IPP PROGRESS REPORT (Draft copy) Borehole Seismic Imaging of Near-Vertical Structures A Joint Project Between Falconbridge Limited and the Geological Survey of Canada David W. Eaton Geological Survey of Canada Continental Geoscience Division 1 Observatory Crescent Ottawa, ON K1A 0Y3 This copy printed June 18, 1999 Introduction: This report summarizes progress to date in the IPP Project Borehole Seismic Imaging of Near-Vertical Structures. This project continues from an earlier IPP project Seismic Imaging of Complex Geological Structures: A New Approach in Exploration Technology. The overall goal of these projects is to develop new seismic technology suitable for mineral exploration in a hard-rock environment where dips are steep (> 60o), a geological setting generally considered to be "hostile" for conventional surface-based seismic methods. The initial phase of this project, summarized by Salisbury et al. (1996), was a feasibility study aimed at determining the densities and acoustic velocities of ores and host lithologies in the Kidd Creek camp using laboratory and in situ measurements from boreholes BH-4509 and BH-4741. The range of acoustic impedance (the product of velocity and density) for the rock types sampled showed that certain lithologic contacts, in particular ore material juxtaposed against felsic host rocks, are characterized by a large change in impedance and should produce strong seismic reflections. With this information, phases 2 and 3 were conducted in 1994 using borehole seismic techniques, which arepreferable to surface-seismic methods in areas of near-vertical dip because a large fraction of the reflected energy stays in the subsurface and is not intercepted by surface receiver arrays. In phase 2, a series of "vertical seismic profiles" (VSP's) were acquired in BH-5171 and BH-5100, using techniques similar to methods used in hydrocarbon exploration1. These tests yielded cross-sectional images of the North Rhyolite contact (Guest and Milkereit, 1994; Eaton et al., 1996) and showed that small explosive charges in water-filled pits provide good seismic sources for this scale of exploration. As a followup to this, a unique "side-scanning seismic" experiment was designed and implemented in June, 1994. This dataset provided a series of images analagous to level plans through the North Rhyolite contact, and showed convincingly that massive sulphide ores can be detected and delineated by seismic methods (Eaton et al., 1995). The current IPP project is a continuation of the side-scanning seismic experiment. The present survey, however, is more ambitious in scope and tests a section of the distal Mine Rhyolite with a more-or-less linear strike (Fig. 1), in contrast with the curvilinear trend of the North Rhyolite contact investigated previously. The level of exploration of this contact (at the time the survey was planned) was also releatively modest, compared with the section of the North Rhyolite contact studied earlier. Survey Concept and Design The ideal borehole for a BSP or side intersects the target contact at a deep level. This permits placement of receivers below the depth level of interest at a location well separated from the contact. BH-4226 was selected by D. Crick because it met these criteria. The objective of the current project is to scan a previously undrilled section of the distal Mine Rhyolite west of its intersection by BH-4226 (Fig. 1). Experience from the previous side-scanning survey showed that to achieve the desired imaging goals the best shot-line orientation is parallel to the contact. Each contact-parallel shotline produces, in principal, a series of closely spaced plan-view image sections. Therefore, in order to scan both shallow and deep portions of the contact, two contact-parallel lines were used (Lines 1 and 2; Fig. 1). In order to provide a tie between the image sections from lines 1 and 2, a perpendicular tie line was also shot (Line 3; Fig. 1). The previous side-scanning experiment was acquired along a curvilinear section of the North Rhyolite contact, which acted as a type of parabolic mirror causing focusing of the reflected energy into the receiver array. As a result, the area of the North Rhyolite contact imaged by the seismic survey was somewhat larger than would normally be the case. In the more typical case of linear strike trend, more distant contact-parallel lines must be longer than shot lines closer to the contact in order to scan a similar along-strike section. For this reason line 1, located further from the contact, consisted of 40 shots evey 30 m, whereas line 2 consisted of 40 shots every 15 m (approximately). The tie line consisted of 20 shots with an intermediate shot spacing of 25 m. Site Preparation Prior to the arrival of the data acquisition crew, the shot lines were prepared by Falconbridge personel and a private contractor. Site preparation consisted of: 1) flagging of lines to be cut; 2) preparation of cut lines through variable terrain (muskeg to parkland conditions) using a backhoe vehicle; 3) drilling of 100 overburden holes to ~ 4 m depth; 4) insertion of 5 m lengths of PVC tubing into each hole to act as casing. The use of casing is intended to facilitate the re-loading of shot holes. Data Acquisition Data acquisition took place in very hot (up to 38o) conditions between June 12-21, 1995. The data acquisition crew was composed of 8 people, split into 4 groups (1 shooting crew, 2 loading crews and 1 recording crew). This large group was necessary in order to acquire an adequate number of shots (200) per day. Holes were loaded with 175 g pentalite boosters that were detonated using