Future changes in rainfall have serious impacts on human adaptation to climate change, but quantification of these changes is subject to large uncertainties in climate model projections. To narrow these uncertainties, significant efforts have been made to understand the intermodel differences in future rainfall changes. Here, we show a strong inverse relationship between present-day precipitation and its future change to possibly calibrate future precipitation change by removing the present-day bias in climate models. The results of the models with less tropical (40° S–40° N) present-day precipitation are closely linked to the dryness over the equatorial central-eastern Pacific, and project weaker regional precipitation increase due to the anthropogenic greenhouse forcing^1,2,3,4,5,6 with stronger zonal Walker circulation. This induces Indo-western Pacific warming through Bjerknes feedback, which reduces relative humidity by the enhanced atmospheric boundary-layer mixing in the future projection. This increases the air–sea humidity difference to enhance tropical evaporation and the resultant precipitation. Our estimation of the sensitivity of the tropical precipitation per 1 K warming, after removing a common bias in the present-day simulation, is about 50% greater than the original future multi-model projection.

未来降雨的变化会严重影响人类对气候变化的适应能力，但对这些变化的量化将受到气候模型预测的巨大不确定性的影响。为了缩小这些不确定性，人们已经做出了巨大的努力来了解未来降雨变化中的模型间差异。在这里，我们显示了当前降水量及其未来变化之间的强反关系，从而可能通过消除气候模型中的当前偏差来校准未来的降水变化。目前降水较少的热带（40°S–40°N）模型的结果与赤道中东部太平洋的干旱密切相关，并且由于人为温室强迫^1,2，导致区域降水增加较弱^，3,4,5,6具有更强的纬向沃克环流。这通过Bjerknes反馈引起印度洋-西太平洋变暖，这将通过在未来的预测中增强的大气边界层混合来降低相对湿度。这增加了海-海湿度差，以增强热带蒸发和由此产生的降水。在当前模拟中消除常见偏差后，我们对每1 K变暖的热带降水敏感性的估计比最初的未来多模型预测高约50％。