Simulation of an Arizona Dust Storm Using a High-resolution WRF-FENGSHA-CMAQ Modeling Framework

Abstract:

Dust storms over south-central Arizona, particularly the region stretching from Tucson to Phoenix, have caused several fatal highway accidents by reducing visibility and have posed serious health risks by degrading air quality conditions. Moreover, the region is highly endemic to Valley fever, an illness caused by a soil-dwelling fungus, that is known to be associated with the regional dust activity. In recent years, it has been noted that the active or abandoned farmlands are the major sources of dust for some of the events. The water cuts from the Colorado River, as well as the climate model projections indicating a drier southwest in the future, can mean that the problem may worsen in the coming years. The previous modeling studies have limited capabilities, to reliably send early warnings and inform decision making, that result from four main issues: low resolution, inaccurate input data, ignoring cropland as a dynamic dust source, and the uncertainties in the wind erosion process. To address some of these issues, we have developed a high-resolution (1km horizontally) dust modeling system by updating the dust emission model FENGSHA within the WRF-FENGSHA-CMAQ modeling framework. FENGSHA is a semi-empirical dust emission model that relies on the observationally constrained threshold friction velocities which dictate the initiation and the amount of dust emission. The major updates in the model include changes in the ways the drag partitioning and soil moisture influence threshold friction velocities, changes in the representation of sediment supply, and fallow cropland as a new dust source. The updates accompanied the implementations of new land-use and soil texture data created using updated and very high-resolution source data. The updated modeling system simulates a past dust storm reasonably well (best performance being over the Phoenix area) when compared with the EPA's ground (point) observations of particulate matter (PM10) and exhibits significant improvements. The simulations indicate cropland can contribute up to 60% of the PM10 during the dust storm, highlighting the importance of including this dust source in the emission budget and dust simulations, in particular in light of potential water cuts in this region.