesota
The Le Sueur River is the primary sediment contributor to the Minnesota River, which is impaired by sediment under the Clean Water Act. The necessary first step toward developing a watershed management plan is identifying sediment sources and sinks throughout the basin. The focus of this study is to quantify the contribution of sediment from different sources in order to understand the broader problems affecting the larger Minnesota Basin. The lower Le Sueur River is actively incising in response to late Pleistocene baselevel fall on the Minnesota River, a result of the catastrophic draining of glacial Lake Agassiz. Potential sediment sources along the Le Sueur include high bluffs composed of highly consolidated Pre-Illinoisian, Illinoisian and Wisconsinin till and alluvial sediment, growing ravines and gullies, sediment exchange between the active channel and floodplains, as well as upland agricultural fields, which account for over 90% of the land surface. We applied side-scanning LiDAR to quantify the amount of sediment being eroded from the high bluffs and banks. To understand meander migration rates as well as ravine and gully retreat rates, and to supplement the understanding of bluff retreat rates, we have used aerial LiDAR along with comparative analysis of historic aerial orthophotos. The aerial LiDAR will also be used to calculate the volume of sediment being stored in the floodplains and terraces along the river, which will be used to create a floodplain exchange model. Preliminary results indicate that bluff retreat rates could be as high as 0.4 meters per year and meander bend migration rates up to 3 meters per year in isolated reaches of the watershed. Data collected from each potential source will be combined to create an overall sediment routing model for the watershed.
H21A-0196
Hydrogeologic Characterization of Fractured Crystalline Bedrock on the Southern Part of Manhattan, New York, Using Advanced Borehole Geophysical Methods
ABSTRACT. Advanced borehole-geophysical methods were used to assess the hydrogeology of fractured crystalline bedrock in 31 of 64 boreholes on the southern part of Manhattan Island, N.Y. The majority of boreholes penetrated gneiss, schist, and other crystalline bedrock, and had an average depth of 591 ft (180 m) below land surface (BLS). In this study we use a combination of advanced and conventional borehole geophysical logs, and hydraulic measurements to characterize the fractured-rock ground-water flow system in southern Manhattan, N.Y. Borehole-geophysical logs collected in this study included natural gamma, single-point-resistance (SPR), short-normal resistivity (R), mechanical and acoustic caliper, magnetic susceptibility, borehole-fluid temperature and resistivity, specific conductance (SpC), dissolved oxygen (DO), pH, redox, heat-pulse flowmeter (at eight selected boreholes), borehole deviation, acoustic and optical televiewer (ATV and OTV), and directional borehole radar (at 23 selected boreholes). A new geophysical probe that collects multiple fluid parameters, included fluid- temperature, SpC, DO, pH, and redox logs; these were used to help delineate transmissive fractures in the boreholes. All boreholes penetrated moderately fractured bedrock that contained medium and large fractures. A total of 208 large fractures were delineated in the 31 boreholes logged with the OTV. Stereonet analysis of the large fractures indicates most are part of a subhorizontal population cluster with a mean orientation of N43 degrees E, 07 degrees SE and a smaller secondary population cluster dipping toward the northwest. A total of 53 faults were delineated with two major population clusters--one with a mean orientation of N12 degrees W, 66 degrees W and the other with a mean orientation of N11 degrees W, 70 degrees E. Foliation was fairly consistent throughout the study area with dip azimuths ranging from northwest to southwest and dip angles ranging from 30 to 70 degrees. A total of 59 bedrock boreholes had specific-capacity test data analyzed for total borehole transmissivity. The bedrock transmissivity ranged from 0.7 to 870 feet squared per day (0.07 to 81 meters squared per day) in the study area. The majority of boreholes (69 percent) had transmissivities less than 100 feet squared per day (9.3 meters squared per day). Heat-pulse flowmeter logging was completed at eight boreholes. A total of 77 transmissive fractures were delineated at the 8 boreholes. Stereonet analysis of these transmissive fractures indicates two population clusters of fractures--one with a mean orientation of N11 degrees E, 14 degrees SE and the other with a mean orientation of N23 degrees E, 57 degrees NW. These data suggest that the fractured-rock ground-water-flow system in southern Manhattan is interconnected, dominated by subhorizontal fractures, and fractures that dip moderately to the northwest.
