Joint application of seismic refraction and vertical electrical sounding fo... http://www.ias.ac.in/currsci/dec251999/articles18.htm 1 of 14 9/13/2008 5:27 PM Joint application of seismic refraction and vertical electrical sounding for the delineation of shallow aquifers Sankar Kumar Nath* and Shamsuddin Shahid Department of Geology and Geophysics, Indian Institute of Technology, Kharagpur 721 302, India Inversion of seismic refraction and vertical electrical sounding data alone may lead to incorrect parameter estimation in complex geological set-up, namely, blind zone in seismics, suppression and equivalence in geoelectrics. A unique solution may be achieved by integrating physically different sets of data into a joint or sequential inversion scheme. In this paper we have introduced one such algorithm, wherein the seismic velocity–depth section from ray inversion guides the evolution of a well-resolved geoelectric section in 2D environment from 1D DC resistivity inversion by evolutionary programming. The synthetic models suffering from the suppression and equivalence problems could be successfully reconstructed using this approach. Three test sites at Keshpal, Tata Metalik and Salboni in Midnapur District, West Bengal were selected through remote sensing and GIS tools. By sequentially interpreting the seismic and resistivity data, we could delineate a shallow aquifer with an average thickness of 6.8 m at Tata Metalik. At Salboni, however, a thin aquifer at 46 m depth needed an additional electrolog/litholog depth control for its delineation by the proposed joint approach. SEISMIC refraction and geoelectric sounding methods can play major roles in evaluation of groundwater resource potential. Although, vertical electrical sounding (VES) is appropriate when geological and hydrogeological units are gently dipping with large lateral extent of minor variation in lithology 1 , it fails to solve the equivalence and suppression problems. Seismic refraction prospecting also has its limitations, namely, the failure to identify thin layers (blind zone) and velocity inversion. The limitations in individual techniques can be reduced to a great extent by adopting either joint inversion or multi-sequential inversion schemes. Vozoff and Jupp 2 introduced one such method by combining DC resistivity and magnetotelluric data and observed improved resolution in the estimated model parameters. Different physical quantities can be integrated into a joint or sequential inversion if, at least, the measured *For correspondence. (e-mail: nath@gg.iitkgp.ernet.in) data are influenced by a subset of common underground parameters. For example, when using seismic and geoelectrical data representing physically different responses of near-surface features, the layer thickness is the only parameter common to both. The boundary between two layers of different resistivities and seismic velocities may not coincide. Thus, it may occur that, the number of layers derived by independent inversion of seismic and geoelectric data are not identical. However, in order to be sure that geoelectric and seismic data can be combined in a joint inversion scheme, layer-interfaces are assumed 3 to be the same for both geoelectric and seismic properties. Sequential seismic and geoelectric surveys in a coal mine enabled Breitzke et al. 4 to correlate the geoelectrical and seismic interfaces within an accuracy of 0.5 m and estimate an improved seismic model using a geoelectrically derived model. Even a coal seam with 24 times higher resistivity than that of the surrounding strata could be unambiguously recognized. In a joint inversion, different measurement sets are combined to one set before initiating the process of inversion as recorded by Hering et al. 3 and Misiek et al. 5 for geoelectric and surface wave seismic data. In the multi-sequential inversion, also termed as joint application, we can use the results of one inversion to guide the input to the other. Requirements of a joint inversion are stricter than that of multi-sequential inversion. In the former, both methods must be estimating the same structure, while in the latter some degrees of overlap are required 6 . There are several reports on both the joint inversion 2,3,5,7,8 as well as multi-sequential inversion 1,4,6,9 . A joint application algorithm, multi-sequential in approach, is proposed here to invert seismic refraction travel time data for the delineation of subsurface velocity and interface depths in the first instant. The depth values are passed on to the direct dissemination of VES curves using Koefoed’s algorithm 10 to generate a starting model for the subsequent resistivity inversion by Evolutionary Programming (EP). Beard and Morgan 11 indicated that 2D subsurface geometry can be approximately delineated using contours based on apparent resistivity and 1D inversion. In the present investigation, seismic ray inversion provides 2D control over the 1D inversion of VES data. For the construction of final subsurface model seismic layer thicknesses and sequentially inverted resistivity values are considered as parameters. One major aim of this investigation being the delineation of aquifers, two synthetic models as well as real field examples from Keshpal and Tata Metalik areas in Midnapur District, West Bengal were chosen for groundwater resource studies. The test sites were selected on the basis of hydrogeological inputs, e.g. litholog, drainage density, slope, soil composition, their proximity to river Kasai and the geological depositional sequence favourable for shallow groundwater potential zone. To validate this integration approach for deeper aquifers facing constraints of blind zone in seismic and equivalence/suppression in electrical method, another test site is selected at Salboni where electrical log or litholog data are used for depth control. However, the present software is applicable to any geological set-up without depth restriction. The limitation is posed mainly by the seismic data due to the limited penetration power of the mechanical seismic sources used for refraction analysis.