CAO, SONG, and IAN LERCHE, University of South Carolina, Columbia, SC

ABSTRACT: Risk Analysis in Basin Modeling

Basin modeling development can be separated into three stages. The first stage is model development during which developing mathematical/computer models is the main theme. As basin modeling techniques are widely used in hydrocarbon exploration, the uncertainty and sensitivity associated with basin modeling become major issues (the second stage). For a given model we have to examine how the model results are influenced by the change (uncertainty) in the model assumptions, parameters of the model, and of errors in and finite sampling of the input data. The third stage is risk analysis in basin modeling. Because of the uncertainties associated with basin modeling, the model results are not absolutely correct, and should be assigned with a certain risk.

In principle, risk analysis in basin modeling should take a Monte Carlo approach, i.e. simulating probability distributions of model results by considering all possible values of parameters and all possible outcomes of the uncertainties. Unfortunately, the Monte Carlo approach is not appropriate in practice, because many uncertainties and sensitive variables are associated with basin modeling, so that it takes too much computing time to complete even a few Monte Carlo simulations. In this paper, we present a numerical method for risk analysis in basin modeling which performs a very quick risk analysis without losing accuracy. The procedure to be exhibited does not require massive Monte Carlo computer runs and so is of importance in providing risk assessments in a timely manner.

AAPG Search and Discovery Article #90987©1993 AAPG Annual Convention, New Orleans, Louisiana, April 25-28, 1993.