1-D Magnetotelluric Data Inversion using Guided Grid Search Algorithm

Abstract:

Inversion is a process to know the earth model from measured data. In Magnetotelluric (MT) case, linearized non-linear approach is often used for inversion. While this approach is producing earth models that fit the measured data, the results are influenced by the starting model. This may causing the model produced from inversion doesn't reflect the true earth layer and a problem arise when bad starting model were used, the inversion may not produce the best result. The developments of technologies, especially in computers, allow us to do a faster computation, allowing more exploration of models. Many new algorithms are created to take advantage of this, one of them is Grid Search Algorithm. This algorithm is simply a set of trial and error from many solution parameters, which will make the computation takes a very long time whenever the search of solution require many parameters. In order to solve this problem a guided search is needed to cut the computation time. In this paper, a guide for grid search algorithm is established by understanding the pattern of 1-D MT response from layered earth. Many cases of 1-D model synthetic data is presented to know the effectiveness of this algorithm in 1-D MT inversion.