An evaluation dataset for the skills of CMIP5 and CMIP6 models in simulating climate of China

General circulation model (GCM) simulations archived by the Coupled Model Intercomparison Project (CMIP) are crucial tools for climate science. However, with various GCM results simulated by different countries and institutions, researchers have difficulty in choosing appropriate models for their unique study area. To this end, this dataset provieds Tayler skill scores of 28 GCMs in simulating temperature and precipitation of 631 reference sites across China under daily, monthly and seasonally scales. These scores are calculated based on the observations of meteorological stations and historical simulations of GCMs during 1970-2005. 

The dataset is very important for researchers to select locally appropriate GCMs. For example, researchers concerned with climate change of Beijing could firstly download the GCMs with sound performance at station 54511 (i.e., NorESM2-LM, INM-CM5-0 and MPI-ESM1-2-LR for temperature and NorESM1-M, IPSL-CM5A-LR and INM-CM4 for precipitation), and then conduct the further works of downscaling.

由耦合模式比较计划（Coupled Model Intercomparison Project, CMIP）归档的大气环流模式（General Circulation Model, GCM）模拟结果，是气候科学研究的关键工具。然而，不同国家与科研机构产出的各类GCM模拟结果纷繁多样，研究者难以针对自身特定研究区域遴选适配的模式。为此，本数据集提供了28个GCM在日、月、季三种时间尺度下，模拟中国境内631个参考站点气温与降水的泰勒技能评分（Taylor skill score）。该评分基于1970年至2005年间的气象站观测数据与GCM历史模拟结果计算得到。

本数据集对于研究者遴选适配区域的GCM具有重要应用价值。例如，关注北京气候变化的研究者可先下载在54511号气象站表现优异的GCM：气温适配模式包括NorESM2-LM、INM-CM5-0与MPI-ESM1-2-LR，降水适配模式包括NorESM1-M、IPSL-CM5A-LR与INM-CM4，随后开展进一步的降尺度研究工作。