Probability and Sensitivity Analysis of Two-Dimensional Basin Modeling Results
Published:January 01, 1999
E. Bagirov, I. Lerche, 1999. "Probability and Sensitivity Analysis of Two-Dimensional Basin Modeling Results", Numerical Experiments in Stratigraphy: Recent Advances in Stratigraphic and Sedimentologic Computer Simulations, John W. Harbaugh, W. Lynn Watney, Eugene C. Rankey, Rudy Slingerland, Robert H. Goldstein, Evan K. Franseen
Download citation file:
A procedure determines the relative importance of uncertainties in input information and in multiple parameter estimation to all outputs from two-dimensional basin modeling codes. The procedure does not rely on Monte Carlo methods, but on some simple properties of the cumulative probability distribution of output variations related to uncertainties. As a consequence, only a couple of computer trials are needed to evaluate the relationship of the variability of outputs to input uncertainties.
The procedure is applied to a two-dimensional cross section with evolution of the section with time. Attention is focused first on mainly geologic input uncertainties, and then on uncertainties of thermal factors and of hydrocarbon kinetic factors. Each group is initially taken separately, and then all three groups of uncertainties are combined and used simultaneously. The influence of each group of uncertainties on a suite of different outputs from the basin model is explored at different times across the evolving section. At each lime step, the relative sensitivity is examined of the uncertainty in a specified output to each group of input uncertainties, as is the relative importance of the uncertainty in a specified input to the suite of all outputs at each time step. In addition, the global relative importance of input uncertainties to output variabilities is considered, thereby providing a measure of output uncertainty effects, no matter where and when they occur, as a consequence of input uncertainties.
This work enables one to assess which inputs need to be more tightly constrained, and also to determine by how large a factor they need to be better constrained if the uncertainties on a suite of specified outputs are to remain within given tolerance limits.
The advantage to this rapid procedure is that one can focus more quickly on those factors of dominance in controlling, say overpressure development or hydrocarbon charge in a basin, without having to spend an inordinate amount of time, effort, or financial or staff resources on providing narrower limits of uncertainty to those input factors that provide but little change in output uncertainties.
Figures & Tables
Numerical Experiments in Stratigraphy: Recent Advances in Stratigraphic and Sedimentologic Computer Simulations
Numerical Experiments in Stratigraphy: Recent Advances in Stratigraphic and Sedimentologic Computer Simulations - This volume presents the results derived from a three-day workshop held at the University of Kansas, Lawrence, Kansas, from May 15 through May 17, 1996. The objectives of the workshop were to document, characterize, demonstrate, and compare different computing procedures that have been utilized in simulating stratigraphic sequences. Both inverse and forward simulation modeling procedures are represented. The results of the workshop and the papers assembled here include: (1) an enhanced understanding of similarities and differences between models and modeling philosophies, (2) increased communication among modeling groups and geoscientists, (3) critical evaluation of applications and assessment of how models have been utilized, and (4) improvements and refinements in techniques for generating and describing model input and output.