水文地质工程地质2026，Vol.53Issue(2)：14-34,21.DOI:10.16030/j.heg.202507054
地质灾害气象预警研究进展与发展趋势
Advances and trends in geo-hazard early warning based on meteorological factors
摘要
Abstract
Geo-hazards triggered by extreme weather events occur frequently in China and worldwide,posing severe risks to lives and property.Geo-hazard early warning based on meteorological factors is a critical means of advancing disaster prevention"ahead of the event"and improving the effectiveness of mitigation efforts.In recent years,the increasing frequency of intense rainfall-induced geo-hazard clusters has highlighted the urgent need to enhance both model accuracy and system efficiency for precise early warning.This study provides a systematic review of progress in geo-hazards early warning models and operational systems,identifies their key challenges,and outlines future development directions.Geo-hazards meteorological early warning models fall into three major categories:statistical models,physical models,and machine learning models.Statistical models are widely used due to their simplicity and ease of implementation,but their accuracy is constrained by the quantity and quality of historical samples.Physical models can reveal failure mechanisms but face limitations in regional applications due to complex instability processes and parameter uncertainties.Machine learning models are rapidly advancing and offer strong potential for extracting multi-source data patterns and improving prediction performance;however,challenges remain in terms of limited samples and complex feature representation in real-world conditions.A global review of early warning practices in 7 countries and 54 regions shows that 90%of operational systems use statistical models,while 10%employ physical models.Research on machine learning models has surged in recent years,with some systems entering pilot testing and demonstrating promising prospects for broader application.Future efforts should focus on integrating multi-source geological-meteorological big data with advanced AI techniques to improve warning accuracy,strengthen response capability,and promote more precise and effective geo-hazard disaster prevention and mitigation.关键词
地质灾害气象预警/阈值模型/成灾机理/机器学习/多源数据融合Key words
Geo-hazards early warning based on meteorological factors/threshold model/disaster mechanism/machine learning/multi-source data integration分类
天文与地球科学引用本文复制引用
刘艳辉,李宇柔,马世伟,肖锐铧,王惠卿..地质灾害气象预警研究进展与发展趋势[J].水文地质工程地质,2026,53(2):14-34,21.基金项目
国家重点研发计划项目(2023YFC3007205) （2023YFC3007205）
国家自然科学基金项目(42077440 （）
41202217) （）
