Predicting Runoff Events with RUSLE2.

RUSLE2 (Revised Universal Soil Loss Equation) provides robust sheet and rill erosion estimates for a wide range of land use, soil, and climatic conditions. However, its original form assumes daily rainfall disaggregation, not producing a reasonable set of runoff events for channel or other runoff-driven calculations.   We herein describe and evaluate a method for predicting a series of representative runoff events whose sizes, durations, and timings are estimated from information in the extensive RUSLE2 database.  This method uses relationships that capture seasonal effects of antecedent soil water content and increased curve numbers during months with intense rainfall to predict mean monthly runoff, annual runoff event frequency, and a gamma distribution parameter characterizing the population of runoff event depths. The runoff event size for any specified return period can then be estimated, allowing RUSLE2 to be used in risk assessment calculations. The event series is determined by assuming that the largest event occurs on the day of the year with the maximum disaggregated runoff value, with this event’s size determined for a user-specified return period, defaulting to 1 year. The ratio of this maximum event depth to the maximum daily disaggregated runoff depth is used to determine both the interval between subsequent events and the resulting event runoff depths.   The approach estimates the same sheet and rill erosion as the original daily values, but the new runoff and sediment load estimates are suitable for linkage with a process-based ephemeral gully channel erosion model.  Comparison with observations made from 1995 to 2002 at Treynor, IA, found that the Nash-Sutcliffe model efficiency coefficient for predicted monthly runoff depths was 0.35 and predicted runoff event depths were within 15% of corresponding observed runoff events for return periods ranging from 1 to 100 years.