seismic caps. These explosives had been stored since the previous winter in a powder magazine located south of the mine complex. The survey was acquired at 13 receiver levels (labelled A-M) at wireline depths approximately every 7.5 m from 944 m to 1030 m. The staging area for the survey was at the collar for BH-4226 (Fig. 2). The acquisition system was supplied by the University of Alberta and consisted of a Geosource 3-component borehole seismometer, an 1100-m 7-conductor analog cable and a Bison 12-channel recording system. Since the downhole tool acquires data one level at a time, the 100 shotholes needed to be reloaded and fired 13 consecutive times. Precise time break for both the recording system and explosives were provided by synchronized GSC shooting boxes designed to fire on the minute (see Fig. 3). This detonation system necessitated constant radio contact between the shooting and acquisition crews. This level of activity on the radio sometimes posed problems, since during the period of acquisition one of the mine radio channels was essentially unavailable for normal mine operations. A thicker PVC casing was used for this survey, compared with the previous one. Unfortunately, the material used to case the holes for the 1995 survey was more brittle than in the previous case and essentially disintegrated after the first level was shot. Thus, the shotholes were re-loaded and fired 12 times in open (uncased) conditions. For holes drilled into clay (all of lines 2 and 3, and shots 35-40 of line 1), this did not pose any problems. The shotholes filled naturally with water and retained their integrity throughout the survey. The remaining shotholes for line 1, however, were dry and collapsed during the first few passes. In order to complete this part of the survey, special arrangements were made for a backhoe to dig small pits along line 1. An ore truck filled with water was brought to the site each morning to fill a water tank mounted on skids. This tank was dragged down the line by the backhoe operator in order to refill the shooting pits for the next two passes. Single station GPS position was used to survey in some of the shot locations. At each site surveyed in this manner, between 100 and 200 individual readings were averaged to compensate for satelite noise. Unfortunately, the accuracy of this method is still no better than 50 m. Relative positions of shots were determined by chaining the shot lines. This was combined with the GPS position to locate all of the shots from the survey. Data Processing Processing of the data was carred out at the Geological Survey of Canada during the summer and fall of 1995. Tests were conducted using a variety of different processing strategies, and it became apparent that some of the techniques used in the previous experiment (Eaton et al., 1995) were not suitable for this dataset (possible reasons are discussed below). The final processing strategy consisted of: 1. Filtering the data to remove 60 Hz contamination. 2. Aligning the P-wave direct arrivals (flattened to 0.15 s). 3. Statistical rotation of horizontal components into radial and transverse orientations (to correct for tool rotation in the borehole). 4. Deterministic minimum-phase deconvolution (design window centred on P-wave direct arrival after enhancement by median filtering). 5. Bandpass filter (30-60-150-200 Hz). 6. Automatic Gain Control. 7. Restoration of traces to original recording time. The main difference between this processing scheme and the one used previously (see Eaton et al., 1995) is the absence of steps to remove P-wave and S-wave direct arrivals. Attempts to include these steps resulted in significant deterioration in the signal. Borehole Logging A suite of borehole logs was collected in October as part of this IPP project. Borehole logging provides a method to callibrate the seismic images and tie with the geology of the borehole. The logs most relevant to seismic interpretation (velocity and density) are shown in Figure 4. The calculated impedance log (the product of velocity and density) is also shown. It was discovered when the logging data became available that some of the receiver levels for the borehole seismic survey were actually situated within a massive sulphide zone. This is unfortunate, since ideally receivers should be placed in a homogeneous environment to achieve the cleanest possible signal. The massive sulphides are clearly discernible in Figure 4 as a zone of high density. The change in velocity from high to low velocity within this zone suggests a change in ore mineralogy from top to bottom, with the lower part mainly depleted in pyrite. Seismic Interpretation The complete three-component dataset from the survey is summarized in Figures 5-7. The data shown have been processed to up to and including process 6 in the list above. Thus, the data are flatenned on the P-wave direct arrival, at a time of 0.12 s. In general, the data show less continuity of reflections from one panel to the next, in comparison with the previous survey. This is due, in part, to the lack of direct-arrival suppression routines in the data processing sequence. However, at least part of this is attributed to less continuity of stratigraphic horizons along this portion of the Rhyolite, in comparison with the North Rhyolite. Although the data show less reflector continuity than before, a number of interesting anomalies have been identified. The anomalies are generally recognized by virtue of their apparent dip in the image, which contrasts with both the overall fabric (due mainly to direct arrivals and associated reverburations). This type of signature