TY  - Data
T1  - A dataset of annual maximum snow depth with 10-meter resolution in typical snow regions of China during 2017–2023 
A1  - Shi Yingjie
A1  - Li Xiaofeng
A1  - Wei yanlin
A1  - pan Xin
A1  - Zheng Zhaojun
A1  - Li Yijing
DO  - 10.12072/ncdc.snow.db7725.2026
PY  - 2026
DA  - 2026-08-25
PB  - National Cryosphere Desert Data Center
AB  - Snow depth is a key parameter for quantifying terrestrial water storage and characterizing snow hydrological processes. As an extreme snow metric, Annual Maximum Snow Depth (SDmax) reflects the peak seasonal snow accumulation and maximum snow water storage, providing essential information for snowmelt runoff simulation, snow disaster risk assessment, and studies of permafrost stability and ecosystem dynamics. However, existing snow depth products are generally limited by coarse spatial resolution and considerable retrieval uncertainties, making it difficult to accurately characterize the spatial heterogeneity of snow distribution across complex terrain and diverse land surface conditions. In this study, we developed a high-resolution SDmax retrieval model by integrating the AlphaEarth Foundations (AEF) geospatial foundation model with in situ snow depth observations from meteorological stations using the Gradient Boosted Decision Trees (GBDT) model. Based on this framework, we produced the first 10 m Annual Maximum Snow Depth dataset for the major seasonal snow-covered regions of China spanning 2017–2023, with all products provided in GeoTIFF format. Independent validation against ground observations demonstrates that the dataset achieves good accuracy and spatial consistency, with a correlation coefficient (R) of 0.71, a root mean square error (RMSE) of 6.71 cm, and a mean absolute error (MAE) of 4.51 cm. This dataset can provide high-precision foundational data support for regional water resource assessment, snow disaster evaluation, and climate change research.
DB  - NCDC
UR  - https://www.ncdc.ac.cn/portal/metadata/7fda555c-ca26-4b83-b46b-b825e9b2adc2
ER  -