Search

Summary The worth of archival, long-term, high spatial and temporal resolution radar-rainfall estimates for hydrologic modeling has not yet been established. In particular, uncertainties remain regarding accuracy at the event time-scale, which is a modeling requirement for mass conservation. To explore this issue, an archival radar-rainfall precipitation dataset produced from the National Reflectivity Composite for the GEWEX Continental-Scale International Project (GCIP) and US National Weather Service, Weather Surveillance Radar 1988-D (WSR-88D) Stage III (SIII) data were compared with data from 982 National Climatic Data Center (NCDC) rain gages throughout the Mississippi River Basin. These two datasets are derived from the same radar-reflectivity observations but they are different. The GCIP dataset was developed from low resolution (5 dBZ bins) national radar mosaic imagery using a systematic procedure that employed one climatologically-derived reflectivity-rainfall relationship, while the SIII radar-rainfall estimates are produced using a processing algorithm developed by the US National Weather Service that employs near-real time adjustment and varied parameterizations in different River Forecast Centers around the Mississippi River Basin. Statistical analyses were performed to compare both radar-rainfall datasets with rain gage data on an event time-scale for the entire year of 1997. The investigation shows that the GCIP dataset, given a change in the equation used to calculate rainfall from reflectivity, has a lower root mean square difference for the majority of gages in four of the five river forecast centers that cover the Mississippi River Basin compared to SIII in 1997.

The need to simulate surface water flows in watersheds with diverse runoff production mechanisms has prompted the Department of Defense to invest in the development of a new, physically based hydrologic model, called the Gridded Surface/Subsurface Hydrologic Analysis (GSSHA) model. GSSHA is a reformulation and enhancement of the two-dimensional, physically based model CASC2D. The GSSHA model simulates stream flow generated by both infiltration-excess and saturation-excess mechanisms, as well as exfiltration, and groundwater discharge to streams. The model employs mass-conserving solutions of partial differential equations and closely links the hydrologic components to assure an overall mass balance. Testing of the model on a watershed with diverse runoff mechanisms indicates that the model is useful for investigating the important streamflow producing processes at the watershed scale and determining the contributions of surface water, saturated groundwater, and unsaturated groundwater to the overall water...

A number of watershed-scale hydrological models include Richards' equation (RE) solutions, but the literature is sparse on information as to the appropriate application of RE at the watershed scale. In most published applications of RE in distributed watershed-scale hydrological modelling, coarse vertical resolutions are used to decrease the computational burden. Compared to point- or field-scale studies, application at the watershed scale is complicated by diverse runoff production mechanisms, groundwater effects on runoff production, runon phenomena and heterogeneous watershed characteristics. An essential element of the numerical solution of RE is that the solution converges as the spatial resolution increases. Spatial convergence studies can be used to identify the proper resolution that accurately describes the solution with maximum computational efficiency, when using physically realistic parameter values. In this study, spatial convergence studies are conducted using the two-dimensional, distributed-parameter, gridded surface subsurface hydrological analysis (GSSHA) model, which solves RE to simulate vadose zone fluxes. Tests to determine if the required discretization is strongly a function of dominant runoff production mechanism are conducted using data from two very different watersheds, the Hortonian Goodwin Creek Experimental Watershed and the non-Hortonian Muddy Brook watershed. Total infiltration, stream flow and evapotranspiration for the entire simulation period are used to compute comparison statistics. The influences of upper and lower boundary conditions on the solution accuracy are also explored. Results indicate that to simulate hydrological fluxes accurately at both watersheds small vertical cell sizes, of the order of 1 cm, are required near the soil surface, but not throughout the soil column. The appropriate choice of approximations for calculating the near soil-surface unsaturated hydraulic conductivity can yield modest increases in the required cell size. Results for both watersheds are quite similar, even though the soils and runoff production mechanisms differ greatly between the two catchments. Copyright © 2003 John Wiley & Sons, Ltd.

