Date of Award
Summer 8-27-2021
Degree Type
Thesis
Degree Name
Master of Science (MS)
Department
Earth Sciences
Advisor(s)
Chipman, Melissa
Keywords
Alaska, Boreal, Ecology, Geospatial, Remote Sensing, Wildfire
Subject Categories
Ecology and Evolutionary Biology | Forest Sciences | Geography | Life Sciences | Remote Sensing | Social and Behavioral Sciences
Abstract
Wildfires are a common occurrence in the boreal ecosystems of the Pacific Northwest. Studies suggest that anthropogenic climate change has fostered more frequent and higher severity fires in recent decades in these forests, which may result in substantial changes in vegetation structure and ecosystem functioning. However, large-scale studies examining the linkages between changing boreal wildfire regimes and vegetation structure have historically been limited in spatial scope due to the broad area and inaccessibility of many boreal regions, including the Alaskan interior. The development and advancement of satellite remote sensing instruments and geospatial analysis techniques provide researchers with unmatched abilities to conduct large-scale studies of boreal fire-vegetation dynamics. This research utilizes publicly available multispectral Landsat imagery, Synthetic Aperture Radar imagery, Digital Elevation Models, wildfire perimeter data, and landcover classification products to gain insights into the linkages between climate, wildfire, and vegetation throughout the entire boreal ecoregion of Alaska. Analyses utilizing existing wildfire and landcover geospatial products suggest significant declines in both fire-adapted black spruce-dominated forests and fire-resistant deciduous forests from 2001 to 2016, of -50.0% and -19.3% landcover area, respectively. However, post-fire recruitment of deciduous forests far exceeds evergreen forest types in regions that had experienced one fire (3.4 times more likely) and two or more fires (4.9 times more likely), between 1970 and 2019. Novel spectral unmixing and fractional coverage analyses using evergreen, deciduous, and early successional endmembers yielded significant monotonic declines in the change in pixel proportion of evergreen forests as a function of wildfire frequency, with -2.36%, -25.35%, and -35.15% for areas of zero, one, and two or more fires, respectively. In contrast, deciduous changes in pixel proportion exhibited higher degradation in non-fire regions (-6.23%) than evergreens, lower magnitude decreases in single-fire areas (-21.59%), and a significant rebound in coverage in regions that burned two or more times over the study period (-10.04%). These findings suggest that deciduous forests are more resistant to wildfire than evergreen-dominated systems, and their recruitment following evergreen-fueled wildfires may moderate the frequency and severity of subsequent local fire regimes.
Access
Open Access
Recommended Citation
Deutsch, Eric John, "Observations of Post-wildfire Landcover Trends in Boreal Alaska Using a Suite of Remote Sensing Approaches" (2021). Theses - ALL. 563.
https://surface.syr.edu/thesis/563
Included in
Ecology and Evolutionary Biology Commons, Forest Sciences Commons, Remote Sensing Commons