Automated Approaches for Snow and Ice Cover Monitoring Using Optical Remote Sensing

Update item information
Publication Type dissertation
School or College College of Social & Behavioral Science
Department Geography
Author Selkowitz, David James
Title Automated Approaches for Snow and Ice Cover Monitoring Using Optical Remote Sensing
Date 2017
Description Snow and ice cover exhibits a high degree of spatial and temporal variability. Data from multispectral optical remote sensing instruments such as Landsat are an underutilized resource that can extend our ability for mapping these phenomena. High resolution imagery is used to demonstrate that even at finer spatial resolutions (below 100 m), pixels with partial snow cover are common throughout the year and nearly ubiquitous during the meltout period. This underscores the importance of higher spatial resolution datasets for snow cover monitoring as well as the utility of fractional snow covered area (fSCA) monitoring approaches. Landsat data are used to develop a fully automated approach for mapping persistent ice and snow cover (PISC). This approach relies on the availability of numerous Landsat scenes, an improved technique for automated cloud cover mapping, and a series of automated postprocessing routines. Validation at 12 test sites suggest that the automated PISC mapping approach provides a good approximation of debris-free glacier extent across the Arctic. The PISC mapping approach is then used to produce the first single-source, temporally well-constrained (2010-2014) map of PISC across the conterminous western U.S. The Landsat-derived PISC map is more accurate than both a previously published dataset based on aerial photography acquired during the 1960s, 1970s and 1980s and the National Land Cover Database (NLCD) 2011 extent of perennial snow and ice cover. Further analysis indicates differences between the newly developed Landsat-derived PISC dataset and the previously published glacier dataset can likely be attributed to changes in the extent of PISC over time. Finally, in order to map mean annual snow cover persistence across the entire landscape, we implement a novel canopy adjustment approach designed to improve the accuracy of Landsat-derived fSCA in forested areas. In situ observations indicate canopy-adjusted snow covered area calculated from all available Landsat scenes can provide an accurate estimate of mean annual snow cover duration. The work presented here lays the groundwork for addressing scientific questions regarding the spatial and temporal variability of snow cover, snow accumulation and ablation processes, and the impact of changes in snow cover on physical and ecological systems.
Type Text
Publisher University of Utah
Subject Geography
Dissertation Name Doctor of Philosophy
Language eng
Rights Management (c) David James Selkowitz
Format Medium application/pdf
ARK ark:/87278/s6q28s42
Setname ir_etd
Date Created 2019-10-22
Date Modified 2019-10-22
ID 1469517
Reference URL