John R. Fanchi, Theodore A. Pagano, Thomas L. DavisExperience from several large projects has improved the oil and gas industry's understanding about what time-lapse seismic monitoring can and cannot do to aid in field development and production management.
Colorado School of Mines
Time-lapse seismic reservoir monitoring is the comparison of 3D seismic surveys at two or more points in time. Commonly known as 4D seismic, time-lapse seismic reservoir monitoring has great potential for increasing our ability to image fluid movement between wells. The oil and gas industry has recognized for some time that the analysis of 4D data volumes can improve the quality of reservoir characterization, identify movement of fluid interfaces, and help operators locate bypassed reserves. The benefits of improved reservoir characterization and increased recovery efficiency do have a price: Anderson et al. have estimated that 4D seismic monitoring can add $1/bbl or more to the cost of producing oil from a field.1
This article assesses the status of 4D seismic monitoring as a practical tool for reservoir development and production. After a brief introduction to the concept of 4D seismic monitoring, several case studies are reviewed to demonstrate the state of the art of 4D seismic monitoring. These case studies show the effective application of 4D seismic monitoring for a range of reservoir management strategies in a variety of rock and fluid systems. For discussion of additional case studies, see Jack.2
What is 4D seismic monitoring?Time-lapse seismic surveying compares one 3D seismic survey with one or more repeat 3D seismic surveys taken in the same geographic location at different times; 4D seismic monitoring is the comparison of changes in 3D seismic surveys as a function of the fourth dimension, time. By comparing the differences in measurements of properties such as travel times, reflection amplitudes, and seismic velocities, changes in the elasticity of the subsurface can be monitored over time.
There are two principal elastic parameters that affect seismic waves: the bulk modulus and the shear (or rigidity) modulus. The bulk modulus is related to rock and fluid compressibility. Shear (S) waves are affected by bulk density and shear modulus, while compressional (P) waves are affected by bulk density and both bulk and shear moduli. The combination of bulk and shear modu* used in the calculation of P-wave velocities is called stiffness and is a measure of the rock-frame stiffness and pore-fluid stiffness.
Reservoir elasticity is affected by lithology, fluid content, and variations in pore pressure. Seismic velocity (V), attenuation (Q), and reflectivity measurements contain information on the fluid distribution in the reservoir. For example, the ratio of compressional to shear velocity (VP/VS) is dependent on the bulk modulus and shear modulus of the rock, which are related to porosity and fluid content in the pore space. Seismic monitoring of changes in reservoir elasticity can be linked to properties associated with the movement of fluids in the reservoir. This link yields information that can be used to improve the validity of fluid-flow models and the reservoir-management decisions that rely on flow-model forecasts.
Technological innovations have made it possible to probe the elastic properties of a reservoir by recording P and S-waves as they pass through the reservoir. The ability to monitor both P and S-waves is referred to as full-vector wavefield, or multicomponent, imaging.3 Time-lapse, multicomponent seismology is a tool for volume resolution-the ability to sense changes in the bulk rock-fluid properties of the reservoir with time.
An example of a time-lapse, multicomponent survey is the 4D, three-component (3C) seismic survey. This survey records one vertical and two horizontal velocity components. The recording procedure is similar to that of earthquake seismologists and facilitates the combined recording of P and S-waves. Comparisons of travel time or seismic velocity measurements, amplitudes, and frequencies of P and S-waves enable the discrimination of rock and fluid properties and their changes over time.
Multicomponent surveys can take advantage of seismic anisotropy. Seismic anisotropy is a measure of the fine scale structure in the reservoir. An anisotropic reservoir exhibits differences in properties as a function of spatial orientation. Horizontal permeability, for example, will be greater in one direction than another. In an anisotropic medium, a shear wave splits into two orthogonally polarized components (S1 and S2). The S1 wave is faster, and its velocity and attenuation are affected by lithology, porosity, and pore saturants. By contrast, the S2 wave is slower, and its velocity and attenuation are affected by reservoir features such as fractures. The different dependencies of the S1 and S2 waves provide information for determining dynamic reservoir properties such as permeability, porosity, and fluid saturations. Multicomponent seismic studies thus have been especially useful in the characterization of reservoir rock properties, including lithology, porosity and fractures.
