AUTOMATED RECOGNITION OF PERMAFROST DISTURBANCES USING HIGH-SPATIAL RESOLUTION SATELLITE IMAGERY AND DEEP LEARNING MODELS

This study examines the impacts of wildfires on permafrost thaw in the Arctic spanning up to ~90 years post-fire and explores the potential of using radar satellite observations to enhance our understanding of environmental change. By assessing diverse high-latitude permafrost landscapes, we show that ground deformation anomalies after fires display similar patterns across regions, while radar backscatter varies depending on predominant ground temperatures.

This paper explores how the new Surface Water and Ocean Topography (SWOT) satellite can improve monitoring of lake ice and seasonal changes in cold regions. Although SWOT observes water surfaces worldwide, how its measurements respond to frozen lakes is not yet well understood. We show that these observations can reveal important ice features, freeze–melt transitions, and changes in snow and ice conditions. The results provide new opportunities for monitoring frozen freshwater environments.

Arctic coastal landscapes, like those in northernmost Alaska, are often characterized by saline permafrost which is prone to thawing. We studied six sediment cores to understand how thawing and salinity affect organic matter breakdown and landscape change. Our results show that salinity accelerates organic matter degradation when permafrost thaws. This highlights the overlooked risk of salinity in shaping Arctic landscapes and carbon mineralization as the climate continues to warm.

Climate warming causes permafrost to thaw, releasing greenhouse gases and affecting ecosystems. We studied sediments from Arctic coastal landscapes, including land, lakes, lagoons, and the ocean, finding that organic carbon storage and quality vary with landscape features and saltwater influence. Freshwater and land areas store more carbon, while saltwater reduces its quality. These findings improve predictions of Arctic responses to climate change and their impact on global carbon cycling.

Snow distribution characterization is essential for accurate snow water estimation for water resource prediction from existing in situ observations and remote-sensing data at a finite spatial resolution. Four different observed snow distribution datasets were analyzed for Gaussianity. We found that non-Gaussianity of snow distribution is a signature of the wind redistribution effect. Generally, seasonal snowpack can be approximated well by a Gaussian distribution for a fully snow-covered area.