Adapting a physically based hydrologic model with a geographic information system for runoff prediction in a small watershed
Abstract
The physically-based, variable source area model of K. J. Beven and M. J. Kirby (1979) was modified to simulate the runoff produced by short term rainfall events at the basin scale. The model was extended to incorporate tile drainage runoff. The modified version of the model was interfaced with a raster based Geographic Information System (GIS). Tools were developed to extract the topographic parameters from the GIS database, and to transfer the model output into GIS layers in order to identify the critical areas within the watershed that contribute to runoff production. The model was applied to the Animal Science Farm, an 835 acre agricultural watershed, located near Purdue University Campus. A Monte Carlo simulation procedure was adopted to determine the most sensitive parameters. An automatic calibration based on a Quasi-Newton approach was then performed using four rainfall events. The verification phase, using another set of four single event storms, produced a satisfactory fit to the measured streamflows. Finally, the topographic and soil spatial variability were found to be important in predicting peak discharge and runoff volume.
Degree
Ph.D.
Advisors
Delleur, Purdue University.
Subject Area
Civil engineering|Hydrology
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.