AUTHOR=Kim Byeongyeon , Kim Taejong 

TITLE=AI in extreme weather events prediction and response: a systematic topic-model review (2015–2024)

JOURNAL=Frontiers in Environmental Science

VOLUME=Volume 13 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/environmental-science/articles/10.3389/fenvs.2025.1659344

DOI=10.3389/fenvs.2025.1659344

ISSN=2296-665X

ABSTRACT=IntroductionClimate change is driving a sharp rise in the frequency and intensity of extreme-weather events, magnifying their social and economic impacts and exposing the limits of conventional physics-based forecasting systems.MethodsTo understand how artificial intelligence (AI) helps meet this challenge, we systematically analyzed 8,642 peer-reviewed articles published between 2015 and 2024 in the Web of Science, applying Latent Dirichlet Allocation (LDA) topic modelling to map the literature.ResultsFive principal research themes emerged: 1) Forecasting and Prediction of Extreme-Weather Events, 2) Flood Prediction and Risk Assessment, 3) Drought Monitoring and Agricultural Risk Assessment Using Machine Learning, 4) Climate Change and Ecosystem Response to Extreme-Weather Events Using Machine Learning, and 5) Multisource Imagery and Deep Learning for Disaster Detection and Damage Assessment. Across these domains, AI-driven models improve forecast skill, fuse heterogeneous hydrometeorological data for real-time warning, and quantify ecological impacts at finer spatial-temporal scales than traditional approaches; recent advances include diffusion models that sharpen rainfall and wind forecasts, recurrent networks that enhance runoff prediction, and transformer-based vision models that automate high-resolution damage mapping.DiscussionThe evidence indicates that AI can increase the reliability of extreme-weather prediction, accelerate disaster-response workflows, and ultimately reduce societal losses. Methodologically, this study offers the first large-scale, quantitative mapping of AI research in extreme-weather prediction and response, capturing both thematic prevalence and temporal evolution—an empirical perspective that extends and strengthens insights from prior qualitative reviews.