Item Details

Print View

Using Hyperspectral Data to Identify Arctic Tundra Plant Communities and Estimate Vegetation Biomass at a Low Arctic Alaska Site

Bratsch, Sara
Format
Thesis/Dissertation; Online
Author
Bratsch, Sara
Advisor
Epstein, Howard
Abstract
Arctic ecosystems have been highly responsive to warming temperatures, with temperature increases in the Arctic occurring at rates almost twice greater than the rest of the globe over the past few decades. Changes to arctic ecosystems include increased albedo and surface heat flux, longer and warmer growing seasons, altered patterns of precipitation and evapotranspiration, decreased snow cover, and increased soil temperature and permafrost thaw. These factors result in vegetation changes such as increased vascular plant coverage, change in species composition, increased shrub coverage, and greater vegetation biomass in the Low Arctic, a region where plant vegetation communities have been historically dominated by mosses and lichens. Remote sensing is a useful tool for tracking these vegetation changes. This is commonly done using broad-band indices such as Normalized Difference Vegetation Index (NDVI), which has been used to identify a 9-15% increase in greenness in the Alaskan Arctic from 1982-2010. However, many changes to Arctic vegetation systems are occurring on a species level where differences may not be identifiable with broad-band remote sensing. Despite this, the use of hyperspectral remote sensing for assessing vegetation dynamics remains scarce. This study uses handheld hyperspectral remote sensing data collected during the 1999 growing season to establish relationships between spectra and vegetation communities and vegetation community biomass and to show how hyperspectral remote sensing methods can improve upon traditional methods for arctic vegetation community classification and biomass estimation. In Chapter 1, I assess the ability of hyperspectral data to differentiate among four vegetation communities in the Low Arctic of Alaska. Ivotuk, Alaska, serves as my primary study site. I then use models from Ivotuk to predict community membership at five other research sites along the North Slope of Alaska. Results show high accuracy when differentiating communities within the Ivotuk site, and mixed accuracy at the other sites. Sites located nearer to Ivotuk have better community differentiation while more northern sites had decreased classification accuracy. In Chapter 2, I use hyperspectral data and harvested biomass data to develop relationships between spectra and biomass quantities for plant tissue types at Ivotuk, Alaska. Biomass-spectra relationships were most significant for shrubs and shrub component parts, supporting other findings that shrubs are dominant controls over reflectance in arctic vegetation communities. This research finds that hyperspectral remote sensing is a valuable tool for identifying vegetation communities and tracking biomass changes occurring with climate change. Both chapters suggest that regions outside of typical NDVI, specifically bands in the blue, green, and red edge, may contain more useful information for differentiating among communities and estimating biomass. Hyperspectral remote sensing is therefore a valuable tool for Arctic research as it allows us to identify finer differences in vegetation communities. This study presents an example of the use of hyperspectral remote sensing to improve upon classification of tundra vegetation communities in the Arctic. Research sites in Alaska are remote and accessing sites is expensive. Ground-based remote sensing studies are crucial as they allow for the development of spectral relationships that can then potentially be extrapolated to hyperspectral satellite remote sensing, and used to track vegetation changes on larger temporal and spatial scales.
Published
University of Virginia, Department of Environmental Sciences, MS (Master of Science), 2016
Published Date
2016-04-26
Degree
MS (Master of Science)
Collection
Libra ETD Repository
In CopyrightIn Copyright
▾See more
▴See less

Availability

Read Online