UC Davis

Soil moisture links hydrologic and atmospheric processes and impacts important properties of the atmospheric boundary layer via turbulent land-atmosphere exchange. Research on land-atmosphere interactions and their impacts on the simulated boundary layer in semi-arid regions with substantial irrigation is relatively sparse. We use the Weather Research and Forecasting (WRF) model to evaluate the influence different land surface models (LSMs) and planetary boundary layer (PBL) schemes have on the performance of simulations through comparisons with multi-scale observations during a fifteen-day summertime period during 2016. The focus region for this study is the Central Valley (CV), California, which receives little to no rain in the summer and relies on widespread irrigation for agriculture. Results demonstrate that the LSM drives the differences between simulations, showing only minor variations with changing the PBL scheme. Simulations using the RUC (Rapid Update Cycle) and PX-NO (Pleim-Xu without soil moisture and soil temperature nudging) LSMs generated better comparisons with observed PBL depths. Contrasting RUC however, PX-NO better simulates surface fluxes and humidity, whereas Noah (Noah Unified) and Noah-MP (Noah Multiparameterization) simulate better temperatures despite relatively poor surface flux performance. For most quantities, indirect soil nudging in PX (Pleim-Xu) did not improve results compared to PX-NO, which may be related to soil moisture initialization, the nudging dataset, or a need for model improvements in arid regions. Despite these variations in performance statistics across simulations and quantities, we show that potential evapotranspiration (ETo) has robust performance statistics across simulations. This suggests that ETo depends more strongly on net radiation, which performs relatively well across simulations, than on wind, temperature, and humidity, and indicates a further disconnect between ETo and latent heat fluxes in WRF simulations. Finally, we suggest strategies to obtain the necessary observations to better understand the multi-scale dynamics in the CV and drive subsequent model development.