Investigation of rainfall-runoff process in relation to soil physical and hydraulic properties

Loading...
Thumbnail Image

Date

Authors

Sajid, Ali H.

Journal Title

Journal ISSN

Volume Title

Publisher

University of Guelph

Abstract

Non-point pollution sources are intermittent and vary in magnitude spatially and temporally, thus making their quantification difficult. The driving force of non-point source pollution of surface waters is the rainfall-runoff process. The present research was planned with the objectives of investigating rainfall-excess mechanism; identifying the important hydrologic parameters affecting the spatial and temporal variation in runoff generation; providing predictive models for description of spatial variability in soil properties and their influence on runoff generation; evaluating the specificity of various fractal models at various spatial scales; and evaluation of two interpolation techniques for soil properties. A two year soil data, originally taken on an 8-m grid, re-sampled on 16-m grid, were interpolated at 4-m and 1-m resolutions by Kriging and inverse distance weighting models. Different cross-validation indices were applied to evaluate interpolated results. Fractal dimensions of soil properties data were computed at different spatial resolutions using variogram, triangular prism and isarithm techniques. Fractal surface was generated to examine its possible use in the modelling of rainfall-runoff process. Finally, rainfall-excess step in runoff generation to identify parts of the field generating rainfall-excess and the soil characteristics associated with it, was investigated and modeled at various spatial resolutions using Green-Ampt model and Philip's sorptivity approach. The results show that Kriging and IDW models are very competitive to each other; however, Kriging was found to be relatively more stable for hydraulic properties. Interpolation models do not reflect spatial variability and make interpolated data smoother. The spatial interpolated soil properties do not follow fractal laws, and fractal models do not provide a universal fit for soil properties at all spatial resolutions. Various soil properties can share the same fractal dimension. Simulated fractal surface cannot be used in spatial lumped or distributed hydrologic models. At a given time, rainfall-excess flux does not generate all over the watershed. Different soil properties and their interaction dominate rainfall-excess mechanism at different spatial resolutions. Initial soil water deficit, topography, spatial connectivity, and rainfall characteristics are dominant factors affecting spatial distribution of observed rainfall-excess areas. Least permeable areas are the most likely to generate rainfall-excess runoff in modeling approach. Rainfall-excess process has been found to be highly sensitive to sorptivity, followed by hydraulic conductivity and rainfall intensity. The GA infiltration model used with Philip's sorptivity is a satisfactory method to predict rainfall-excess generating area, flux and rate.

Description

Keywords

rainfall-runoff, soil, physical properties, hydraulic properties, spatial variability

Citation