FINER-SCALE MONITORING AND ANALYSIS OF 2016–2021 DUST STORMS DISASTERS IN CHINA BASED ON HIMAWARI-8 SATELLITE IMAGERIES
Keywords: Himawari-8, Dust Storm, Classification, Near Real-time Detection, Analysis
Abstract. Today, remote sensing is one of the important techniques for natural disasters monitoring. Using the Himawari-8 new generation of geostationary meteorological satellite image data, we identified and monitored 2016–2021 dust storm disasters in China at the finer scale, and compared the object-oriented classification, supervised and unsupervised classification methods. Taking the dust storm on May 5, 2016 as example, the average accuracy and kappa coefficient of the supervised classification were 0.78 and 0.68, while the object-oriented classification were 0.81, 0.73. We used the object-oriented classification method to classify the 2016-2021 dust storm incidents across China. In 2021, the largest area affected by sandstorms was about 68.33km2, the smallest area was about 2.08km2, and the average area was about 22.25km2. During 2016–2021, the most severe and wide-spreading sandstorm occurred on March 14–18 2021. The dust storm first appeared in northwestern Mongolia on March 14, and then spread southeast to China. The dust area covered Beijing on the 15th, and then spread from eastern China to the western regions of North China and Northeast China, including most of the Hexi Corridor. On the 16th and 17th, the influence of the dust storms narrowed to central and western Inner Mongolia, northern Gansu and southern Mongolia. The dust storm gradually disappeared on the 18th.