12 Jun 2026

Automated Snow Cover Detection on Mountain Glaciers Using Spaceborne Imagery and Machine Learning

Abstract: Tracking the extent of seasonal snow on glaciers over time is critical for assessing glacier vulnerability and the response of glacierized watersheds to climate change. We present an automated snow detection workflow for mountain glaciers using supervised machine-learning-based image classifiers and Landsat 8 and 9, Sentinel-2, and PlanetScope satellite imagery. We develop the image classifiers by testing numerous machine learning algorithms with training and validation data. The workflow produces daily to twice monthly time series of several glacier mass balance and snowmelt indicators from 2013 to present. Workflow performance is assessed by comparing automatically classified images and snow lines to manual interpretations at each glacier site. The image classifiers exhibit over-all accuracies of 92 %–98 %, κ scores of 84 %–96 %, and F scores of 93 %–98 % for all image products. The median difference between automatically and manually delineated median snow line altitudes is −31 m across all image products. The Sentinel-2 classifier produces the most accurate glacier mass balance and snowmelt indicators and distinguishes snow from ice and firn the most reliably. Although they are less accurate, the Landsat- and PlanetScope-derived estimates greatly enhance the temporal coverage of observations. The transient accumulation area ratio produces the least noisy time series, making it the most reliable indicator for characterizing seasonal snow trends. The temporally detailed accumulation area ratio time series reveal the timing of minimum snow cover conditions varies by up to a month between Arctic and midlatitude sites, underscoring the potential for bias when estimating glacier minimum snow cover conditions from a single late-summer image. Widespread application of our automated snow detection workflow has the potential to improve regional assessments of glacier mass balance, land ice representations within Earth system models, water resources, and the impacts of climate change on snow cover across broad spatial scales.

12 Jun 2026

Spatiotemporal Patterns of Accumulation and Surface Roughness in Interior Greenland with a GNSS-IR Network

Abstract: The dry-snow zone is the largest region of the Greenland Ice Sheet, yet temporally and spatially dense observations of surface accumulation and surface roughness in this area are lacking. We use the global navigation satellite system interferometric reflectometry (GNSS-IR) technique with a novel, low-cost GNSS network of 12 stations in the vicinity of the ice sheet summit to reveal temporal and spatial patterns of accumulation of the upper snow layer. We show that individual measurements are highly precise, while the aggregate of hundreds of daily measurements across a large spatial footprint can detect millimeter-level surface changes and is biased by −2.7 ± 3.0 cm com-pared to a unique validation data set that covers a similar spatial extent to the instrument sensing footprint. Using the validation data set, we find that the reflectometry technique is most sensitive to the surrounding 4–20 m of the surface, with the GNSS antenna at a height of 1–2 m above ground level. Along with an exceptionally high accumulation rate at the beginning of the study, we also detect an across-slope dependence in accumulation rates at yearly timescales. For the first time, we also validate GNSS-IR sensitivity to meter-scale surface heterogeneities such as sastrugi, and we construct a time series of surface roughness evolution that suggests a seasonal pattern of heightened wintertime roughness features in this region. These surface accumulation and rough-ness measurements provide a novel data set for these critical variables and show a statistically significant relationship with occurrences of both high winds and precipitation events but only moderate correlations, suggesting that other processes may also contribute to accumulation and enhanced surface roughness in the interior region of Greenland.

2 Apr 2026

Analyzing Historical Snow Trends in Interior Alaska

Abstract: This study examines 40 years (water years 1982–2021) of snowpack characteristics to consider its hydrological implications in the 5350 km² Chena River basin. Using observations and a fine-scale physics model, we analyzed trends of snow water equivalent (SWE), snow onset and disappearance, and snow cover duration (SCD). New hydrological insights for the region: Results indicate a decline in SWE across the modeled domain, averaging a decrease of 3 mm per decade, with larger decreases (up to 10 mm per decade) at lower elevations. While domain-averaged SWE trends were not statistically significant, observed SCD showed statistically significant decreases: - 5.2, - 5.0, and - 4.4 days per decade at Teuchet Creek, Fairbanks F.O., and Little Chena Ridge, respectively. Notably, observations at SNOTEL stations and modeling revealed no statistically significant change in domain-averaged Rain-on-Snow (ROS) events over the 40-year period, contrasting some regional future estimates of increased ROS frequency. Peak streamflow did not consistently correlate with peak SWE levels, suggesting that other environmental factors such as ROS events and rapid temperature increases (e.g., a 10◦C spike observed in 1992) are key drivers of hydrological outcomes. These findings improve understanding of complex subarctic hydrological processes impacting permafrost and highlight the need for adaptive water resource management to mitigate multi-factor risks like flooding and wildfire, requiring proactive planning.