This paper presents representative applications and models that can take advantage of spatially distributed data in a geographic information system (GIS) format for watershed analysis and hydrologic modeling purposes. The intention is to inform hydrologic engineers about the current capabilities of GIS, hydrologic analysis modules, and distributed hydrologic models, and to provide an initial guide on implementing GIS for hydrologic modeling. This paper also discusses key implementation issues for individuals and organizations that are considering making the transition to the use of GIS in hydrology. Widespread use of GIS modules and distributed watershed models is inevitable. The controlling factors are data availability, GIS-module development, fundamental research on the applicability of distributed hydrologic models, and finally, regulatory acceptance of the new tools and methodologies. GIS modules and distributed hydrologic models will enable the progression of hydrology from a field dominated by techniques that require spatial averaging and empiricism to a more spatially descriptive science.

Physically based, two-dimensional, distributed parameter Hortonian hydrologic models are sensitive to a number of spatially varied parameters and inputs and are particularly sensitive to the initial soil moisture field. However, soil moisture data are generally unavailable for most catchments. Given an erroneous initial soil moisture field, single-event calibrations are easily achieved using different combinations of model parameters, including physically unrealistic values. Verification of single-event calibrations is very difficult for models of this type because of parameter estimation errors that arise from initial soil moisture field uncertainty. The purpose of this study is to determine if the likelihood of obtaining a verifiable calibration increases when a continuous flow record, consisting of multiple runoff producing events is used for model calibration. The physically based, two-dimensional, distributed, Hortonian hydrologic model CASC2D [Julien et al., 1995] is converted to a continuous formulation that simulates the temporal evolution of soil moisture between rainfall events. Calibration is performed using 6 weeks of record from the 21.3 km 2 Goodwin Creek Experimental Watershed, located in northern Mississippi. Model parameters are assigned based on soil textures, land use/land cover maps, and a combination of both. The sensitivity of the new model formulation to parameter variation is evaluated. Calibration is performed using the shuffled complex evolution method [Duan et al., 1991]. Three different tests are conducted to evaluate model performance based on continuous calibration. Results show that calibration on a continuous basis significantly improves model performance for periods, or subcatchments, not used in calibration and the likelihood of obtaining realistic simulations of spatially varied catchment dynamics. The automated calibration reveals that the parameter assignment methodology used in this study results in overparameterization. Additional research is needed in spatially distributed hydrologic model parameter assignment methodologies for hydrologic forecasting.

Results from five physical hydraulic model studies of riverside water intakes situated along the Missouri River reach between Sioux City, Iowa, and St. Louis, Missouri, are presented. Movable-bed, undistorted Froude-scale models are used to determine the effectiveness of structural modifications in the vicinity of the intake to limit the influx of bed-load sediments. Solutions developed in each case include a series of submerged flow-turning vanes located on the riverward side of the intake. A sediment-barrier wall between the vanes and intake increases the streamwise velocity component, enhancing the effectiveness of flow-turning vanes in maintaining a deep scour trench. Effective solutions determined using physical hydraulic models are verified at the prototype scale, as demonstrated by years of trouble-free operation at locations where the recommended sediment-control measures have been installed. Results presented in this paper provide design guidelines for bed-load sediment control at riverside water-intake structures on sand-bed rivers.

The proliferation of watershed databases in raster Geographic Information System (GIS) format and the availability of radar-estimated rainfall data foster rapid developments in raster-based surface runoff simulations. The two-dimensional physically-based rainfall-runoff model CASC2D simulates spatially-varied surface runoff while fully utilizing raster GIS and radar-rainfall data. The model uses the Green and Ampt infiltration method, and the diffusive wave formulation for overland and channel flow routing enables overbank flow storage and routing. CASC2D offers unique color capabilities to display the spatio-temporal variability of rainfall, cumulative infiltrated depth, and surface water depth as thunderstorms unfold. The model has been calibrated and independently verified to provide accurate simulations of catchment response to moving rainstorms on watersheds with spatially-varied infiltration. The model can accurately simulate surface runoff from flashfloods caused by intense thunderstorms moving across partial areas of a watershed.

