Effect of antecedent rainfall on runoff during low-intensity rainfall

Abstract

The objective of this study was to determine the relative significance of several rainfall characteristics to the occurrence and magnitude of runoff from small (< 300 ha) watersheds during very low-intensity rainstorms (maximum rainfall intensities < 15 mm/hr). This was done by making hourly measurements of rainfall and runoff at 1-hr intervals for four years on five watersheds in western Oregon, U.S.A., an area with a modified marine climate. Normal annual precipitation at the site is 1.02m and occurs primarily as rain during the winter months. Rainfall intensities only rarely exceed 15 mm/hr. The watersheds (0.46 to 285 ha in size) were fall-planted (November) to wheat. The soils are predominantly Willakenzie series (Ultic Haploxeralfs, fine-silty, mixed, mesic). Rainfall characteristics were computed for each rainfall event, defined as a period of consecutive rainy hours separated by six or more consecutive dry hours. Included were magnitude, average intensity, maximum intensity and duration of rainfall during rainfall events, magnitude of rain during the 12, 48 and 120 hrs preceding events, the magnitude of rain since October 1 of each year and number of dry hours preceding the event. The amount of runoff was computed for each rainfall event. Stepwise discriminant analysis was used to determine those values and combinations of rainfall-event characteristics that were significant in predicting the occurrence and amount of runoff. Between 44% and 93% of the rainfall events caused no measurable runoff. Discriminant analysis indicated that the magnitude of rainfall during the 12 and 120 hrs preceding the events was more significant in determining the occurrence of runoff than either rainfall magnitude, average intensity or maximum intensity. The majority of rainfall events on all of the watersheds resulted in < 25% runoff. Discriminant analysis indicated that the amount of runoff during events was determined more by the magnitude of rainfall in the 12, 48 and 120 hrs preceding the events than by the other event characteristics. The results suggest that in climates characterized by low-intensity rainfall, antecedent rainfall is more important in controlling the occurrence and amount of runoff than rainfall magnitude or intensity. The results also suggest that more accurate predictions of runoff can be achieved by using values of antecedent-rainfall characteristics.