Search

SWAT+ is a completely restructured version of the Soil and Water Assessment Tool (SWAT) that was developed to face present and future challenges in water resources modeling and management and to meet the needs of the worldwide user community. It is expected to improve code development and maintenance; support data availability, analysis, and visualization; and enhance the model's capabilities in terms of the spatial representation of elements and processes within watersheds. The most important change is the implementation of landscape units and flow and pollutant routing across the landscape. Also, SWAT+ offers more flexibility than SWAT in defining management schedules, routing constituents, and connecting managed flow systems to the natural stream network. To test the basic hydrologic function of SWAT+, it was applied to the Little River Experimental Watershed (Georgia) without enhanced overland routing and compared with previous models. SWAT+ gave similar results and inaccuracies as these models did for streamflow and water balance. Taking full advantage of the new capabilities of SWAT+ regarding watershed discretization and landscape and river interactions is expected to improve simulations in future studies. While many capabilities of SWAT have already been enhanced in SWAT+ and new capabilities have been added, the model will continue to evolve in response to advancements in scientific knowledge and the demands of the growing worldwide user community.

A spatially distributed representation of basin hydrology and transport processes in hydrologic models facilitates the identification of critical source areas and the placement of management and conservation measures. Floodplains are critical landscape features that differ from neighboring uplands in terms of their hydrological processes and functions. Accordingly, an important step in watershed modeling is the representation of floodplain and upland areas within a watershed. The aim of this study is (1) to evaluate four floodplain-upland delineation methods that use readily available topographic data (topographic wetness index, slope position, uniform flood stage, and variable flood stage) with regard to their suitability for hydrological models and (2) to introduce an evaluation scheme for the delineated landscape units. The methods are tested in three U.S. watersheds ranging in size from 334 to 629 km2 with different climatic, hydrological and geomorphological characteristics. Evaluation of the landscape delineation methods includes visual comparisons, error matrices (i.e. cross-tabulations of delineated versus reference data), and geometric accuracy metrics. Reference data was obtained from SSURGO (Soil Survey Geographic database) and FEMA (Federal Emergency Management Agency) flood maps. Results suggest that the slope position and the variable flood stage method work very well in all three watersheds. Overall percentages of floodplain and upland areas allocated correctly were obtained by comparing delineated and reference data. Values range from 83 to 93 % for the slope position and from 80 to 95 % for the variable flood stage method. Future studies will incorporate these two floodplain-upland delineation methods into the subwatershed-based hydrologic model SWAT to improve the representation of hydrological processes within floodplain and upland areas. This article is protected by copyright. All rights reserved.

Hydrologic interaction between surface and subsurface water systems has a significant impact on water quality, ecosystems and biogeochemistry cycling of both systems. Distributed models have been developed to simulate this function, but they require detailed spatial inputs and extensive computation time. The SWAT model is a semi-distributed model that has been successfully applied around the world. However it has not been able to simulate the two way exchanges between surface water and groundwater. In this study, the SWAT-LU model – based on a catena method that routes flow across three landscape units (the divide, the hillslope and the valley) – was modified and applied in the floodplain of the Garonne River. The modified model was called SWAT-LUD. Darcy's equation was applied to simulate groundwater flow. The algorithm for surface water level simulation during flooding periods was modified and the influence of flooding on groundwater levels was added to the model. Chloride was chosen as a conservative tracer to test simulated water exchanges. The simulated water exchange quantity from SWAT-LUD was compared with the output of a 2D distributed model, 2SWEM. The results showed that simulated groundwater levels in the LU adjoining the river matched the observed data very well. Additionally, SWAT-LUD model was able to reflect the actual water exchange between the river and the aquifer. It showed that river water discharge has a significant influence on the surface-groundwater exchanges. The main water flow direction in the river/groundwater interface was from groundwater to river, water flowed in this direction accounted for 65 % of the total exchanged water volume. The water mixing occurs mainly during high hydraulic periods. Flooded water was important for the SW-GW exchange process, it accounted for 69 % of total water flowed from the river to the aquifer. The new module also provides the option of simulating pollution transfer occurring at the river/groundwater int

Integrated river basin models should provide a spatially distributed representation of basin hydrology and transport processes to allow for spatially implementing specific management and conservation measures. To accomplish this, the Soil and Water Assessment Tool (SWAT) was modified by integrating a landscape routing model to simulate water flow across discretized routing units. This paper presents a grid-based version of the SWAT landscape model that has been developed to enhance the spatial representation of hydrology and transport processes. The modified model uses a new flow separation index that considers topographic features and soil properties to capture channel and landscape flow processes related to specific landscape positions. The resulting model is spatially fully distributed and includes surface, lateral and groundwater fluxes in each grid cell of the watershed. Furthermore, it more closely represents the spatially heterogeneous distributed flow and transport processes in a watershed. The model was calibrated and validated for the Little River Watershed (LRW) near Tifton, Georgia (USA). Water balance simulations as well as the spatial distribution of surface runoff, subsurface flow and evapotranspiration are examined. Model results indicate that groundwater flow is the dominant landscape process in the LRW. Results are promising, and satisfactory output was obtained with the presented grid-based SWAT landscape model. Nash–Sutcliffe model efficiencies for daily stream flow were 0.59 and 0.63 for calibration and validation periods, and the model reasonably simulates the impact of the landscape position on surface runoff, subsurface flow and evapotranspiration. Additional revision of the model will likely be necessary to adequately represent temporal variations of transport and flow processes in a watershed.