[1] Development of a generally applicable rainfall-runoff model and identification of associated model parameters require understanding of connections between physical processes at disparate scales and hydrological similarities between catchments. In this study, we test the hypothesis that understanding of geomorphometric scaling relations can reduce uncertainty when transferring model parameters between catchments when applying the TOPMODEL concept. Scaling relations on contributing area, slope, and contour length were successfully used to scale the topographic index distribution in watersheds located in vastly different regions of the world: Japan, Nepal, Panama, and the United States. Model parameters were identified through calibration of TOPMODEL in the 210-km2 Kamishiiba catchment in Japan. These parameters were transferred to two Sun Koshi River subcatchments in Nepal, namely, the 850-km2 Likhu and 620-km2 Balephi catchments, the 414-km2 Upper Rio Chagres catchment in Panama, and the 37-km2 Town Brook catchment in the United States. Results show how a priori estimates of the most sensitive model parameters can be used to make predictions in poorly gauged or ungauged basins with some degree of confidence provided that scale effects are considered. This result hints at the potential universality of the topographic index distribution scaling relations in catchments where runoff is dominated by subsurface flow.

The Gridded Surface/Subsurface Hydrologic Analysis (GSSHA) model has been developed with research funding from the US Army, the US Environmental Protection Agency (EPA), US Army Corps of Engineer (USACE) civil research programs (LMS and SWWRP), and with funding for civil and military applications to fill an existing need; GSSHA has the ability to explicitly simulate spatially-varied hydrologic processes to solve a variety of common engineering problems. The GSSHA model features two-dimensional overland flow, sediment and water quality transport, coupled to one-dimensional stream flow, sediment and water quality transport, integrated with two-dimensional groundwater flow. GSSHA simulation times can be minimized by running GSSHA on parallel architectures, such as 64-bit multi-processor PCs, or any shared memory resource. Model setup and post-processing are greatly aided by the DoD Watershed Modeling System (WMS). Recent additions/improvements to the GSSHA model include: subsurface pipe networks for urban and agricultural drainage, improved sediment erosion and transport, simulation of wetlands, lakes, and reservoirs, and coupling of GSSHA to the Nutrient Sub-Module (NSM) to allow complex interaction among nitrogen and phosphorous species, with the intent to add more contaminants as modules are developed. GSSHA has been used for analysis and prediction of a wide range of issues and measures including: surface water runoff, soil moisture, groundwater recharge, transport of sediments and associated contaminants, transport of volatile contaminants, management of military training lands, and effects of urbanization on runoff and sediment loading. This paper focuses on improvements of the GSSHA erosion source and sediment transport routines.

The need to simulate surface water flows in watersheds with diverse runoff production mechanisms has led to the development of the physically-based hydrologic model Gridded Surface Subsurface Hydrologic Analysis (GSSHA). GSSHA is a reformulation and enhancement of the two-dimensional, physically based model CASC2D. The GSSHA model is capable of simulating stream flow generated by a variety of sources, including runoff due to infiltration excess and saturated sources areas and seeps, as well as direct interaction between streams and the saturated groundwater. The model employs mass-conserving solutions of partial differential equations. The hydrologic components are closely linked, assuring an overall mass balance. The model has been applied to a diverse variety of projects and has been proven useful for analysis of hydrologic and sedimentation processes, and can provide information needed for designed systems and the potential effects of projects, land-use change, environmental restoration, best management practices, climate change, and related issues. This manual describes the model formulation, model input, and provides information on the practice of hydrologic modeling with GSSHA, and hydrologic modeling in general.

Runoff production in tropical watersheds is governed by a wide variety of potential sources and there have been few rigorous studies to date. The 414 km2 upper Rio Chagres basin offers a unique opportunity to better understand the runoff production mechanisms in tropical watersheds through data analysis and modeling with rainfall and runoff data. Flow data and tipping bucket rain gage data are available at both the basin outlet (Chico gage) and for an 80.6 km2 internal basin location (Piedras gage). Modeling is performed using the Sacramento Soil Moisture Accounting Model (SAC-SMA), calibrated using data from 2000 and verified using data from 2001. The flood event of 28–31 December 2000 was examined in detail. Data analysis and modeling reveal critical threshold storages in the catchment, and anomalously high runoff production at the start of the wet season. This conclusion is supported by field studies that reveal evidence of high storage capacity and dry season water repellency. Observation of discrete quasi-stable baseflows in the upper Rio Chagres is not seen in the internal Rio Piedras drainage, which is shown to exhibit ephemeral behavior year-round. New data collection and monitoring is proposed for the upper Rio Chagres catchment, including measurements of rainfall above canopy, cloud stripping, stemflow, throughfall, soil moisture, interflow, and overland runoff measurements.

