Loading...
Evaluating the use of hyperspectral imagery to calculate raster-based wetland vegetation condition indicator
Journal Title
Aquatic Ecosystem Health & Management,
Readers/Advisors
Journal Title
Term and Year
Publication Date
2021
Book Title
Publication Volume
Publication Issue
Publication Begin
Publication End
Number of pages
Research Projects
Organizational Units
Journal Issue
Abstract
Field observations and measurements of wetland plants have traditionally been used to monitor
and evaluate wetland condition; however, there has been increasing use of remote sensing applications
for rapid evaluations of wetland productivity and change. Combining key aspects of field- and remote
sensing-based wetland evaluation methods can provide more efficient or improved biological indices.
This exploratory study set out to develop a raster-based Wetland Vegetation Condition Indicator system
that used airborne hyperspectral imagery-derived data to estimate plant-community quality (via wetland
classification and Coefficient of Conservatism) and vegetation biomass (estimated using the Normalized
Difference Vegetation Index). The Wetland Vegetation Condition Indicator system was developed for three
Lake Ontario wetland areas and compared to a field-based floristic quality index and a dominant-plant
based Floristic quality indexdom. The indicator system serves as a proof-of-concept that capitalized on the
spatial and spectral attributes of high-resolution imagery to quantify and characterize the quality and
quantity of wetland vegetation. A Pearson correlation analysis showed moderate r values of 0.59 and 0.62
for floristic quality index and floristic quality indexdom, respectively, compared to the indicator method.
The differences are potentially due to the spatial resolution of the imagery and the indicator method only
accounting for the dominant plants within each assessment unit (pixel), therefore disregarding understory
plants or those with low abundance. However, the multi-metric Wetland Vegetation Condition Indicator
approach shows promise as an indicator of wetland condition by using remotely sensed data, which could
be useful for more efficient landscape-scale assessments of wetland health, resilience, and recovery.
Citation
Aquatic Ecosystem Health & Management, 24(4): 100–114, 2021.
Description
Suir, Glen M.: 1U.S. Army Engineer Research and Development Center, Geospatial Data Analysis Facility, ATTN: CEERD-EE-C,
3909 Halls Ferry Road, Vicksburg, MS 39180-6199
Wilcox, Douglas A.: 2Department of Environmental Science and Ecology, SUNY Brockport, NY