The world's first 4D surface-gravity surveillance of a waterflood has been implemented at Prudhoe Bay, Alaska. This monitoring technique is an essential component of the surveillance program for the Gas Cap Water Injection (GCWI) project. A major factor in the approval process for the waterflood was to show that we could monitor water movement economically where a very limited number of wells penetrated the waterflood area. The drilling of numerous surveillence wells to monitor water movement adequately would have been cost-prohibitive. Field surveys now show conclusively that density changes associated with water replacing gas are being detected readily with high-resolution surface-gravity measurements. The gravity methods used to monitor the waterflood include time-lapse (4D) measurement of surface gravity over the reservoir followed by inversion of the 4D signal for mass-balance calculation and flood-front detection.
This paper will focus on field results of time-lapse surface-gravity surveys. Differences in the gravity field over time reflect changes in the reservoir-fluid density. The inversion procedure was formulated and coded to allow for various constraints on model parameters such as density, total mass, and moment of inertia. The gravity survey was designed to permit the inversion for reservoir mass distribution, with resolution on the order of hundreds of meters in the presence of uncorrelated noise of reasonable magnitude (12-µGal standard deviation).
Time-differenced gravity-survey results clearly show an increase in surface gravity that is a result of the injected-water mass. Density-change maps deduced from measured gravity change show that water movement is reasonably similar to the reservoir simulations and to the water detected in observation wells. The overall ultimate gravity signal is predicted to increase to approximately 250 µGal, ultimately resulting in accurate maps of the water movement.
This paper discusses the use of surface-gravity measurements as a reservoir-surveillance technique, specifically to monitor the gas-cap water injection in the Prudhoe Bay oil-field. The fundamental problem of monitoring the gas-cap water-injection project is the small number of monitoring wells and the lack of producing wells in the gas-cap area of Prudhoe Bay. Distances between some monitoring wells are greater than 10,000 ft (3048 m), and years will be required for the injected water to propagate to these distances. Too few wells exist to monitor the water movement adequately with conventional downhole-logging techniques. To address this problem, the Prudhoe Bay surveillance program uses a combination of conventional downhole logging in existing wells and 4D surface-gravity monitoring. The major monitoring concern with the waterflood is ensuring that water added in the gas cap does not flow downdip prematurely into the oil-producing portions of the field in which it could interfere with a highly efficient gravity-drainage mechanism.
Surface-gravity instruments measure the Earth's gravitational field at a specific point or station. With an array of these measurements, local structural traps, stratigraphic traps, or fluid movement can be identified, provided that there is a sufficient density contrast between the feature of interest and the surrounding rock. The surface-gravity technique can be applied to any field, depending upon the reservoir thickness, size, depth of burial, porosity, and the density contrast between the fluids. The surface-gravity technique requires that several time-lapse gravity surveys be made over the life of the field. The first survey should be performed before any change in the fluid volumes to obtain baseline data. The baseline survey can be subtracted from future gravity surveys to obtain the gravity anomaly associated with the change in fluid volumes. The technique assumes that any other time-dependent gravity changes can be accounted for either by measurement or by modeling and that noise caused by the measurement process and unmodeled (near-surface) density changes have tolerable characteristics.
The GCWI project at Prudhoe Bay produces an increasing positive gravity anomaly because of the added mass over time caused by water replacing gas in the pore space. Density variations from local geology and topography that do not change with time are effectively canceled when gravity data from different time epochs are differenced. The time-differenced, or 4D, gravity signal is then inverted to obtain a reservoir-density-change model. This change in reservoir density represents the waterflood progression.
The gravity signal of interest is the observed gravity corrected for instrument drift, solid Earth and ocean tides, and polar motion and atmospheric-pressure changes. Topographic changes in the vicinity of the stations are likely on permafrost and bay ice over periods of years (1-cm elevation equals 3-µGal gravity) so that the elevation difference contributes free air and Bouguer (i.e., mass) correction terms to the gravity difference. The 4D survey is possible only through the use of high-accuracy global positioning system (GPS) and a µGal-precision gravimeter. The Micro-g Lacoste A-10 absolute-gravity meter is used to measure the acceleration of a falling mass in a self-contained experiment, which can be tied rigorously to standards of length and time. Each gravity observation is the result of approximately 1,000 repetitions of this simple experiment. Each station observation is independent of all other stations and instrument calibrations, unlike conventional relative-gravity-meter surveys. Survey parameters must be duplicated as closely as possible from year to year in order to minimize survey error. The stability of permanent-station monuments is poor in permafrost environments and is impractical on bay ice; therefore, station recovery must be accomplished by navigation using real-time submeter GPS (Parkinson and Enge 1996). Relocating stations to within 1 m permits neglecting a latitude correction (the gravitational latitude effect at 70° north is less than 1 µGal for northing differences of less than 1 m) and to avoiding an interpolation operation before time-differencing the gravity data. The centimeter-precision location can be obtained by using real-time kinematic or post-processed, carrier-phase ambiguity resolution ("rapid static") GPS methods. Multiple base stations can be used in network-averaged solutions for increased accuracy. Both the GPS and gravity data are usually obtained within a 20-minutes-long station occupation.
Extensive gravity modeling has been performed using reservoir simulations that includes simulated noise with various magnitudes and characteristics to determine tolerable noise levels to ensure a successful monitoring program at Prudhoe Bay (Hare et al. 1999). In addition to this, four field test surveys have been completed since 1994 to verify the accuracy of both the time-lapse gravity data and the GPS gravity-station-location data. It has been established that time-lapse gravity data can be obtained, at a sufficiently low noise levels, to ensure that the injection water can be monitored properly. The Prudhoe Bay reservoir is buried at approximately 8,200 ft (2500 m) in the gas-cap region of the field and has a maximum gas-water-density contrast of 0.12 g/cm3. Even at this depth, Prudhoe Bay is a good candidate for the surface-gravity monitoring technique because of a reservoir thickness as great as several hundred feet and a high porosity.
Previous publications have included a description of the reservoir-simulation models and the resultant density contrast; inversion of the simulated surface-gravity anomaly to determine the degree to which water-movement progression can be monitored; and a review of the four gravity- and GPS-data-acquisition field tests that were performed on the bay ice and tundra in March of 1994, 1997, 2000, and 2001 (Hare et al. 1999; Brady et al. 1995a, 1995b, 2002a, 2002b, 2005).
This paper will discuss the results of the two baseline surveys performed in the winters of 2002 and 2003 and the results of the first time-lapse gravity survey of water movement after water injection began. This 2005 survey clearly shows that the gravity anomaly created by injecting 340 million bbl of water can be detected and mapped.