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 Department of Defense (DoD) is also charged with managing approximately 200,000 km 2 of land within the United States on military installations and flood control and river improvement projects. The WSG provides the Army with predictions of stream flow and stage, inundated areas, saturated areas, soil moistures, groundwater levels, and contaminant fate and transport. Predictions are provided for anticipated changes in weather conditions, project alternatives and land-use changes. The WSG group uses a variety of models that are supported by the DoD graphical user interfaces (GUI) Watershed Modeling System ( WMS ), Groundwater Modeling System ( GMS ), and Surface water Modeling System ( SMS ). These GUIs are commonly referred to the XMS system. The XMS interfaces support a variety of model classes, from simple lumped-parameter runoff models, to 2-D overland, and 3-D unsaturated groundwater models.

We performed a detailed research comparison in an urban watershed between The U.S. Army Corps of Engineers, Engineer Research and Development Center (ERDC), Gridded Surface/Subsurface Hydrologic Analysis (GSSHA) model and the U.S. Army Corps of Engineers Hydrologic Engineering Center (HEC) HEC-1 watershed model. GSSHA represents a reformulation and enhancement of the CASC2D model. GSSHA has been successfully applied in a number of nonurbanized watersheds. This paper compares the performance of the two-dimensional, distributed parameter model GSSHA against the standard practice conceptual, lumped parameter model HEC-1 for runoff predictions in a small, urbanized watershed. Specific focus is on the calibration and verification results using data from nearly 40 rainfall events in an urbanized watershed. One longstanding criticism of distributed models, such as GSSHA, is that despite more detailed representation of the spatial variability of watershed characteristics and precipitation, their performance has not shown significant improvement over lumped parameter models. Only through marked improvements in runoff prediction can the benefits of the distributed approach be shown superior to the simpler to apply lumped parameter models. Urbanized watersheds provide the best opportunity for this demonstration because they contain a modified drainage network that can have a large effect on the runoff as seen at the watershed outlet. Further, lumped parameter model formulations cannot adequately describe the amount of heterogeneity and the unique spatial organization of an urban watershed. For this reason, accurate urban modeling with widely used standard-practice, conceptual rainfall-runoff models in the United States (i.e. HEC-1, TR20) can be difficult. Statistical analysis of simulation results compared with observed hydrographs clearly demonstrate that GSSHA is superior to standardpractice lumped-conceptual rainfall-runoff models, because of its distributed hydrodynamic formulation. This result has important implications for the development of advanced techniques for runoff modeling and flood prediction in urbanized areas.

[1] The application of physically based hydrologic models implies they properly simulate processes at the computational scale. A chief criticism is that model predictions are compared only to discharge data. The physically based, hydrologic model CASC2D is reformulated such that soil moistures and fluxes can be computed using Richards' equation. The gridded surface subsurface hydrologic analysis (GSSHA) model is calibrated and verified against outlet discharge measurements during the growing season. The verified model is used to simulate an extended period during which measurements of soil moisture are available. Though soil moisture data are not used in the calibration and verification efforts, the model reproduces both the trends and the magnitude of soil moisture during the growing season. With additional formulation enhancements, soil moistures during the nongrowing season are also reproduced within a root-mean-square error of 0.1. However, more work is needed to understand the underprediction of runoff during the nongrowing season.

CASC2D is a distributed-parameter, physics-based hydrologic model. The original single-event formulation of CASC2D includes two-dimensional diffusive-wave overland flow routing, Green and Ampt infiltration, and one-dimensional diffitsive-wave routing in rectangular channels. CASC2D has been extensively enhanced over the past several years. The US Army Corps of Engineers (USACE) is currently using version 1.19. This version has a continuous formulation that includes the effect of soil moisture redistribution on infiltration, calculates evapo-transpiration, and performs soil moisture accounting between rainfall events. Version 1.19 also performs upland erosion based on three sediment sizes, transports the sediment overland to the channel, and routes the sediment through the channel as wash or bed load. Improved overland flow routing routines have been added to increase model stability and decrease run times. Other additions in version 1.19 include a three-layer infiltration algorithm, a variety of channel cross-sections, and integration with the Dept. of Defense Watershed Modeling System (WMS). This version of CASC2D has been rigorously tested, verified, and applied at numerous US Army locations. CASC2D version 1.19 is applicable only in Hortonian watersheds. The model formulation is now being extensively revised to promote its applicability in non-Hortonian or mixed watersheds. The new version includes two-dimensional saturated groundwater flow and one-dimensional vadose zone flow linked to the current overland flow and channel routing routines. Improved dynamic-wave channel routing with hydraulic structures, lakes, wetlands, and detention basins are being incorporated into the model. Future versions of CASC2D will also have a more robust channel erosion and transport scheme, a bank failure algorithm, and the ability to transport sediment through hydraulic structures (i.e., culverts, bridges, etc.). It is anticipated that this new version will be completed in the year 2000. The new model will retain all the features of version 1.19. After testing and verification, the revised model, which is named GEISSHA, will replace CASC2D version 1.19 as the USACE’s working distributed, physics-based watershed hydrodynamics model.