16 Sep 2025

Standard Operating Procedures for the Design, Maintenance, and Operation of Arctic and Subarctic Winter Roads

Abstract: Operations in cold regions require vehicular maneuvering across snowpacks or frozen surfaces. Winter roads and their route determination, construction, and monitoring are widely studied. This report analyzes historical and current literature on winter road construction and operations, reviews risk assessment techniques, examines the impact of uncertain weather on road reliability, and provides a standard operating procedure for design, maintenance, and use. Winter roads, snow roads, ice roads, and ice bridges enable seasonal access in Arctic and Subarctic regions. They allow cross-country maneuverability over terrain like wetlands and bogs, which are impassable in summer. These roads are critical for training, logistics, and construction in areas without all-season access. When combined with ice bridges they can provide near-unlimited travel. Effectiveness depends on proper planning, construction, and monitoring. Snow roads require controlled compaction for strength, while ice roads require sufficient ice thickness to support loads. Both rely on tools like visual inspections, ground-penetrating radar, and unmanned aerial systems to ensure safety. With extreme seasonal variability, adaptive strategies are essential. Shortened seasons and unpredictable freeze–thaw cycles demand modern technologies, predictive weather modeling, and improved reinforcement. This report integrates historical knowledge with engineering advancements to improve winter road durability, reduce risks, and support cold-region operations.

15 Sep 2025

Standard Operating Procedures for the Design, Construction, and Maintenance of Linear Infrastructure in Fens in Cold Regions

Abstract: In Alaska and across the Arctic and Subarctic, winter conditions can enable the expansion of linear infrastructure across the frozen landscape of fen wetlands. This expands military training opportunities into lowland wet, boggy, mostly impassable terrain. However, there are personnel, civilian, and environmental risks from using fens as travel corridors and drop zones. The effective design, construction, operation, and maintenance of such infrastructure on fens supports the dual mandate of troop training to fulfill the mission and protect the environment. This Technical Report (TR) addresses the risks of the establishment and use of linear infrastructure on the DoD lands in Alaska and in other austere cold environments where the DoD operates. This TR is founded on a review of methods used by US Army Installations, focusing primarily on Fort Wainwright in Interior Alaska. It establishes basic standard operating procedures (SOPs) by drawing on federal agency and international best practices and emerging research in circumpolar regions and beyond. This TR serves as a reference document for military land and infrastructure planners and unit leadership to create and maintain linear infrastructure on fens as environmental challenges evolve and opportunities develop to further the Army mission in high latitude environments.

10 Sep 2025

Snow Depth Measurements from Arctic Tundra and Boreal Forest Collected During NASA SnowEx Alaska Campaign

Abstract: Boreal forest and Arctic tundra environments collectively hold the largest percentage of global terrestrial seasonal snow cover. Тhe in-situ snow measurement network is sparse and costly in these remote northern regions. Here, we complement existing snow depth monitoring in Arctic tundra and boreal forest by presenting an extensive (64°N–70°N) snow depth dataset and description of ground-based snow depth measurements collected during the NASA SnowEx Alaska intensive field campaign, March 7–16, 2023. We also report the accuracy of snow depth measurements in shallow boreal forest and Arctic tundra snowpack and share considerations in developing the consistent and repeatable snow depth data collection procedures. Snow depth measurements and technical validation described in this paper can serve as a robust product for testing snow remote sensing techniques, and for providing a reference dataset for climatological and hydrological studies.

