A semi-automatic procedure for the interpretation of geophysical data

Abstract

A procedure for the semi-automatic interpretation of geophysical data is described. A simple model, such as the geophysical response from a layer or a point source, is first constructed. Parameters of the model are then varied simultaneously while being constrained in that the model?s geophysical response must fit the observed data. Each observed data point contributes a ?spray? of solutions in parameter space. The region of parameter space with the greatest density of solutions corresponds to the model that best fits the data, and error bounds can be obtained by examining the distribution of the solutions about that point. The procedure is demonstrated on gravity, magnetic, and seismic refraction models. The method is simple and fast to apply, and can be used as a filtering procedure rather than an interpretation procedure by replacing the measured data with the forward model response of the most probable set of model parameters.