Delineation of potential hot spots of aeolian dust using normalized difference water index

Abstract

Due to the drop in water levels during the monsoon season, the soil particles of the exposed riverbed of Zhuoshui River are easily suspended in the air, causing serious river dust problems and endangering the quality of life and health of local residents. Exposed riverbed is the main source of dust, and it affects PM10 concentration in air quality monitoring stations. In addition to meteorological conditions, the moisture content of the source dust is also an important factor. This study used satellite images from the lower river section (Bridge Ziqiang to Bridge Xibin) to extract the exposed riverbed, used the cluster analysis based on normalized difference water index (NDWI) to further classify the bare land into three categories, and then combined with meteorological factors to construct multi-regression models of PM10 (PM10 = 45.89W + 18.44T − 373.85; R2 = 0.64, p < 0.001 and PM10 = 21.80T + 39.70W + 37.2M2 − 646.25; R2 = 0.74, p < 0.001; W, wind speed; T, air temperature; M2, bare land with medium NDWI). The results showed that the meteorological factor interpretation rate was about 64%, and if the index of bare land with medium NDWI was included, the model interpretation rate could be increased to 74%, hence the M2 factor had a high influence on the PM10 concentration.