Guidelines for applying 4D seismicWang has identified several criteria that can be used to identify good candidate reservoirs for time-lapse seismic reservoir monitoring.4 One criterion is the magnitude of bulk and shear moduli. Reservoirs with relatively small moduli have weak elastic frames and are good candidates for time-lapse seismic monitoring. Examples include reservoirs with unconsolidated sands or fractured reservoirs. Wang also observed that reservoirs experiencing large compressibility changes in either the rock or pore fluids can exhibit a significant seismic response over time. Examples include reservoirs in which the gas phase is either appearing or disappearing.
Integrated flow model studies have found that 4D seismic should be most effective in regions where the gas phase is appearing or disappearing.5 This observation is in agreement with Wang's criteria and has been substantiated in the field. For example, Kelamis et al. noted the importance of similar gas saturation behavior in their study of a clastic oil reservoir with a large gas cap in the Persian Gulf.6 The contrast between fluid compressibilities is greatest when one survey images only liquid, while a subsequent survey images liquid and gas.
Very small amounts of gas-as low as 1% saturation-are sufficient to change the total compressibility of a reservoir system by an order of magnitude or more. The change in fluid compressibility with time can generate observable differences in seismic response when comparing one survey with another. The differences in seismic response appear in measurements of attributes like compressional and shear velocities. The ratio of compressional to shear velocities is an important parameter for providing information that can be used to improve reservoir characterization and fluid-flow models.
The necessity to compare two surveys that have been conducted at different points in time over the same region leads to the issue of survey repeatability.7 The signal in a 4D seismic survey is the magnitude of the change in acoustic response of the reservoir between two surveys taken over the same region at different times. Detection of the signal requires that the differences between time-lapse surveys should be due to actual reservoir changes and not to differences in data acquisition, processing, or interpretation.
Fluid-phase behavior, such as the appearance or disappearance of a free gas saturation during the life of the field, can be used to schedule sequential 3D seismic surveys.5 8 From a reservoir-management perspective, the scheduling of time-lapse seismic surveys should optimize the acquisition of reservoir-engineering information.
Anderson et al. have observed that 4D seismic is most effective in offshore fields where high quality 3D seismic surveys exist.1 They also noted that 4D seismic works better in soft, unconsolidated sands rather than in hard, carbonate reservoirs. The Vacuum field case study presented below shows that 4D, multicomponent seismology can be applied effectively in carbonates.
A common factor in most screening criteria is the identification of significant changes to properties that directly influence seismic response. For example, if a reservoir is subjected to large pressure or temperature changes, the resulting change in petrophysical properties can lead to an observable change in seismic response. The case studies described below substantiate the screening criteria outlined here. They demonstrate that 4D seismic can be an effective tool for reservoir management. It is important to keep in mind, however, the caution pointed out by Lumley and Behrens: "...This new technology is not a panacea but rather an exciting emerging technology that requires very careful analysis to be useful."9
Case studiesMany of the current time-lapse seismic reservoir-monitoring projects are in their development stages, and 4D seismic influences on reservoir management have not yet been fully realized. The following case studies illustrate the technical success of 4D seismic monitoring and the subsequent impact on reservoir-management decisions.
Gullfaks fieldThe Gullfaks field is operated by Statoil and is located approximately 50 miles east of the North Shetland Platform in the Norwegian North Sea. Production began in 1986 from Jurassic, offshore marine to fluvial reservoir sands. In 1995, a 4D seismic study was performed over the Tarbert formation, the upper-most producing unit.10 The Tarbert formation is a clean, fluvial-deltaic sandstone with an approximate thickness of 180 ft and an average porosity of 34%. The Tarbert reservoir is currently being waterflooded for pressure maintenance.11
Acoustic reservoir signature from the 1995 3D seismic survey was compared to a baseline survey taken in 1985. Differences in the acoustic reservoir signature enabled the operator to determine saturation changes and potentially bypassed pay. Although the appearance or disappearance of gas is often more observable,4 5 the replacement of oil by water in the Tarbert reservoir resulted in a detectable 9% change in seismic character.
