Subsidence Prediction and Inversion of Subsidence Data

Abstract This paper presents a forward model for subsidence prediction and an inversion model to calculate the reservoir behaviour from subsidence data. The forward model employs combinations of analytic solutions to the elasticity equations, which approximate the boundary conditions. There are only few free parameters and consequently the calculation times are limited. Still, the model is applicable to a multi-layer subsurface with elasticity parameters changing per layer. The inverse problem is usually ill-conditioned or ill-posed and has therefore been regularized by introducing a priori data about the pressure field or by imposing a certain amount of smoothness. The strength of the regularization parameters are best determined using the model resolution of the inversion and synthetic cases. These also can be used to optimize the experimental design. Inversion of a synthetic case with a "wrong" forward model shows the importance of obtaining reliable subsurface elasticity data. Without such data, the quality of the inversion can be highly compromised.