A historical and future view of elevation-dependent precipitation change in CMIP6 models

Abstract

Elevation-Dependent Precipitation Change (EDPC) is increasingly recognised as a key feature of mountain climate change, yet its global characteristics and representation in climate models remain poorly understood. Here, we assess the ability of current-generation global climate models to capture EDPC and to provide projections of its future evolution. We analyse elevational trends in annual mean, heavy, and extreme daily precipitation for the historical (1951–2020) and future (2015–2099) periods across five mountain regions—Tibetan Plateau, US Rockies, Greater Alpine Region, northern Andes, and southern Andes—using 28 CMIP6 models and the ERA5 reanalysis as a reference. Model simulations are analysed using a clustering approach to identify groups of models exhibiting similar EDPC patterns. Results show inter-model variability, with cluster composition differing across regions and precipitation indices. During the historical period, distinct EDPC behaviours emerge: in some regions (e.g. the Tibetan Plateau), at least one cluster closely reproduces the ERA5 profile, whereas in others (e.g. the Andes and the US Rockies) no cluster matches ERA5. Projections reveal region- and cluster-dependent EDPC signals, precluding a robust characterisation of future EDPC. Models that best reproduce ERA5 EDPC behaviour do not constrain projected future responses. Nevertheless, cluster membership remains stable across emission scenarios (SSP2−4.5 and SSP5−8.5), with EDPC magnitude amplified under the higher-emission scenario, indicating modulation by climate change intensity. Finally, while the multi-model ensemble mean reproduces ERA5 reasonably well in the historical period, it tends to overestimate precipitation changes in projections, highlighting the value of clustering for assessing EDPC in climate model projections.