Spectroscopic and hyperspectral evaluation of possible hydrocarbon occurrences in estuarine sediments, North Charleston, South Carolina, USA

Reflectance spectroscopy is becoming an important tool in remote identification and mapping of surface hydrocarbons, which may be very useful in determining past and present introduction of hydrocarbons into the environment. Using surface sediment samples in a complex urban/estuarine setting of North Charleston, SC it is possible to define the natural background of the area. By introducing common hydrocarbon pollutants (e.g. crude, motor oil, diesel) into typical background, a spectral library of background spectra, affected by a particular pollutant can be produced. Using the lab-derived hydrocarbon pollutant spectra as an input, it is possible to carry out classification of airborne hyperspectral data as a response to past events and baselining or current response and mapping of a potential spill. The results show that introduction of hydrocarbons produces observable effects, which can be used to delineate presence of affected areas. Relatively simple spectral-similarity algorithms (SAM, Spectral Unmixing) can be successfully used to extract the similar spectral matches in the hyperspectral image data on the basis of the library spectra produced.