Seismic anisotropy, arising from the crystallographic or lattice-preferred orientation of anisotropic minerals or the shape-preferred orientation of melts or cracks, can establish a critical link between Mars's past evolution and its current state. So far, although seismic anisotropy in Mars has been proposed due to different velocities of vertically and horizontally polarized shear waves in the Martian crust, obtained from crustal converted waves, multiples, and surface waves recorded by the InSight seismometer, the evidence is plausible. Notably, the shear wave splitting, which stands out as a straight indicator of seismic anisotropy, has not been reported using marsquake records. In this study, we employ Low-frequency marsquakes detected by the InSight seismometer to reveal shear wave splitting in Mars. We find that the direct S waves of three marsquake recordings (S0173a, S0235b, and S1133c) with high signal-to-noise ratios exhibit the splitting pheonmenon. We rule out the possibility of apparent anisotropy through synthetic tests, affirming the presence of seismic anisotropy in Mars. The delay time (about 1.33 s on average) measured from the direct S wave splitting is too large to be solely attributed to the seismic anisotropy in the upper crust (0 - 10 km) beneath the InSight. Thus, seismic anisotropy in the deeper region of Mars is indispensable. Combined with other geophysical evidence near the InSight landing site, the strong seismic anisotropy observed in this study implies the porous crust with aligned cracks being greater than 10 km beneath the InSight and/or the presence of an active mantle plume underneath the Elysium Planitia of Mars.
The optimization of well locations and controls is an important step in the design of subsurface flow operations such as oil production or geological CO2 storage. These optimization problems can be computationally expensive, however, as many potential candidate solutions must be evaluated. In this study, we propose a graph network surrogate model (GNSM) for optimizing well placement and controls. The GNSM transforms the flow model into a computational graph that involves an encoding-processing-decoding architecture. Separate networks are constructed to provide global predictions for the pressure and saturation state variables. Model performance is enhanced through the inclusion of the single-phase steady-state pressure solution as a feature. A multistage multistep strategy is used for training. The trained GNSM is applied to predict flow responses in a 2D unstructured model of a channelized reservoir. Results are presented for a large set of test cases, in which five injection wells and five production wells are placed randomly throughout the model, with a random control variable (bottom-hole pressure) assigned to each well. Median relative error in pressure and saturation for 300 such test cases is 1-2%. The ability of the trained GNSM to provide accurate predictions for a new (geologically similar) permeability realization is demonstrated. Finally, the trained GNSM is used to optimize well locations and controls with a differential evolution algorithm. GNSM-based optimization results are comparable to those from simulation-based optimization, with a runtime speedup of a factor of 36. Much larger speedups are expected if the method is used for robust optimization, in which each candidate solution is evaluated on multiple geological models.