6 May 2025

Applications of the CRREL–-Geometric Optics Snow Radiative Transfer (GOSRT) Model: Incorporating Diffraction and Simulating Detection of Buried Targets

Abstract: Radiative transfer through a snow surface within the visible and near infrared (NIR) spectra is complicated by the shape, size, and configuration of the snow grains that comprise the snow surface. Ray-tracing and photon-tracking techniques combined with 3D renderings of snow resolved at the microscale have shown promise as a means to directly simulate radiative transfer through snow with no restrictions on the snow grain configuration. This report describes and evaluates the US Army Cold Regions Research and Engineering Laboratory (CRREL) Geometric Optics Snow Radiative Transfer (GOSRT) model. In particular, we describe the incorporation of the diffraction process into the photon-tracking framework and evaluate how accurately the model simulates the spectral albedo of targets buried within the snow. We find that the model simulated spectral albedo is little affected by the incorporation of diffraction for most applications. However, there are nonnegligible impacts on simulated albedo for small grains in the NIR due to a reduction in forward scattering. We conclude by recommending that diffraction is neglected in CRREL–GOSRT for most cases, as including it substantially increases the computational expense with minimal impacts on the result. Finally, we show that buried targets are only distinguishable for very shallow snowpacks.

14 May 2024

Snow-Impacted National Inventory of Dams by GAGESII Watershed

Abstract: This Engineering Research and Development Center (ERDC) Technical Note describes the development of a set of locations within the contiguous United States (CONUS) where snowmelt is a component of the annual streamflow. The locations are selected from the US Geological Survey (USGS) Geospatial Attributes of Gages for Evaluating Streamflow II (GAGESII) and National Inventory of Dams (NID) data sets. The 30-year normal snow regimes were used to identify all GAGESII watersheds that have any of the basin delineated as transitional (rain/snow), snow dominated, or perennial snow zones. NID dams that are within snow affected GAGESII watersheds are included in the data set. The purpose of this ERDC Technical Note is to describe the development of a comprehensive data set of CONUS GAGESII and dam infrastructure affected by snow changing regimes.

29 Sep 2023

Microbial Activity in Dust-Contaminated Antarctic Snow

Abstract: During weather events, particles can accumulate on the snow near the Pegasus ice and Phoenix compacted-snow Runways at the US McMurdo Station in Antarctica. The deposited particles melt into the surface, initially forming steep-sided holes, which can widen into patches of weak and rotten snow and ice. These changes negatively impact the ice and snow runways and snow roads trafficked by vehicles. To understand the importance of microbes on this process, we examined deposited dust particles and their microbial communities in snow samples collected near the runways. Snow samples were analyzed at the Cold Regions Research and Engineering Laboratory where we performed a respiration study to measure the microbial activity during a simulated melt, isolated microorganisms, examined particle-size distribution, and performed 16S rRNA gene sequencing. We measured higher levels of carbon dioxide production from a sample containing more dust than from a sample containing less dust, a finding consistent with viable dust-associated microbial communities. Additionally, eleven microorganisms were isolated and cultured from snow samples containing dust particles. While wind patterns and satellite images suggest that the deposited particles originate from nearby Black Island, comparisons of the particle size and chemical composition were inconclusive.

28 Sep 2023

Incorporating Advanced Snow Microphysics and Lateral Transport into the Noah-Multiparameterization (Noah-MP) Land Surface Model

Abstract: The dynamic state of the land surface presents challenges and opportunities for military and civil operations in extreme cold environments. In particular, the effects of snow and frozen ground on Soldier and vehicle mobility are hard to overstate. Current authoritative weather and land models are run at global scales (i.e., dx > 10 km) and are of limited use at the Soldier scale (dx < 100 m). Here, we describe several snow physics upgrades made to the Noah-Multiparameterization (Noah-MP) community land surface model (LSM). These upgrades include a blowing snow overlay to simulate the lateral redistribution of snow by the wind and the addition of new prognostic snow microstructure variables, namely grain size and bond radius. These additions represent major upgrades to the snow component of the Noah-MP LSM because they incorporate processes and methods used in more specialized snow modeling frameworks. These upgrades are demonstrated in idealized and real-world applications. The test simulations were promising and show that the newly added snow physics replicate observed behavior with reasonable accuracy. We hope these upgrades facilitate ongoing and future research on characterizing the effects of the integrated snow and soil land surface in extreme cold environments at the tactical scale.