A review of observation, simulation, and prediction of extreme climate events
极端气候事件的观测、模拟与预测综述

摘要

极端气候事件是指发生概率极低但影响巨大的气象异常现象，包括极端降水、干旱、高温、寒潮、风暴等。这些事件不仅对自然生态系统造成严重影响，还对农业生产、水资源管理、城市基础设施及社会经济安全带来显著威胁。随着全球气候变化的加剧，极端事件的频率和强度呈现上升趋势，科学监测、精确模拟及可靠预测已成为应对气候风险的重要手段。近年来，观测技术不断进步，卫星遥感、地面气象站网络及再分析数据提供了高分辨率、长时间序列的数据支持；同时，全球和区域气候模式的发展，使得对极端事件的形成机制、时空演变及未来趋势的研究更加深入。统计学方法、动力学模式及人工智能技术的结合，为极端事件预测提供了新的思路。本文综述了极端气候事件的观测方法、模拟技术及预测手段，分析了不同研究方法的优势与局限，并讨论了极端事件对生态系统、农业、城市环境及社会经济的综合影响。最后，提出了未来研究的关键方向，包括多源数据融合、多模式集合预报及跨学科协同研究，以期为科学决策和风险管理提供理论依据和技术支撑。

Abstract

Extreme weather events refer to meteorological anomalies with extremely low probability but great impact, including extreme precipitation, drought, high temperature, cold wave, storm, etc. These events not only have a serious impact on natural ecosystems, but also pose a significant threat to agricultural production, water resources management, urban infrastructure and social and economic security. With the aggravation of global climate change, the frequency and intensity of extreme events show an upward trend. Scientific monitoring, accurate simulation and reliable prediction have become important means to deal with climate risks. In recent years, observation technology has been continuously improved, satellite remote sensing, ground weather station network and reanalysis data provide high-resolution, long-time series data support; at the same time, the development of global and regional climate models makes the formation mechanism, spatio-temporal evolution and future trends of extreme events more in-depth research. The combination of statistical methods, dynamic models and artificial intelligence technology provides new ideas for extreme event prediction. This paper summarizes the observation methods, simulation techniques and prediction methods of extreme weather events, analyzes the advantages and limitations of different research methods, and discusses the comprehensive effects of extreme weather events on ecosystems, agriculture, urban environment and social economy. Finally, the key directions of future research, including multi-source data fusion, multi-model ensemble forecasting and interdisciplinary collaborative research, are proposed to provide theoretical basis and technical support for scientific decision-making and risk management.