A study of 107 summer events finds that heat driven by tropical weather patterns is easier to predict than heat linked to two high-latitude wave patterns.
Researchers identified 107 rapidly escalating heat extremes in North China during the summers of 1940–2024 and grouped them into four spatial types. They found that the rapid warming involved different combinations of warm air moving horizontally into the region and sinking air warming as it was compressed.
The four types were associated with different large-scale weather patterns involving tropical thunderstorms, wave-like patterns in the middle and high latitudes, or interactions between them. Subseasonal forecasts showed a consistent difference in predictability among the types: tropical-driven events were the most predictable, while events linked to two high-latitude wave patterns had the lowest forecast skill.
Four paths to sudden heat
The study identified 107 rapidly escalating heat extremes in North China over 85 summers, from 1940 through 2024, and classified them into four spatial types. The types differed in how much rapid warming came from horizontal warm-air transport and from sinking air, which produces warming through compression.
Each type was linked to a different intraseasonal weather pattern. Some involved tropical convection, meaning organized tropical thunderstorms; others involved wave trains in the middle and high latitudes, or interference between those patterns. Forecast systems that predict weather from weeks to months ahead showed a clear hierarchy: tropical-driven events were the most predictable, while events associated with two extratropical wave trains were the least predictable.
Why the types matter
Rapidly escalating heat extremes can leave less time for communities, farmers and businesses to prepare. Showing that these events have different physical pathways—and different levels of predictability—provides a process-based basis for improving forecasts of high-impact heat in North China.
The findings also indicate that forecast performance may depend on which type of heat event is developing. Tropical-driven events may offer more useful subseasonal warning information than events tied to two high-latitude wave patterns.
Evidence and limits
The researchers combined analyses of observed events from 1940–2024 with thermodynamic and atmospheric-circulation analyses, and examined forecasts from subseasonal-to-seasonal dynamical prediction systems. The abstract also refers to climate model experiments.
The findings are specific to rapidly escalating summer heat extremes in North China and to the four types identified in this study. The abstract does not provide numerical forecast-skill scores, details of the prediction systems, or comparisons with other regions, so it does not establish how well the results generalize beyond this setting.
// Source
Communications Earth & Environment · 2026 · DOI: 10.1038/s43247-026-04031-8
Authors: delete, Wenqi Zhang, Yixuan He, Jing Yang, Qing Bao, Shuai Hu, Huayu Chen, Qinmei Han, Chen Xiaochen
Institutions: Chinese Academy of Sciences, Nanjing University of Information Science and Technology, Beijing Normal University, Institute of Atmospheric Physics, China Meteorological Administration, Max Planck Institute for Meteorology, International Max Planck Research School on Earth System Modelling