Gullfaks field is dominated by fault-block geology. Thus, a thorough understanding of fault control on oil migration and reservoir compartmentalization is needed to optimize field development. With the aid of 4D seismic, a new method for fault-seal analysis was tested. Saturation changes observed in the 4D seismic were coupled with the fault network. If the 4D data set displayed a saturation change across a fault that was originally thought to be sealing, the transmissibility across the fault had to be redefined to allow flow across the fault. Fig. 1 [112,016 bytes] is a map of seismic saturation changes and a display of how these changes were used to define the sealing character of the Gullfaks fault network. Analysis of the 4D data set led to a better understanding of reservoir compartmentalization and subsequent reservoir development.
As a result of the reservoir-monitoring project at Gullfaks, two fault-block compartments with initial oil saturation were identified. A long horizontal well was completed through the two fault-block compartments and through a third region with slightly diminished oil saturation. The new well confirmed the oil saturation distribution predicted by the 4D seismic project. A second well, previously abandoned, was sidetracked and recompleted in areas where 4D predicted no oil drainage. Upon completion, this well produced 6,000 b/d of oil.
Eugene Island 330/338Eugene Island Block 330 in the Gulf of Mexico is operated by PennzEnergy, and Eugene Island 338 is operated by Texaco. The fields are adjacent and located approximately 50 miles south of the Louisiana coast. The LF sand reservoir of Eugene Island 330/338 has been the subject of a 4D seismic reservoir study aimed at identifying bypassed pay.12 13 The 240-acre LF sandstone reservoir has an average porosity of 27%, permeability of 500 md, and an average water saturation of 35%. The LF sand occurs at a depth of 6,900-7,000 ft subsea. Cumulative production from the LF reservoir was 1.2 million bbl oil equivalent during the period 1974-88.
A 3D data set of the LF sand reservoir was collected in 1988. It was normalized and compared to an earlier data set collected in 1985. Differential pressure (lithostatic minus reservoir pressure) effects on seismic response appear as hydrocarbons are depleted, but effects due to changes in fluid phases and fluid properties can dominate 4D seismic differences.4 5 This was substantiated in the LF sandstone. The appearance of gas due to pressure depletion changed the reservoir acoustic response and enabled the operator to identify where gas-oil ratios had increased. A change in acoustic response along the reservoir boundary made it possible to identify the updip migration of the reservoir water-oil contact.
The operator identified bypassed pay by recognizing regions in the 3D data set that remained constant between 1985 and 1988. This led to a 1994 completion of a 1,200 ft horizontal well in the suspected zone. A seismic profile of the targeted zone and the horizontal well is displayed in Fig. 2 [117,796 bytes]. Initial production from this well exceeded 1,500 b/d and by 1996 had added a total of 1 million bbl to the cumulative oil production from the field.
A similar technique was used on the IC sand reservoir of Eugene Island 330 to identify bypassed oil. It led to the completion of a successful well in 1991.
Additional 3D seismic data sets were gathered at Eugene Island 330/338 in 1992 and 1994. Each is being coupled with the two earlier data sets and currently serves as the basis for multiple, ongoing time-lapse studies. Eugene Island 330/338 demonstrates that a baseline 3D seismic survey (a survey taken before the commencement of production) is not necessary.
South Timbalier 295The South Timbalier 295 field is located in the Gulf of Mexico 120 miles south of New Orleans. Limited production began in 1989. 4D seismic monitoring was implemented on the K8/K16 reservoir, which is a late Pliocene turbidite channel system with complex porosity and permeability connectivity.14 The K8/K16 reservoir is the uppermost producing horizon at a depth of 10,200 ft subsea.
A baseline 3D seismic survey was conducted in 1988. In 1994, a second 3D seismic survey was completed. By this time, the K8/K16 reservoir of South Timbalier 295 had produced 4 million bbl of oil and 15 bcf of gas, about half of its reserves. During primary recovery, the K8/K16 reservoir dropped below bubble point pressure, and a considerable volume of gas began to appear. As a result, differences between the two 3D seismic surveys were easily observed and mapped. Probable drainage pathways were determined through an iterative process that involved matching actual 3D seismic responses to reservoir simulation-predicted 3D seismic responses by changing the flow pattern in the reservoir simulation.
