Distributed acoustic sensing (DAS) is a rapidly growing technology for seismic acquisition, with the potential to sample rarely available wavefield components in reservoir settings. How to best utilize the information DAS supplies to estimate reservoir properties is an open question. Full waveform inversion (FWI) of DAS data, alone or in combination with geophone data, is a natural possibility to pursue. A mixed formulation must accommodate particle velocity data and single component measurements of strain or strain-rate in the direction tangent to a fiberoptic cable, which itself may take on some characteristic shape. Expecting that these amplitude and directionality properties of DAS data will impact parameter resolution in FWI, especially when incorporating finite gauge lengths, we develop two appraisal methods. The first is an analytic description of the relationship between the spatial period and the elastic wave sensitivity within a helical-wound fiber (which builds on a symmetry class of fibers insensitive to shear-strains). The second is an extension of scattering radiation pattern analysis to DAS sensors of arbitrary geometry. We then numerically analyze the FWI response. Using 2D simulations and several simple models and the Marmousi2 we analyze the effect that shaping of the DAS fiber has on parameter estimations, by comparing inversion results derived from straight and various coiled fibers in a horizontal well. Fiber geometry is observed to have important implications for the accuracy and fidelity of DAS-FWI parameter estimates. It is also clear that the complementary features of DAS and standard geophone data impact FWI. Simultaneous inversions of surface geophone and DAS data from a horizontal wells convincingly out-perform inversions from either dataset alone.
