中国区域CMIP6降尺度降水温度风速数据集（1979-2100）

中国区域0.25°日降水量、最高温度、最低温度和近地表全风速驱动数据集，由六个CMIP6模式（CanESM5、FGOALS-g3、GFDL-CM4、IPSL-CM6A-LR、MPI-ESM1-2-HR和MRI-ESM2-0），以中国区域地表气象要素驱动数据集（CMDF）格点降水和温度为实测参考，在将模式双线性插值至0.25°网格的基础上，用等距离累积分布函数法，对模式的日降水、最高温度和最低温度进行降尺度偏差校正。数据时间范围为历史期1979-2014年，和未来SSP2-RCP4.5和SSP5-RCP8.5两种情景下的2015-2100年。
统计降尺度后模式月降水量、最高温度和最低温度的平均绝对误差大幅度下降，与实测值相关性接近于1。驱动可变下渗容量（VIC）水文模型能得到较好的径流模拟结果，可用于中国区域的未来气候分析与流域水文预估研究。

A China-region 0.25° driving dataset of daily precipitation, maximum temperature, minimum temperature, and near-surface total wind speed was developed based on six CMIP6 models (CanESM5, FGOALS-g3, GFDL-CM4, IPSL-CM6A-LR, MPI-ESM1-2-HR, and MRI-ESM2-0). Using the gridded precipitation and temperature from the China-region surface meteorological element driving dataset (CMDF) as in-situ reference, the models were first bilinearly interpolated to a 0.25° grid, then subjected to downscaling bias correction for their daily precipitation, maximum temperature, and minimum temperature via the equidistant cumulative distribution function method. The dataset covers the historical period from 1979 to 2014, as well as two future scenarios: SSP2-RCP4.5 and SSP5-RCP8.5, spanning 2015 to 2100.
After statistical downscaling, the mean absolute error (MAE) of the monthly precipitation, maximum temperature, and minimum temperature of the models decreased substantially, with their correlation coefficients with in-situ observations approaching 1. When used to drive the Variable Infiltration Capacity (VIC) hydrological model, the dataset yields satisfactory runoff simulation results, making it suitable for future climate analysis and watershed hydrological projection studies across China.

搜集汇总

背景与挑战

背景概述

该数据集是一个覆盖中国区域的高分辨率气候数据集，基于六个CMIP6模式，通过统计降尺度方法生成了1979年至2100年的日降水量、最高温度、最低温度和近地表全风速数据。数据包括历史期和未来两种气候情景（SSP2-RCP4.5和SSP5-RCP8.5），空间分辨率为0.1°至0.25°，经过偏差校正后误差大幅降低，适用于未来气候分析和流域水文预估研究。