Enjang Jaenal Mustopa¹, Supriadi Supriadi², Doddy Sutarno³, and Keisuke Ushijima⁴
(1) Exploration Geophysics Laboratory Dept. of Earth Resources System Engineering, Kyushu University, 6-10-1 Hakozaki, Higashi-ku, Fukuoka, 812-8581, Japan, phone: +81-92-6423643, fax: +81- 92-6423614, [email protected]
(2) Physics Department, Semarang State University (UNNES), Jl. Raya Sekaran, Gunung Pati, Semarang
(3) Physics Department, Bandung Institute of Technology, Jl. Gannesha 10, Bandung, 40132
(4)Exploration Geophysics Laboratory, Department of Earth Resources System Engineering, Kyushu University, Fukuoka, 812-8581, Japan

In the past time, some authors have published papers about 1-D inversion method of Magnetotelluric (MT) data through least-square (LS) method. Ls method assumes simple Gaussian statistic. However estimation procedure based on LS would not be statistically optional, as outliers (abnormal data) are frequently superimposed on a normal ambient MT noise field which is approximately Gaussian. In this situation can be seriously misleading.

This paper introduces an inversion method by a robust procedure, which is commonly based on the least square (LS) approach, by minimizing the separation between observed data and computed apparent resistivity of the parameter model involved in the inversion problem. Huber hybrid method, one of technique for getting weighting factor, is used in this robust procedure. The first solution in the iterative method involves LS procedure. Singular Value Decomposition (SVD) is also used in which produces significant improvement in computational precision. Nevertheless, inversion technique does not result a unique solution.

This procedure is applied to interpret Controlled Source Audio frequency (CSAMT) far-field data from Merapi volcano which the final result is determination of resistivity upper structure of Merapi Volcano.

AAPG Search and Discovery Article #90913©2000 AAPG International Conference and Exhibition, Bali, Indonesia