Gravity inversion for rifted margin deep structure using extension and isostatic constraints

Citation
Fj. Condi et al., Gravity inversion for rifted margin deep structure using extension and isostatic constraints, GEOPHYS J I, 138(2), 1999, pp. 435-446
Citations number
21
Categorie Soggetti
Earth Sciences
Journal title
GEOPHYSICAL JOURNAL INTERNATIONAL
ISSN journal
0956540X → ACNP
Volume
138
Issue
2
Year of publication
1999
Pages
435 - 446
Database
ISI
SICI code
0956-540X(199908)138:2<435:GIFRMD>2.0.ZU;2-7
Abstract
We present a gravity inversion method for determining the deep structure of rifted margins. It uses a 2-D earth model parametrized as multiple, irregu larly shaped, polygonal bodies, each of uniform density. The method has thr ee novel features. First, it links parameters in the shallow parts of the m odel to those in the deep parts by using a uniform extension model as a con straint in the inversion. The shallow structure will typically be known wit h greater certainty than the deep structure from shallow seismic and boreho le data. Second, it provides for variable weighting of prior information on densities, shapes, the extension model and smoothing to find geologically reasonable models. Third, it estimates densities and shapes simultaneously. The first two features are used to compensate for the inherent deficiencie s of poor depth resolution and non-uniqueness in gravity modelling. The las t two make the method an efficient way to explore a range of models. Synthe tic tests of sensitivity to noise indicate that the isostatic extension con straint promotes the recovery of the short-wavelength Moho topography and e liminates spatial undulations in deep structure due to noise in the data. S ynthetic tests of sensitivity to untrue prior information show that the iso static extension constraint increases the range of acceptable recovered mod els over no isostatic extension constraint. The range of recovered Moho pos itions suggests a vertical resolution of about 2 km. Although many recovere d models fit the data, the results imply a methodology for choosing a best set of models, and we suggest guidelines for applying the method to real ma rgins.