The fundamental assumption of many successful geochemical and geomicrobial technologies developed in the last 80 years is that hydrocarbons leak from subsurface accumulations vertically to the surface. Driven by buoyancy, the process involves sufficiently large volumes directly measurable or indirectly inferable from their surface expressions. Even when the additional hydrocarbons are not measurable, their presence slightly changes the environment, where complex microbial communities live, and acts as an evolutionary constraint on their development. Since the ecology of this ecosystem is very complicated, we propose to use the full-microbiome analysis of the shallow sediments samples instead of targeting a selected number of known species, and the use of machine learning for uncovering the meaningful correlations in these data. We achieve this by sequencing the microbial biomass and generating its DNA fingerprint, and by analyzing the abundance and distribution of the microbes over the dataset. The proposed technology uses machine learning as an accurate tool for determining the detailed interactions among the various microorganisms and their environment in the presence or absence of hydrocarbons, thus overcoming data complexity. In a proof-of-technology study, we have taken more than 1000 samples in the Neuqu謠Basin in Argentina over three distinct areas, namely, an oil field, a gas field, and a dry location outside the basin, and created several successful predictive models. A subset of randomly selected samples was kept outside of the training set and blinded by the client operator, providing the means for objectively validating the prediction performance of this methodology. Uncovering the blinded dataset after estimating the prospectivity revealed that most of these samples were correctly predicted. This very encouraging result shows that analyzing the microbial ecosystem in the shallow sediment can be an additional de-risking method for assessing hydrocarbon prospects and improving the Probability Of Success(POS) of a drilling campaign.
Mapping hydrocarbon microseepage prospect areas by integrated studies of ASTER processing, geochemistry and geophysical surveys in Assam-Arakan Fold Belt, NE India