rel="nofollow" href="#fb3_img_img_357e15cb-83cd-548a-bc5b-85622a225068.png" alt="ModifyingAbove d With right-arrow"/> along the E‐W, N‐S, and Vertical directions, and l x , l y , l z are the direction cosines of the look vector.
Given our knowledge of the satellite orbit, the line of sight of the radar antenna is known, as are the corresponding direction cosines of the velocity vector r a and r d . It is thus possible to write the following system of equations:
Figure 2.2 Example of motion decomposition combining ascending and descending acquisition geometry.
(2.5)
where l x, a , l y, a , l z, a , and l x, d , l y, d , l z, d are the direction cosines of the satellite line of sight for both ascending and descending acquisitions. The problem is poorly posed if we now want to invert for the full three‐dimensional velocity vector, as there are three unknowns (d x , d y , and d z ) and only two equations. However, because the satellite orbit is almost circumpolar, the sensitivity to possible motion in the north‐south direction is usually very small (the direction cosines l y, a and l y, d are close to 0). This allows us to rewrite the system in the following form:
(2.6)
an equation that may be solved for d x and d z .
2.3. DATA INTERPRETATION AND INVERSION METHODS
When a volume of fluid is introduced under pressure at some depth within the Earth, it changes the state of rock‐fluid system around the injection site. The nature of the change is determined by the temperature, composition, pressure, and flow rate of the fluid and the initial conditions within the host formation. Generally, the injected fluid will displace the in‐situ fluid, leading to pressure and, consequently, volume changes that depend on the compressibility of the system as a whole. These volume changes will lead to strain within the Earth that will be transmitted outward from the injection site, ultimately reaching the Earth's surface. Given sufficient strain at depth, the resulting surface displacement can be large enough to produce a significant signal, observable by modern geodetic instruments such as a SAR satellite. Such signals can produce valuable information concerning the source of the deformation.
Given an observable pattern of surface deformation, one can attempt to infer properties of the source generating the deformation. That is, one can invert the observed data to estimate parameters describing the source. One of the most important aspects of an inversion for volume change is how the source model is parameterized. In order to effectively represent the source, one must know its basic geometrical properties and boundary conditions. For example, if the injected fluid is confined to a porous layer, then it is important to include the impermeable boundaries of the layer. Or, if the injected fluid induces the opening of a tensile fracture, then the strike and dip of the feature will need to be specified, perhaps by a data‐fitting procedure. For a tensile fracture, the volume change is completely determined by the aperture change over the fracture. Thus, the nature of the elemental source will be different from the volume expansion of a grid block. In fact, the expansion of a grid block can be modeled by three orthogonal tensile fractures, one along each axis. Other, more complicated source combinations are possible, such as slip induced along a tensile fracture. A correct formulation of the problem requires knowledge of the geology as well as of the general stress conditions at depth. Furthermore, one must be willing to modify the formulation in light of new information. For example, at In Salah, a double‐lobed pattern of surface deformation indicated a tensile feature at depth, giving rise to a modified source model.
Under favorable circumstances, surface deformation may be used to image the migration of the injected and displaced fluid at depth and the geological features that are controlling it. Needless to say, making such inferences will require knowledge of the properties of the fluids involved and of the overburden. However, the nature of the injected fluid is usually well known and the structure and properties of the overburden are also constrained by seismic data and well logs. One does not typically inject fluid blindly into the Earth. For the sake of our discussion, we will assume that the Earth behaves elastically at a sufficient distance from the source and for the time intervals in question, a few months to a few years. Furthermore, we will assume that the elastic properties can be reasonably estimated from the available data. Other rheological models, such as viscoelasticity, are certainly possible and are not a barrier to the approach that we shall lay out here. However, for the conditions associated with the geological storage of greenhouse gases, an elastic overburden is a reasonable model.
Geodetic methods do not really allow us to image fluid flow directly. In particular, displacements at the surface and within the overburden are responding to the effective volume changes induced by fluid injection and withdrawal. Imagine a box around that part of the reservoir influenced by fluid injection or withdrawal. If we cut out the reservoir and simply applied the same displacements as those caused by the fluid flow to the boundaries of the box, the resulting deformation in the overburden, and on the surface, would be the same. So, in reality, one can solve only for such effective displacements or volume changes. Other assumptions, such as constitutive laws for the reservoir, are required if we are to interpret the volume changes in terms of fluid pressure changes, thermal expansion, or nonelastic processes. Therefore, we will formulate our inversion in terms of reservoir fractional volume change in order to minimize the number of assumptions that must be invoked.
In an elastic Earth, the displacement at the surface, and in the overburden, is linearly related to the volume changes within the source region (Aki & Richards, 1980; Vasco et al., 1988). Thus, one can write the calculated displacements as an integral over with source volume V as
where Δ v(y) is the fractional volume change at source location y. The quantity G i (x, y) is a Green's function representing the i‐th displacement component at location x that results from a point volume increase at y. The Green's function encapsulates the physics of the propagation of elastic deformation from the source to the observation point, obtained by solving the governing equation with a point source. For a simple medium with sufficient symmetry, such as a homogeneous half‐space, it is possible to produce an analytic Green's function. For a