Evaluation of the temperature downscaling performance of PRECIS to the BCC-CSM2-MR model over China

In recent years, the Providing REgional Climates for Impacts Studies (PRECIS) developed by the UK Met Office Hadley Centre has been used widely for climate downscaling and has demonstrated superior performance in reproducing the historical climatology in many regions around the world. However, all previous studies based upon PRECIS were solely driven by the GCMs from the Hadley Centre (i.e., HadCM3 and HadGEM2-ES). This inevitably limits its capability of exploring the uncertainties of regional climate projections caused by different boundary conditions from many other GCMs developed by different research centers around the world. Therefore, in this study, we attempt to investigate the downscaling performance of PRECIS to other GCMs for the first time. In particular, we use the CMIP6 boundary conditions from the BCC-CSM2-MR model from Beijing Climate Center to drive the PRECIS model in order to generate 25 km high-resolution climate scenarios for China. The downscaled temperature simulations from PRECIS for the baseline period (i.e., 1979–2004) are compared to the observational data (CN05.1) to help evaluate the downscaling performance. The results show that PRECIS can provide an obvious improvement in simulating the annual mean temperature over China, particularly in the east of China, with the bias of ~ 1 °C. Future changes in mean temperature over China under the SSP245 and SSP585 emission scenarios are further investigated with PRECIS. The downscaled projections by PRECIS show a smaller increase in temperature compared to its driving BCC-CSM2-MR, except for the late of this century under the SSP585. This study is the first attempt to use PRECIS to downscale a non-Hadley-Centre GCM and the evaluation results suggest that PRECIS does present a superior downscaling performance. The results from this study will provide a scientific basis for using PRECIS to downscale other GCMs in support of the exploration of regional climate projection uncertainties associated with different boundary conditions.