well-logging and near-surface seismic methods for aquifer ... · near surface 2008 – 14th...

Near Surface 2008 – 14th European Meeting of Environmental and Engineering Geophysics Kraków, Poland, 15 - 17 September 2008

Well-logging and near-surface seismic methods for aquifer detection Gabriela M. Suarez*, Soo-Kyung Miong, Joe Wong, Robert R. Stewart, Alejandro D. Alcudia, Hanxing Lu, and Khaled Al Dulaijan Summary Well-logs, high resolution shallow seismic, and VSP surveys were conducted in an experimental well near Priddis, Alberta, to test their feasibility in characterizing the near-surface stratigraphy and hydrogeological characteristics of the Paskapoo Formation. According to the analysis of well logs, the well penetrates interbeds of sandstone-shale layers and five porous zones were identified at depths of 28, 39, 50, 62 and 120 m, ranging from 0.34 to 0.58. Among these porous zones, there were three water-bearing zones at depths of 28, 50 and 120 m. Based on the P-wave velocities calculated from the sonic log, surface-wave analysis was carried out to estimate S-wave velocities. Up to depths of 40 m, the respective P-wave velocities for the sandstone and shale units were 2250 m/s and 2080 m/s, and 1220 and 1150 m/s for the S-wave, respectively. Using a microphone-geophone analysis, the airwave was characterized by high frequency and broadband. The integral interpretation of all the datasets and well-logs show a good correlation between all of them. The water-bearing zones (28 and 50 m) and the strong reflector at 76 m were resolved by the surface seismic and VSP data. Introduction

Near-surface seismic reflection surveys, VSP, and well-logs were acquired near Priddis, Alberta. The objectives were to develop techniques to characterize and improve the image of the near-surface, in addition to testing whether the groundwater saturated zones of the Paskapoo formation, the largest single source of groundwater in Canadian Prairies (Grasby, 2006), can be characterized by these methods. There can be resolution limitations in seismic reflection techniques when mapping water saturated zones (Steeples and Miller, 1998). This study incorporates high-resolution near-surface seismic methods, vertical seismic profiling (VSP), and the well-logs to enhance near-surface resolution.

The study was completed in two parts. The first part was undertaken during the University of Calgary’s field school and included well logging (open-hole and cased-hole), VSP, high-resolution 2D seismic refraction and a 3D seismic survey (Figure 1). The second part was completed in March 2008, and included a side-by-side comparison of a multicomponent 2D line, a land streamer line, and a microphone line for air-pressure measurements.

The well-logs and drilling cuttings were used for a formation log description and identifying porous zones near the well. Data from the refraction survey was used for surface-wave analysis to estimate S-wave velocity in the area. The microphone measurements were used to characterize the air-wave. The 3-D and 2-D surface seismic and VSP datasets were processed to obtain stacked sections, which are then correlated with 1-D synthetic seismogram generated from the well-logs.


Figure 1. a) Location of the test well (green circle) and the seismic survey configuration of at the Priddis site (land streamer-red line), 3-D (blue area) and refraction 2-D (purple line): Survey geometry and the acquisition parameters of b) land streamer, c) 2D refraction line and 3D, and d) VSP with lithologic description (yellow-sandstone, brown-shale) and the well-logs. Description of the surveys The surveys were all undertaken at the University of Calgary’s geophysical test site near Priddis, Alberta (40 Km SW of Calgary).

Refraction survey (Figure 1b): The 403 m vertical geophone 2-D line consisted of a fixed array with a 180 m spread and 72 receivers. The receiver and source spacing were 2.5 m and 12.5 m, respectively. The source used for this 2-D seismic line was a five-pound sledgehammer. The record length was 600 ms with 0.125 ms sampling rate.

Surface seismic reflection and microphone surveys (Figure 1c): The 3-D survey consisted of 11 North-South source lines and 7 East-West receiver lines every 50 m, with vertical-component receivers and a vibroseis source every 10 m. The side-by-side 2-D multicomponent seismic line consisted of a 200 m planted-geophone line, a 20 m land streamer system, 32 microphone locations and a 400 m source line. For the 2D survey, 200 3-C planted-geophones at 1 m spacing and microphones every 5 m were used. Both of the surveys employed an 18000 lb Envirovibe source with 4 times vertical stack sweeping from 10 to 250 Hz sweep with an 11 seconds listen time.

