Snow water equivalent (SWE) variability and its drivers over different regions remain uncertain due to lack of representativeness of point measurements and deficiencies of existing coarse‐resolution SWE products. Here, for the first time, we quantify and understand the snowpack change from 1982 to 2016 over conterminous United States at 4‐km pixels. Annual maximum SWE decreased significantly (p < 0.05) by 41% on average for 13% of snowy pixels over western United States. Snow season was shortened significantly by 34 days on average for 9% of snowy pixels over the United States, primarily caused by earlier ending and later arrival of the season over western and eastern United States, respectively. October–March mean temperature and accumulated precipitation largely explain the temporal variability of 1 April SWE over western United States, and considering temperature alone would exaggerate the warming effect on SWE decrease. In contrast, temperature plays the primary role in the 1 April SWE variability over eastern United States.

In the semiarid interior western USA, where a majority of surface water supply comes from mountain forests, high‐resolution aerial lidar‐based surveys are commonly used to study snow. These surveys provide rich information about snow depth, but they are usually not accompanied with spatially explicit measurements of snow density, which leads to uncertainty in the estimation of snow water equivalent (SWE). In this study, we use a novel approach to distribute ~300 field measurements of snow density with artificial neural networks. We combine the resulting density maps with aerial lidar snow depth measurements, bias corrected with a very large and precisely geolocated array of field‐measured snow depths (~4,000 observations), to create and validate maps of snow depth, snow density, and SWE over two sites along Arizona's Mogollon Rim in February and March 2017. These maps show differences between midwinter and late‐winter snow conditions. In particular, compared to that of snow depth, the spatial variability of snow density is smaller for the later snow survey than the earlier snow survey. These gridded data also show that the representativeness of Snow Telemetry and other point measurements is different for the midwinter and late‐winter snow surveys. Overall, the lidar artificial neural network SWE estimates can be as much as 30% different than if Snow Telemetry density were used with lidar snow depths to estimate SWE.

Accumulation of snowfall during winter and snowmelt in the subsequent spring or earlier summer provides a dominant water source in alpine regions. Most land surface and hydrological models use near‐surface air temperature (T_a) thresholds to partition precipitation into snow and rain, underestimating snowfall over drier regions. We developed a snow‐rain partitioning scheme using the wet‐bulb temperature (T_w), which is closer to the surface temperature of a falling hydrometeor than T_a. T_w becomes more depressed in drier environments as derived from T_w depression equation using T_a and surface air humidity, resulting in a greater fraction of snowfall. We implemented this new T_w scheme in the Noah‐MP land surface model and evaluated the model against a high‐quality ground‐based snow product over the contiguous United States. The results suggest that the new T_w scheme substantially improves the model skill in simulating snow depth and snow water equivalent over most snow‐covered grids, especially the higher and drier continental mountain ranges in the Western United States, while it retains the modeling accuracy over the more humid Eastern United States.

It is critically important but challenging to estimate the amount of snow on the ground over large areas due to its strong spatial variability. Point snow data are used to generate or improve (i.e., blend with) gridded estimates of snow water equivalent (SWE) by using various forms of interpolation; however, the interpolation methodologies often overlook the physical mechanisms for the snow being there in the first place. Using data from the Snow Telemetry and Cooperative Observer networks in the western United States, we show that four methods for the spatial interpolation of peak of winter snow water equivalent (SWE) and snow depth based on distance and elevation can result in large errors. These errors are reduced substantially by our new method, i.e., the spatial interpolation of these quantities normalized by accumulated snowfall from the current or previous water years. Our method results in significant improvement in SWE estimates over interpolation techniques that do not consider snowfall, regardless of the number of stations used for the interpolation. Furthermore, it can be used along with gridded precipitation and temperature data to produce daily maps of SWE over the western United States that are comparable to existing estimates (which are based on the assimilation of much more data). Our results also show that not honoring the constraint between SWE and snowfall when blending in situ data with gridded data can lead to the development and propagation of unrealistic errors.

There is a large uncertainty of snow water equivalent (SWE) in reanalyses and the Global Land Data Assimilation System (GLDAS), but the primary reason for this uncertainty remains unclear. Here several reanalysis products and GLDAS with different land models are evaluated and the primary reason for their deficiencies are identified using two high-resolution SWE datasets, including the Snow Data Assimilation System product and a new dataset for SWE and snowfall for the conterminous United States (CONUS) that is based on PRISM precipitation and temperature data and constrained with thousands of point snow observations of snowfall and snow thickness. The reanalyses and GLDAS products substantially underestimate SWE in the CONUS compared to the high-resolution SWE data. This occurs irrespective of biases in atmospheric forcing information or differences in model resolution. Furthermore, reanalysis and GLDAS products that predict more snow ablation at near-freezing temperatures have larger underestimates of SWE. Since many of the products do not assimilate information about SWE and snow thickness, this indicates a problem with the implementation of land models and pinpoints the need to improve the treatment of snow ablation in these systems, especially at near-freezing temperatures.

Accurate mapping of snow amount in the mountains is critical as mountain snowpacks are water supply for millions of people. Satellite remote sensing has been largely unable to reliably detect the amount of snowpack in these areas. Recently, C‐band Synthetic Aperture Radar (SAR) data from the Sentinel‐1 (S1) satellites have shown potential for measuring snow depth in the mountains. However, their spatiotemporal coverage is incomplete, and their evaluation with robust, aerial snow depth data is limited. Here, we evaluate two S1 snowpack datasets with some of the best available gridded snowpack data over the Colorado Rockies and Sierra Nevada mountains in the western US: the Airborne Snow Observatory (ASO) and the University of Arizona (UA) snowpack datasets. Compared to ASO and UA data, the S1 data are biased high when snow is shallow, and biased low when snow is deep (particularly later in spring when there is wet snow), though these biases are reduced for deep snow areas when wet snow pixels are removed. We then apply corrections based on machine learning that account for physiographic characteristics to improve the accuracy of the S1 data. Furthermore, we fill gaps in the S1 data by using snow persistence, but also account for potential snow accumulation and ablation, to generate temporally complete snow depth maps over mountainous areas. Corrected and gap‐filled S1 snow depth mapping could be especially important for snow monitoring in remote mountain areas where other techniques for snow mapping do not work or are logistically infeasible or cost‐prohibitive.