Report on Processing of Aeromagnetic Data Using the VariZ Susceptibility Technique From the Southeast Part of the Athabasca Basin, Saskatchewan

Transcription

1 Open File Report Report on Processing of Aeromagnetic Data Using the VariZ Susceptibility Technique From the Southeast Part of the Athabasca Basin, Saskatchewan A Report Prepared for the Saskatchewan Geological Survey by Condor Consulting, Inc. 2010

2 Open File Report Report on Processing of Aeromagnetic Data Using the VariZ Susceptibility Technique From the Southeast Part of the Athabasca Basin, Saskatchewan A Report Prepared for the Saskatchewan Geological Survey by Condor Consulting, Inc Printed under the authority of the Minister of Energy and Resources

3 Although the Saskatchewan Ministry of Energy and Resources has exercised all reasonable care in the compilation, interpretation, and production of this product, it is not possible to ensure total accuracy, and all persons who rely on the information contained herein do so at their own risk. The Ministry of Energy and Resources and the Government of Saskatchewan do not accept liability for any errors, omissions or inaccuracies that may be included in, or derived from, this product. This report is available for viewing at: Saskatchewan Ministry of Energy and Resources 2101 Scarth Street, 3rd floor Regina, SK S4P 2H9 (306) FAX: (306) and at: Parts of this publication may be quoted if credit is given. It is recommended that reference to this product be made as follows: Condor Consulting, Inc. (2010): Report on processing of aeromagnetic data using the VariZ Susceptibility technique from the southeast part of the Athabasca Basin, Saskatchewan; a report prepared for the Saskatchewan Geological Survey, Sask. Ministry of Energy and Resources, Open File , URL < openfiles>, 47p and digital files. Manuscript received March 2010 Released June 2010 Saskatchewan Geological Survey Saskatchewan Geological Survey ii Open File

4 Contents page Summary... 1 Magnetic Survey... 2 Introduction... 2 Processing Technique and Products... 4 Processing... 4 Mag3D... 4 Unconstrained and Constrained Modelling... 4 Products... 6 Discussion of Results... 8 Data Quality... 8 Unconformity Model... 8 Mag3D Outcomes... 8 Grid Outcomes... 9 Conclusions and Recommendations References Appendix A: Notes on Condor Magnetic Processing and ZS Processing Appendix B: Drill Hole Database Used for Unconformity Model Figures 1. Focus area for VariZ Susceptibility study Image of mag_lev and location map for study area Examples of unconstrained and constrained Mag3D models for the Athabasca Basin Examples of depth slices (defined by unconformity surface) extracted from the two Mag3D models Location of drill holes provided to define unconformity surface-coloured image shows the inferred depth Unconstrained and constrained models viewed from the top Side on view of unconstrained and constrained models (view to east) showing focus area Detailed look at feature highlighted in Figure Slice at -210 m msl through the unconstrained and constrained Mag3D models in vicinity of McArthur River deposit Saskatchewan Geological Survey iii Open File Report

5 10. From the upper left going counter clock-wise are images of the TMI-Tilt on the mag_lev grid, constrained model at the unconformity, -100 m below and -50 m below Same set of images as in Figure 10 but zoomed in around the McArthur River deposit a) Comparison of unconstrained and constrained models at unconformity surface (0 depth). b) Comparison of unconstrained and constrained models at unconformity surface (0 depth) in the vicinity of McArthur River deposit a) Comparison of unconstrained and constrained models at -50 m depth. b) Comparison of unconstrained and constrained models at -50 m depth in the vicinity of McArthur River deposit a) Comparison of unconstrained and constrained models at -100 m depth. b) Comparison of unconstrained and constrained models at -100 m depth in the vicinity of McArthur River deposit Tables 1. Survey products Example of spreadsheet provided... 8 Saskatchewan Geological Survey iv Open File

