Dust storm detection and monitoring using multi-temporal INSAT-3A-CCD data

Sand and dust on being agitated by winds often trigger huge dust storms which are characterized by high aerosol optical depth (AOD), tau(a), lower Angstrom exponent (alpha) and near-zero visibility condition. Geostationary platforms are most suitable for studying the dynamic behaviour of such events. In the present study we have used multi-temporal Indian National Satellite System (INSAT-3A-CCD) data on an hourly basis to characterize a massive dust storm that occurred during 15-16 October 2008 over the northern Arabian Sea. An algorithm has been developed to estimate AOD and the Angstrom exponent using the near-infrared (NIR) channel of INSAT-3A-CCD. AOD at 550 nm and the Angstrom exponent, alpha(810, 550 nm) were computed during 14-18 October 2008. The mean value of tau(a)(550 nm) was found to be 1.03 during the dust event, almost two to three times higher than in dust-free conditions. Similarly, the mean alpha(810, 550 nm) value reduced almost by half during the dust event, indicating the presence of larger particles. In general an increasing trend of AOD values was noticed till early afternoon and a decreasing trend was observed thereafter. INSAT-derived atmospheric parameters were compared with Moderate Resolution Imaging Spectroradiometer-derived AOD and alpha for similar observation period. A good correlation for tau(a) was found with R-2 = 0.92 and root mean square error (RMSE) = 0.054. However, a R-2 value of 0.74 with RMSE of 0.13 was obtained for alpha. Analysis of air mass back-trajectory indicates the source of this dust event originates from deserts of Afghanistan and Pakistan.