Dust storm detection and classification in the Middle East using the combination of reflectance and thermal infrared properties of MODIS sensor

Dust storm events carry a huge amount of aerosols to the atmosphere. These particles reduce air quality and can cause breathing, allergic, cordial problems. This environmental disaster increasingly outbreaks in arid and semiarid areas (e.g. Middle East) in recent years. Monitoring from space using remote sensing is one of the most effective and extensive techniques for dust storm detection. Terra MODerate resolution Imaging Spectroradiometer (MODIS) data have been utilized in this study for dust detection and classification in the Middle East. Dust particles emissivity differences in thermal infrared signals can discriminate the dusty from non-dusty pixels. D-parameter is calculated by combination of visible reflectance properties and Brightness Temperature Difference (BTD) between 11 and 12 µm channels. D-parameter is examined were using 28 satellite images (expanded in 2008 to 2009) and showed that it is very effective approach for dust area detection. Relation between D parameter and visibility (obtained from 41 synoptic stations) was investigated. The results showed that there is an exponential relation with R=0.68 between them. This relationship can retrieved the visibility from D parameter with Mean Percent Absolute Error=45.6%. According to the complicacy of problem and vast study region, R=0.68 and MPAE=45.6% seems very good. Hence we utilized this developed exponential relationship to retrieved visibility maps in Middle East from D-parameter maps. Then these generated visibility maps were classified to the different dust storm classes. This classification method was developed using the combination of different dust classification methods according to the visibility values. In this classification method, the dust storms were classified to 5 different classes (Sever Dust Storm¡ Medium Dust Storm¡ Light Dust Storm¡ Blowing Dust and Dust in Suspension).