H21A-0197
Is the Relationship Between Peak Runoff Discharge and Land-Use Pattern Convex? A Numerical Experiment with the IHLUO Model
This study presents a modeling methodology to understand the impacts of spatial land use patterns on watershed hydrology and the characteristics of their interrelationship. An integrated hydrological and land-use optimization (IHLUO) model was used to delineate optimal land patterns that minimize peak runoff under various land use configurations. The solutions (i.e., peak runoff) obtained from the IHLUO model were analyzed using a Weibull distribution, in order to estimate the global optimum value and its confidence intervals and to assess their convergence toward the global optimum. The result supports the case for the global optimality, suggesting convexity of the hydrological function with optimized land-use variables. The convexity of the function is further argued, examining the physics of hydrological processes (the SCS-CN number) that are used in the hydrological simulation.
H21A-0198
Calibration Of The Variable Infiltration Capacity (VIC) Model For Streamflow Estimation Along The Upper Assiniboine River Basin
The Canadian Prairie is characterized by slightly complex hydrological processes that lend much of the explanation to the previous unsuccessful hydrological modelling studies that had been undertaken in this region. In order to comprehend the dynamics of drought occurrence, persistence and termination on the Prairie, there is a need for an accurate quantification of the energy and water budgets within this region. Methods currently exist to assess basin budget studies based on atmospheric analysis; however, it is important that land surface budgets support these results in order to understand surface influences on drought evolution. Hydrological models equipped with energy closure solutions are suited to this role and currently, there are a number of these models available (e.g. MESH (formally WATCLASS), TOPLATS, VIC) and it is the goal of this research to evaluate a number of these to determine those most suited to the analysis of prairie drought. One such model, that is gaining considerable attention, generates runoff based on the Xinanjiang/Arno/VIC variable infiltration capacity concept. The Variable Infiltration Capacity (VIC) model was selected for simulating streamflow and low flow over the 13,000km2 drainage catchment of the Upper Assiniboine Basin with its outlet at Kamsack. Pre-calibration stages of the VIC model revealed more fluxes being produced than required with frequent shifts in the amplitudes of peak flows when undertaking qualitative comparisons of the simulated and observed hydrographs for the various periods under investigation. To enhance the performance of this model, a Multi- Objective Complex Evolution (MOCOM-UA) scheme was applied to achieve its calibration. This technique is essentially based on three fundamental concepts; population, rank-based selection and competitive evolution (Yapo et al, 1998). The results obtained subsequent to the calibration of the VIC model support the conclusion that, with better parameter estimation of a distributed model; there is a significant improvement in its streamflow forecast accuracy.
H21A-0199
Integrating Thermokarst Features in the UpperKuparuk and Imnvait Basin Using Remote Sensing and Ground-Truth Data into TOPOFLOW
Surface runoff fluxe in the Arctic fresh water hydrological cycle can be modified by the presence of thermokarst; the visible surficial manifestation of permfrost degradation including ground subsidence, thermokarst ponds, beaded streams and appearance of water tracks. Predicting and characterizing potential hydrological response is an important component for engineering infrastructure appropriate for the climatic conditions. The Upper Kuparuk and Imnavait basins north of the Brooks Range in Alaska are part of a long-term monitoring effort, and provide an opportunity to pair hydrological studies and high-resolution topography models with remotely sensed data, to create a quality spatial distributed perspective. Imagery from EO-1's Advanced Land Imager (ALI) captured in August 2004 was used to contrast techniques for identifying and quantifying thermokarst features. Principle component analysis and a MNF transformation using a combination of existing, field and lab-derived spectra of vegetation and land-cover were used to classify water tracks, and compared to ground based surveys from 2007. Thermokarst ponds and the beaded stream formation were quantified using airphotos taken in 2004 and 2007. A geospatial database was constructed to examine the relationships between the imagery, existing digital elevation model and field data collected in 2007. The spatial extent of the thermokarst features and associated characteristics will be merged into the processed-based, spatially-distributed hydrological model TopoFlow, based upon the joining of ARHYTHM (ARctic HYdrologic and THermal Model), and a D8-based rainfall- runoff model.