Mapping and characterization of the API gravity of offshore hydrocarbon seepages using multispectral ASTER data

The objective of this work is the qualitative remote characterization (API gravity degree) of oil seepages on the ocean surface. In order to achieve this goal, multispectral data acquired by the Advanced Spaceborne Thermal Emission and Reflection (ASTER) sensor are employed. ASTER registered tracts of oil along oceanic sectors of the Campos Basin (Brazil) and the Bay of Campeche (Gulf of Mexico) in several occasions. Numerous evidences indicate that these oil patches bear a straight link to oceanic seepages. The delimitation and segmentation of these seepages in the ASTER imagery are accomplished making use of an unsupervised, neural network fuzzy-clustering algorithm. Spectra representative of the seepages are extracted from atmospherically-corrected ASTER data pixels (9 bands spanning from visible to shortwave infrared wavelengths) contained in the classified segments. The ASTER spectra are checked against a predictive °API partial least square regression model. This model is established on the basis of oil spectra of known ºAPI varying from 13 to 47 yielded by laboratory measurements. Considering this model, API gravity degrees of 19.6 +/− 1.37 and 15.9 +/− 2.9 are remotely estimated for the seepage in the Campos Basin and the Bay of Campeche, respectively. Oils produced from Campos and Campeche fields typically show °API varying from 17-24 and 12-16.5, correspondingly. These results indicate the potential of the methodology proposed and of ASTER data and alike to remotely infer physical-chemical properties of hydrocarbons, since a close match was verified between predicted and true API gravity degrees for both study areas. The data, methods and experience gained in this research can be operationally tested in offshore oil exploration and, likewise, be adapted to environmental monitoring of oil spills in coastal regions.

► We aim to map oils seepages on the ocean and to estimate their API degree remotely. ► ASTER imagery processed through neural nets can isolate the oil from other sea features. ► An API regression model is used to classify ASTER pixels with oil spectral signatures. ► API predicted with remotely data closely match true API observed in the oil fields.