257

Dust detection with a new dust index using MODIS data

Aojie Di1 ,4, Yong Xue1 ,2, Xihua Yang3, John Leys 3, Lu She1 ,4
1Key Laboratory of Digital Earth Science, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing, China, 2Faculty of Life Sciences and Computing, London Metropolitan University, London, UK, 3New South Wales Office of Environment and Heritage, New South Wales, Australia, 4University of Chinese Academy of Sciences, Beijing, China

Dust is one of the main types of atmospheric aerosol. It changes the atmospheric-radiation budget of terrestrial systems through the scattering and absorption, and produces indirect radiation force by changing the atmospheric composition and properties of cloud. Sand dust disaster generally refers to disaster caused by sand and dust weather namely floating dust, blowing dust and dust storm. The study on dust storms has an important significance to local economic development and global climate change research. Satellite remote sensing is one of the ideal means for monitoring large regional distribution and intensity of dust and studying the interaction of dust aerosols and regional climate. Moderate resolution Imaging Spectroradiometer (MODIS) can provide extensive earth observation with as much as 36 spectral bands. In this paper, a new dust index is proposed for MODIS data. This index take advantage of red band (band1, 0.65μm), short wave infrared band (band 6, 1.65μm), mid infrared band (band20, 3.75μm) and thermal infrared band (band31, 11.03μm) as well as the MODIS AOD. AOD is firstly derived using SRAP (Synergetic Retrieval of Aerosol Properties algorithm) algorithm (Xue et al., 2014) from MODIS. The proposed dust index can not only detect the distribution of dust, but also indicate its intensity. A good exponential correlation is established between in situ measured visibility obtained from an atmospheric station in Karshgar, and MODIS AOD. The result shows a squared correlation coefficient of 0.67, which demonstrates the rationality of identifying and quantifying dust storm intensity with satellite retrieved AOD. Two case studies of dust storms are displayed in this paper. One occurred on September 23th, 2009 in Australia and the other one occurred on April 24, 2014 in Taklimakan Desert, one of the predominant dust origin in China. Both of them show a good performance of the proposed index.