VSP survey and well-logs (Figure 1d): Two offset VSP datasets at 1.5 m and 15 m were acquired. Both of them employed a 3C downhole sonde spaced at 0.5 m. A sledge hammer and a vibroseis truck were used as seismic sources. The depth coverage of the VSP surveys was 4 m to 63 m for the 1.5 m and 90 m for the 15 m one. A suite of well-log measurements including spontaneous potential, gamma-ray, resistivity, caliper, temperature, density and neutron porosity logs were acquired in the open-hole before the well was cased. The full-waveform sonic log, acquired after casing the well, is shown in Figure 1d.

Well-log analysis: Stratigraphic characterization

According to the driller’s lithologic description, the well penetrates mainly sand and shale layers. The natural gamma-ray, resistivity and neutron-neutron logs delineated these layers and the lithologic information drawn from that agree with the driller (Figure 1d). We


have identified five porous zones (zones I to V) at depths of 28 m, 39 m, 50 m, 62 m, and 120 m. Out of these five zones, three were saturated with water at depths of 28m (zone I), 50m (zone III) and 120 m (zone V) as indicated by the driller. After the conventional quantitative analysis (Asquith and Krygowski, 2004), the porosities in the five zones ranged from 0.34 to 0.58 with the highest porosities observed in the top three zones (I, II and III). The water saturations ranged from 0.12 to 0.79. Surface-wave analysis and S-wave velocity

In seismic surveys more than two thirds of the seismic energy can be imparted into Rayleigh waves when a P-wave source is used (Park et al., 1999). These waves are non-dispersive in the case of propagation along the surface of an isotropic homogeneous half-space. In more complicated layering, dispersion properties of surface-waves can be useful for geotechnical engineering and S-wave reflection seismology purposes, as the S-wave velocity of near-surface can be estimated from surface-wave analysis. The method applied for surface-wave analysis in the Priddis site involved recording Rayleigh waves on vertical-component geophones, estimating phase-velocity dispersion curves for Rayleigh waves, and inverting these dispersion curves to estimate S-wave velocity as a function of depth.

Figure 2. Results of the surface-wave analysis: Example of a shot gather (left), dispersion curve (center), S-wave velocity profile (right).

Figure 2a shows a shot gather where most of the energy is surface-wave with no obvious body waves, with a velocity ranging from 270 m/s to 890 m/s. From the dispersion curve for this shot (Figure 2b), the strongest trend was picked as the fundamental mode and used for obtaining the 5 layer initial model of the inversion. The P-wave velocities for the initial model were calculated from the sonic log and they ranged from 2250 to 2080 m/s at depths shallower than 40 m. A Poisson's ratio of 0.405 was used for the S-wave velocity model, producing after the inversion velocities ranging from 200 m/s to 1400 m/s increasing with depth (Figure 2c). The vertical variation of S-wave velocities near the test well is in rough agreement with the lithologies drawn from the drilling cuttings. Air blast on microphones

Source-generated noise in shallow seismic surveys is a potential problem in recording reflection signals. Stewart (1998) proposed a noise-reducing multi-sensor method for seismic land operations. It basically consists of a dual-sensor (microphone-geophone) instrument that could record air-pressure and be used to attenuate some of the air-noise coupled into the geophone records.

