Modeling wind-blown desert dust in the southwestern United States for public health warning: A case study

• Published in: Atmospheric environment (1994) Vol. 39; no. 33; pp. 6243 - 6254
• Main Authors: Yin, Dazhong, Nickovic, Slobodan, Barbaris, Brian, Chandy, Beena, Sprigg, William A.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.10.2005; Elsevier Science
• Subjects: Air. Soil. Water. Waste. Feeding; Applied sciences; Atmospheric pollution; Biological and medical sciences; Desert dust cycle modeling; Dust particulate matter pollution; Earth, ocean, space; Environment. Living conditions; Exact sciences and technology; External geophysics; Medical sciences; Meteorology; Model evaluation; Particles and aerosols; PM2.5 and PM10 concentrations; Pollution; Public health warning system; Public health. Hygiene; Public health. Hygiene-occupational medicine; USA, Southwest; Desert dust cycle modeling; Dust particulate matter pollution; PM2.5 and PM10 concentrations; Model evaluation; Public health warning system; Severe weather warning; aerosols; Air quality; Forecast model; troposphere; Coarse particle; dust storms; deserts; Health and environment; Fine particle; algorithm performance; Atmospheric pollution forecasting; Air pollution; Public health; wind erosion
• ISSN: 1352-2310; 1873-2844
• DOI: 10.1016/j.atmosenv.2005.07.009

A model for simulating desert dust cycle was adapted and applied for a dust storm case in the southwest United States (US). This is an initial test of the model's capability as part of a future public health early warning system. The modeled meteorological fields, which drive a dust storm, were evaluated against surface and upper-air measurement data. The modeled dust fields were compared with satellite images, in situ surface PM2.5 and PM10 data, and visibility data in the areas affected by the dust event. The model predicted meteorological fields reasonably well. The modeled surface and upper-air field patterns were in agreement with the measured ones. The vertical profiles of wind, temperature, and humidity followed closely with the observed profiles. Statistical analyses of modeled and observed meteorological variables at surface sites showed fairly good model performance. The modeled dust spatial distributions were comparable with the satellite-observed dust clouds and the reduced visibility patterns. Most encouragingly, the model-predicted and observed PM2.5 peak hours matched reasonably well. The model produced better PM2.5 peak hours than PM10 peak hours. The temporal varying trends of daily and hourly PM2.5 and PM10 concentrations at most of the measurement sites were similar to those observed. Discrepancies between the values of the modeled and the measured surface PM2.5 and PM10 concentrations differed with time and location. Sometimes the modeled and measured concentrations can have one order of magnitude differences. These revealed there were possible deficiencies in the simulation of the dust production strength and location, and the representation of dust particle size in the modeling. Better land surface data and size representation of the dust production are expected to further improve model performance.