Geodetic imaging of thermal deformation in geothermal reservoirs - production, depletion and fault reactivation

Abstract

We investigate thermally induced surface deformation in geothermal systems. To define source mechanisms at depth, we assess the mechanical process of subsurface deformation by assuming a spherically cooled fractured reservoir in an infinite medium and derive relations that define magnitudes of thermal contraction, stress change and permeability evolution. The magnitude of thermal deformation in typical geothermal system is larger than anticipated and suggests two different modalities of surface subsidence – thermal contraction and fault reactivation. Here, surface deformation (vertical displacement, surface tilt and horizontal strain) induced by the two different modalities are assessed with Mogi (contraction) and Okada (slip) models and compared with instrumental sensitivity of high precision surface geodetic tools. We show that 1year of geothermal operation at 10MW with a power plant conversion efficiency of 12% can yield ~3.0×104m3 of subsurface volume change. For a reservoir at 2000m depth, this induces ~1.7mm of vertical surface displacement, ~800 nano-radians of surface tilt and ~900 nano-strains of surface strain. This result implies that typically observed magnitudes of surface subsidence (order of cm/year) are naturally expected in massive (100MW scale) geothermal operations and observed surface subsidence may largely be the result of thermal contraction. Conversely, thermal unloading can trigger fault reactivation. Analysis with an Okada slip model shows these shear offsets on pre-existing faults can also result in surface deformations of considerable magnitude. Our analysis of field operational data from various geothermal projects suggests that both thermal contraction and slow fault reactivation may contribute to the observed large surface deformation. Comparison of predicted deformation with instrumental sensitivity of high precision surface tools confirms that geodetic signals, especially tilt and strain, are indeed sufficiently large to describe reservoir evolution and to potentially deconvolve reservoir parameters of interest, such as permeability.