Time-frequency analysis using the Gabor transform was done to analyze the amplitude variations with time and frequency of the 10-250 Hz linear sweep recorded. The highest amplitudes occurred at the highest frequencies and latest times. The characteristics of the airwave source seem to be nonstationary since the airblast amplitude decays with distance, even though the frequency content and waveform shape seem to be consistent. This consistency could be highly attributed to the gentle topography of the site. 2D land streamer sections The few reflections existent in this area are imaged on both stacked sections (Figure 3 a and b); however, the land streamer sections looks more contaminated with noise. There is a


considerable difference in amplitude between the sections, but the same events between 75 and 300 ms can be observed. Comparing these two sections with the corresponding 3-D cross-line (Figure 3), we observed that the 3-D sections provide a better image after 300 ms, but can not resolve the shallower reflectors that are present in the land streamer datasets. Data correlation

The synthetic seismogram which was generated using well logs, the processed 1.5 m and 15 m offset VSP data, land streamer seismic section and the well-logs are integrated for interpretation (Figure 4). The previously outlined porous zones at depths of 28, 39, 50 and 62 m are identified in the synthetic seismogram and the two VSP corridor stacks at respective two-way-travel times of 54, 60, 66 and 74 ms; two of these reflectors represent the position of the water-saturated zones (see blue lines in Figure 4). An additional strong reflection is observed at 86 ms (at a depth of 76 m) in the middle of a relatively clean sand zone. This reflector is also imaged by the land streamer data. Positions of the strong reflections at ~54 ms and 70 ms observed from the surface seismic data also correlate well with those from the VSP stacks. The discrepancy in the position of the porous zones imaged in the synthetic seismogram, VSP stacks and the surface seismic stack can be attributed to differences in geometry, source frequencies, trace signal-to-noise ratios, instrument timing errors (Gochioco, 1998), processing simplifications and vertical resolution.

Conclusions

High resolution near-surface seismic, VSP data and various well-logs were acquired and analyzed to characterize the groundwater bearing zones of the Paskapoo formation near Priddis, Alberta. Based on the P-wave velocities calculated from the sonic log, surface-wave analysis was carried out to estimate S-wave velocities. Up to depths of 40 m, the respective P-wave velocities for the sandstone and shale units were 2250 and 2080 m/s: The S-wave velocities for these units were 1220 and 1150 m/s, respectively. The air-wave noise generated from the seismic source was characterized by microphone-geophone analysis. The airwave at the survey site was characterized by high frequency (150-250 Hz) and broadband: the highest amplitudes are in the highest frequencies.

The well-logs were used to delineate the sandstone and shale unit up to depths of 130 m and five porous zones were identified at depths of 28, 39, 50, 62, and 120 m. The porosities for these zones ranged from 0.34 to 0.58 with the water saturations ranging from 0.12 to 0.79. The integrated study of the processed land streamer, synthetic seismogram, two VSP data sets, and the well-logs show good correlation between the various data sets. The water-bearing zones at depths of 28 m and 50 m were resolved by the surface seismic and VSP data. Also both datasets resolved the strong reflector occurring at a depth of 76 m. The surface seismic, VSP and the well-logs presented in this study provided useful information regarding the near-surface stratigraphy and hydrogeological characteristics of the Paskapoo Formation near Priddis, Alberta.

Figure 3. (a) Comparison of stacked sections from the 2D land streamer line,(b) planted-geophone line and (c) coincident 3-D cross-line with identification of the common reflectors .


Figure 4. Integrated interpretive display plotted in two-way traveltime. From left to right: Land streamer seismic, synthetic seismogram, VSP corridor stacks of 1.5 m and 15 m offset VSP, normal-move-out corrected up-going primary fields of 15 m VSP data tied to the depth indexed sonic and gamma-ray logs. The porous zones are indicated by the yellow and blue lines (water-bearing zones). References Grasby, S., 2006. Paskapoo Formation – Alberta’s most heavily used aquifer: The Fossil Water report, 1, No. 4, pp.1-3. Park, C. B., R. D. Millar, and J. Xia, 1999, Multichannel analysis of surface waves: Geophysics, 64, 800-808. Steeples G.W. and Miller, R.D., 1998, Avoiding pitfalls in shallow seismic reflection surveys: Geophysics, 63, No.4, pp. 1213-1224. Stewart, R.R., 1998, Air-noise reduction on geophone data using microphone records: CREWES Research Report, 10, 5.1-5.8.

well-logging and near-surface seismic methods for aquifer ... · near surface 2008 – 14th...

Documents