积雪变化对全球地表能量平衡及水循环有着显著影响，但目前雪深数据还存在较大误差，各数据在不同时空条件下表现各异，因而限制了雪深数据在气候变化及水文模型中的应用。本研究拟对四种北半球典型的被动微波雪深数据及再分析资料雪深数据开展长时间序列的数据融合。通过对时空分组的多源雪深数据进行累积分布函数（CDF）匹配，降低数据整体的系统误差，通过Multiple Collocation估算数据的随机误差，并以此为参考分配权重，融合生成北半球1980-2018年的0.25°分辨率的逐日雪深融合数据集。根据地面观测数据和分类及回归树（CART）对融合数据开展逐像元的误差估计。本研究旨在制备一套高质量的北半球长时间序列融合雪深数据集，并提供逐像元的误差评估，以期为雪深在气候变化、水文水资源等领域的应用提供高精度的数据支持。

The significant seasonal and interannual variability of snow cover have notable influences on surface energy budget and global water balance. However, current snow depth datasets are still with large uncertainties and inconsistencies, which limits their applications on climate change projection and hydrological process simulation. In this study, four representative snow depth datasets will be fused over the Northern Hemisphere. The Cumulative Distribution Function (CDF) scale will be used to decrease the systematic error of the datasets, and the random error will be calculated based on Multiple Collocation, and the weight of each dataset can be assigned based on the random error. Then, the fused snow depth datasets can be generated over the Northern Hemisphere during 1980-2018. The fused data will be validated against ground observations, and uncertainty can be calculated through Classification And Regression Tree (CART) for each pixel. The purpose of this study is to generate a long-term high-quality snow depth dataset with error information, and to provide a reliable snow depth data source for climate change projection and hydrological simulations.