首页|期刊导航|中国科学数据(中英文网络版)|基于多模式融合的中国西南地区地表蒸散发数据集

基于多模式融合的中国西南地区地表蒸散发数据集OA

A dataset of surface evapotranspiration based on multi-model fusion in southwest China

中文摘要英文摘要

蒸散发(ET)是联系陆地生态系统和气候系统的重要纽带,明晰其时空变化对于评估区域水资源承载力和生态系统响应气候变化具有重要意义.为提升未来情景下西南地区 ET 估算的可靠性,本研究基于高精度的 SiTHv2 数据系统评估了 28 个 CMIP6 模式的 ET 模拟能力,并遴选出 5 个最优模式;随后采用分位数增量映射方法对其历史期与未来情景进行偏差校正,并利用贝叶斯模型平均(BMA)方法融合多模式结果,最终构建了西南地区 0.1°分辨率的 ET 数据集,涵盖 CMIP6 历史全强迫模拟(1982-2020)及未来情景(2026-2100)下的 3 种典型共享社会经济路径情景(SSP1-2.6、SSP2-4.5、SSP5-8.5).历史期验证结果表明,融合后的 BMA-ET 能较好反映区域东高西低的空间分布格局,相较于单一模式表现出更高精度,相关系数提升至 0.93,RMSE 和 MAE 分别降至 8.63 和 6.05 mm/month.在西南地区 15 个通量观测站点的独立验证中,BMA-ET 亦表现出良好一致性(R²=0.72,RMSE=19.24 mm/month,MAE=13.65 mm/month),进一步证明数据集的可靠性.本数据集在严格控制模式偏差和多模式不确定性的基础上,具有覆盖时段长、空间分辨率高、历史与未来统计特征一致等优势,可为中国西南地区水资源演变研究、生态系统响应分析和气候变化影响评估提供科学可靠的数据支撑.

Evapotranspiration(ET)is a crucial link between terrestrial ecosystems and the climate system,and understanding its spatiotemporal variations is essential for assessing regional water resource carrying capacity and ecosystem responses to climate change.To improve the reliability of ET estimates over Southwest China under future scenarios,this study evaluated the ET simulation performance of 28 CMIP6 models using the high-accuracy SiTHv2 dataset and selected the five best-performing models.Subsequently,quantile delta mapping(QDM)was applied to bias-correct both the historical and future simulations,and Bayesian model averaging(BMA)was used to integrate the multi-model outputs.A continuous ET dataset at 0.1° resolution for Southwest China was thus developed,covering the CMIP6 historical all-forcing simulations(1982-2020)and three representative Shared Socioeconomic Pathway scenarios(SSP1-2.6,SSP2-4.5,SSP5-8.5)for the future period(2026-2100).Validation for the historical period shows that the fused BMA-ET effectively captures the spatial pattern of higher ET in the east and lover ET in the west and achieves higher accuracy than any individual model,with the correlation coefficient increasing to 0.93 and RMSE and MAE decreasing to 8.63 and 6.05 mm/month,respectively.Independent validation at 15 flux observation sites across Southwest China further confirms the dataset's reliability,with BMA-ET showing strong good agreement with observations(R²=0.72,RMSE=19.24 mm/month,MAE=13.65 mm/month).By rigorously controlling model bias and multi-model uncertainty,the dataset offers long temporal coverage,high spatial resolution,and consistent statistical properties between historical and future periods.It provides a scientifically robust foundation for studies of water resource dynamics,ecosystem responses,and climate-change impact assessments in Southwest China.

赵清梅;王根绪;孙菊英;林杉;胡兆永

四川大学水利水电学院山区河流保护与治理全国重点实验室,成都 610065四川大学水利水电学院山区河流保护与治理全国重点实验室,成都 610065四川大学水利水电学院山区河流保护与治理全国重点实验室,成都 610065四川大学水利水电学院山区河流保护与治理全国重点实验室,成都 610065四川大学水利水电学院山区河流保护与治理全国重点实验室,成都 610065

蒸散发中国西南地区贝叶斯模型平均CMIP6

EvapotranspirationSouthwest ChinaBayesian Model AveragingCMIP6

《中国科学数据(中英文网络版)》 2026 (3)

344-356,13

科技基础资源调查专项(2022FY100205)国家自然科学基金(42201146).Scientific and Technological Basic Resources Survey Project(2022FY100205)National Natural Science Foundation of China(42201146).

10.11922/11-6035.nesdc.2025.0229.zh

评论