Verification and Validation of Desert Dust Forecasts and Their Impact on Respriatory Health Applications in the Southwestern United States

The paper describes the performance of the Dust Regional Atmospheric Model (DREAM) after being nested within the National Centers for Environmental Prediction (NCEP/eta) operational weather model. The model system is designed to simulate dust entrainment, transport, and concentration under changing atmospheric and terrestrial conditions to better forecast dust episodes that impose health outcomes on populations at risk. The approach was to: (1) benchmark the model system’s performance after it was modified to fit a new domain in the southwest U.S.; (2) verify and validate model outputs statistically by comparing outputs from ground station observations; and (3), replace selected baseline terrain parameters in DREAM by assimilating comparable satellite data and comparing the results. Having base-lined, verified, and validated the model system’s performance, Earth observation data were assimilated sequentially into a series of new model runs. Terrain parameters were assimilated for barren ground dust sources, fraction of photosynthetically active radiation; digital topography, aerodynamic surface roughness, and soil moisture. Replacement of the baseline parameters with these assimilated parameters improved dust model performance without imposing negative impacts on observed meteorological fields. Major gains were made in modelling the onset of dust storms, the timing, of peak hour concentration, and duration of near-surface high dust concentrations. There was no verifiable improvement in measuring the magnitude of dust concentrations, even though the enhanced model predicted accurately the occurrence of dust storm events at most locations in the model domain.