Climate variability plays a crucial role in the annual fluctuations of crop yields, posing a substantial threat to food security. Maize, the main cereal in sub-Saharan Africa, has shown varied yield trends during increasingly warmer growing seasons. Here we explore how sub-seasonal dry–wet spell patterns contribute to this variability, considering the spatial heterogeneity of crop responses, to map weather-related risks at a regional level. Our results show that shifts in specific dry–wet spell patterns across growth stages influence maize yield fluctuations in sub-Saharan Africa, explaining up to 50–60% of the interannual variation, which doubles that explained by mean changes in precipitation and temperature (30–35%). Precipitation primarily drives the onset of dry spells, while the influence of temperature increases with event intensity and peaks at the start of the growing season. Our large-scale, data-limited analysis approach has the potential to inform climate-smart agriculture in developing regions.

气候变化对农作物产量的年度波动起着至关重要的作用，对粮食安全构成重大威胁。玉米是撒哈拉以南非洲地区的主要谷物，在日益温暖的生长季节中，其产量趋势出现了变化。在这里，我们探讨次季节干湿期模式如何导致这种变化，考虑作物反应的空间异质性，以绘制区域层面的天气相关风险。我们的结果表明，不同生长阶段特定干湿期模式的变化影响撒哈拉以南非洲的玉米产量波动，解释了高达 50-60% 的年际变化，是降水和温度平均变化解释的两倍（30 –35%）。降水主要推动干旱期的到来，而温度的影响则随着事件强度的增加而增加，并在生长季节开始时达到峰值。我们的大规模、数据有限的分析方法有潜力为发展中地区的气候智能型农业提供信息。