ABSTRACT

A multi-offset, azimuthal, three-component VSP survey was acquired over a deep Unayzah sandstone gas reservoir in central Saudi Arabia. The objective was to confirm the direction for a proposed horizontal sidetrack to a vertical delineation well. A production test at the vertical well indicated a possible reservoir boundary less than 100 feet away from the well. It was hoped that some of the offset sections might image a small fault, which may represent such a boundary. In addition to the normal zero-offset VSP conducted at the vertical well, two multi-offset VSPs at orthogonal azimuths, and four single-offset VSPs at other azimuths were acquired.

Due to excellent cased-hole conditions for the receivers and uniform sandy terrain for the sources, the VSP records were nearly noise-free. The VSP data had virtually twice the dominant frequency of the surface seismic data. The offset VSP sections (both compressional and shear) succeeded in better resolving the reservoir properties of the gas-bearing Unayzah-A sandstone. They showed that its quality deteriorates to the northeast, whereas both its thickness and porosity remain uniform to the northwest, the proposed direction of the horizontal sidetrack. The shorter offset sections in each quadrant surrounding the well showed no clear disruptions that may indicate faulting. This suggests that the boundary detected from the well test may be due to facies change, cementation, or mineralized fractures associated with a major fault bounding the accumulation to the south.

INTRODUCTION

The Wudayhi field was discovered in 1998 in central Saudi Arabia, about 40 km west of the supergiant Ghawar field (Figure 1). Vertical exploration and delineation wells tested gas at a rate of 10–15 million standard cubic feet per day and condensate at up to 2,000 barrels per day from the upper Carboniferous – mid-Permian Unayzah Formation. The formation is about 580 ft thick and divided into the Unayzah-A (UNZA), Unayzah-B (UNZB) and Unayzah-C (UNZC) reservoir zones (Figure 2). The porosity varies between 3 and 9% and the permeability is less than 10 mD. To achieve greater gas production rates, a highly deviated sidetrack was proposed from a vertical well. The sidetracked section would have an inclination of over 85° in order to traverse the whole gas pay zone. Also during the testing of this well, the pressure build-up analysis indicated that a probable reservoir boundary existed at a distance of less than 100 ft away from the well. A major bounding fault zone to the south may be associated with this boundary.

In order to address these uncertainties various seismic techniques were used. Initially, based on the inversion of the 3-D surface data, it was concluded that the approximate two-way seismic time interval corresponding to the reservoir has a lower average acoustic impedance to the west and northwest of the well (Figure 3). Based on sonic and density logs from the well, reservoir porosity was found to strongly and inversely correlate with acoustic impedance (Figure 4). However, the synthetic seismogram, which adequately matched the seismic data, showed that the predominant wavelength of the 3-D data is greater than the thickness of the gas pay zone (Figure 5). In order to more closely represent the gas pay zone, the dominant frequency would have to be doubled, from 30 to 60 Hertz (Hz). To achieve the required resolution, an azimuthal offset VSP around this well was therefore acquired, and the results are described in this paper.

WUDAYHI FIELD AND UNAYZAH RESERVOIR

The Wudayhi field is a gently dipping anticlinal structural trap. It is part of a forced fold related to draping of the sedimentary cover over a basement fault-bounded block. The reservoir units, Unayzah-A (130 ft), Unayzah-B (100 ft) and Unayzah-C (350 ft), consist primarily of eolian and fluvial sandstones with a large degree of variation in permeability and porosity due to extensive cementation. Surface 3-D seismic data indicates that the Unayzah gas reservoirs are highly faulted, and reservoir evaluation indicates a high level of vertical and lateral heterogeneity. Fracture analysis from azimuthal VSP and 3-D surface seismic azimuthal AVO data indicates an EW-trending fault zone consistent with the regional tectonic trend (Al-Hawas et al., 2003; Neves et al., 2003; Nebrija et al., 2003). The top of the productive Unayzah-A reservoir lies at a depth of 13,727 ft below the surface and its structure exhibits four-way closure with a longitudinal axis oriented in the E-W direction. To the south and northwest, the gas accumulation is bounded by faults (Figure 3).

VSP DATA ACQUISITION AND PROCESSING

In a VSP field configuration, the seismic source is located at the surface and the closely spaced three-component (3C) receivers are clamped at depth in the borehole. This configuration provides the VSP with a better temporal and spatial resolution compared to the surface seismic technique, mainly for the following two reasons. Firstly, the seismic reflections travel only once through the highly attenuating weathered zone and, secondly, the size of the Fresnel zone is smaller because the receivers are located much closer to the target. When the source is located away from the borehole the configuration is called an offset VSP (OVSP) as distinct from zero-offset VSP (ZVSP), where the source is located almost directly above the receivers.