6 Summary This report covers the processing of magnetic data covering the southeast corner of the Athabasca Basin, northern Saskatchewan. The purpose of the work was to demonstrate a new form of processing for magnetic data developed by Condor Consulting, Inc. termed VariZ Susceptibility. This processing is intended to enhance the ability to extract geological information from magnetic data sets where there is a significant amount of non-magnetic cover over top of the magnetic rocks. This situation occurs in the Athabasca Basin where the Athabasca sandstone unit of NeoProterozoic age covers Archean basement. Included with this report is a brief outline of the technique and a short discussion of the outcomes. These results should aid in the better understanding of lithologies, structure, and alteration in the basement rocks within the study area. The latter two factors are recognized as likely important in formation of unconformity uranium deposits. Saskatchewan Geological Survey 1 Open File Report

7 Magnetic Survey Introduction The magnetic data used in the present study was made available to Condor Consulting, Inc. (Condor) by the Saskatchewan Northern Geological Survey (Survey). The outline of the study area is shown in Figure 1. Figure 1 Focus area for VariZ Susceptibility study. These data were acquired as part of an Athabasca Basin-wide initiative by the Survey to produce a uniform aerial geophysical coverage over the Basin using a standard set of acquisition parameters. The data provided to Condor has not been published so for the purposes of this report, will be referred to as the magnetic data. Figure 2 shows an image of the mag_lev field used in the processing superimposed over top of a location map for the study area. A total of 22, 242 lkm of magnetic data were used in the processing. Saskatchewan Geological Survey 2 Open File Report

8 Figure 2 Image of mag_lev and location map for study area. Saskatchewan Geological Survey 3 Open File Report

9 Processing Technique and Products Processing The processing was broken up into three parts; modelling of the magnetic data using the Mag3D code with no constraint, modelling the data with a constraint defined by the unconformity, and then extraction of grids from the 3D models to show the various outcomes. The unconformity model was provided by the Saskatchewan Geological Survey in the form of a Microsoft Excel database of drill holes with their locations and depth to the unconformity. These data will be discussed further. Mag3D Mag3D is a voxel-style inversion program developed at the University of British Colombia (Li and Oldenburg, 1996). This program can be run in either an unconstrained or constrained mode. A constraint is a priori information about the earth that is used to bias the inversion model towards a particular outcome. Details on Condor s Mag3D modelling are provided in Appendix A. Unconstrained and Constrained Modelling Figure 3 shows sections through two Mag3D models from the Athabasca Basin. The upper model has no constraint applied whereas the lower model has a constraint applied derived from the drill-hole defined depths to unconformity. The effect of the constraint is to in effect force the modelled susceptibility into the basement rocks since experience has shown there is generally little magnetically susceptible rock present in the Athabasca sandstone sequences. From a modelling perspective, forcing the susceptibility into the basement in effect sharpens the appearance of the sources in the basement. This is illustrated in Figure 4. In this figure, depth slices through the Mag3D outcomes are shown for the two models at a depth of 25 m below the unconformity surface. This is not a constant value in absolute terms (i.e., above sea level or below ground surface), but a variable defined by the available drilling constraints on the depth to the unconformity. The constrained model provides a much sharper image of the magnetic sources. Saskatchewan Geological Survey 4 Open File Report

10 Figure 3 Examples of unconstrained and constrained Mag3D models for the Athabasca Basin. Saskatchewan Geological Survey 5 Open File Report

11 Figure 4 Examples of depth slices (defined by unconformity surface) extracted from the two Mag3D models. Products Table 1 lists the products provided. Other products can be prepared from the existing data set, if required. All products were created using the following parameters: Projection Description: Datum: NAD83 Ellipsoid: GRS 1980 Projection: UTM (Zone: 14N) Central Meridian: 105ºW False Northing: 0 False Easting: Scale Factor: Saskatchewan Geological Survey 6 Open File Report

