Geomorphologic characteristics and responses of Indiana watersheds

Daniel Joseph Schuller, Purdue University

Abstract

The purpose of this study is to gain insight into the relationship between the geomorphology and the hydrology of Indiana watersheds. This relationship is illustrated in the prediction of runoff hydrographs based on the hillslope and stream channel configuration within a watershed. Twelve Indiana basins ranging in size from 3.0 square miles to 58.0 square miles were selected for this study. Stream network maps were obtained for these watersheds by scanning existing paper maps and then manually vectorizing the resulting raster based images. Using the vectorized networks as input, a geographic information system, Fortran codes, and Matlab routines were employed to extract and compute the geomorphologic characteristics of the study watersheds. Common geomorphologic characteristics such as drainage density, mainstream length, total stream length, width function, and Horton ratios were determined for each watershed. Horton's laws were found to be valid for the study watersheds. Furthermore, it was learned that simplified networks would provide nearly identical Horton ratio values. Fractal dimensions for individual streams and stream networks were determined using several methods. The variability of these results precluded the selection of one accurate and robust method for computing each of the fractal dimensions of interest. The self-similarity and self-affinity characteristics of the networks were also investigated. It was determined that Indiana stream networks tend to be self-affine. Geomorphologic instantaneous unit hydrographs (GIUHs) were developed for each watershed using a width function based method. These resulting GIUHs were convolved with derived effective rainfall hyetographs to determine direct runoff hydrographs. The direct runoff hydrographs were compared to the observed runoff hydrographs to determine the utility of the GIUH method. The GIUH model contained three parameters: dispersion, stream celerity, and hillslope roughness. It was found that the dispersion parameter had little effect on the shape of the resulting hydrograph. The stream celerity and the hillslope roughness were significant factors in determining the shape of the resulting hydrograph. Due to differences in topography, it is impossible to conclude that either hillslope effects or network effects are more pronounced in different size watersheds. It was determined that the GIUH method produces reasonable results but that it cannot be applied to ungaged basins due to the wide variability in parameter values observed in the study. The last part of the investigation indicated that researchers and practitioners can use simplified width functions in the GIUH model to reduce the computational intensity of the method.

Degree

Ph.D.

Advisors

Rao, Purdue University.

Subject Area

Civil engineering|Hydrology|Geography

COinS