Predicting soil thickness and dust content in upland watersheds of the Mojave Desert

Soil thickness is an important input parameter for hydrologic models. As the spatial distribution of soil thickness is strongly variable at hillslope scales, modeling soil thickness at the watershed scale is needed if we are to accurately model hydrologic and geomorphic processes that depend on soil thickness. Recent studies have demonstrated that numerical models which assume a long-term balance between soil production and soil erosion can accurately predict the spatial distribution of soil thickness in semiarid and relatively humid areas, where soil thickness varies gradually along hillslope profiles. Yet, these models have not been tested in arid regions, where soil thickness varies abruptly and where the influence of dust accretion on soil thickness may be much more significant. In this study, we developed and tested a numerical model for the prediction of soil thickness and dust content in an arid mountainous area of the Mojave Desert, using field measurements of soil thickness, geochemical analyses of soil and bedrock samples, and quantitative analyses of LIDAR data. The study site is characterized by thin soils (<1-2 m) and extreme variability in soil thickness at essentially all spatial scales. Soil production rates in the model were quantified using exponential and humped soil production functions, and soil erosion rates were quantified using the nonlinear depth- and slope-dependent transport function, assuming that only the upper soil horizons (A+B) can be transported. We calibrated the soil production and dust accretion rates in various plutonic lithologies (granite, diorite, and quartz monzonite) using 1) field data of soil thickness, and 2) dust content of the soil estimated by analysis of immobile element concentrations. We validated the model results using measured data on soil thickness and the presence/absence of soil in an attempt to adequately represent the great spatial heterogeneity in soil thickness at our study sites. We found good agreement between observed and predicted soil thickness and dust content in soil. Our results show that we can simultaneously predict the soil thickness and the fraction of eolian dust in the soil at different sites across the landscape.