Debris flow prediction models based on environmental factors and susceptible subarea classification in Sichuan, China

Abstract

Environmental factors account for the occurrence of debris flow, as well as different weights of subareas with different risk levels. Considering the relationship between debris flow and rainfall (including the intraday rainfall and the effective rainfall of the previous 10 days), seven environmental factors, including elevation, slope, aspect, flow accumulation, vegetation coverage, soil, and land use, were added in this study. The whole area of Sichuan Province was divided into subareas according to different risk levels. Debris flow prediction models were then established by using a logistic regression model. Results showed that the prediction accuracy was decreased approximately by 3 % after the environmental factors were introduced to the entire study area. The prediction accuracy of the prediction models that comprised the introduced environmental factors was increased by 22.2, 9.7, and 14.3 % in different susceptible areas (moderately susceptible, highly susceptible, very highly susceptible), respectively, compared with that of the prediction models in which rainfall was only considered. Therefore, the research method that introduced the environmental factors may be used to improve the accuracy of debris flow prediction models based on susceptible area classification.