Item

Abstract

From poroelasticity theory we know that fluid diffusion will induce matrix deformation and vice versa. In practice, well known phenomena for such coupled processes are, for example, occurrence of seismo-tectonically induced groundwater fluctuations, land subsidence as a result of fluid extraction from subsurface reservoirs, production-induced surface strain near the vicinity of wells, reservoir- or injection-induced seismicity. Modeling of deformation and pore-pressure data that have been observed near the surface can help to image the dynamics and to assess the hydraulic properties of subsurface aquifers. We here present a semi-analytical Haskell propagator method to fully handle linear poroelastic problems in a multilayered half-space. Our method is a powerful tool for various reasons: (1) It is faster than traditional numerical schemes when respective discretization of the object region is chosen and solutions are sought for single locations only; (2) a problem is easily formulated, as only a set of five poroelastic parameters per layer plus the layers' thicknesses need to be specified; (3) the method is highly flexible, as forcing functions of point injection, single force (e.g., surface loading), double couple dislocation (earthquakes), etc. may be readily incorporated; (4) the so-called loss-of-precision problem of the original propagator algorithm has been fully overcome using the orthonormalization technique. The effectiveness of the new tool has been demonstrated by modeling pump-induced near-surface tilt data obtained at a test site near Sopron in western Hungary. The results show that the hydraulic diffusivity of the shallow subsurface aquifer can be assessed with an accuracy better than half an order of magnitude, if other elastic parameters and the geometry (depth and thickness) of the water-bearing formations are sufficiently known from, for example, bore-log records. Moreover, the present method can be applied to model induced seismicity based on the Coulomb failure criterion.