12 Table 1 Survey products. Mag3D Modelling Outcomes The following products are provided as part of the Mag3D modelling. UBC mesh and sus files for constrained and unconstrained models Magnetic models X, Y, and Z format for constrained and unconstrained models Notes on processing - Appendix A Grid Files 1) Standard grids: The mag_lev grd file is provided along with a set of filtered products. Included are a set of what is termed the ZS suite of derived products including the TMI-Tilt product-a descriptive note on the ZS suite is attached in Appendix A. 2) Grids from VariZ Susceptibility processing: Variz-mag3d-const-D0belowUC.grd Variz-mag3d-const-D50belowUC.grd Variz-mag3d-const-D100belowUC.grd Variz-mag3d-unconst-D0belowUC.grd Variz-mag3d-unconst-D50belowTopo.grd Variz-mag3d-unconst-D50belowUC.grd Variz-mag3d-unconst-D100belowTopo.grd Variz-mag3d-unconst-D100belowUC.grd Processing Report (one hard copy + PDF) Saskatchewan Geological Survey 7 Open File Report

13 Discussion of Results Data Quality The magnetic data quality was deemed satisfactory, but a decorrugation procedure was performed prior to performing the Mag3D modelling. A useable topographic field was not provided in the provided GDB (Geosoft database). The field DGPSZ was examined, but found to be incomplete. Given time constraints on the processing, Condor elected to use the Shuttle Imaging Radar DTM data for the terrain model. Unconformity Model As noted previously, Condor was provided with a Microsoft Excel spreadsheet of the form shown in Table 2. The location of the drill holes and the depth to the unconformity are shown in Figure 5. Table 2 Example of spreadsheet provided. HOLE_ID Easting Northing Elevation UC Depth UC Elevation UEX Due to the irregular distribution of control points for the unconformity surface, padding of the data set was required to generate a surface which could then be used in the inversion processing. A complete listing of the drill holes provided is in Appendix B. Mag3D Outcomes A series of images showing the two Mag3D models; the unconstrained and constrained are shown in Figures 6 to 8. The first image (Figure 6) shows the two models from the top down view. The constrained model shows very little character except around the edges where the unconformity model shows there to be basically no sandstone cover and hence the basement susceptibility response is visible. Figure 7 shows the location of anomalous zone with the two models on edge. For purposes of the display in 3D, the base of the model has been trimmed so the upper part of the model (unconformity boundary) can be properly displayed. Figure 8 shows the close-up view of the responses of the two models. While the strong feature is reasonably well resolved in both, there is a zone of more subtle response that is significantly enhanced in the constrained inversion. Figure 9 shows a zoom in on the 3D models around the McArthur River deposit. The slice is at a mean sea level of ~210 m. The constrained model shows considerably more detail than the unconstrained model. Saskatchewan Geological Survey 8 Open File Report

14 Figure 5 Location of drill holes provided to define unconformity surface-coloured image shows the inferred depth. Grid Outcomes A series of grid products were produced from the unconstrained and constrained Mag3D models. The first image, Figure 10 shows four products; the TMI-Tilt product of the mag_lev grid and three grids of the surface of the constrained Mag3D model taken at depths below the unconformity surface of 0, -50 m, and -100 m. The TMI-Tilt product is a very sensitive filter product which highlights trends in the magnetic data not always apparent in TMI images and was deemed a good standard product which to compare the new outcomes. This surface for the three constrained grids was not a constant elevation, but rather a constant value below the unconformity surface. Also shown in each image are contours of the depth to the unconformity surface. Figure 11 shows the same set of images, but zoomed in around the McArthur River deposit. This set of images show the grids from the constrained model are providing more resolution than the standard high resolution (TMI-Tilt) grid. The series of Figures 12a, 13a, and14a show comparisons between the unconstrained and constrained models for the three depths noted above; at the unconformity (defined as 0 depth), -50 m and -100 m. There are differences between the two grids, but not as much as possibly when the voxel models themselves are examined; i.e., Figure 8 and Figure 9. Part of these differences in relative enhancement can have to do with the depth of the unconformity; the shallower the unconformity the closer the unconstrained and constrained models will appear. In deeper parts of the Basin, greater differences are expected. Saskatchewan Geological Survey 9 Open File Report

15 Figure 6 Unconstrained and constrained models viewed from the top. Saskatchewan Geological Survey 10 Open File Report

16 Figure 7 Side on view of unconstrained and constrained models (view to east) showing focus area. Saskatchewan Geological Survey 11 Open File Report

