Hydrologic and water quality modeling of agricultural watersheds equipped with tile drains using a geographic information system and fractal concepts

Sanghyun Kim, Purdue University


The main purpose of the present study is to extend the physically based rainfall-runoff simulator TOPMODEL (Beven and Kirkby, 1979) to agricultural watersheds equipped with tile drains. Nine possible flow generation scenarios in tile drained basins were identified corresponding to the initial soil moisture deficit, the relative position of the water table with respect to the tile and the drawdown caused by the tile drain. These flow generation mechanisms were used in the development of the model. The capability of the model to simulate the infiltration excess and the saturation excess surface runoff, the tile flow and the subsurface flow responses through the various possible flow generation mechanisms provides a better insight into the hillslope hydrology of the tile drained watersheds. The modeling of the flow through the unsaturated zone was modified using a transfer function instead of a storage-time approach. A regionalized sensitivity analysis (RSA, Hornberger and Spear, 1979) showed that a good performance of the transfer function methodology for the simulation of the flow in the unsaturated zone was obtained. The seasonal variation of the hydrologic characteristics was also investigated by the RSA. A better understanding of the spatial characteristics of the topographic index map, which is important in TOPMODEL simulation, was achieved using the fractal concepts (Mandelbrot, 1982). The fractal dimension analysis was introduced in the digital elevation model (DEM). Scale issues in hydrologic modeling were studied using the two- and three-dimensional analysis of the fractal dimension in three watersheds which have different topographic features. The formation and development of the saturation excess zone was exhibited by the evolution of the boundary of map of successively decreasing values of the topographic index. The variation of the runoff water quality was investigated by time series analysis. The causal relationships between the hydrologic variable, which govern the flow paths and hence the residence times, and water quality parameters were found using the cross correlation function between the prewhitened hydrologic and water quality time series. The good simulation results of the transfer function from the hydrologic parameters to water quality parameters support the causality of this relationship.




Delleur, Purdue University.

Subject Area

Hydrology|Civil engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server