基于组合权重法的云南省地质灾害灾情年度评价及预测
Annual Evaluation and Prediction of Geological Disasters in Yunnan Province Based on Combination Weight Method
刘静 1杨迎冬 2魏蕾 2陈安1
作者信息
- 1. 昆明理工大学 国土资源工程学院,昆明 650093;自然资源部高原山地地质灾害预报预警与生态保护修复重点实验室,昆明 650216;云南省高原山地地质灾害预报预警与生态保护修复重点实验室,昆明 650216
- 2. 云南省地质环境监测院,昆明 650216;自然资源部高原山地地质灾害预报预警与生态保护修复重点实验室,昆明 650216;云南省高原山地地质灾害预报预警与生态保护修复重点实验室,昆明 650216
- 折叠
摘要
为合理评价云南省地质灾害灾情年度等级,探寻地质灾害的时序发展规律,并预测未来年份地质灾害灾情等级,本文选取死亡失踪人数、直接经济损失、灾情数量3个影响因素作为权重指标,采用层次分析法与熵权法相结合的组合权重法对1991-2020年地质灾害灾情年度等级进行评价.利用差分自回归积分移动平均模型(autoregressive inte-grated moving average method,ARIMA)和BP(back propagation)神经网络模型2种方法,对2021年、2022年、2023年灾情等级进行对比预测和验证.研究表明:地质灾害灾情年度具有5年的周期性变化特征,BP神经网络预测结果准确性更高,根据BP神经网络预测2023年为较轻灾年,从2019年开始地质灾害灾情年度等级有加重的趋势.
Abstract
To reasonably evaluate the annual level of geological disaster in Yunnan Province,explore the temporal development rules of geological disasters,and predict the disaster levels for future years,three influencing factors,namely,death and missing persons,direct economic losses,and the number of disasters,were selected as weight indicators.The combination weighting method of Analytic Hierarchy Process and entropy weight method was used to evaluate the annual level of geological disasters from 1991 to 2020.By using the autoregressive inte-grated moving average method(ARIMA)and back propagation(BP)neural network model,a comparative prediction and validation of the disaster levels for the years 2021,2022 and 2023 were conducted.The study revealed that the annual levels of geological disasters exhibit a periodicity of 5 years.The BP model showed higher accuracy in predicting disaster levels.Prediction based on BP neural network,2023 is expected to be a year with a relatively light disaster level.Since 2019,there has been a trend of increasing severity in the annual levels of geological disasters.
关键词
地质灾害/灾情/组合权重法/BP神经网络/预测Key words
geological hazards/disaster impact/combination weighting method/BP neural network/forecast引用本文复制引用
基金项目
云南省地质灾害气象风险预警业务能力建设(云财资环[2022]4号)
&&(云财资环[2020]68号)
&&(云财资环[2021]23号)
&&(云自然资地勘[2020]445号)
出版年
2024
NETL
NSTL
万方数据