Abstract
Northeast Asia is one of the Northern Hemisphere mid‐ to high‐latitude regions most sensitive to global warming, where high temperatures, drought and wildfires have become increasingly frequent and concurrent, yet their cascading linkages remain poorly quantified within a unified daily‐scale framework. We integrate daily observations from 829 meteorological stations, soil moisture from a land‐surface data assimilation product and MODIS active‐fire detections and use a boosting‐based ensemble machine‐learning model to substantially improve the simulation of 0–10 cm relative soil moisture. On this basis, we derive triggering thresholds along a heatwave‐drought‐wildfire (HDW) cascading hazard chain. The 6‐day moving mean of daily maximum temperature (AUC = 0.74) defines the key temporal window for drought triggering, with thresholds in spring and autumn generally lower than those in summer. The mean triggering rate of the HDW hazard chain during 2017–2023 (0.18) is 157% higher ( p −1 ), pointing to a potential risk of irreversible change. By contrast, under progressively strengthened ignition control and fire management systems, the triggering rate in Northeast China shows a decreasing trend (−0.002 year −1 ), suggesting that policy interventions may partially offset the amplifying effect of climate warming on the hazard chain. This study links, at the daily scale, cumulative heat exposure, soil‐drought thresholds and wildfire occurrence along a cascading pathway and provides a quantitative basis for improving compound hazard risk assessment and designing regionally differentiated fire management strategies.