Runoff Prediction by GIS Using Optimal Number of Rain Gauges

Finding optimal number of rain gauges to predict runoff without sacrificing any accuracy is important. Jackknife is an iterative technique of procedural removal of the data and finding its effect on final calculation. In this study, Jackknife method is used to remove rain gauges sequentially to observe their impacts on runoff. Precipitation surface is created by interpolation using available rain gauge data to calculate the runoff. Three interpolation techniques; Inverse Distance Weighting (IDW), Spline and Thiessen Weighting are used to create the precipitation surface. The effect of rain gauge removal on the runoff prediction is calculated by a statistical measurement which is a function of simulated and observed runoff. Methodology is applied both on uniform and non-uniform types of precipitation distributions. It is found that while the required minimum number of rain gauges varies with the interpolation technique for non uniform precipitation distribution the number is constant for uniform precipitation distribution. Walnut Gulch in southern Arizona is considered as the study area.