17 Figure 8 Detailed look at feature highlighted in Figure 7. Saskatchewan Geological Survey 12 Open File Report

18 Figure 9 Slice at -210 m msl through the unconstrained and constrained Mag3D models in vicinity of McArthur River deposit. Saskatchewan Geological Survey 13 Open File Report

19 Figure 10 From the upper left going counter clock-wise are images of the TMI-Tilt on the mag_lev grid, constrained model at the unconformity, -100 m below and -50 m below. Figure 11 Same set of images as in Figure 10 but zoomed in around the McArthur River deposit. Saskatchewan Geological Survey 14 Open File Report

20 Figure 12a - Comparison of unconstrained and constrained models at unconformity surface (0 depth). Figure 12b Comparison of unconstrained and constrained models at unconformity surface (0 depth) in the vicinity of McArthur River deposit. Figures 12b, 13b, and14b show the same themes as the a series but are zoomed in around the McArthur River deposit. From these images it is apparent that there is possibly some minor instability in the inversion outcomes for the 0 depth (defined at the unconformity surface). While these effects are not consistent throughout the image, either the -50 m or -100 m grid is deemed more reliable. Saskatchewan Geological Survey 15 Open File Report

21 Figure 13a Comparison of unconstrained and constrained models at -50 m depth. Figure 13b Comparison of unconstrained and constrained models at -50 m depth in the vicinity of McArthur River deposit. Saskatchewan Geological Survey 16 Open File Report

22 Figure 14a Comparison of unconstrained and constrained models at -100 m depth. Figure 14b Comparison of unconstrained and constrained models at -100 m depth in the vicinity of McArthur River deposit. Saskatchewan Geological Survey 17 Open File Report

23 Conclusions and Recommendations A portion of an aeromagnetic data set in the southeast corner of the Athabasca Basin has been processed so as to enhance the magnetic character coming from the basement rocks; this processing called VariZ Susceptibility incorporates the thickness of the overlying sandstone cover as a control on the inversion modelling of the magnetic data. A brief examination of the outcomes in both voxel format and as grids shows there is a sharpening of response from the constrained magnetic model. The grid generated at the actual unconformity surface for the constrained model appeared to show some instability believed caused by how the reference model interacts with the inversion processing. As a result the two grids just below the unconformity at -50 m and -100 m are recommended for general use. Saskatchewan Geological Survey 18 Open File Report

24 References Li, Y. and Oldenburg, D.W. (1996): 3-D inversion of magnetic data; Geophys., v 61, no.2, p Saskatchewan Geological Survey 19 Open File Report

25 Appendix A: Notes on Condor Magnetic Processing and ZS Processing Variz Survey 3D Magnetic Inversion Specifications For Saskatchewan Geological Survey The University of British Columbia 3D magnetic inversion program, version 4.0, was used for the inversions. The inversion was performed with only the topography and a regional unconformity surface as model constraints along with the normal UBC style objective function. Area Inverted: The area to be inverted was roughly defined by a polygon provided by the following coordinates. Easting Northing The area that was inverted, although not saturated with data was defined by the rectangular block: E to E, N to N. The coordinates are in WGS84, UTM Zone 13N. The area of interest above was selected so that it encompassed the part of the survey of interest, but avoided large magnetic features to the north. Magnetic Data: The original TMI data came from a VTEM survey, provided as a Geosoft GDB file, Variz Focus Area.gdb. Geosoft grid: The mag_lev column in the Variz Focus Area.gdb was gridded using Geosoft s minimum curvature at 100 m cells, appropriate for the 400 m line spacing, to create mag_lev.grd. This column appears to show a clean leveled data set and is positioned within the GDB in a manner to suggest it was final data. The grid was further decorrugated by Condor Consulting to remove remnant visible errors between f light lines. No additional data cleaning was deemed required prior to inverting. A regional trend was removed by performing a high-pass filter employing a Butterworth ramp with a 50 km center wavelength. A DC shift was then applied to the windowed data of 200 nt, which was chosen based on a judgment of which lows are negative due to dipole effects in the data. No further preparation of the data was required. Saskatchewan Geological Survey 20 Open File Report

