基于机器学习的湖南省干旱灾害风险评价研究

Research on Drought Disaster Risk Assessment Based on Machine Learning in Hunan Province

旱灾是制约农业发展的主要因素,为提高湖南省应对干旱灾害的能力,基于机器学习构建湖南省农业旱灾风险评价模型,探究湖南省农业干旱危害的时空分布特征。结果显示:湖南省农业旱灾综合风险由西向东减小,整体风险处于中低等级,波动小且稳定性好。旱灾受灾率与农业旱灾综合风险分布相一致。采用Pearson检验,将旱灾综合风险与绝收率、受灾率进行对比发现,湖南省农业旱灾综合风险与绝收率和受灾率存在显著相关性,证明本研究旱灾风险评价模型的有效性。

Drought is the main factor restricting agricultural development.In order to improve the ability of Hunan Province to respond to drought disasters,a machine learning based agricultural drought risk assessment model is constructed to explore the spatiotemporal distribution characteristics of agricultural drought hazards in Hunan Province.The results show that the comprehensive risk of agricultural drought in Hunan Province has decreased from west to east,and the overall risk is at a medium to low level,with small fluctuations and good stability.The disaster rate of drought is consistent with the comprehensive risk distribution of agricultural drought.This article uses Pearson's test to compare the comprehensive risk of drought with the failure rate and disaster rate,and finds a significant correlation between the comprehensive risk of agricultural drought in Hunan Province and the failure rate and disaster rate,proving the effectiveness of the drought risk assessment model in this study.