Sediment delivery from hillslopes to rivers is spatially variable and may lead to long-term delays between initial erosion and related sediment yield at the watershed outlet. Consideration of spatial variability is important for developing sound strategies for water quality improvement and soil protection at the watershed scale. Hence, the Soil and Water Assessment Tool (SWAT) was modified and tested in this study to simulate the landscape transport capacity of sediment. The study area was the steeply sloped Arroio Lino watershed in southern Brazil. Observed sediment yield data at the watershed outlet were used to calibrate and validate a modified SWAT model. For the calibration period, the modified model performed better than the unaltered SWAT2009 version; the models achieved Nash–Sutcliffe efficiency (NSE) values of 0.7 and −0.1, respectively. Nash–Sutcliffe efficiencies were less for the validation period, but the modified model’s NSE was higher than the unaltered model (−1.4 and −12.1, respectively). Despite the relatively low NSE values, the results of this first test are promising because the model modifications lowered the percent bias in sediment yield from 73 to 18%. Simulation results for the modified model indicated that approximately 60% of the mobilized soil is deposited along the landscape before it reaches the river channels. This research demonstrates the modified model’s ability to simulate sediment yield in watersheds with steep slopes. The results suggest that integration of the sediment deposition routine in SWAT increases accuracy in steeper areas while significantly improving its ability to predict the spatial distribution of sediment deposition areas. Further work is needed regarding (i) improved strategies for spatially distributed sediment transport measurements (for improving process knowledge and model evaluation) and (ii) extensive model tests in other well instrumente

Accurate simulation of landscape processes in natural resource models requires spatial distribution of basin hydrology and transport processes. To better represent these processes, a landscape version of the SWAT model has been developed to simulate the runoff, run-on, and infiltration processes that typically occur in different parts of the landscape. The model addresses flow and transport across hydrologic response units prior to concentration in streams, and is capable of simulating flow and transport from higher landscape positions to lower positions. The SWAT landscape model was tested using data collected from a heavily vegetated riparian buffer system in the Atlantic Coastal Plain near Tifton, Georgia. Simulations of surface runoff, lateral subsurface runoff, and groundwater flow for an upland field, a grass buffer, and a sub-divided forested buffer floodplain were generated. Model results and field data indicate that surface runoff was dominant in the upland field, while groundwater flow was dominant in the grass buffer and the floodplain. While average annual surface runoff agreed satisfactorily with observations from the site, annual and monthly simulated values varied considerably fromobserved values. Simulated surface runoff tracked general trends in the observed data, but winter months and extreme events were overestimated while summer months were underestimated. Annual surface runoff predictions at the edge of the upland field varied from the observed data by 11% to 44%. The Nash-Sutcliffe efficiency for annual estimates of surface runoff at the field edge was 0.83 for the three-year calibration period. The results demonstrate the ability of the model to simulate the surface runoff and enhanced infiltration typically associated with riparian buffer systems. Additional revision of the model will likely be necessary to adequately represent redistribution of water between surface, lateral subsurface flow, and groundwater flow.

River basin management requires a spatially distributed representation of basin hydrology and nutrient transport processes. To accomplish this, the Soil and Water Assessment Tool (SWAT) watershed model was enhanced to simulate water flow across discretized landscape units. The model structure more closely reflects the complex controls on infiltration, runoff generation, run-on, and subsurface flow without requiring large computational resources or detailed parameterization. Four landscape delineation methods were compared: lumped, hydrologic response units (HRUs) or hydrotope, catena, and grid. The lumped method using dominant soil and land use and the HRU delineation do not consider landscape position when computing runoff. The catena method routes flow across a representative catena with divide, hillslope, and valley units. The distributed methoddivides the watershed into cells (1 ha each) for routing. All methods were calibrated and validated for the USDA-ARS Brushy Creek watershed (17.3 km2) near Riesel, Texas. The calibration results indicate that measured flow at the basin outlet is similar (daily N-S around 0.65) for all four models, or conversely, the new models (catena and grid) do as well as the existing models (lumped and HRU based) in predicting daily flow at the basin outlet. The advantage of the catena and grid models is that the impacts of spatial changes in land use and BMPs on the hillslope valley continuum can now be more realistically assessed.

The Normalized Difference Vegetation Index (NDVI) derived from satellite data has been applied to various vegetation studies. The objective of this study was to assess the feasibility of using the NDVI response to plant water content to predict ground water level over a watershed located in the Edwards Aquifer of Texas, USA. Results showed that the precipitation data collected inside the watershed were not highly correlated to ground water depth within 10 d of the event, though a 60-foot sinkhole in the study site was expected to collect rainfall and recharge ground water in a short time. Alternatively, the NDVI derived from SPOT-VEGETATION satellite data and potential evapotranspiration (PET) based on the Hargreaves PET model were significantly correlated to ground water depth. Moreover, the stream flow measurements were correlated to ground water level as well. Two simple models were developed for estimating ground water levels in the artesian and recharge zones. Independent validations were performed to verify both models. All three variables (NDVI, PET, and stream flow) were directly or indirectly related to the precipitation. The PET was mainly controlled by air temperature, and the temperature was negatively related to precipitation. The NDVI values were affected by both temperature and precipitation, and the amount of rainfall was strongly correlated to the stream flow. This study initiated a unique approach to surveying ground water level based on satellite information and meteorological data.