thumbnail

Oil detection in the coastal marshes of Louisiana using MESMA applied to band subsets of AVIRIS data

Remote Sensing of Environment
By: , , , , and 
https://doi.org/10.1016/j.rse.2014.12.009

Links

Abstract

We mapped oil presence in the marshes of Barataria Bay, Louisiana following the Deepwater Horizon oil spill using Airborne Visible InfraRed Imaging Spectrometer (AVIRIS) data. Oil and non-photosynthetic vegetation (NPV) have very similar spectra, differing only in two narrow hydrocarbon absorption regions around 1700 and 2300 nm. Confusion between NPV and oil is expressed as an increase in oil fraction error with increasing NPV, as shown by Multiple Endmember Spectral Mixture Analysis (MESMA) applied to synthetic spectra generated with known endmember fractions. Significantly, the magnitude of error varied depending upon the type of NPV in the mixture. To reduce error, we used stable zone unmixing to identify a nine band subset that emphasized the hydrocarbon absorption regions, allowing for more accurate detection of oil presence using MESMA. When this band subset was applied to post-spill AVIRIS data acquired over Barataria Bay on several dates following the 2010 oil spill, accuracies ranged from 87.5% to 93.3%. Oil presence extended 10.5 m into the marsh for oiled shorelines, showing a reduced oil fraction with increasing distance from the shoreline.

Study Area

Publication type Article
Publication Subtype Journal Article
Title Oil detection in the coastal marshes of Louisiana using MESMA applied to band subsets of AVIRIS data
Series title Remote Sensing of Environment
DOI 10.1016/j.rse.2014.12.009
Volume 159
Year Published 2015
Language English
Publisher Elsevier Inc.
Contributing office(s) Crustal Geophysics and Geochemistry Science Center
Description 10 p.
First page 222
Last page 231
Country United States
State Louisiana
Other Geospatial Barataria Bay
Online Only (Y/N) N
Additional Online Files (Y/N) N
Google Analytic Metrics Metrics page
Additional publication details