Modeling and Data Analysis of 2011 Phoenix Dust Storm

Despite the frequency of dust storms in the southwest region of United States, studies pertaining to extreme dust events across Arizona are quite limited. Massive dust events, like the July 5, 2011 Haboob across Phoenix largely affect visibility and damage properties in Phoenix and Maricopa. This highlights the need to model and predict the occurrence of dust storms in the region. Unlike dust storms in Africa and Asia, little is known about the role of dust (and haboobs) in altering atmospheric chemistry, southwest monsoon and local precipitation patterns in Arizona. Here, we investigate the key processes driving the haboobs in Arizona through modeling and satellite data analysis. In particular, we evaluate the general performance of the Weather Research and Forecasting model with chemistry (WRF-Chem) in simulating the July 5, 2011 Phoenix haboob using retrievals of aerosol optical properties and mass concentrations from satellite instruments such as Moderate Resolution Imaging Spectroradiometer (MODIS) and Cloud-Aerosol Lidar and Infrared Pathfinder Satellite (CALIPSO), in conjunction with available surface PM10 and PM2.5 concentration measurements from in-situ observation sites. We use a nested modeling domain covering Utah, California and Arizona at a horizontal resolution of 5.4 km (outer) and 1.8 km (inner). This configuration is based primarily on the operational weather forecast currently being carried out in the University of Arizona. Initial results using the dust-only option in WRF-Chem and Global Forecast System (GFS) boundary conditions, show that the model was able to capture the spatio-temporal pattern of the haboob but largely underestimated the magnitude of surface dust concentration and aerosol optical depth. Model results using Goddard Chemistry Aerosol Radiation and Transport (GOCART) aerosol scheme provide relatively better magnitudes for the fine-mode aerosols. Transport of dust is also clearly discernible from Terra-MODIS on the 6th of July. However, our evaluation of the haboob event is limited by the availability of retrievals since they are greatly affected by thick cloud cover during the monsoon season. We also present results from different experiments that we carried out to better understand the sensitivity of the model performance to various modeling components such as dust emission, transport, deposition, initial and boundary conditions, and other meteorological drivers. Finally, we note that in addition to dust storm forecasting, the coupled weather-air quality modeling configuration also provides a unique opportunity to explore the impact of incorporating aerosols in numerical weather prediction through direct and indirect forcing from aerosols. Our future goal is to extend our operational weather forecast to include size-resolved dust and other aerosols.