Remote Sensing of 2000–2016 Alpine Spring Snowline Elevation in Dall Sheep Mountain Ranges of Alaska and Western Canada

Satellite mapping of red snow on North American glaciers

Red snow caused by blooms of microalgae darkens the surface of summer snowfields, increasing snowmelt. To assess the contribution of red snow to supraglacial snowmelt in northwestern North America, we systematically mapped the 2019-2022 distribution of blooms by applying supervised classification to 6158 satellite images. Blooms occurred on 5% of the total glaciated area, heavily affecting many glaciers in years of prolonged snow cover duration. Individual glaciers had up to 65% of their surface area affected by bloom in one melt season, which we estimate caused as much as 3 cm of snow meltwater equivalent averaged across the glacier surface. These results demonstrate appreciable snowmelt caused by red snow albedo over vast areas of North American glaciers.

Automated Glacier Snow Line Altitude Calculation Method Using Landsat Series Images in the Google Earth Engine Platform

Glacier snow line altitude (SLA) at the end of the ablation season is an indicator of the equilibrium line altitude (ELA), which is a key parameter for calculating and assessing glacier mass balance. Here, we present an automated algorithm to classify bare ice and snow cover on glaciers using Landsat series images and calculate the minimum annual glacier snow cover ratio (SCR) and maximum SLA for reference glaciers during the 1985–2020 period in Google Earth Engine. The calculated SCR and SLA values are verified using the observed glacier accumulation area ratio (AAR) and ELA. We select 14 reference glaciers from High Mountain Asia (HMA), the Caucasus, the Alps, and Western Canada, which represent four mountainous regions with extensive glacial development in the northern hemisphere. The SLA accuracy is ~73%, with a mean uncertainty of ±24 m, for 13 of the reference glaciers. Eight of these glaciers yield R2 > 0.5, and the other five glaciers yield R2 > 0.3 for their respective SCR–AAR relationships. Furthermore, 10 of these glaciers yield R2 > 0.5 and the other three glaciers yield R2 > 0.3 for their respective SLA–ELA relationships, which indicate that the calculated SLA from this algorithm provides a good fit to the ELA observations. However, Careser Glacier yields a poor fit between the SLA calculations and ELA observations owing to tremendous surface area changes during the analyzed time series; this indicates that glacier surface shape changes due to intense ablation will lead to a misclassification of the glacier surface, resulting in deviations between the SLA and ELA. Furthermore, cloud cover, shadows, and the Otsu method limitation will further affect the SLA calculation. The post-2000 SLA values are better than those obtained before 2000 because merging the Landsat series images reduces the temporal resolution, which allows the date of the calculated SLA to be closer to the date of the observed ELA. From a regional perspective, the glaciers in the Caucasus, HMA and the Alps yield better results than those in Western Canada. This algorithm can be applied to large regions, such as HMA, to obtain snow line estimates where manual approaches are exhaustive and/or unfeasible. Furthermore, new optical data, such as that from Sentinel-2, can be incorporated to further improve the algorithm results.

Habitat selection by Dall's sheep is influenced by multiple factors including direct and indirect climate effects

Arctic and boreal environments are changing rapidly, which could decouple behavioral and demographic traits of animals from the resource pulses that have shaped their evolution. Dall's sheep (Ovis dalli dalli) in northwestern regions of the USA and Canada, survive long, severe winters and reproduce during summers with short growing seasons. We sought to understand the vulnerability of Dall's sheep to a changing climate in Lake Clark National Park and Preserve, Alaska, USA. We developed ecological hypotheses about nutritional needs, security from predators, energetic costs of movement, and thermal shelter to describe habitat selection during winter, spring, and summer and evaluated habitat and climate variables that reflected these hypotheses. We used the synoptic model of animal space use to estimate parameters of habitat selection by individual females and calculated likelihoods for ecological hypotheses within seasonal models. Our results showed that seasonal habitat selection was influenced by multiple ecological requirements simultaneously. Across all seasons, sheep selected steep rugged areas near escape terrain for security from predators. During winter and spring, sheep selected habitats with increased forage and security, moderated thermal conditions, and lowered energetic costs of movement. During summer, nutritional needs and security influenced habitat selection. Climate directly influenced habitat selection during the spring lambing period when sheep selected areas with lower snow depths, less snow cover, and higher air temperatures. Indirectly, climate is linked to the expansion of shrub/scrub vegetation, which was significantly avoided in all seasons. Dall's sheep balance resource selection to meet multiple needs across seasons and such behaviors are finely tuned to patterns of phenology and climate. Direct and indirect effects of a changing climate may reduce their ability to balance their needs and lead to continued population declines. However, several management approaches could promote resiliency of alpine habitats that support Dall's sheep populations.

Seasonal influence of snow conditions on Dall's sheep productivity in Wrangell-St Elias National Park and Preserve

Dall's sheep (Ovis dalli dalli) are endemic to alpine areas of sub-Arctic and Arctic northwest America and are an ungulate species of high economic and cultural importance. Populations have historically experienced large fluctuations in size, and studies have linked population declines to decreased productivity as a consequence of late-spring snow cover. However, it is not known how the seasonality of snow accumulation and characteristics such as depth and density may affect Dall's sheep productivity. We examined relationships between snow and climate conditions and summer lamb production in Wrangell-St Elias National Park and Preserve, Alaska over a 37-year study period. To produce covariates pertaining to the quality of the snowpack, a spatially-explicit snow evolution model was forced with meteorological data from a gridded climate re-analysis from 1980 to 2017 and calibrated with ground-based snow surveys and validated by snow depth data from remote cameras. The best calibrated model produced an RMSE of 0.08 m (bias 0.06 m) for snow depth compared to the remote camera data. Observed lamb-to-ewe ratios from 19 summers of survey data were regressed against seasonally aggregated modelled snow and climate properties from the preceding snow season. We found that a multiple regression model of fall snow depth and fall air temperature explained 41% of the variance in lamb-to-ewe ratios (R2 = .41, F(2,38) = 14.89, p<0.001), with decreased lamb production following deep snow conditions and colder fall temperatures. Our results suggest the early establishment and persistence of challenging snow conditions is more important than snow conditions immediately prior to and during lambing. These findings may help wildlife managers to better anticipate Dall's sheep recruitment dynamics.

