Owner Profile

基于ECMWF、GFS、ICON、UKMO等8大全球数值预报引擎，融合历史误差动态加权，输出概率化最高温预测结果。覆盖全球20+重点城市，数据每日更新。

This system leverages eight global numerical weather prediction models including ECMWF, GFS, ICON and UKMO, and adopts dynamic weighting based on historical forecast errors to generate probabilistic prediction outputs (mean value μ and uncertainty σ). It delivers superior accuracy compared with single-model forecasts. Covering over 20 major global cities, the dataset is updated daily.