is consistent with the main anomaly interpreted from the 1994 dataset. Other anomaly attributes to note are amplitude and consistency from one panel to the next. In order to determine the locations of anomalies, accurate propagation velocities and positional inforamtion are required. To assess these, the direct arrival picks used to process the data are displayed in Fig. 8 (P waves) and Fig. 9 (S-waves - line 1 only). In Figure 8, it is apparent that several populations exist. The points shaded in red are from the near part of line 1 (offsets less than 1300 m) and all of line 3. These points yield a velocity of 6753 m/s by linear regression (consistent with the logging results), with an interecept (shot static) of 6.2 ms. The points in green are from the far offsets of line 1 (shots 1-20). These points yield a similar velocity, but a larger intercept value, consistent with the thicker overburden for this area. The shots from line 2 give an unrealistic velocity of < 5000 m/s; this suggests that there are errors in the locations for these shots from the GPS surveying technique. Analysis of the S-wave picks from line 1 (Fig. 9) yields an estimated velocity of 3326 m/s. At this point, only anomaly 1 has been examined in detail. This anomaly appears primarily on the "vertical" component, and has an arcuate shape (Fig. 5). Note that the vertical component is actually oriented along the axis of the borehole. Fig. 10 shows anomaly 1 after processing step 7, restoration of the data to the original traveltime. An additional mild coherency filter has also been applied here. It is now clear that anomaly 1 has a nearly hyperbolic shape, with its apex near shot 21. This implies that it is produced by an isolated scattering point situated closest to shot 21 from the survey. A second scattering anomaly has also been identified (not visible in Fig. 5), with an apex at ~ 0.34 s near shot 24. Figure 11 shows the results from forward modelling of this anomaly. The location of the scattering point used to generate this model is shown at the lower right. This location was determined using traveltime inversion of the picks from scattering anomalies 1 and 2. The fit here is reasonable, but not perfect. The traveltime inversion scheme used to obtain these results is a least-squares method. The sqaured traveltime errors were minimized and an additional penalty function was included to satisfy the observation that the scattering point is closest to shot 21. A plot of the inversion error is shown in Fig. 12, which gives some idea of the position uncertainty in easting and northing. Another potentially useful plot is the inversion depth (Fig. 13), which seems to show a "flat spot" at about the same location. This inversion location is plotted with respect to the shot lines in Fig. 14. Figure Captions Fig. 1. Surface geology for Kidd Creek mine and environs, showing location of the 1994 and 1995 borehole seismic surveys. The side-scanning survey for 1994 was conducted in BH-5171 and investigated a curvlinear portion of the North Rhyolite contact. Ancilliary logging results were collected in BH-4509. the 1995 borehole seismic survey was conducted in BH-4226, using 100 shotholes drilled along lines 1, 2 and 3. This survey investigated a more linear Rhyolite trend closer to the main Kidd Creek orebody. Fig. 2. a) Line 1, looking west, at its intersection with line 3. b) Main staging area at the collar of BH-4226. the white truck houses the recording equipment from the University of Alberta. Fig. 3. a) Shooting crew with additional helper from the loading crew. Arrow shows GSC shooting box used to detonate the charges. b) Shot fired along line 1. The best coupling was achieved when there was no blowout; this was not a particularly good shot. Fig.4. Borehole logging data (velocity and density logs) from BH-4221. Mafic units show up as high velocity zones, while felsic units show a much lower velocity. A mineralized zone from 990-1000 m wireline depth is indicated by high densities in this interval. Fig. 5. Processed seismic data from line 1, showing borehole logging data from the same depth interval for reference (top). Fig. 6. Processed seismic data from line 2, showing borehole logging data from the same depth interval for reference (top). Fig.7. Processed seismic data from line 2, showing borehole logging data from the same depth interval for reference (top). Fig. 8. Direct arrival pick times (P waves only). Fig. 9. Direct arrival pick times (S waves only). Fig. 10. Enlargement of anomaly 1, after restoration of times to original recording time. Fig. 11. Synthetic modelling of anomaly 1. Fig. 12. Traveltime inversion error plot for anomaly 1. Fig. 13. Estimated depths from traveltime inversion for anomaly 1. Fig. 14. Location of scattering point from traveltime inversion. References: Eaton, D. 1995. Seismic Imaging of Complex Geological Structures: A New Approach in Exploration Technology, IPP Report, Phase 3. Eaton, D., Guest, S., Milkereit, B., Bleeker, W., Crick, D., Schmitt, D. and Salisbury, M. 1996. Seismic Imaging of Massive Sulphide Deposits, Part III: Borehole Seismic Imaging of Near-Vertical Structures. Submitted to Economic Geology. Guest, S. and Milkereit, B. 1994. Seismic Imaging of Complex Geological Structures: A new Approach in Exploration Technology, IPP Progress Report, Phase 2. Salisbury, M., Bleeker, W., Eaton, D. and Milkereit, B. 1996. Integrated Seismic and Rock Properties Studies at Kidd Creek. manuscript in preparation. 1 Since boreholes drilled for mineral exploration are rarely vertical, the alternative term "borehole seismic profile", or "BSP", is suggested. 8