Evaluation of snow depth and snow cover represented by multiple datasets over the Tianshan Mountains: Remote sensing, reanalysis, and simulation.

Mountain snowpacks play important roles in water resource and ecological system. However, existing snow depth products show great uncertainties across the Tianshan Mountains, Central Asia. This study evaluated and compared four snow depth datasets over the Tianshan Mountains, including snow depth datasets from ERA5, ERA5‐Land, passive microwave, as well as a dynamically downscaled simulation by Weather Research and Forecasting (WRF) model. The snow depth was evaluated against in situ observations while the snow cover extent was evaluated by interactive multisensor snow and ice mapping system (IMS) snow cover. Furthermore, the snow‐related metrics, such as annual mean snow depth and snow cover days, were compared among the four datasets. The results showed that the spatial patterns of snow‐related metrics were relatively consistent among datasets, although discrepancies existed in magnitude. Additionally, the temporal variations in snow‐related metrics depended on the dataset employed. The simulated snow depth from WRF got the lowest bias in daily snow depth value against the in situ observations (mean bias = 0.12 cm, root mean square error = 2.08 cm), additionally, it had the best performance in classifying correct snow grids compared with IMS (probability of detection = 0.766). It also outperformed in annual mean snow depth and snow cover days by 47.2 and 13.3% compared with ERA5‐Land snow depth. This study highlights the confidence in characterizing both the spatial pattern and temporal variations of snow depth products based on WRF model simulation. [ABSTRACT FROM AUTHOR]