环境科学
生物圈
绿化
气候变化
全球变暖
植被(病理学)
优势(遗传学)
降水
气候学
温室气体
生态系统
闪光灯(摄影)
生产力
大气(单位)
浪涌
自然(考古学)
全球变化
碳通量
强迫(数学)
温室效应
荒漠化
暴发洪水
大气科学
生物圈模型
作者
Li J,Yao Zhang,Emanuele Bevacqua,Shijie Jiang,Xing Yuan,Sha Zhou,Jinghao Qiu,ZhaoLi Wang,Jakob Zscheischler,K Wang,Shilong Piao
出处
期刊:Science Advances
[American Association for the Advancement of Science]
日期:2026-05-27
卷期号:12 (22): eaea8452-eaea8452
标识
DOI:10.1126/sciadv.aea8452
摘要
Flash droughts have become more frequent, yet their underlying mechanisms of such changes remain unclear. Using an explainable artificial intelligence-based clustering framework with reanalysis products and Earth system models, we show that the major driving mechanisms of global flash droughts have shifted from precipitation dominance to a compound effect of temperature, transpiration, and precipitation. This transition drives a surge in flash droughts over the past decade, with more rapidly developing, severe, and enduring soil droughts. Anthropogenic warming and vegetation greening are the main drivers of the observed surge. The transition, which has emerged beyond natural climate variability since 2017 and is mainly detected in Eurasia, Amazon, and Africa, exposes ~650 million people under threat and reduces gross primary productivity by 0.15 ± 0.1 petagrams of carbon per year. Our results demonstrate that biosphere and atmosphere responses to anthropogenic forcing have altered flash drought drivers, emphasizing the urgent need for targeted mitigation strategies.
科研通智能强力驱动
Strongly Powered by AbleSci AI