CMIP6情景下中国自然径流预测数据集（2041-2100年）

该数据集为中国未来多气候情景逐月自然径流(深)栅格数据，涵盖EC-Earth3模式下中等排放(SSP245)与高排放(SSP585)两种情景，空间分辨率为0.25°×0.25°,时间分辨率为月尺度，数据单位为mm，时间包含近期(2041-2050)，中期(2061-2070)，远期(2091-2100)未来三个时间段。数据由团队构建的自注意力神经网络(SAANN)模型生成，该模型在测试集表现良好，决定系数R²达到0.86，相关研究成果已在SCI期刊上发表出版，可以为未来气候变化影响水循环过程的研究提供本底数据支持。This dataset is a gridded dataset of monthly natural runoff (deep layer) over China under multiple future climate scenarios. It incorporates two greenhouse gas emission scenarios: SSP245 (medium-emission) and SSP585 (high-emission), as simulated by the EC-Earth3 global climate model. The dataset has a spatial resolution of 0.25° × 0.25° and a monthly temporal resolution, with a data unit of millimeters (mm), and covers three future time periods: the near term (2041–2050), mid-term (2061–2070), and long term (2091–2100). This dataset was generated using a self-attention artificial neural network (SAANN) model developed by the research team. This model delivered excellent performance on the test dataset, with a coefficient of determination (R²) reaching 0.86. Relevant research findings have been published in SCI-indexed journals, and this dataset can serve as baseline data to support studies on the impacts of future climate change on water cycle processes.

搜集汇总

背景与挑战

背景概述

该数据集是中国未来多气候情景下的逐月自然径流栅格数据，涵盖EC-Earth3模式下中等排放(SSP245)与高排放(SSP585)两种情景，空间分辨率为0.25°×0.25°，时间分辨率为月尺度，数据单位为mm。数据集由自注意力神经网络(SAANN)模型生成，模型决定系数R²达到0.86，适用于气候变化影响水循环过程的研究。