Ratios of Erosive Wind Energies on Dry Days and All Days in the Western United States

Abstract

Simulated wind speeds and precipitation events from stochastic weather generators often are not correlated. This study was undertaken to determine (1) if wet and dry day distributions of hourly wind speeds were different, and (2) if different, would using distributions for all days cause errors in predictions of erosive wind energy and wind erosion on dry days. Hourly weather data were obtained from the National Climatic Data Center SAMSON data set at 46 stations in the western U.S. Wind speeds were sorted into 25 classes and a calm class (0 to 0.5 m s-1). After removal of calm periods, distributions were created for all days, dry days, and wet days. The wet days comprised wind speeds from the initial hour of precipitation and the 23 succeeding hours. Among 552 pairs of wet-day and dry-day cumulative monthly wind speed distributions, 87% of the distributions were significantly different (0.10 level or less based on a Kolmogorov-Smirnov test). To determine the importance of these differences, monthly ratios of erosive wind energy were calculated from dry-day and all-day distributions. Over much of the area, the erosive wind energy was lower on dry days than on all days. The eastern Great Plains and eastern Washington had the lowest ratios. Hence, use of an all-day wind speed distribution at these locations likely overestimates potential soil loss from wind erosion. Limited wind erosion simulations using the WEPS model tended to support this conclusion. In contrast, a few of the low-precipitation areas in the west had ratios that were consistently greater than 1. In summary, accuracy of predicted wind erosion from physically based models can be modestly improved by accounting for differences in wind speed distributions on wet and dry days.