A Comparison of Regional - Residual Separation Techniques for Gravity Surveys

Abstract

Tasmania lies on a triangular extension of the Australian continental crust, surrounded on two sides by deep ocean basins. Strong regional effects from these features obscure much of the short wavelength character in the observed gravity data. Interpretation of the data requires reliable separation of the residual and regional components of the gravity field. Bouguer gravity data covering onshore and offshore Tasmania have been used to test a variety of methods of regional ? residual separation, including trend surfaces, filtering, upward continuation and forward modelling. Forward modelling was used to calculate a regional field from a three-dimensional model for Tasmania consisting of continental crust, oceanic crust, water and the mantle. The main advantage of modelling is that it enables the incorporation of geological and geophysical constraints into the regional field. Automated separation techniques are simple, fast and purely numerical but do not allow the incorporation of geological constraints. Quantitative interpretation of the resultant residual maps is difficult and uncertain. All automated methods tested produce undefined dc shifts which limit the usefulness of the residuals. Of the automated methods, trend surfaces produce residual maps which show the best correlation with known geology. The preferred method of regional ? residual separation is forward modelling, although it is both time consuming and subjective. Quantitative interpretation of residuals from modelling can be carried out with a greater degree of confidence since the regional field has a well-defined geological basis.