Abstract. The main results obtained during the last 5-8 yr in the solution of forward and inverse problems of 3D induction studies are summarized. The up-to-date status of 3D modelling is presented and prospective improvements in the formulation and numerical solution of forward problems are discussed. Approximate techniques and practical aspects of 3D modelling are specially considered.The general scheme of 3D interpretation of electromagnetic geophysical observations is outlined and realistic formalized approaches to solving 3D inverse problems, namely direct inversion and formalized model fitting, are studied.
O. IntroductionThe significant progress in the development of electromagnetic (EM) array observations achieved during the last ten years made it necessary to elaborate the system of EM data interpretation in 3D environments. Theoretical studies of 3D problems have been carried out during the last twenty years or more, but not until a few years ago did these works become of practical importance and serious applied researches were begun. Now many publications concerned with 3D induction studies are well known. Still, we have no generalized reviews on this subject (except brief ones like Ward, 1980), but in many special papers meaningful discussions of the problems and outlines of 3D techniques have been presented (Weidelt, 1975a;Vasseur and Weidelt, 1977;Hohmann, 1978;Banks, 1979;Weaver, 1979;Dawson and Weaver, 1979, Fainberg, 1980;Lee et al., 1981;Stodt et al., 1981;Berdichevsky et al., 1982; etc).It looks quite unrealistic to consider in detail the significant achievements in all of the numerous branches of 3D induction studies in a single review. Therefore the subject of this paper is restricted to the study of interpretation techniques of time-harmonic EM field anomalies caused by local and regional 3D conductivity structures. Moreover, attention is paid mainly to the theoretical and computational aspects of the rational formulation and effective solution of 3D forward and inverse problems. These problems form the basis of the interpretation theory and, as it can be seen from the experience of 2D induction studies, their intensive development produces the evident progress in elaboration of the whole interpretation system. It is not surprising that these problems are prevalent in the majority of publications of recent years.It should be noted that general ideas of inhomogeneous media interpretation theory, well illustrated in 2D applications, can be expressed in the main in 3D formulations. This has been shown clearly in the last monographs of Zhdanov (1981, 1983). However, we have usually enough difficulties to overcome in the theoretical analysis and at the stage of numerical considerations. Really, in most 3D