Three‐dimensional joint inversion of traveltime and gravity data across the Chicxulub impact crater

Abstract

[1] In 2005 an extensive new seismic refraction data set was acquired over the central part of the Chicxulub impact crater, allowing us to image its structure with much better resolution than before. However, models derived from traveltime data are limited by the available ray coverage and the nonuniqueness that is inherent to all geophysical methods. Therefore, many different models can fit the data equally well. To address these issues, we have developed a new method to simultaneously invert traveltime and gravity data to obtain an integrated model. To convert velocity to density, we use a linear relationship derived from measurements on core from the Chicxulub impact basin, thus providing a reliable conversion equation that is typical for lithologies of the central part of this crater. Prior to utilizing the inversion on the observed data, we have run a suite of tests to establish the optimum weighting between traveltime and gravity constraints, using a synthetic model of central crater structure and the real experimental geometry. These synthetic tests indicate which inversion parameters lead to the best recovery of subsurface structure, as well as which parts of the model are well resolved. We applied the method to all existing gravity data and to seismic refraction data acquired in 1996 and the new, higher-resolution seismic refraction data acquired in 2005. We favor the traveltime model wherever we have sufficient ray coverage and the joint model where we have no ray coverage.

Structural uplift beneath the Chicxulub impact structure

“…Menke and Abbott, 1991]. Our velocity models have been tested for reliability and robustness [Vermeesch, 2006;Yhap, 2007;Vermeesch et al, 2008] and, unlike other geophysical models,…”

“…1) to convert between velocity and density, Vermeesch [2006] and Vermeesch et al [2008] discovered that a combined travel-time and gravity inversion tended to resolve the upper part of the structural uplift less well than the travel-time inversion alone, but could partially resolve the deeper velocity structure, as well as velocities in other areas with limited ray coverage. We have thus run two separate inversions on the real data: the first uses travel-times only, and the second uses both the travel-time and gravity data.…”

“…Prior to inverting the 1996 and 2005 data, Vermeesch [2006] and Vermeesch et al [2008] ran a suite of inversions on a synthetic velocity model of the central uplift at Chicxulub. This involved forward modeling of the seismic waves, using the same shot and receiver locations 6 6 as in the real data, and inverting the resulting travel-times to recover the original velocity model.…”

“…We have adapted the FAST code to include constraints from the gravity data [Vermeesch, 2006;Vermeesch et al, 2008]. The gravity data were filtered to remove the regional gradient and then sampled on a 4 km grid spacing.…”

“…The data misfit is calculated using both the gravity and travel-time misfit, and there is an extra step to convert between slowness (1/velocity) and density. We have used a suite of measurements [Vermeesch and Morgan, 2004] on core from drill holes within the post-impact sedimentary rocks, Cretaceous target rocks and impact breccias, to determine a best-fit equation for the conversion: density (in kg m -3 ) = -3478 × slowness (in 10 3 s m -1 ) + 3326 [Vermeesch, 2006;Vermeesch et al, 2008]. We have used a relatively large uncertainty of 1 mgal for the gravity (~5 % of the total anomaly) as we do not wish to over-fit the gravity data,…”