The VSP layout for this study is shown in Figure 6. Two orthogonal shot lines were laid out such that one is parallel to the south bounding fault and the other perpendicular to it. Ray-tracing, using the velocity model derived from the zero-offset VSP (ZVSP) at the discovery well, indicated that the critical angle of incidence at the Unayzah Formation is approached at offsets beyond 9,000 ft. Thus, along each of two orthogonal azimuths, four shots were acquired at offsets of 3,000 ft, 5,000 ft, 7,000 ft and 9,000 ft away from the well over a receiver depth range of between 12,400–14,480 ft. The multiplicity of shots resulted from a separate objective of measuring the anisotropy of the reservoir from the splitting of shear waves generated above the reservoir.

In addition to the ZVSP, which was recorded from total depth (TD = 14,540 ft) to the surface, four shots at equidistant offsets of 2,500 ft from the well were acquired at various azimuths around the well to provide high-resolution images away from the borehole. Due to the numerous shot locations and the requirement to reduce lost rig-time, five Vibroseis trucks were used to sweep a 14-second linear signal from 10–120 Hertz (Hz). The survey was acquired with an 8-level, 3C downhole receiver array with 50 ft level spacing.

The Offset VSP (OVSP) data was processed using a 3-component processing flow designed to separate and enhance the upgoing compressionalP and converted P-to-SV (C-wave) wavefields. Due to the excellent cased borehole conditions and uniform sandy terrain, the VSP data were practically noise-free. Figure 7 shows the data for shots C and D after tool rotation correction; notice the strong mode-converted shear-wave reflection from the top of the Unayzah-A unit on the radial component.

The separation of the P and C wavefields was performed using a parametric-inversion approach similar to the one developed by Esmersoy (1990). This technique utilizes the moveout and polarization information to separate the vertical and radial wavefields into down and up P and C wavefields. Figure 8 shows the up P and C wavefields for shots C and D after the parametric decomposition. The C-wave reflection is very prominent and complements the interpretation of the P-wave. Analysis of the spectrum indicates maximum frequencies of up to 50 Hz for both reflected P and C waves before deconvolution of the data (Figure 9).

RESULTS

Both the upcoming P and C waves were deconvolved with a zero-phase operator derived from the downgoing P wavefield. The imaging was performed using a VSP-CDP mapping procedure with a velocity-depth model derived from the VSP data.

Figure 10 shows the tie between the synthetics, ZVSP and OVSP images at the well. Changes in the reservoir are evident in the change of the seismic waveform away from the well. The top of the gas zone corresponds to the zero-crossing, while the top of the Unayzah-B sandstone (peak) is partially resolved in Shot C. One major advantage of the VSP over surface seismic is that the recorded downgoing wavelet at the reservoir level provides an opportunity for a deterministic deconvolution of the recorded upgoing wavefield. The high signal-to-noise (S/N) of the raw VSP data and stability of the images is due to the excellent cased-hole conditions for the receivers and the uniform sandy terrain for the sources. Also there is practically no variation in the downgoing source wavelet within the receiver depth interval for both azimuths. This means that the waveform variations observed in the reflection images cannot be attributed to near-surface geology, but rather to changes within the reservoir interval.

Figures 11a and 11b show the high-resolution VSP reflection image for shots C and D, respectively, spliced into the corresponding surface seismic section. Notice the enhanced definition and clarity of the gas pay in the VSP. An f-k spectral comparison of the VSP and surface seismic data (Figure 12) indicates that the VSP indeed has a higher bandwidth.

INTERPRETATION OF RESULTS

Along the orthogonal azimuths, the offset reflection images, now plotted in depth, are shown in Figure 13 radiating away from the well. The gamma-ray and porosity logs are shown in between, and plotted in the same scale to show the correspondence between the seismic response and the reservoir stratigraphy.

The Unayzah-A gas reservoir, when followed to the northeast along the Shot D section, shows significant waveform variation indicating very variable reservoir quality. Also the distinction between the Unayzah-A and Unayzah-B, and the intermediate siltstone separating the reservoirs, is not clear. Along the Shot C profile, however, both the sandstone gas reservoirs and the siltstone remain virtually unchanged in the northwest direction.

A seismic inversion of the same two offset VSP sections (Figure 14), using a 1-D initial impedance model on the sonic and density logs at this well, indicates uniformly lower acoustic impedance and therefore higher porosity along Shot C. These results support the plan to drill the horizontal sidetrack well in the WNW direction. This direction points to the best-developed reservoir, as initially interpreted from the inversion of the lower resolution surface seismic data.

A comparison between the P-wave and C-wave offset seismic sections (Figures 15a and 15b) clearly demonstrates the much higher vertical and spatial resolution of the shear-wave data. Vertical resolution refers to the ability of the seismic waves to distinguish two closely spaced reflectors and is determined by the dominant wavelength of the seismic data; smaller wavelengths increase the vertical resolution. Spatial resolution is defined by the size of the Fresnel radius; the smaller the radius the better the ability to distinguish two laterally separated geological features. For the same frequency, the Fresnel radius of the C-wave is smaller than the corresponding P-wave (Eaton et al., 1991).