Physically-based distributed parameter hydrologic models are very sensitive to rainfall input. Inaccuracies in the space-time description of rainfall are typically amplified through runoff predictions. Rain gages do not represent ground truth for rainfall at the watershed scale because they are truly point samples while watersheds are more sensitive to the spatial distribution of rainfall. The use of rain gage data necessitates spatial interpolation of the rainfall field. Weather radars provide an indication of areally-averaged precipitation within the radar pulse-volume. Unfortunately, there are no unique relations between the radar observables and the rainfall rate. Furthermore, radar-rainfall estimates are contaminated by a host of random and systematic error sources. In this study, we examine the impact of each separate source of radar-rainfall estimation error on runoff predictions from the hydrologic model CASC2D. Two different approaches are used. First, S-band single- and dual-polarization radar observations of the extreme warm-rain convective storm that impacted Fort Collins, Colorado, on July 28, 1997, are converted to rainfall rates using published algorithms. Secondly, a physically based atmospheric model of convective rainfall is coupled with an active microwave radiative transfer model to simulate radar observations of a thunderstorm. This methodology allows direct assignment of radar-rainfall errors. Results address the relative importance of random and systematic errors, including range effects and vertical profile of reflectivity, on runoff predictions. Additionally, simulations driven by rain gage and radar-rainfall input are compared to illustrate the impact of the different sampling characteristics on model performance.

This final report documents the results obtained through the final phase of the research study. It consists of two papers that have been submitted for publication based on the research. The objective of this project is to quantitatively evaluate the worth of radar-rainfall estimates for physically based hydrologic modeling. These two papers contain valuable results. To summarize, our study has shown that the physical relationship between radar reflectivity and rainfall rate begins to disappear at a range of approximately 60 km from the radar site for radars with approximately 1 degree bean width. At farther ranges, radar returns are significantly affected by the horizontal and vertical gradients of cloud water. This result was verified through careful testing using a simulation methodology wherein convective rainfall is simulated using the Advanced Regional Prediction Model (ARPs), a radar simulator, and the two dimensional hydrologic model CASC2D. Techniques to correct runoff predictions for likely errors using a Bayesian approach are suggested.

[1] The primary advantage of radar observations of precipitation compared with traditional rain gauge measurements is their high spatial and temporal resolution and large areal coverage. Unfortunately, radar data require vigorous quality control before being converted into precipitation products that can be used as input to hydrologic models. In this study we coupled a physically based atmospheric model of convective rainfall with an active microwave radiative transfer model to simulate radar observation of thunderstorms. We used the atmospheric model to simulate a well-documented tornadic supercell storm that occurred near Del City, Oklahoma, on 20 May 1977. We then generated radar observations of that storm and used them to evaluate the propagation of radar rainfall errors through distributed hydrologic simulations. This physically based methodology allows us to directly examine the impact of radar rainfall estimation errors on land-surface hydrologic predictions and to avoid the limitations imposed by the use of rain gauge data. Results indicate that the geometry of the radar beam and coordinate transformations, due to radar-watershed-storm orientation, have an effect on radar rainfall estimation and runoff prediction errors. In addition to uncertainty in the radar reflectivity versus rainfall intensity relationship, there are significant range-dependent and orientation-related radar rainfall estimation errors that should be quantified in terms of their impact on runoff predictions. Our methodology provides a tool for performing experiments that address some operational issues related to the process of radar rainfall estimation and its uses in hydrologic prediction.