Geologic applications of magnetic data and using enhancements for contact mapping

Transcription

1 Geologic applications of magnetic data and using enhancements for contact mapping M. Pilkington ( 1 ), P. Keating ( 1 ) ( 1 ) Geological Survey of Canada, Ottawa, Canada Summary Magnetic field data are fundamental to geophysical approaches to geologic mapping. Both qualitative and quantitative methods are used to extract lithologic and structural information in the horizontal dimension while quantitative methods provide details in the third (depth) dimension. Recent research in magnetic methods continues to expand the capability of data measuring systems, efficiency of forward modeling algorithms, effectiveness of data transformations and the versatility of inversion approaches. Vector and gradient measurements are becoming more widespread, and consequently, interpretation methods must be geared to such expanded data sets. Forward algorithms are exploiting 3D frequency-domain representations and volume to line integral conversions to reduce calculation time. This is particularly timely because of the widespread use of 3D inversion codes. 3D inversion methods have also moved from simple unconstrained problems towards more constrained scenarios incorporating geologic, physical property and possibly other geophysical data. Recent years have also seen a rapid increase in the number of data enhancements that can be used to emphasize certain characteristics in the data resulting in a more easily interpreted representation. The question arises of whether they are all equally useful and effective. Here, we evaluate the utility of ten enhancements based on theoretical arguments, synthetic data modelling, and real data examples. One use of these transforms of particular interest for mapping in the horizontal dimension is locating lateral magnetization contrasts, which we equate with geologic contacts. We find that four of the methods considered are redundant, giving practically identical results for contact mapping. Two others produce similar levels of detail, so both are not needed. Data quality adversely affects two other enhancements, leading to poor continuity of the mapped contacts. The practical utility of each method depends on its performance in the presence of noise, non-vertical dips and non-vertical magnetizations. Introduction Geologic mapping of the Canadian landmass is a vast undertaking. Reconnaissance level mapping covers most of the country but this level of detail is insufficient to directly assist in mineral exploration, which is crucial to continued economic development. Field mapping at the 1:250,000 scale is not economically feasible so a practical alternative is to focus new geologic mapping programs on prospective areas based on interpretations of aeromagnetic data. One simple but effective tool based on magnetic data is pseudo-geologic contact mapping, i.e., locating lateral magnetization contrasts, which can be done using many different enhancement techniques. Here, we compare different contact mapping methods, addressing the effects of noise, interference, non-induced magnetizations and non-vertical structural dips on the estimated contact locations. An aeromagnetic data set from the Abitibi greenstone belt is used to support the comparison. Methodology The ten enhancements considered here are listed in Table 1. Others do exist, mostly in the form of higher-order versions of some of the quantities listed. Some enhancements exhibit a maximum value

2 over magnetization contrasts (contacts), whereas others change from negative to positive, and so the zero value contour maps the contact. Maxima can be located using the algorithm of Blakely and Simpson (1986). The synthetic model (Figure 1) is made up of three bodies: a thin SE-dipping dyke, thick SE-dipping dyke, and vertical rectangular prism. Figure 1 shows that all the enhancements map the edges of the prism except for the magnitude transform, MT, which only exhibits a single maximum at its southern edge. The analytic signal (AS), local wavenumber (LW), total field horizontal gradient magnitude (TF-hgm), pseudogravity horizontal (PSG-hgm) and tilt horizontal (THDR) functions (Table 1) all produce maxima well located over the true edge of the prism while the tilt (TI), balanced horizontal (TDX), theta (TH), and vertical gradient (VG), functions are displaced slightly outside the body. Figure1 Synthetic model and calculated contact locations For the thin dyke, LW, AS and MT all give a single maximum coincident with the dyke location. THDR produces a set of three maxima trends, one coincident with the true location, but two parallel false trends caused by the dipolar nature of the source (Pilkington and Keating, 2004). All other enhancements erroneously map two edges, with TF-hgm having the smallest separation between them and PSG-hgm, TI, TDX, TH and VG giving an equally larger apparent dyke width. These latter six methods always produce two apparent contact locations because the magnetic field on either side of the dyke will change from negative to positive and will exhibit an inflection point, equivalent to a maximum in the horizontal gradient. For the thick dyke, only MT gives a single maximum since the deflection of the maximum location towards the body interior is greatest for this enhancement and the two (expected) maxima merge into one. All other enhancements are able to resolve the two edges of the dyke. LW does the best job with nearly perfect locations on both sides followed by AS with both maxima displaced slightly inwards towards the dyke centre. TF-hgm gives maxima outside the dike edges and displaced down dip. PSG-hgm, TI, TDX, TH and VG map edges further away from the true sides that are also shifted down dip. THDR accurately maps both edges of the dyke with no dip effects but again the maxima are accompanied by another parallel false maxima trend. The results of Figure 1 show that no single enhancement is capable of providing accurate contact or edge locations for all types of magnetic sources. From theory, MT, AS, LW and THDR map isolated contacts perfectly, even if they are dipping or remanent magnetization is present (see references in Table 1). In contrast, locations calculated from the remaining enhancements are compromised by both

