The assessment of scale of spatial and temporal water erosion parameters.

Abstract

This study assessed the effects of scale on distributed water erosion parameters such as effective hydraulic conductivity, interrill and erodibility parameters. To accomplish this, the watershed was split into one, two, three, six, eight and ten hillslope configuration using geostatistical analysis on data collected on a 20 m grid at Kendall 112 in Walnut Gulch Experimental Watershed located near Tombstone, Arizona. Ordinary and universal block kriging were used to split the watershed into cascading planes composed of hillslopes and overland flow elements. The manipulation of data by increasing its spatial variability restored the variability which was smoothed during block estimation. Version 95.7 of Water Erosion Prediction Project (WEPP) was used to simulate runoff, peak discharge and sediment yield from each of the watershed configurations. Based on the results, block kriging can be used to define hillslopes and overland flow elements required in the simulation of runoff and sediment yield using WEPP. From the results, distribution of vegetation parameters by multiple hillslopes did not improve the runoff and sediment yield at the watershed outlet. However, averaging vegetation estimates on a single watershed configuration gave poor results of predicted runoff and sediment than on higher hillslope watershed configurations. The high nugget observed from the vegetation sample variograms reduces the estimation technique to an arithmetic averaging. The model produced plausible results of runoff and peak discharge when the number of hillslopes is increased from one up to eight hillslope watershed configuration. No further significant improvements were realized, beyond an eight hillslope configuration watershed. The erratic nature of predicted sediment yield is explained by the fact that the model does not update rill and interrill parameters during continuous simulation.