A. van Griensven

SWAT: Model Use, Calibration, and Validation

SWAT (Soil and Water Assessment Tool) is a comprehensive, semi-distributed river basin model that requires a large number of input parameters, which complicates model parameterization and calibration. Several calibration techniques have been developed for SWAT, including manual calibration procedures and automated procedures using the shuffled complex evolution method and other common methods. In addition, SWAT-CUP was recently developed and provides a decision-making framework that incorporates a semi-automated approach (SUFI2) using both manual and automated calibration and incorporating sensitivity and uncertainty analysis. In SWAT-CUP, users can manually adjust parameters and ranges iteratively between autocalibration runs. Parameter sensitivity analysis helps focus the calibration and uncertainty analysis and is used to provide statistics for goodness-of-fit. The user interaction or manual component of the SWAT-CUP calibration forces the user to obtain a better understanding of the overall hydrologic processes (e.g., baseflow ratios, ET, sediment sources and sinks, crop yields, and nutrient balances) and of parameter sensitivity. It is important for future calibration developments to spatially account for hydrologic processes; improve model run time efficiency; include the impact of uncertainty in the conceptual model, model parameters, and measured variables used in calibration; and assist users in checking for model errors. When calibrating a physically based model like SWAT, it is important to remember that all model input parameters must be kept within a realistic uncertainty range and that no automatic procedure can substitute for actual physical knowledge of the watershed.

Fit-for-purpose analysis of uncertainty using split-sampling evaluations

Abstract An uncertainty assessment method for evaluating models, the Sources of UNcertainty GLobal Assessment using Split SamplES (SUNGLASSES), is presented, which assesses predictive uncertainty that is not captured by parameter or other input uncertainties. The method uses the split sample approach to generate a quantitative estimate of the fit-for-purpose of the model, thus focusing on the purpose for which the model is used. It operates by comparing the output to be used for decision making to its observed counterpart and the associated uncertainty. The described method is applied on a Soil Water Assessment Tool (SWAT) model of Honey Creek, a tributary of the Sandusky catchment in Ohio, USA. Water flow and sediment loads are analysed. In this case study the uncertainty estimated by the proposed method is much larger than the typically estimated parameter uncertainty.

Distributed computation of large scale SWAT models on the Grid

The increasing interest in larger spatial and temporal scale models and high resolution input data processing comes at a price of higher computational demand. This price is evidently even higher when common modeling routines such as calibration and uncertainty analysis are involved. Likewise, methods and techniques for reducing computation time in large scale socio-environmental modeling software is growing. Recent advancements in distributed computing such as Grid infrastructure have provided further opportunity to this effort. In the interest of gaining computational efficiency, we developed generic tools and techniques for enabling the Soil and Water Assessment Tool (SWAT) model application to run on the EGEE (Enabling Grids for E-science projects in Europe) Grid. Various program components/scripts were written to split a large scale hydrological model of the Soil and Water Assessment Tool (SWAT), to submit the split models to the Grid, and to collect and merge results into single output format. A three-step procedure was applied to take advantage of the Grid. Firstly, a python script was run in order to split the SWAT model into several sub-models. Then, individual sub-models were submitted in parallel for execution on the Grid. Finally, the outputs of the sub-basins were collected and the reach routing process was performed with another script executing a modified SWAT program. We conducted experimental simulations with multiple temporal and spatial scale hydrological models on the Grid infrastructure. Results showed that, in spite of computing overheads, parallel computation of socio-environmental models on the Grid is beneficial for model applications especially with large spatial and temporal scales. In the end, we conclude by recommending methods for further reducing computational overheads while running large scale model applications on the Grid.

Linking SWAT and SOBEK Using Open Modeling Interface (OpenMI) for Sediment Transport Simulation in the Blue Nile River Basin

Computer models assist basin-scale decision making by taking into account upstream-downstream interdependencies. The SWAT (hydrological) model code was developed into an OpenMI-compliant version and linked with the SOBEK (hydrodynamic) model to extend SWATs simulations of basin-scale streamflow and sediment transport. The development of an OpenMI-compliant version of SWAT involved reorganizing the SWAT model code and wrapping it with the OpenMI wrapper utility. The modified SWAT model was linked to the SOBEK model and applied to simulate sediment transport in the Blue Nile River basin. The SWAT model simulated the streamflow and soil erosion in the upstream catchment, while the SOBEK model routed the streamflow and sediment downstream to the basin outlet. Prior to the linking, both the SWAT and SOBEK models were individually calibrated. The results showed that the coupled models simulated the observed hydrodynamics and sediment deposition due to backwater effects, which would not be possible with the SWAT model alone. The developed OpenMI-compliant SWAT model can further be linked to groundwater, climate change, and socioeconomic models to address integrated water resources management needs.