Hydroclimatology of soil erosion in the High Plains and the Western and Southwestern United States.

摘要

The objective of this dissertation is to study aspects of the hydroclimatology of soil erosion in the High Plains and the Western and the Southwestern United States. Both rain and wind are important agents of soil erosion. Rainfall erosivity, which is an indicator of the erosive power of rainfall, is affected by interannual variability due to large-scale climatic anomalies. The variability of winter rainfall erosivity in the Western and the Southwestern United States has a close association with the patterns of the El Niño Southern Oscillation (ENSO), which is shown by several analyses.; The analysis of wind erosivity concentrates on the High Plains where aeolian erosion is the dominant process of soil removal and transport. In particular, this study puts an emphasis on the hydrologic mechanisms controlling the rate of soil erosion and their modeling within a probabilistic framework. Thus, this research studies how the estimation of wind erosion can be affected by the rainfall-dependence of wind speed, as well as by the dependence on relative humidity and on air temperature. These hydrometeorological variables affect the surface moisture condition and the susceptibility of the soil to wind erosion and need to be adequately studied and modeled in the simulation of wind erosion.; Therefore, this study develops a stochastic multivariate model for the purpose of simulating the weather variables at short (hourly) time scales, including wind, temperature, atmospheric humidity and solar radiation. The model takes into consideration the cross correlation between these hydroclimatic variables as well as their autocorrelation. For some stations across the High Plains, tests are performed to verify the adequacy of the model and comparison analyses are made between model-generated series and data records to check if hydroclimatic variables are well reproduced.; In conclusion, it is shown that the model is able to replicate weather variables that fairly agree, in the statistical sense, with the observed time series at the considered site.