Sensitivity of low-level winds simulated by the WRF model in California's Central Valley to uncertainties in the large-scale forcing and soil initialization

The sensitivity of the Weather and Research Forecasting (WRF) model-simulated low-level winds in the Central Valley (CV) of California to uncertainties in the atmospheric forcing and soil initialization is investigated using scatter diagrams for a 5-day period in which meteorological conditions are typical of those associated with poor-air-quality events during the summer in the CV. It is assumed that these uncertainties can be approximated by two independent operational analyses. First, the sensitivity is illustrated using scatter diagrams and is measured in terms of the linear regression of the output from two simulations that differ in either the atmospheric forcing or the soil initialization. The spatial variation of the sensitivity is then investigated and is linked to the dominant low-level flows within the CV. The results from this case study suggest that the WRF-simulated low-level winds in the northern CV [i.e., the Sacramento Valley (SV)] are more sensitive to the uncertainties in the atmospheric forcing than to those in the soil initialization in the typical weather conditions during the summer that are prone to poor air quality in the CV. The simulated low-level winds in the southernmost part of the San Joaquin Valley (SJV) are more sensitive to the uncertainties in the soil initialization than they are in the SV. In the northern SJV, the simulated low-level winds are overall more sensitive to the uncertainties in the large-scale upper-level atmospheric forcing than to those in the soil initialization. This spatial variation in sensitivity reflects the important roles that the large-scale forcing, specified by the lateral boundary conditions and the local forcing associated with the soil state, play in controlling the low-level winds in the CV.