Predicting RUSLE’s erosivity factor with limited rainfall data

Abstract

One of the most widely used indicators of potential water erosion risk is the rainfall erosivity factor (R) of the Revised Universal Soil Loss Equation (RUSLE). R is traditionally determined by calculating a long-term average of the annual sum of the product of a storm’s energy (E) and its maximum 30-minute intensity (I30), known as the EI30. The original method used to calculate EI30 requires pluviograph records for at most 30 minute time intervals. Such high resolution data is difficult to obtain in many parts of the world, and its processing is laborious and time-consuming. The aim of this study was to evaluate nine different procedures for estimating EI30 based on annual, monthly, daily and hourly rainfall records. EI30 values were calculated using the original method for 16 stations in the South-Central region of Chile that have recording rain gauges. The stations were distributed along 800 km (north to south), and spanned a precipitation gradient of 200 to 3500 mm/yr. More than 270-year data were used, and 5400 storms were analyzed. Parameter values were estimated by minimizing the squared error of differences between the estimates and the annual EI30 calculated from the pluviograph data. Both station-specific and regional calibrations were performed. The results show that the empirical procedures evaluated in this study predicted the annual rainfall erosivity well, with r2 values above 0.80. Simple procedures based only annual rainfall provided estimates that are as precise as those based on hourly, daily or monthly records. The results demonstrate that empirical procedures offer a convenient and acceptable means of estimating rainfall erosivity when time and/or data is limiting. Finally, the simple annual precipitation procedure proposed by Renard and Freimund (1994) might prove to be a good tool for land-use planners in the South-Central region of Chile and other areas with similar rainfall characteristics.