University of Tehran Press Desert 2008-0875 27 1 2022 06 01 Detection of dust storms overnight in the South West of Iran using satellite images 35 53 88508 10.22059/jdesert.2022.88508 EN M. Deiravipour Marine Environment, Department of Marine Natural Resources, Khorramshahr University of Marine Science and Technology, Khuzestan, Iran H. Mohammad Asgari Department of Marine Natural Resources, Khorramshahr University of Marine Science and Technology, Khuzestan, Iran S. Farhadi Marine Environment, Department of Marine Natural Resources, Khorramshahr University of Marine Science and Technology, Khuzestan, Iran Journal Article 2022 08 16 Dust concentration, as the level of particulate matter (PM10), has become an important indicator of air pollution, and has attracted a great deal of attention from environmental agencies and organizations, public health, and science worldwide. Over recent years, dust storms with intense drought have had numerous adverse effects on human health and socio-economic situation in arid and semi-arid countries. Despite the inevitability of their occurrence, natural and human activities could exacerbate this phenomenon. Imagery data analyses have improved our understanding of dust detection and monitoring. Previous research has extensively studied the dust storms in day time. Meanwhile, there are a few studies investigating dust detection over-night. For dust detection over-night, several algorithms were utilized herein, including brightness temperature difference (BTD) for 20, 23, 31, and 32 MODIS bands and artificial neural network (ANN). The obtained results revealed that BTD indices have ood performance for dust detection in the southwest of Iran and their accuracy will be better with an increase in the concentration and density of dust and a reduction in cloud cover in the region. The BTD and ANN methods were evaluated using different indices. Our findings revealed that ANN method was more accurate than BTD indices. This finding is probably attributed to the complex properties of dust; artificial neural network is an appropriate method to model nonlinear and complex dust and surface properties. https://jdesert.ut.ac.ir/article_88508_bc0dc25205851ed76da714956468e76e.pdf