Estimation of Near-surface Density Based on Bayesian Method over the Delaunay Tessellation with Second-order Smooth Prior.
Abstract
We propose a Bayesian method to simultaneously estimate near-surface density and Bouguer gravity anomaly over the Delaunay tessellation with second-order smooth prior. Firstly, the objective function is constructed from the observed free-air gravity data, near-surface density and Bouguer gravity anomaly. Then we give the likelihood by assuming that the observed free-air anomaly contains noise with Gaussian distribution. Secondly, with the prior assumption of smooth distribution of Bouguer gravity anomaly, we give the prior probability density function by regarding the flatness and smoothness of Bouguer gravity anomaly as normal distribution. Flatness and smoothness are obtained through the quadratic polynomial fitting over the Delaunay tessellation. Then, the hyper-parameters in our Bayesian model are determined by Akaike's Bayesian Information Criteria (ABIC). Finally, the optimal estimated results are obtained by maximizing the posterior estimation of model parameters. Estimated near-surface density can be used to indicate shallow geologic structure and estimated Bouguer gravity anomaly can provide more reliable data onto the study of regional deep structure and tectonics.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2021
- Bibcode:
- 2021AGUFMNG25A0497N