The frequency content of the upgoing compressional and converted shear waves are practically the same for all offsets (Figure 9). Moreover, the shear wave velocity is about one-half the compressional wave velocity within the reservoir interval thereby resulting in a shorter wavelength (about one-half the P-wave wavelength) for the C-waves. These factors, coupled with the fact that the converted wave is not affected by the presence of the gas, results in a higher vertical and spatial resolution of the C-wave reflection image. Notice that the resolution and details provided by the C-waves is comparable to that of the 60 Hz P-wave synthetic shown in Figure 5. The top of the Unayzah-A reservoir (at about 13,727 ft true vertical depth, TVD) coincides precisely with a trough in the shear section, but the zero-crossing in the P section. Furthermore, the uppermost porous sandstone reservoir between 13,727–3,777 ft TVD matches the width of the seismic trough in the shear section, allowing for an accurate measurement of the isopach of this interval by measuring the isochron of its corresponding trough.

Lower in the section, the tight portion of the Unayzah interval (the siltstone unit at 13,927 ft TVD) and the base of the Unayzah-C (at 14,327 ft TVD) are both characterized by a seismic peak. In other words, a more accurate stratigraphic interpretation of the Unayzah interval is possible with the shear offset section because of the higher vertical and lateral resolution compared to the compressional offset VSP section. This shows the importance of processing the VSP data for shear wave in addition to the compressional wave.

The shorter offset VSP sections acquired around the well (Figure 16) depict the continuity of the reservoir in the four quadrants. At the well, the reservoir structure is practically flat. In none of these offset sections, including Shot 2, which is in the direction of the major bounding fault to the south, is there a clear indication of small-scale faulting, which may indicate the presence of a reservoir boundary. There is however strong evidence of stratigraphic changes, particularly along Shot 1 and Shot 2. Thus, the boundary could be a facies change (gradation into siltstone or cementation of the reservoir) or a mineralized fracture zone. With respect to the latter, the reservoir has been analyzed to be anisotropic based on the azimuthal variation of its AVO response (Neves et al., 2003) with a major axis of anisotropy in the E-W direction, sub-parallel to the south-bounding fault. It is possible that fractures oriented parallel to the fault exist; however, no horizontal wells at a significant angle with respect to the direction of maximum horizontal stress have been drilled to confirm the presence and density of these suspected fractures.

Also the reservoir appears to be more uniform and presumably better developed along the northwest (shots 3 and 4), which again supports the plan to drill the horizontal well along that direction.

CONCLUSIONS

The higher seismic resolution of offset VSP sections has more clearly resolved a deep gas sandstone reservoir compared to its image that was derived from surface 3-D seismic data. These sections show the trend in the distribution of the porous reservoir, thereby reducing the uncertainty in a proposed horizontal well sidetrack that will be drilled to enhance gas production rates. Furthermore, these offset sections along various azimuths around the well have reduced the number of probable causes for the presence of a nearby reservoir boundary, which was indicated by the well test. Offset VSP most certainly have a role to play in obtaining a clearer image of the reservoir, particularly where careful penetration or geosteering within the reservoir using horizontal or deviated wells is involved.

ACKNOWLEDGEMENTS

The authors wish to thank the Ministry of Petroleum and Minerals, Kingdom of Saudi Arabia, and the Saudi Arabian Oil Company (Saudi Aramco) for permission to publish this paper. They also thank two anonymous reviewers for their useful comments, and GeoArabia designer Nestor Niño Buhay for designing some of the final graphics.

ABOUT THE AUTHORS

John C. Owusu is a Geophysical Specialist with Saudi Aramco in Dhahran, Saudi Arabia. He received a PhD (1991) from Texas A&M University, MSc (1984) in Geophysics from the University of Manitoba, and BSc (1979) in Physics from KNUST, Ghana. John worked for Amoco Production Company in Houston from 1989 to 1999 and Texaco Inc. in Houston from 1999-2001 as a Seismic Interpreter, Processing and Operations Geophysicist. He is currently the Leader of the Borehole Seismic Group in Aramco engaged in various activities including 2-D/3-D VSP imaging and seismic anisotropy/fracture studies for reservoir characterization. He is a member of the SEG and EAGE.

john.owusu@aramco.com

Edgardo L. Nebrija is a Consultant Geophysicist with Saudi Aramco. He has a BSc in Physics from the University of the Philippines and a PhD in Geophysics from the University of Wisconsin at Madison. From 1979 to 1992, he worked for Shell Offshore in New Orleans, Louisiana, in various capacities as Marine Seismic Party Chief, Explorationist, and Reservoir Geophysicist. Since 1992 Ed has been a staff member of the Reservoir Characterization Department of Saudi Aramco and has been responsible for the 3-D seismic interpretation and characterization of several Saudi Arabian oil and gas fields, both onshore and offshore. Among these are Marjan, Abu Safah, Shaybah, as well as Ghawar and some of its satellite fields. Ed is a member of the Dhahran Geoscience Society, EAGE, and SEG. He has presented papers at the GEO Conferences and has published several papers in the AAPG Bulletin and GeoArabia.

edgardo.nebrija@aramco.com