Using cloud-based remote sensing to estimate Hurricane Impacts on Vegetative Cover in Carolina Bays in North Carolina
No Thumbnail Available
Date
2018-05
Authors
Advisors
Journal Title
Series/Report No.
Journal ISSN
Volume Title
Publisher
Abstract
Schager, Sara, Masters of Environmental Assessment, Using cloud-based remote sensing to estimate
Hurricane Impacts on Vegetative Cover in Carolina Bays in North Carolina
Carolina Bays are unique geographically isolated wetlands that provide critical habitat to several
species of plants and animals in the southeastern United States. These areas, due to their inland
location, unique vegetation, and shallow inundation of water, can experience strong effects
from hurricanes. Hurricanes’ high winds, and large amounts of precipitation could affect these
shallow, inland wetlands. Carolina Bays have not been well studied, and the impacts of
hurricanes on them have been even less evaluated. I utilized a cloud-based remote sensing
approach to examine the effects of hurricanes on the vegetative cover of Carolina Bays. Using
Climate engine, I utilized the normalized difference vegetation index (NDVI) as a measure of
vegetation status. I evaluated seven hurricanes and 32 sites to determine if hurricanes affected
vegetative cover in Carolina Bays. Out of the 32 sites and seven hurricanes examined, five sites
showed statistically significant effects to vegetation from hurricanes. Site 6 had a higher NDVI
value (0.909) before the hurricane than it did during the same time period the year after the
hurricane occurred (0.755). Additionally, Site 6 was 60% agricultural field. This site showed that
Climate Engine is capable of detecting changes in NDVI values on Carolina Bay sites. Four sites
affected by Hurricane Fran had a significantly lower NDVI the month after the hurricane
occurred than they did the month before the hurricane. Hurricane Fran was a larger hurricane
event and probably had bigger vegetative impacts than the smaller hurricane events. My
research suggests that NDVI values obtained from Climate Engine are capable of determining
vegetative impacts on Carolina Bays, especially changes to NDVI of Carolina Bays during bigger
hurricanes. While Climate Engine makes large amounts of remote sensing data easily accessible,
in our case the remote sensed data did not show very strong impacts of smaller hurricane
events on Carolina Bays. This is probably due to the fact that the actual impacts on the ground
are limited due to the dynamic nature of the Carolina Bays themselves, rather than limitations
of Climate Engine. Limitations such as cloud cover during evaluation windows, resolution of
satellite photography (30 x 30 m pixels), and the fact that NDVI cannot pick up on species
composition changes or changes caused by fallen debris may have hindered evaluation. Future
research on Carolina Bays should focus on changes in depth of water due to precipitation and
fallen debris, which may uncover additional impacts to these ecosystems from hurricanes.
Description
Keywords
Carolina Bays, hurricane impacts, climate engine, NDVI, wetland vegetation, LANDSAT