Abstract:

 The lack of separate statistics on wind and rain damage by disaster-causing factors in the typhoon disaster data in Guangdong is a contributing factor to the existence of typhoon catastrophe index insurance basis risk. To mitigate this basis risk, this paper proposes a method to separate typhoon wind and rain damage. An analysis was conducted on the typhoon disaster data from 2014 to 2022 for 36 counties (districts) in nine coastal cities in Guangdong, focusing on wind and rain indicators. A multiple regression method was employed to construct a typhoon wind and rain disaster loss separation model, which separated the disaster loss due to wind and rain in the typhoon disaster statistics of each county (district). Furthermore, the typhoon observation data and direct economic loss data from Jiangmen Taishan city and Xiangqiao district of Chaozhou were presented as case studies to illustrate the model's validation. Linear correlation analysis showed that eliminating or reducing the effect of multicollinearity on the regression equation at the beginning of modeling may result in the loss of some valid information. The Gamma distribution was found to appropriately fit the distribution pattern of typhoon wind, rain, and damage in Guangdong. The results indicated that 22 counties (districts) passed the significance test at the 0.05 level. The model successfully separated wind- and rain- related damage in Guangdong's typhoon disaster loss statistics, providing a valuable reference for categorizing and analyzing typhoon disaster data by cause.

Key words: Typhoon disaster, Disaster damage separation, Multiple regression, Guangdong