Since 1975 Shell Internationale Petroleum Maatschappij has used a Monte Carlo simulation model for worldwide prospect appraisal. The input parameters to this model describing charge (oil and gas available for trapping and retention), structure, reservoir, and retention (seal characteristics) are given in the form of probability distributions. For the estimation of charge and retention, the model follows a scheme of Bayesian update, using equations derived from calibration studies, i.e., statistical analysis of extensive data sets with a worldwide distribution.

Comparison of predrilling predictions with postdrilling results shows that the underlying calibration procedure is statistically sound and also demonstrates the importance of assessing geologic uncertainty in a quantitative manner. Geologists appear to have been fairly successful in describing the geologic setting of prospects in respect to hydrocarbon charge and retention (the calibrated parts of the system), but serious overestimation/overconfidence have occurred in respect to reservoir parameters and such risks as existence of trap (the uncalibrated parts of the system).

End_of_Article - Last_Page 528------------