Mining the Features of Environmental Physical Field Influencing Trajectories of Mesoscale Convective Systems Based on Spatial Clustering Analysis

Zhongyang Guo,Xiaoyan Dai,Jianping Wu
DOI: https://doi.org/10.1109/fskd.2008.31
2008-01-01
Abstract:The forecasting of disaster weather using spatial data mining technique is still at an initial stage presently. Recent researches have indicated that intensive precipitation in the Yangtze River Basin in China is closely related to the activity of Mesoscale Convective Systems (MCS) moving out of the Tibetan Plateau in China, however, the factors that influence MCS trajectories are very complex. To discover the trajectories of MCS and features of environmental physical field favoring MCS origination and development over the Plateau, in this paper, MCS are automatically tracked over the Plateau using GMS infrared black-body temperature data. Based on these, spatial clustering method, CLARANS method, is applied to analyzing the characteristics of dynamical field, which influence MCS move eastward out of the Plateau in summer, using environmental physical field forecasting values. The results reveal that the methods are effective and valuable approaches to studying the conditions of environmental physical field favoring the trajectory and propagation of MCS, and improving the predictability of intensive convective weather.
What problem does this paper attempt to address?