Changes in Vegetation Phenology and Productivity in Alaska Over the Past Two Decades

Understanding trends in vegetation phenology and growing season productivity at a regional scale is important for global change studies, particularly as linkages can be made between climate shifts and the vegetation’s potential to sequester or release carbon into the atmosphere. Trends and geographic patterns of change in vegetation growth and phenology from the MODerate resolution Imaging Spectroradiometer (MODIS) satellite data sets were analyzed for the state of Alaska over the period 2000 to 2018. Phenology metrics derived from the MODIS Normalized Difference Vegetation Index (NDVI) time-series at 250 m resolution tracked changes in the total integrated greenness cover (TIN), maximum annual NDVI (MAXN), and start of the season timing (SOST) date over the past two decades. SOST trends showed significantly earlier seasonal vegetation greening (at more than one day per year) across the northeastern Brooks Range Mountains, on the Yukon-Kuskokwim coastal plain, and in the southern coastal areas of Alaska. TIN and MAXN have increased significantly across the western Arctic Coastal Plain and within the perimeters of most large wildfires of the Interior boreal region that burned since the year 2000, whereas TIN and MAXN have decreased notably in watersheds of Bristol Bay and in the Cook Inlet lowlands of southwestern Alaska, in the same regions where earlier-trending SOST was also detected. Mapping results from this MODIS time-series analysis have identified a new database of localized study locations across Alaska where vegetation phenology has recently shifted notably, and where land cover types and ecosystem processes could be changing rapidly.

Arctic Vegetation Mapping Using Unsupervised Training Datasets and Convolutional Neural Networks

Land cover datasets are essential for modeling and analysis of Arctic ecosystem structure and function and for understanding land–atmosphere interactions at high spatial resolutions. However, most Arctic land cover products are generated at a coarse resolution, often limited due to cloud cover, polar darkness, and poor availability of high-resolution imagery. A multi-sensor remote sensing-based deep learning approach was developed for generating high-resolution (5 m) vegetation maps for the western Alaskan Arctic on the Seward Peninsula, Alaska. The fusion of hyperspectral, multispectral, and terrain datasets was performed using unsupervised and supervised classification techniques over a ∼343 km2 area, and a high-resolution (5 m) vegetation classification map was generated. An unsupervised technique was developed to classify high-dimensional remote sensing datasets into cohesive clusters. We employed a quantitative method to add supervision to the unlabeled clusters, producing a fully labeled vegetation map. We then developed convolutional neural networks (CNNs) using the multi-sensor fusion datasets to map vegetation distributions using the original classes and the classes produced by the unsupervised classification method. To validate the resulting CNN maps, vegetation observations were collected at 30 field plots during the summer of 2016, and the resulting vegetation products developed were evaluated against them for accuracy. Our analysis indicates the CNN models based on the labels produced by the unsupervised classification method provided the most accurate mapping of vegetation types, increasing the validation score (i.e., precision) from 0.53 to 0.83 when evaluated against field vegetation observations.

Navigating snowscapes: scale-dependent responses of mountain sheep to snowpack properties

Winters are limiting for many terrestrial animals due to energy deficits brought on by resource scarcity and the increased metabolic costs of thermoregulation and traveling through snow. A better understanding of how animals respond to snow conditions is needed to predict the impacts of climate change on wildlife. We compared the performance of remotely sensed and modeled snow products as predictors of winter movements at multiple spatial and temporal scales using a data set of 20,544 locations from 30 GPS-collared Dall sheep (Ovis dalli dalli) in Lake Clark National Park and Preserve, Alaska, USA from 2005 to 2008. We used daily 500-m MODIS normalized difference snow index (NDSI), and multi-resolution snow depth and density outputs from a snowpack evolution model (SnowModel), as covariates in step selection functions. We predicted that modeled snow depth would perform best across all scales of selection due to more informative spatiotemporal variation and relevance to animal movement. Our results indicated that adding any of the evaluated snow metrics substantially improved model performance and helped characterize winter Dall sheep movements. As expected, SnowModel-simulated snow depth outperformed NDSI at fine-to-moderate scales of selection (step scales < 112 h). At the finest scale, Dall sheep selected for snow depths below mean chest height (<54 cm) when in low-density snows (100 kg/m³ ), which may have facilitated access to ground forage and reduced energy expenditure while traveling. However, sheep selected for higher snow densities (>300 kg/m³ ) at snow depths above chest height, which likely further reduced energy expenditure by limiting hoof penetration in deeper snows. At moderate-to-coarse scales (112-896 h step scales), however, NDSI was the best-performing snow covariate. Thus, the use of publicly available, remotely sensed, snow cover products can substantially improve models of animal movement, particularly in cases where movement distances exceed the MODIS 500-m grid threshold. However, remote sensing products may require substantial data thinning due to cloud cover, potentially limiting its power in cases where complex models are necessary. Snowpack evolution models such as SnowModel offer users increased flexibility at the expense of added complexity, but can provide critical insights into fine-scale responses to rapidly changing snow properties.