26 Since the data was TMI, not RTP, the magnetic field information was provided to the inversion program. For both the inducing magnetic field and the induced magnetic field, an inclination of 79.6 degrees and a declination of 10.7 degrees were used. The Earth s field strength of nt was used. These are the IRGF values for the center of the survey area. A magnetic sensor height above topography of 140 m was used, which was taken as the mean sensor height in the provided database for the radar altimeter. The inversion results are not considered to be highly dependent on the survey height and a 10 m error would not be significant. Topographic Data: The Variz Focus Area.gdb data file contained a DGPSZ column, but it was incomplete over the entire data set and could not be used. Instead, the topography came from the Shuttle Radar Topography Mission (SRTM) 90 m data set. Model Design: The finite element mesh for the inversion em ployed a cell size of 200 m by 200 m in the east and north dimensions, and the cell height varied wrt depth. There were 669 cells in the easting, 452 in the northing. In the vertical direction, there were six air cells, each being 50 m thick and the ground cells started at 50 m thick and progressively increased in thickness through 50 vertical cells. Depth of mesh, excluding rim & air cells, was 6000 m. Although this depth seems excessive, it is used to help model long wavelength features in the magnetic data. In addition to the cells specified above, seven rim (buffer) cells required by the finite element algorithm were used on the sides of the mesh and five added to the bottom. The inversion was of sufficient size that it had to be split into 18 tiles, each of which was inverted separately. The results from each tile were merged to constitute the complete model. Based on the average horizontal gradient, the initial error estimate provided to the inversion program was 2.39 nt for all data, which was times the standard deviation of the data plus 0.1 times the average gradient of the data plus 1.0 nt. Inversion Results: Three inversion passes were run, the first with a starting and reference model of 0.0 nt. The model from the first inversion was sharpened then used as the reference for the second inversion. The third inversion pass was constrained by the unconformity surface. The third pass was constrained by a regional unconformity surface defined by drilling results. The Proterozoic sediments above the surface are assumed to have very little magnetic susceptibility, while the Archean rocks below the surface are assumed to contain the bulk of the magnetic material. This surface intersected the topographic surface along the east side of the model. To constrain the inversion, the starting and reference model above the surface was set to 1.0e-5 SI units. In addition to using the starting and Saskatchewan Geological Survey 21 Open File Report

27 reference model, the smallness weights were set to above the surface, 10.0 at the surface, and 1.0 below the surface. The gradient weights in the X, Y, and Z directions were not changed. These weights are considered soft constraints, in that they make it harder for the inversion to put magnetic features in the model above the surface, but do allow it to do so to some extent. Hard constraints, i.e., restricting the range of values each cell is allowed to have, were not employed. The values used in the reference model and the smallness weights are subjective and therefore the susceptibility values above the model cannot be construed as accurate in an absolute sense. The final inversion finished with a data misfit ranging from 0.00 nt to 95.9 nt and an average of 1.25 nt. The standard deviation of the misfits was 1.52 nt. The average maximum misfit for the 18 tiles was nt. These statistics are typical and represent a good fit to the observed data. For all the inversion tiles, length scales (used in the inversion objective function) of 400 m was used in the east-west, north-south directions and 200 m in the vertical direction. The final finite element voxel model was trimmed by seven cells on the east and west, seven cells on the north and south, and five cells from the bottom to removed the finite element buffer cells. The voxel model was further cleaned by nulling the voxels where there was no data. Smooth Model Inversion The UBC 3D magnetic inversion is a smooth model inversion, producing fuzzy objects in the resulting block model. This can be mitigated to some extent by sharpening techniques and by constraints. However, the model will still contain a large spatial uncertainty, which manifests as indistinct boundaries. Therefore, the magnetic susceptibility values in this model in general underestimate the actual susceptibility of the magnetic objects being imaged, since the susceptibility is smeared over a greater volume than the actual object. Depth Weighting: It is a well known fact that static magnetic data have no inherent depth resolution. A numerical consequence of this is that when an inversion is performed the constructed susceptibility is concentrated close to the observation locations. In order to overcome this, the inversion introduces a depth weighting to counteract this natural decay. The weighting approximately cancels the sensitivity decay and gives cells at different locations equal probability to enter into the solution with a non-zero susceptibility. This weighting is approximately the depth of the finite element cell to the 3 rd power, to counter a sensitivity drop of that is approximately 1/z 3. Scott Thomas Condor Consulting, Inc. March 25, 2010 For notes on new enhancement filters and ZS processing, the reader is referred to: Shi, Z. and Butt, G. (2004): New enhancement filters for geological mapping; in ASEG 17th Geophysical Conference and Exhibition, Sydney, Ext. Abstr., 5p. Saskatchewan Geological Survey 22 Open File Report

