George A. Riggs

Snow reanalyses that combine process based models and remote sensing observations of snow provide estimates of seasonal snow water equivalent (SWE) evolution that surpass the accuracies of traditional modeling approaches. However, snow reanalyses are only available over smaller subregions and sometimes use computationally expensive modeling approaches. We investigate whether 1 km‐resolution and daily SWE from a popular reanalysis could be learned by connecting only trusted meteorological fields (multidecadal precipitation patterns and daily air temperature) and remotely sensed snow cover using a deep learning model. Relative to point observations of SWE evolution in the western United States, the deep learning model was able to reproduce the spatial and temporal evolution estimated by the snow reanalysis. Further, we found that the deep learning model was efficient and could be expanded geographically to estimate SWE evolution in the European Alps, demonstrating a high average coefficient of correlation (0.81) and low peak‐SWE bias (<1%) versus point estimates of SWE in seasonally snowy locations, with no statistical difference between regions at different elevations and with different forest cover. This study demonstrates how deep learning approaches could be used to mine connections between daily SWE evolution, snow cover remote sensing, and limited meteorological information to generate and expand the geographical extent of fine‐resolution historical snow estimates in complex terrains.

Snow reanalyses that combine process-based models and remote sensing observations of snow provide estimates of seasonal snow water equivalent (SWE) evolution that surpass the accuracies of traditional modeling approaches. However, snow reanalyses are only available over smaller subregions, and sometimes use computationally expensive modeling approaches. We investigate whether 1 km-resolution and daily SWE from a popular reanalysis could be learned by connecting only trusted meteorological fields (multidecadal precipitation patterns and daily air temperature) and remotely sensed snow cover using a deep learning model. Relative to point observations of SWE evolution in the western United States, the lightweight deep learning model was able to reproduce the spatial and temporal evolution estimated by the snow reanalysis. Further, we found that the deep learning model could be trained in the western United States and then reused to estimate SWE evolution in the European Alps, demonstrating a high average coefficient of correlation (0.81) and low peak-SWE bias (< 1%) versus point estimates of SWE. SWE from the deep learning model also outperformed SWE estimates from physically based land surface simulations, capturing elevation-driven impacts on SWE spatial heterogeneity and interannual differences in seasonal SWE magnitudes important for water resources, climate regulation, and local ecology. This study demonstrates how deep learning approaches could be used to mine connections between daily SWE evolution, snow cover remote sensing, and limited meteorological information to generate and expand the geographical extent of fine-resolution historical snow estimates in complex terrains.

The snowfall in the Baltimore/Washington metropolitan area during the winter of 2009/2010 was unprecedented and caused serious snow‐related disruptions. In February 2010, snowfall totals approached 2 m, and because maximum temperatures were consistently below normal, snow remained on the ground the entire month. One of the biggest contributing factors to the unusually severe winter weather in 2009/2010, throughout much of the middle latitudes, was the Arctic Oscillation. Unusually high pressure at high latitudes and low pressure at middle latitudes forced a persistent exchange of mass from north to south. In this investigation, a concerted effort was made to link remotely sensed falling snow observations to remotely sensed snow cover and snowpack observations in the Baltimore/Washington area. Specifically, the Advanced Microwave Scanning Radiometer onboard the Aqua satellite was used to assess snow water equivalent, and the Advanced Microwave Sounding Unit‐B and Microwave Humidity Sounder were employed to detect falling snow. Advanced Microwave Scanning Radiometer passive microwave signatures in this study are related to both snow on the ground and surface ice layers. In regard to falling snow, signatures indicative of snowfall can be observed in high frequency brightness temperatures of Advanced Microwave Sounding Unit‐B and Microwave Humidity Sounder. Indeed, retrievals show an increase in snow water equivalent after the detection of falling snow. Yet, this work also shows that falling snow intensity and/or the presence of liquid water clouds impacts the ability to reliably detect snow water equivalent. Moreover, changes in the condition of the snowpack, especially in the surface features, negatively affect retrieval performance. Copyright © 2011. This article is a U.S. Government work and is in the public domain in the USA.