Date of Award:


Document Type:


Degree Name:

Doctor of Philosophy (PhD)


Watershed Sciences

Committee Chair(s)

Gerald F. Gifford


Gerald F. Gifford


R. H. Hawkins


G. E. Hart


K. Johnson


Grazing is a primary land use in much of the western United States, but little is known about grazing impacts on water quality. The most sensitive water quality indicators of grazing are the fecal indicator bacteria. The objective of this study was to develop a general transport model describing the movement of fecal indicator bacteria from upland sources to channel systems.

Model development was done using simulated rainfall and a runoff surface 30.48 m by 1.83 m.

Initially the runoff surface was smooth concrete and was used to examine the effects of distance from the outlet on coliform counts by locating fecal material at various distances for five replications. Afterwards, the surface was covered by clay soil. Total and fecal coliforms were determined by the multiple-tube method.

Overland flow was described by the kinematic wave equations. Bacterial transport was modeled with a random ordinary differential equation. Initial conditions and assumptions allowed solution for the probability density function (pdf), means, and variances.

The pdf at the slope outlet was found to be normal for the assumed conditions. Solutions for the means and variances were different because initial conditions differed for the relationship between equilibrium and travel time. Three parameters were fitted, a mean retention and two variance terms. The retention parameter appeared to be constant for all cases. The variance terms were obtained only for the rising hydrograph.

For the concrete surface, comparison of predicted and observed means and variances indicated poor fits during initial stages of simulation. Observed values attained steady state rapidly.

There was no replication on the soil surface, and an initial run found high background counts. Background counts were considered to be constant and incorporated into the mean equation. A numerical solution to the mean equation was required because of the unsteady rainfall excess. The background counts and clay content of the soil prevented detection of impacts from a single source.