Remote sensing techniques for prediction of watershed runoff

Abstract

Hydrologic parameters of watersheds for use in mathematical models and as design criteria for flood detention structures are sometimes difficult to quantify using conventional measuring systems. The advent of remote sensing devices developed in the past decade offers the possibility that watershed characteristics such as vegetative cover, soils, soil moisture, etc., may be quantified rapidly and economically. Experiments with visible and near infrared data from the LANDSAT-1 multispectral scanner indicate a simple technique for calibration of runoff equation coefficients is feasible. The technique was tested on 10 watersheds in the Chickasha area and test results show more accurate runoff coefficients were obtained than with conventional methods. The technique worked equally as well using a dry fall scene. The runoff equation coefficients were then predicted for 22 subwatersheds with flood detention structures. Predicted values were again more accurate than coefficients produced by conventional methods.