Towards imaging 3D crust and mantle structure by means of ocean tidal loading tomography

Ocean tide loading (OTL) brings about recurring deformation of the Earth’s surface. Some of the OTL harmonics, e.g. M2, O1, Mf, cause sufficiently large surface displacement to be registered by the Global Navigation Satellite Systems (GNSS). These displacements are sensitive to the interior structure of the planet in a broad range of temporal and spatial scales making them a potentially unique source of information about the planet’s response at low frequencies. Comparison between observations and predictions for 1D elastic Earth models result in discrepancies of up to 3 mm (Bos et al., 2015, Martens et al., 2016). Spatial coherency of these discrepancies hints to 3D interior structure as one of the main sources of such residuals.
In this context, we present a framework to invert OTL observations for 3D crustal and mantle structure based on a trust-region Newton-type iterative algorithm. Furthermore, we resort to the adjoint approach as an efficient means of computing the gradient for the high-dimensional model space. Focusing on the design of the inverse algorithm, we constrain ourselves to deformations of an isotropic elastic planet, which are governed by a self-adjoint forward operator. In order to assess the robustness of the method, we perform a suite of 3D synthetic inversions that mimic the distribution of the GNSS stations in South America. Preliminary results indicate enhanced sensitivities to the crustal and upper mantle density and elastic properties in the vicinity of the coastlines.