The seismic response remained the same in a region downdip from the producers during the period from 1988 to 1994. In conjunction with the reservoir simulation, it was determined that suboptimal oil migration occurred in this area. The poor oil migration suggested the need to repressurize the reservoir downdip so that this bypassed oil could be recovered.
A pressure-support program was initiated in 1997, and a water injection well was drilled downdip of the original oil-water contact. The integration of a comprehensive 4D seismic data set and a thorough reservoir-simulation study made it possible for the field operator to pinpoint an injection location that would best serve K8/K16 reservoir pressure maintenance. If the hydrocarbon-evolved gas is successfully forced back into solution, the disappearance of the free gas phase should be observable as a seismic response. A subsequent 3D seismic survey should detect the results of reservoir repressurization and associated fluid movement.
Vacuum fieldVacuum field is located 20 miles west of Hobbs, N.M., on the northwestern shelf of the Permian basin. Discovered in 1929, it produces oil from Permian age, San Andres carbonates at an average depth of 4,500 ft. The field is part of a large anticlinal structure formed by drape of sediments over a basement-controlled fault block. The field has produced 500 million bbl of oil from 678 producing wells. A waterflood is maintaining reservoir pressure above the bubble-point pressure to prevent the development of a free gas phase. A carbon dioxide (CO2) huff-and-puff was initiated in well CVU97 in late 1995.
The first ever 4D, 3C seismic survey was acquired in 1995 by the Colorado School of Mines Reservoir Characterization Project over the Texaco-operated CO2 huff-and-puff in the Central Vacuum Unit.15-17 The 4D, 3C seismic survey consisted of two 3D, 3C surveys. Each 3D, 3C survey was acquired with three source components recorded by three-component receivers. These surveys were staged approximately 8 weeks apart. During this period, approximately 50 MMscf of CO2 was injected into well CVU97 and allowed to soak.
Oil swelling and viscosity changes associated with the mixing of oil and CO2 affected the elastic properties of the pore-fracture system. Shear-wave data showed a 4D anomaly coincident with the location of high CO2 concentrations. The anomaly is shown south of well CVU-97 near well CVU-200 in Fig. 3 [66,151 bytes]. The high CO2 concentration away from the injection well led to the interpretation that this was a high-permeability zone.
The 4D, 3C identification of permeability distribution and heterogeneity in Vacuum field has enabled the operator to identify more attractive areas for CO2 flood implementation. This led to the drilling of a dual-lateral well that has been one of the most productive wells in the field. In addition, the 4D, 3C program is successfully being used to monitor the CO2 flood process.
Duri fieldThe Duri field is operated by P.T. Caltex Pacific Indonesia.18 It is located in the eastern coastal plain of Central Sumatra. The field produces around 300,000 b/d of high-viscosity oil from Miocene sandstones at depths of 200-900 ft. Porosity ranges from 30% to 38%, and permeability is typically greater than 1,500 md. In 1985, continuous steamflooding was employed to reduce the oil viscosity and increase recovery by approximately 50%.
Once steam communication has been established between an injection and production well, resistance to steam flow between the two wells falls. This results in channeling between the two wells and prohibits efficient vertical and horizontal sweep, thereby lowering ultimate oil recovery.
Duri field displays high porosity and a low dry-bulk modulus due to its shallow, unconsolidated nature. In addition to steam-induced thermal effects, the high porosity and low dry-bulk modulus are favorable for 4D seismic applicability.4 In 1996, a 4D seismic pilot program was initiated at Duri field to monitor horizontal and vertical steam distribution in the reservoir. After only 2 months of steam injection, gas that developed due to pressure depletion was forced back into solution by the increase in reservoir pressure from steamflooding. As explained by Fanchi5 and Wang,4 the effect of disappearing gas on seismic character created observable differences in 4D seismic signature.
