An Analysis of Conglomerate CLIGEN Climate Files in WEPP for Climate Stations of Chile

Abstract

Soil erosion rates have risen worldwide due to human activities including but not limited mining and agriculture. Nowhere is this truer or more pressing of an issue than in developing countries. However, these also tend to be the countries where data necessary for accurate soil erosion modeling is unavailable. The USDA-developed Water Erosion Prediction Project has been shown capable of accurately modeling soil erosion for sites around the world, but requires at least hourly rainfall data. Using five climate stations from three different locations in Chile where hourly data is available, this study compares the accuracy of erosion modeling results based on measured hourly rainfall values and those based on borrowed hourly rainfall data from U.S. climate stations. Inputs for WEPP were created both from real data and from a combination of real and substitute data. The model was run based on all inputs and the results were statistically compared. It was found that use of borrowed rainfall data yields unreliable results. For some locations and some borrowed data, modeled erosion rates are comparable to those based exclusively on measured data. However, this is not universally true, and no simple method was found to relate daily data with hourly data so as to know which borrowed data sets may yield statistically significant results and which may not. These considerations may assist soil erosion modelers and researchers outside the U.S. to more accurately model soil erosion in WEPP and aid policy makers to decide whether to invest in data-gathering infrastructure.