Monotonic Increase of Extreme Precipitation Intensity With Temperature When Controlled for Saturation Deficit.

Past studies based on univariate scaling analyses at the weather time scale documented a negative scaling of extreme precipitation intensity (EPI), which prevents EPI extrapolation from past climate. Here we present a bivariate scaling analysis and show that, contrary to the univariate scaling results, EPI monotonically increases with temperature and shows no negative scaling when controlled for saturation deficit. The observed EPI‐temperature relationship in saturated atmosphere is surprisingly similar among different regions and closely follows the Clausius‐Clapeyron scaling; climate models produce greater regional dependence of the scaling relationship with a wide range of scaling rate. For extratropical regions, the model‐simulated EPI‐temperature relationship under saturation shows a past‐to‐future continuity, which could potentially support extrapolation to a warmer climate. The scaling at saturation bridges the EPI‐temperature relationship between weather and climate time scales and may enable potential prediction of future precipitation extremes via extrapolation from past observations. Plain Language Summary: Predicting changes of extreme precipitation intensity (EPI) and their regional dependency is a crucial but challenging task. Extrapolation based on past observations has been hindered by the apparent decrease of EPI on hot days, a phenomenon often referred to as "negative scaling". Here we propose a bivariate scaling analysis to account for the impact of both temperature and atmospheric saturation deficit (SD), and show that under a constant SD, EPI increases monotonically with temperature and shows no negative scaling, which contradicts the findings from conventional scaling analysis. In saturated atmosphere, the EPI‐temperature relationship based on observations is surprisingly similar among different regions and is close to the thermodynamics‐based Clausius‐Clapeyron scaling; climate models produce a larger disparity among different regions with wide ranging scaling rates. As climate changes, the model simulated EPI‐temperature relationships in saturated atmosphere over extratropical regions show a certain degree of continuity between historical and warmer climates, contrasting the strong discontinuity in conventional analysis. The past‐to‐future continuity of the relationship makes it possible to derive future extreme precipitation through extrapolation from past observations. The scaling relationship in saturated atmosphere bridges the weather and climate time scales, and provides a new emergent constraint for climate model performance. Key Points: Under a constant saturation deficit, extreme precipitation intensity increases monotonically with temperature, showing no negative scalingThe observed precipitation scaling in saturated atmosphere shows little regional dependence and is close to the Clausius‐Clapeyron ratioClimate models produce larger regional dependence, but show some past‐to‐future continuity of the scaling relationship [ABSTRACT FROM AUTHOR]