--> Abstract: Quantifying Undiscovered Oil and Gas: A Probabilistic Predictive Model and Evaluation Using GIS Compatible Tools; Weights of Evidence, Logistic Regression, an Artificial Neural Network and Fuzzy Logic, by Beatrice Maré-Jones; #90033 (2004)
[First Hit]

Datapages, Inc.Print this page

Quantifying Undiscovered Oil and Gas: A Probabilistic Previous HitPredictiveNext Hit Model and Evaluation Using GIS Compatible Tools; Weights of Evidence, Logistic Regression, an Artificial Neural Network and Fuzzy Logic

Beatrice Maré-Jones
Victoria University, School of Earth Sciences
Wellington, New Zealand
[email protected]

Using a variety of Geographical Information Systems (GIS) compatible tools a play-specific probabilistic estimation methodology, to quantify undiscovered oil and gas is being developed. The methodology predicts the distribution and volumes of undiscovered oil and gas based on spatial analysis of petrophysical characteristics of known discoveries. The methodology is being developed for the Taranaki Basin in New Zealand, which has a complex depositional and structural history, and will be applicable to other areas of the world.

The methodology has two key components: a geologic model and a Previous HitpredictiveNext Hit model. The geologic model establishes key petroleum system components and processes, specific to the basin of interest, using a three-dimensional sediment maturity and petroleum generation and migration modelling. The mass of petroleum expelled from the basin’s kitchen areas is estimated using a deterministic maturity model. The expelled amounts are used, in a 3-D petroleum migration model, to estimate hydrocarbon accumulations at specific times during the evolution of the basin.

The Previous HitpredictiveTop model is an integration of data-driven and knowledge driven systems. Exploration of the use of newly available to GIS tools, namely weights of evidence, logistic regression, an artificial neural network and fuzzy logic, and spatial analysis of the basin’s geologic model (evidential data) and discovered oil and gas characteristics (occurrence data) provides an innovative way to predict undiscovered oil and gas.

AAPG Search and Discovery Article #90033©2004 AAPG Foundation Grants-in-Aid