28 Appendix B: Drill Hole Database Used for Unconformity Model HOLE_ID Easting Northing Elevation UC Depth UC Elevation UEX UEX 001A UEX 001B UEX CLC CLC CLC CLC9N CLC9N CLC9S CLC9S CLC9S CLC9S CLC9S WQ WQ WQ WQ WQ WQ WQ WQ WQS ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML Saskatchewan Geological Survey 23 Open File Report

29 HOLE_ID Easting Northing Elevation UC Depth UC Elevation ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML D Saskatchewan Geological Survey 24 Open File Report

30 HOLE_ID Easting Northing Elevation UC Depth UC Elevation D D D D D D D D D D D D D D WB WB WB WB BK BK BK BK BK BK BK BK BK D D D D D D D D D D D D D ML ML ML ML ML ML Saskatchewan Geological Survey 25 Open File Report

31 HOLE_ID Easting Northing Elevation UC Depth UC Elevation ML WBE WBE SM SM SM ZF ZF ZF SM SM SM SM SM SM CR BF HG HG HG HG HG R R R R R R R R R R R R ZM ZM ZM ZM ZM ZM ZM ZO ZO EL EL EL Saskatchewan Geological Survey 26 Open File Report

32 HOLE_ID Easting Northing Elevation UC Depth UC Elevation ZK ZK ZK ZM ZO ZO ZO ZO ZRE ZRE ZRE MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC Saskatchewan Geological Survey 27 Open File Report

33 HOLE_ID Easting Northing Elevation UC Depth UC Elevation MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC AH MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MAC MC MC CL CL CL CL HG AL AL EL EL Saskatchewan Geological Survey 28 Open File Report

34 HOLE_ID Easting Northing Elevation UC Depth UC Elevation EL EL EL EL EL EL EL EL EL EL LN LN LN LN MH MH MH MH MH MH WHB WHC WHC WL WL WL WL WL AL ML ML ML ML ML BF BF BF BF BF BF BF BF BF MK MK MK Saskatchewan Geological Survey 29 Open File Report

35 HOLE_ID Easting Northing Elevation UC Depth UC Elevation MK MK MK MK MK MK MK MK MK MK MK MK MK MK MK MK MK BF BF BF BF BF BF BF BF BF BF BF PP PP PP PP BF BF BF BF BF BF BF BF BF BF BF BF BF BF Saskatchewan Geological Survey 30 Open File Report

36 HOLE_ID Easting Northing Elevation UC Depth UC Elevation BF BF BF BF BF BF BF MH MH MH MH MH MH MH MH MH MH MH MH MH MH ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML BF BF BF ML ML Saskatchewan Geological Survey 31 Open File Report

37 HOLE_ID Easting Northing Elevation UC Depth UC Elevation ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML ML AL AL AL MH MH MH MH MH MH ML ML ML ML ML ML ML ML ML EL 096A WL P BF BF BF BF BF BF ML Saskatchewan Geological Survey 32 Open File Report

38 HOLE_ID Easting Northing Elevation UC Depth UC Elevation ML Saskatchewan Geological Survey 33 Open File Report