3 of these conditions. All techniques will suffer from interference effects, i.e., when sources are close enough that their anomalies overlap, since the combined anomaly shape will differ from the theoretical response used in synthetic calculations. The kinds of effects from interference can be seen in Figure 1 where the prism comes close to the thin dyke. Overall, the TI, TDX, TH and VG maps show the best continuity, however these four have the greatest sensitivity to non-vertical dips. Figure 2 (top right) shows the magnetic field covering a portion of the Abitibi Greenstone Belt, Quebec, along with mapped contacts using the same procedures used for the synthetic example in Figure 1. The relative performance of the enhancements is much different now that real data, containing the combined effects of tie-line levelling, gridding and other corrections, are used. The most strongly affected are AS and LW, which show poor continuity along geologic strike. This effect is predominantly due to gridding, which leads to well-defined maxima in both the AS and LW functions at locations where a flightline crosses a magnetization boundary, but a lack of maxima at gridded locations between the flightlines. Figure 2 Data from Abitibi and calculated contact locations The TI and VG maps are actually equivalent, as are the TDX and TH maps. From Table 1 we see that the zero value of VG and TI are equivalent, since if the VG (df/dz) is zero, then TI (tan -1 (df/dz/[(df/dx) 2 + (df/dy) 2 ] 1/2 )) will also be zero. From the expression for TDX, it follows that maxima from the TDX enhancement are equivalent to a zero value of TI (θ) and VG. There are more maxima (TDX, TH) than zero values (TI, VG) in the real data example because of interference between anomalies. Thus, in terms of contact locations we can replace all four enhancements with just one (either TDX or TH). TF-hgm produces a distribution of contacts similar to the previous four enhancements. Although, theoretically, it is susceptible to false maxima trends, for steeply dipping

4 Precambrian Shield regions, most maxima will correspond to real contacts. The last set of results is given by the THDR function, which similarly to its behaviour in the synthetic model, leads to the greatest number of solutions. Many of these contact locations are false trends that parallel the real contacts and are caused by interference from adjacent anomalies and dipolar effects. Conclusions A wide range of potential field enhancements are available to assist in geologic mapping. Many have been introduced recently, with seven out of the ten listed in Table 1 post-dating Nonetheless, several enhancements appear to be redundant for contact mapping, with the TI, VG, TDX and TH methods providing almost identical solutions. Either the TDX or TH maps can be used in place of all four techniques. For PSG-hgm and MT, the similarity of the mapped contacts can be considered a practical redundancy, so both maps are not needed. The AS and LW methods are susceptible to flightline effects, leading to breakdown of the continuity of mapped contacts. For THDR, the presence of extra, false contact locations can complicate mapping, especially in areas of complex source distributions. Of the ten listed enhancements, eight exhibit maxima and two reach a value of zero over a contact. Therefore, all of them have a physical meaning that can be exploited in interpretation. How well this relationship fares in the presence of noise, non-vertical dips and non-vertical magnetizations ultimately determines their practical utility. References Blakely, R.J., and Simpson, R.W Approximating edges of source bodies from magnetic or gravity anomalies. Geophysics, 51, Cooper, G.R.J., and Cowan, D.R Enhancing potential field data using filters based on the local phase. Computers & Geosciences, 32, Cordell, L.E., and Grauch, V.J.S Mapping basement magnetization zones from aeromagnetic data in the San Juan Basin, New Mexico. In The utility of regional gravity and magnetic anomaly maps. Edited by W.J. Hinze. Society of Exploration Geophysicists, Tulsa. pp Grauch, V.J.S., Hudson, M.N., and Minor, S.A Aeromagnetic expression of faults that offset basin fill, Albuquerque basin, New Mexico. Geophysics, 66, Hood, P.J Gradient measurements in aeromagnetic surveying. Geophysics, 30, Miller, H.G., and Singh, V Potential field tilt - a new concept for location of potential field sources. Journal of Applied Geophysics, 32, Pilkington, M., and Keating, P Contact mapping from gridded magnetic data - a comparison of techniques. Exploration Geophysics, 35, Roest, W.R., Verhoef, J., and Pilkington, M. 1992, Magnetic interpretation using the 3-D analytic signal. Geophysics, 57, Stavrev, P., and Gerovska, D Magnetic field transforms with low sensitivity to the direction of source magnetization and high centricity. Geophysical Prospecting 48, Thurston, J.B., and Smith, R.S Automatic conversion of magnetic data to depth, dip, and susceptibility contrast using the SPI method. Geophysics, 62, Verduzco, B., Fairhead, J.D., Green, C.M., and MacKenzie, C New insights to magnetic derivatives for structural mapping. Geophysics - The Leading Edge, 23, Wijns, C., Perez, C., and Kowalczyk, P Theta map: Edge detection in magnetic data. Geophysics, 70, L39-L43.

5 Table 1. Enhancement methods for field f, having components X, Y, and Z. Method Abbreviation Formula Reference Vertical gradient VG df/dz Hood 1965 Pseudogravity horizontal PSG-hgm [(dpsg/dx) 2 + (dpsg/dy) 2 ] 1/2 Cordell & Grauch 1985 Total field horizontal TF-hgm [(df/dx) 2 + (df/dy) 2 ] 1/2 Grauch et al Balanced horizontal TDX -1 tan ([(df/dx) 2 + (df/dy) 2 ] 1/2 Cooper & Cowan 2006 / df/dz ) Analytic signal AS [(df/dx) 2 + (df/dy) 2 + (df/dz) 2 ] 1/2 Roest et al Tilt TI -1 tan (df/dz/[(df/dx) 2 + (df/dy) 2 ] 1/2 ) Miller & Singh 1994 Theta map TH [(df/dx) 2 + (df/dy) 2 ] 1/2 /AS Wijns et al Tilt horizontal THDR [(dti/dx) 2 + (dti/dy) 2 ] 1/2 Verduzco et al Local wavenumber LW [d 2 f/dxdz df/dx + d 2 f/dydz df/dy Thurston & Smith d 2 f/dz 2 df/dz ] /AS 2 Magnitude transform MT [X 2 + Y 2 + Z 2 ] 1/2 Stavrev & Gerovska 2000