The problem of kimberlite pipes localization using seismic method face such problems as weak velocity contrast and vertical orientation of target object. Thus, traditional seismic approach seems to be useless to resolve this task, and low-cost aeromagnetic method appears to be more effective. But when kimberlite rocks have low magnetic properties or are covered by magnetic layers, aeromagnetic method fails. Solution is to use regular seismic data having it processed via special tomography-based processing graph, using special seismic tomography technology.

Technology includes following: ray geometry calculation and ray tracing within different measurement schemas; methodology of complex wave times and amplitudes usage for velocity and effective absorption-dissipation factor determination; ray tomography algebraic reconstruction methods such as filtered back projection algorithm, adapted to non-regular geophysical conditions with a prior information accounting; diffraction tomography reconstruction within single-scattering approximation technique.

Seismic tomography method is divided into ray tomography and diffraction tomography. In turn, ray tomography is divided into transmission tomography, refraction tomography and reflection tomography. Let us examine what kinds of tomography are convenient for exploration of pipes. Transmission tomography is applicable for cross-hole data and is for use on detailed exploration stage (not considered). Refraction 2D and 3D prospecting and tomography is valuable when pipe is cut off with low velocity layer that is not covered with any high velocity one. Reflection 2D and 3D tomography is applicable when some convenient reflection boundary is present below the pipe body. Diffraction 2D and 3D tomography is convenient in all cases.

Figure 1. Different schemas of applying seismic tomography to kimberlite pipes prospecting.

When considering any kind of ray tomography, there are several steps of processing. First step is geometry calculation and ray tracing. Second step is determination of travel times and wave amplitudes. Last step is final tomography reconstruction. When processing amplitudes, there is one additional step before reconstruction. It takes in account the dependency of amplitude on absorption and dissipation. This additional step produces effective absorption-dissipation by each ray using initial amplitudes of all available amplitude data. The trick is that all these steps are not completely isolated. For example, wave amplitudes determination and transfer amplitudes into effective absorption-dissipation depends on calculated geometry. In turn, geometry may depend on velocity reconstruction. This leads to possible iterations over the processing graph between different steps. There are several approaches to perform reconstruction within ray tomography. It was found that methods based on iteration technique such as simultaneous iteration reconstruction tomography are useless when target object is small enough. That means, reconstruction technique has to have enough resolving ability to determine pipes small as scores of meters in diameter. Such resolving ability was obtained with of porting classic filtered back projection technique to irregular seismic data case. Porting is based on linearity of ray tomography equations. Indeed, initial unidimensional convolution (Gilbert transform) followed with bidimensional back projection is replaced with initial bidimensional pack projection followed with bidimensional convolution. When talking about diffraction tomography, the starting point is Born's weak scattering approach. To be exact, not weak scattering but single scattering is considered, indeed. It means that we can use back wave field propagation in frequency domain to obtain such geological media characteristic as complex dissipation factor. This factor consolidates velocity-caused scattering (in terms of classic wave linear equation) with nonlinear absorption. The resolving ability of this technique depends on spatial aperture and discretization (steps between sources and receivers) and frequency of seismic signal bandwidth both.

Technology and algorithms are tested on physical and mathematical models and approved by exploring Yakutia kimberlite fields.

Figure 2. Results of seismic tomography prospecting in Yakutia kimberlite field. a - regular CDP stack section, b - geological section with number of vertical tectonic faults and kimberlite pipe in center, c - tomography section of absorption-dissipation factor, d - tomography section of velocity.

Results of applying ray tomography graph to regular seismic CDP data are presented. In whole, vertical, horizontal and inclined seismic ray tomography sections allow faults and kimberlite pipes localization much better than regular seismic CDP stack sections. Absorption-dissipation factor sections, obtained with this technology allow better localization of desired object than any other method. Velocity tomography sections also give an opportunity to localize kimberlite pipe area. Diffraction tomography sections allow obtaining better resolution of internal pipe structure. It is a fact, that such strong reflectors as reflection boundaries may obscure such weak inhomogeneity as kimberlite pipe. In this case, the mentioned single scattering approach becomes invalid. Within this issue, the prior subtraction of reflection waves is quite useful procedure.

Figure 3. Results of tomography processing of model seismic data. Top left - model section of zero offset seismic records (pseudo-acoustic section). Bottom left - model of kimberlite pipe as heterogeneous vertical structure with covering and underlying horizontal boundaries. Top right - initial complex dissipation factor section. Bottom right - complex dissipation factor section obtained after previous subtraction of reflection waves.

Seismic tomography method seems to be a good solution for kimberlite pipes prospecting. It gives more precise information than other methods and it becomes unique when kimberlite rocks are non-magnetic or are covered with magnetic layers. It may be used within different kinds of measurement geometry and it allows combining any geometry with another. Additional advantage is that regular field data gathering schemas are convenient and even old pre-stack CDP data may be processed afresh to obtain ultimately new geological information. Full graph of seismic tomography processing is supported with "Geotomo" software system and presented on the World Wide Web at http://www.webstructor.net/geotomo/.

Kolonin A.G., Possibility of using transmission seismic waves for local inhomogeneities localization, Geology and Geophysics, Novosibirsk, 1988, V.3, p.101-110.

Wu R.S., Toksos M.N., Diffraction tomography and multisource holography applied to seismic imaging, Geophysics, 1987, V.52, p.11-25.