The Duri field operators have found that conventional steam monitoring techniques, such as temperature and tracer surveys, often provide incomplete data that may lead to erroneous steam-management decisions. However, when conventional techniques are coupled with 4D seismic, a more thorough understanding of steam distribution and channeling is established. This information has been used to optimize remedial workover design, well constraints, and production strategies to maximize the effect of steamflooding. The 4D pilot program at Duri field was so successful that time-lapse seismic surveys have been shot in five additional areas of varying steamflood maturity.19 New surveys will be designed and implemented as undeveloped areas are placed on steam injection.
State of the artThe case studies support the belief that time-lapse seismic monitoring can improve the quality of reservoir characterization, identify movement of fluid interfaces, and help locate bypassed reserves in systems where changes in seismic response over time are detectable. Routine integration of 4D seismic monitoring in the reservoir management process continues to be hampered by gaps in technology that slow the transfer of data from one discipline (geophysics) to another (reservoir engineering).
Tobias has pointed out that the development and implementation of 3D, model-centric methods based upon computer-generated 3D representations of the earth are changing the way the industry characterizes reservoirs.20 The integration of data from different disciplines will be further enhanced by the use of flow models that include petrophysical calculations.5 Integrated flow models make it possible to work directly with seismically generated data at any point during the life of the reservoir. They simplify the data-transfer process between disciplines, enhance consensus-building, and provide performance predictions in a format that is familiar to reservoir managers.
Integrated flow models are a natural extension of the model-centric methods that are now being used in the geosciences. The extension of petrophysical algorithms to include data generated by multicomponent, 3D seismic surveys should add even more data for accurate reservoir characterization and optimized reservoir management.
- Anderson, R.N., Guerin, G., He, W., Boulanger, A., Mello, U., Watson, T.J., "4-D Seismic Reservoir Simulation in a South Timbalier 295 Turbidite Reservoir," The Leading Edge, October 1998, pp. 1416-1418.
- Jack, I., Time-Lapse Seismic in Reservoir Management, 1998 Distinguished Instructor Short Course, Society of Exploration Geophysicists.
- "Seismic Technology: Evolution of a Vital Tool for Reservoir Engineers," Journal of Petroleum Technology, February 1999, pp. 22-28.
- Wang, Z-J., "Feasibility of Time-Lapse Seismic Reservoir Monitoring: the Physical Basis," The Leading Edge, September 1997, pp. 1327-1329.
- Fanchi, J.R., "Flow models time 4D seismic surveys," OGJ, Mar. 15, 1999, pp. 46-51.
- Kelamis, P.G., Uden, R.C., and Dunderdale, I., "4D Seismic Aspects of Reservoir Management," Paper No. OTC 8293, Offshore Technology Conference, Houston, 1997.
- Ross, C.P., and Altan, S., "Time-Lapse Seismic Monitoring: Repeatability Processing Tests," Paper No. OTC 8311, Offshore Technology Conference, Houston, 1997.
- Fanchi, J.R., Principles of Applied Reservoir Simulation, Gulf Publishing, 1997.
- Lumley, D., and Behrens, R., "Practical Issues for 4D Reservoir Modeling," Journal of Petroleum Technology, September 1997, pp. 998-999.
- Soenneland, L., Signer, C., Veire, H.H., Johansen, R.L., and Pedersen, L., "Four-Dimensional Seismic on Gullfaks," Paper No. OTC 8290, Offshore Technology Conference, Houston, 1997.
- Veire, H., Reymond, S., Signer, C., Tenneboe, P.O., and Soenneland, L., "New Methods Boost 4D seismic Role in Reservoir Management," OGJ, Sept. 14, 1998, pp. 41-46.
- Anderson, R.N., Boulanger, J., He, W., Sun, Y.F., Xu, L., Sibley, D., Austin, J., Woodhams, R., Andre, R., and Rinehart, K., "4D Seismic Helps Track Drainage, Pressure Compartmentalization," OGJ, Mar. 27, 1995, pp. 55-58.
- Anderson, R.N., Boulanger, J., W. He, Sun, Y.F., Xu, L., Sibley, D., Austin, J., Woodhams, R., Andre, R., and Rinehart, K., "Method Described for Using 4D Seismic to Track Reservoir Fluid Movement," OGJ, Apr. 3, 1995, pp. 70-74.
