--> On the Benefits of Seeps Detection in Offshore Frontier Areas based on Multitemporal Satellite Sar Data and Manual Interpretation: From Northern African to Southern African Promontories, Levantine and Natal Basins, Selected Historical and Recent Sar Data

2018 AAPG International Conference and Exhibition

Datapages, Inc.Print this page

On the Benefits of Seeps Detection in Offshore Frontier Areas based on Multitemporal Satellite Sar Data and Manual Interpretation: From Northern African to Southern African Promontories, Levantine and Natal Basins, Selected Historical and Recent Sar Data

Abstract

For years, Remote Sensing has been more and more widely accepted by Geoscientists as a useful tool in the early stages of Exploration campaigns to unveil natural oil seepages. Satellite data has much to offer in today’s restrained contexts of tight exploration budgets. Acquisition and interpretation are very fast and provide a valuable additional information to be combined with other geological and geophysical datasets. The technology is perfectly suited to Frontier ventures where dense seismic might be lacking, and where pollution, one of the major argument brought against seeps studies, might be less hostile or diffuse. Remote Sensing stands out as a very affordable asset compared to other Exploration expenditures and will strengthen the knowledge on a petroleum system within a basin or province. Seeps interpretation heavily relies on the man seating behind the computer. In a time of worshipped Big Data, where the industry gets thrived by the progress of Artificial Intelligence, there’s still nevertheless a point to make by discussing the benefits of manual interpretation. Seeps detection has little to do with binary concerns, it’s quite never all true or all wrong: it’s often “in-between” and in the end, it’s all in the eyes of someone that has seen seeps around the world in several different contexts. Seeps have their own signatures, but they can easily be confused with “lookalikes” (algae, bathymetrical or atmospheric artefacts). Surely a Machine Learning algorithm can recognize an oil leakage on a million of images through the internet, but can a program make the distinction between a natural leakage and a lookalike? If we’re not ready enough to trust computers to make automatic interpretations, how can humans become more accurate? One of the answers might be found in the multitemporal approach of the methodology that allows to highlight recurrences and spatial proximity of active seepages. Just like basalts sills or salt canopies render seismic interpretation tricky underneath surface, Remote Sensing has its own hazards above surface. Like successful or unsuccessful drillings, Exploration remain an art enhanced by the experiences and skills of the analysts.