As the Earth warms, the spatial and temporal response of seasonal snow remains uncertain. The global snow science community estimates snow cover and mass with information from land surface models, numerical weather prediction, satellite observations, surface measurements, and combinations thereof. Accurate estimation of snow at the spatial and temporal scales over which snow varies has historically been challenged by the complexity of land cover and terrain and the large global extent of snow-covered regions. Like many Earth science disciplines, snow science is in an era of rapid advances as remote sensing products and models continue to gain granularity and physical fidelity. Despite clear progress, the snow science community continues to face challenges related to the accuracy of seasonal snow estimation. Namely, advances in snow modeling remain limited by uncertainties in modeling parameterization schemes and input forcings, and advances in remote sensing techniques remain limited by temporal, spatial, and technical constraints on the variables that can be observed. Accurate monitoring and modeling of snow improves our ability to assess Earth system conditions, trends, and future projections while serving highly valued global interests in water supply and weather forecasts. Thus, there is a fundamental need to understand and improve the errors and uncertainties associated with estimates of snow. A potential method to overcome model and observational shortcomings is data assimilation, which leverages the information content in both observations and models while minimizing their limitations due to uncertainty. This article proposes data assimilation as a way to reduce uncertainties in the characterization of seasonal snow changes and reviews current modeling, remote sensing, and data assimilation techniques applied to the estimation of seasonal snow. Finally, remaining challenges for seasonal snow estimation are discussed.

随着地球变暖，季节性降雪的空间和时间响应仍然不确定。全球雪科学界利用地表模型、数值天气预报、卫星观测、表面测量及其组合的信息来估计积雪和质量。历史上，由于土地覆盖和地形的复杂性以及全球积雪覆盖区域的广泛性，在雪量变化的空间和时间尺度上准确估计雪量一直受到挑战。与许多地球科学学科一样，随着遥感产品和模型不断获得粒度和物理保真度，雪科学正处于快速发展的时代。尽管取得了明显进展，但雪科学界仍然面临与季节性降雪估算准确性相关的挑战。也就是说，雪模拟的进步仍然受到建模参数化方案和输入强迫的不确定性的限制，遥感技术的进步仍然受到可观测变量的时间、空间和技术限制的限制。准确的雪监测和建模提高了我们评估地球系统状况、趋势和未来预测的能力，同时服务于供水和天气预报方面的全球高度重视的利益。因此，迫切需要了解和改进与降雪估计相关的误差和不确定性。克服模型和观测缺点的一种潜在方法是数据同化，它利用观测和模型中的信息内容，同时最大限度地减少由于不确定性而造成的局限性。本文提出数据同化作为减少季节性降雪变化特征的不确定性的一种方法，并回顾了当前应用于季节性降雪估计的建模、遥感和数据同化技术。最后，讨论了季节性降雪估算的剩余挑战。