|Beskow, S -|
|Mello, C -|
|Avanzi, J -|
Submitted to: American Geophysical Union
Publication Type: Abstract Only
Publication Acceptance Date: June 4, 2010
Publication Date: August 8, 2010
Citation: Norton, L.D., Beskow, S., Mello, C.R., Avanzi, J.C. 2010. A Simple Hydrological Simulation Tool for Watershed Planning and Application to a Brazilian Watershed [abstract]. American Geophysical Union. Meeting of the Americas. August 8-13, 2010, Foz do Iguacu, Parana, Brazil. 2010 CD ROM. Technical Abstract: Hydrological modeling of watersheds is a convenient and easy way to evaluate the effects of changing land-use or management strategies on erosion, stream flow and sediment yield for various purposes such as designing downstream structures or impoundments, or implementing strategies to control soil erosion. Geographic information systems (GIS) easily allow hydrological modeling both spatially and temporally. However, many areas of world have limited detailed data sets on rainfall, stream flows and/or soils to drive many existing complex soil erosion or hydrological models. In order to be able to model watersheds with limited data, a simple hydrological tool in a GIS framework was developed to be able to model hydrological parameters in watersheds. The tool contains three modules: 1) the first includes surface runoff, sub-surface flow, base flow and capillary rise, 2) the second module includes the a routing method to move water from adjacent cells, and 3) a module based on the Muskingum-Cunge method to linearly move flows into a channel. Being a grid based tool within a GIS, the soil water balance over time is computed for each grid cell based on topography, soil characteristics, climate and land-use. Sensitivity analysis to adjust input parameter ranges and uncertainty analysis of parameters was conducted using data from an experimental watershed in Minas Gerais State, Brazil. A Monte-Carlo simulation of the 13 input parameters was conducted to optimize the input parameters for use in prediction of discharge. Four parameters proved to be the most sensitive and included two related to base flow, the surface runoff and the initial soil moisture. These four parameters can not be easily measured in the field for the entire watershed and therefore, need to be estimated experimentally for a given watershed though optimization. The tool was calibrated using data from the experimental watershed on precipitation and stream flow. The data yielded a Nash-Sutcliffe (NS) coefficient of 0.82 for the calibration period while the validation period gave a NS of 0.76. The tool was deemed satisfactory for estimating stream flow for similar watersheds where it can be calibrated with limited existing data. This tool may be coupled with an erosion model such as the Universal Soil Loss Equation (USLE) or the Water Erosion Prediction Project (WEPP) model or a water quality model, to easily predict downstream flow, sediment yield or contaminate loading from a watershed both spatial and temporally. The impact of this research is that a simple tool, requiring limited data on soil, land-use, topography and climate can be used to identify problem areas within a watershed where land-use may be changed to improve downstream loadings of sediment and other contaminates of concern.