- Anderson, R.N., Boulanger, A., Amaefule, J., Forrest, M., Howell III, J.I., Nelson Jr., H.R., and Rumann, H.A., "Quantitative Tools Link Portfolio Management with Use of Technology," OGJ, Nov. 30, 1998, pp. 48-54.
- Davis, T.L., "Reservoir Characterization Project - Phase VII: Dynamic Reservoir Characterization," in Time-Lapse Seismic in Reservoir Management by Ian Jack, 1998 Distinguished Instructor Short Course, Tulsa, Society of Exploration Geophysicists.
- Benson, R.D. and Davis, T.L., "4-D, 3-C Seismic Put to Task in CO2 Huff-And-Puff Pilot," American Oil & Gas Reporter, February 1999, pp. 87-94.
- Talley, D.J., Davis, T.L., Benson, R.D., and Roche, S.L., "Dynamic Reservoir Characterization of Vacuum Field," The Leading Edge, October 1998, pp. 1396-1402.
- Jenkins, S.D., Bee, M.F., Lyle, J.H., and Murhantoro, E., "4-D Seismic: Powerful New Technology For Monitoring Steam Movement in Duri field - Central Sumatra," presented at the 23rd Ann. Indonesian Petroleum Assoc. Mtg., 1994.
- Waite, M.W., and Rusdinadar, S., "Seismic Monitoring of the Duri Steamflood: Application to Reservoir Management," The Leading Edge, September 1997, pp. 1275-1278.
- Tobias, S., "From G&G to S&S: Watershed Changes in Exploration-Development Work Flow," OGJ, Nov. 30, 1998, pp. 38-47.
BibliographyHe, W., Anderson, R.N., Xu, L., Boulanger, A., Meadow, B., and Neal, R., "4D Seismic Monitoring Grows as Production Tool," OGJ, May 20, 1996, pp. 41-46.
Landro, M., and Pedersen, L., "4D Seismic Monitoring at Gullfaks: Advances and Challenges to Date," Extended Abstracts, EAGE 59th Conference and Technical Exhibition, Geneva, 1997.
Waite, M.W., Rusdinadar, S., Rusdibiyo, A.V., Susanto, T., Primadi, H., and Satriana, D., "Application of Seismic Monitoring to Manage an Early-Stage Steamflood," Paper No. SPE 37564, International Operations & Heavy Oil Symposium, Bakersfield, Calif., 1997.
Waite, M.W., Rusdinadar, S., Jenkins, S.D., and Bee, M.F., "Using 4D Seismic to Monitor and Improve Steamflood Efficiency," World Oil, November 1998, pp. 49-56.
Williams, P., "Time-Lapse Seismic," Oil and Gas Investor, May 1998, pp. 30-38.
John R. Fanchi is a professor in the Petroleum Engineering Department at the Colorado School of Mines. Prior to joining CSM in 1998, he worked in the technology centers of three major oil companies. His oil and gas industry responsibilities have revolved around reservoir modeling, both in the areas of simulator development and practical reservoir management applications. His publications include simulation software, three books, and numerous articles. He has a PhD in physics from the University of Houston.
Theodore A. Pagano is a graduate student in the Petroleum Engineering Department at the Colorado School of Mines. He is pursuing a master of engineering degree with a specialization in reservoir engineering. In 1997, he received a BSc in geological sciences from the School of Engineering at the University of Notre Dame.
Thomas L. Davis is a professor in the Geophysics Department at the Colorado School of Mines, where he is codirector of the Reservoir Characterization Project. Except for a 3-year sojourn at the University of Calgary, he has been a faculty member at CSM since 1974. He was second vice- president of the Society of Exploration Geophysicists in 1989-90 and was SEG's spring 1995 distinguished lecturer. Davis holds a bachelor's degree in engineering from the University of Saskatchewan, a master's degree in geophysics from the University of Calgary, and a PhD in geophysical engineering from CSM.
Copyright 1999 Oil & Gas Journal. All Rights Reserved.