Detectability by Electromagnetic Depth Sounding - a data mining tool aiding interpretation of shallow sediments altered by oil and gas seeps

Abstract

SUMMARY Electromagnetic depth sounding (EMDS) data are computed for simulated multi-layer earth models at different frequencies and geometries using digital linear filters. Detectability of sub-surface layers is computed as point-to-point difference of data instances between threelayer - homogeneous layer and two-layer - homogeneous layer media for similar range of frequencies. H and Ktype earth models are also considered in the computations. Response curves, computed for two different layer earth media models are superimposed, thus, the separation between curves is the direct indication of the involved “detectability effect”. The degree of separation among response curves between two geometrically or parametrically changing data properties, called resolution, has direct impact on detectability effect. Though detectability does not provide any qualitative or quantitative interpretation of data attributes, computed detectability effect significantly changes EMDS response resolution at varying layer-earth data attributes. Knowledge on strength of detectability and scalable properties among layer-earth media are interpreted based on resolution and coherency between two model response curves. Because of change in layered earth properties, varying horizontal and vertical resolution and coherency attributes between response curves provide considerable detectability effect. This process, termed as data-mining, facilitates extraction of knowledge of layer properties within multi-layer earth media. This detectability effect provides knowledge of n-layer-earth simulation, which can effectively respond to and aid the interpretation of actual geological models deduced from experimental data. These studies could prove to be useful for investigating shallow petroleum oil and gas seeps and their associated sediment alterations in the basin margin areas.