Thermodynamically constrained joint inversion of seismic refraction, surface elevation and gravity data: Crustal structure and composition in the Porcupine Basin (North-East Atlantic)

The Porcupine Basin is a failed rift offshore SW of Ireland, formed in response to several rifting and subsidence phases during the Late Paleozoic to Cenozoic. Extension in the Porcupine Basin led to a dramatic crustal thinning that increases from north to south, and is potentially associated with an increasing degree of mantle serpentinization towards the center of the basin. However, the tectono-magmatic processes involved in the basin formation are still poorly understood. To tackle this issue we need first to better constrain the composition of the crust and the uppermost mantle.

We developed a probabilistic joint inversion of seismic traveltimes, surface elevation and gravity data. The unknowns of the inverse problem are the proportions of metastable mineral phases within the crust, geometry of the Moho and interfaces between different petrologies. We also invert for bound water content in the uppermost mantle to account for serpentinization. For a given petrology, elastic moduli and density of the composite are computed as functions of pressure and temperature. The temperature is found as numerical solution of the 3D steady-state heat equation. This petrological forward problem provides seismic velocities and density distributions on finite-difference mesh, which are used to solve eikonal equation for traveltimes, isostatic balance and Poisson equation for the gravitational potential. A Markov chain Monte Carlo algorithm is used to sample the posterior probability density function of the model parameters.

We use the RAPIDS4 wide-angle seismic data gathered in 2002 using 65 ocean-bottom seismometers across the central feature of the Porcupine Basin known as the Porcupine Arch that is characterized by a conspicuous positive free-air gravity anomaly. First arrivals are detectable up to the offsets of about 100 km. First, we perform traveltime tomography with more data weight on shorter offsets to resolve the P-wave velocity of the sedimentary cover; we predict density from the velocity and fix the resulting model of the sediments during the Monte Carlo inversion. We invert the first-arrival traveltimes jointly with surface elevation and free-air gravity anomaly from satellite altimetry.