Sensitivity Analysis Methods for Basin Modeling

Bryan Travis
Earth and Environmental Sciences Division, Los Alamos National Laboratory, Los Alamos, NM USA

Computing sensitivities of processes in basin models with respect to spatially or temporally distributed model parameters can provide important information for site assessment, field monitoring and data inversion, but can be computationally expensive. There are, however, a few mathematical approaches that are very efficient and readily adapted to a wide range of mechanical and fluid dynamical system models, linear or nonlinear. Examples are given to illustrate the utility of differential (adjoint) sensitivity analysis, as well as of representers and interval analysis, for fluid flow in porous media with chemical and biological activity. These can be extended to multi-scale and stochastic modeling approaches. Further, the relationship of these to the problem of defining the null space in the case of data inversion is considered.

AAPG Search and Discover Article #90066©2007 AAPG Hedberg Conference, The Hague, The Netherlands