GRSG Conference 2022: Orbit to Outcrop
Title: Spectral Characterization of the Categories of the Bonn Agreement Oil Appearance Code
Author: Rebecca Del Papa Moreira Scafutto
Oil spills in the marine environment result from human activity. Due to the severe impact on the marine ecosystem, quick action is essential to map and classify the spill to deliver effective remediation responses. The Bonn Agreement Oil Appearance Code (BAOAC) is one of the oldest and most commonly used approaches to classify oil spills.
The BAOAC stablishes a correlation between the visual appearance of the oil at the sea surface to the thickness of the oil layer, following a code with 5 categories: sheen, rainbow, metallic, discontinuous true oil colour and continuous true oil colour. Remote sensing techniques are largely used to map and monitor oil spills in the ocean.
Satellite, airborne and more recently droneborne sensors operating in a broad range of wavelengths offer the possibility to detect and characterize oil at the sea surface in different scenarios. Here, we developed a controlled laboratory experiment to evaluate the spectral characteristic of the 5 categories of the BAOAC in the VNIR-SWIR range.
In the experiment, samples of oil film over water with thickness varying from 0.3 to 300 µm were imaged with the NEO-HySpex VNIR-1800 (400 – 1000 nm) and SWIR-384 (930-2500 nm) hyperspectral cameras, and measured with an ASD FieldSpec spectrometer. Images were acquired every 2 hours in the first 12 hours after sample preparation and then after 24, 48 and 72 hours.
Images were processed with a wavelet function to remove noise and enhance oil features. A mean spectrum was extracted from each sample. Results indicate that in the VNIR-SWIR range: (i) main spectral features of the hydrocarbons (HC) located at 1700 nm and 2300 nm can be identified in all categories of the BAOAC; (ii) there is a direct relationship between each category of the BAOAC (i.e. thickness of the oil film) and the depth of HC features; (iii) the geometry and definition of the HC feature at 1700 nm improves over time. The research demonstrates that oil films can be qualitatively distinguished based on VNIR-SWIR image data, with the potential to be used for the estimation of oil spill thickness based on the BAOAC parameters.