New technology and methodologies promise to reduce costs through repeat seismic surveys.

Field developments often imply large investments, particularly offshore. Billions of dollars frequently are put into production facilities and pipelines. For a field producing 100,000 b/d of oil, a few percentage points increase in recovery factor may imply a large improvement in the return on investment.
To optimize recovery, the reservoir and physical changes that occur during production - in the reservoir's pore fluids, pressure and temperature - must be understood. Hence, geoscientists need to learn how fluids move in the reservoir, how different zones or compartments communicate, which zones are properly drained and where pressure is maintained.
Seismic monitoring, time-lapse, repeated 3-D seismic and 4-D seismic are some of the terms used to describe the monitoring of a reservoir by geophysical methods. The term "baseline survey" often is used to denote the last 3-D survey acquired before production started, and "monitor surveys" for later surveys. 4-D seismic is about detecting changes between these surveys. Changes can be found directly by subtracting two seismic datasets of different vintages, or they can be found indirectly by calculating and comparing a series of attributes. The seismic difference often is called the 4-D signal.
Imagine a situation where well and seismic measurements are recorded continuously. Wellbore instruments measure the fluid flow, pressure, temperature and composition. Permanently installed seismic sources and receivers provide seismic on request. The unprocessed seismic is transferred directly to a facility with very high processing capacity, and the results are available to integrated teams of geoscientists and reservoir engineers a day or two later. This is the ideal 4-D setting - a close-to-real-time picture of the reservoir is always available, and new information can be applied immediately to manage the reservoir. Today's situation is far from this. Many 4-D surveys are done as standalone jobs, and the lapse since the previous seismic survey is normally 2 to 5 years. Still, a series of successful 4-D cases has been documented during the past years, demonstrating that existing seismic technology already has added value to reservoir management. Most of these cases are from the North Sea and Gulf of Mexico, but they also come from onshore fields in North America and Indonesia, where combustion is the drive mechanism.
Planning a 4-D project
Careful feasibility studies comprising seismic and reservoir modeling are necessary to evaluate if 4-D seismic is a good investment.
The questions that must be addressed before starting a 4-D project include:
• Are the physical properties of the reservoir such that production-related changes may be detected by current seismic technology?
• Can seismic acquisition of a new survey be done in such a way that differences caused by the acquisition are significantly less than the seismic differences generated by oil or gas production?
• Can two or more seismic surveys be processed so that the differences caused by production-related changes are significantly less than processing-related artifacts?
• Finally, can the information derived from a 4-D survey be used in a timely manner?
Very good tools are available for feasibility studies. The reservoir simulator predicts the likely changes that will occur in fluid composition, pressure and temperature as a function of production time. These changes can be translated to changes in elastic properties through rock physics modeling. Seismic modeling provides the likely seismic response production will cause.
The first requirement for successful monitoring is that the rock comprising the reservoir must have a nature that makes fluids visible on seismic. The strength of the 4-D signal reflects the change in the reservoir's acoustic impedance caused by production, and this change must be larger than the nonrepeatable noise recorded. Since acoustic impedance is the product of velocity (V) and density (r), the 4-D signal is strongly influenced by the incompressibility (Krock) of the reservoir rock and the production-generated change in incompressibility of the pore fluids
(? Kfluids). This implies that the reservoir rock must be so compressible that there is a strong contribution from the pore fluids in the seismic response. Hence, soft rocks like unconsolidated sands are perfect for 4-D seismic, while incompressible rocks like carbonates are difficult. With respect to the fluids, a reservoir where oil is produced by gas injection is a good candidate. Other good candidates are fields where gas goes out of solution and forms a gas cap, or where gas is replacing water. Heavy oil replaced by water may be quite a challenging case.
Acquiring 4-D
When the 4-D seismic signal is sufficiently strong to be detected, geoscientists can plan how to acquire the monitor 3-D. During the past 10 years, the development within electronics, vessel design and towing technology has enabled use of an increasing number of streamers (up to 16), spaced as closely as 165 ft (50 m). As a result, geoscientists can sample the subsurface very densely and still acquire seismic five to 10 times more efficiently than 10 years ago. Thus, even if the main objective of a new survey is reservoir monitoring, most oil companies also want the highest possible seismic quality. This may prevent acquisition with the same configuration as the baseline survey and affect the repeatability. A possible solution to this problem is to use high-density 3-D (HD3D) acquisition combined with a wider spread than used in the baseline survey. The overlap and densely sampled wavefield achieved by this type of survey allow us to match the base survey's offset and azimuth in the monitor survey. In addition, HD3D technology provides excellent base surveys due to higher fold and lower noise.
A common worry for repeatability is navigation accuracy. However, global positioning systems technology has improved the determination of source and receiver positions greatly, making it possible to find the position of a subsurface reflection point with an accuracy of about 15 ft (5 m).
Repeatability still may be degraded due to:
• variations in equipment performance;
• failures in equipment;
• differences in weather conditions;
• differences in positioning between monitor and baseline survey;
• changes in the acquisition system from one survey to next; and
• changes in acquisition geometry from one survey to next (line direction, source depth, streamer depth, fold, number of streamers or spread width).
Change in line direction may give large differences in the amplitudes of two datasets, increasingly so if the reservoir is dipping. Two surveys shot with large differences in spread width (spread width equals the number of streamers times the distance between them) also may give data that differ significantly in amplitude. Hence, a monitor and baseline survey should have the same shooting direction and not vary much in spread width. HD3D technology offers a flexible solution to this problem.
To a certain extent, the lack of repeatability can be avoided by using permanent equipment on the sea floor. But this also has a few drawbacks:
• it is an expensive initial investment, especially if high fold is wanted and many sensors must be placed (the cost and risk of deploying permanent equipment increase with water depth);
• the response of electronic equipment degrades over time, and eventually the equipment fails;
• the maintenance of buried equipment is impossible or expensive, especially in deep waters; and
• the opportunity to take advantage of the technological development is reduced.
Processing 4-D data
Differences in acquisition configuration will degrade the 4-D signal somewhat. This can be mended partly when processors try to minimize all effects other than those related to oil and gas production.
The selected seismic surveys in a 4-D project should be processed in parallel. The same processing sequence ought to be used for all datasets except for deterministic processing to remove known differences between the surveys. A common way of processing the data is to pick velocities on the dataset that was last acquired and apply these on all vintages of the 4-D seismic.
Because 4-D projects employ seismic of different vintages, it is preferable to have processing algorithms that can be used to regularize the data. Regularization techniques can be used to suppress changes in the data due to changes in nominal acquisition geometry such as variations in number of streamers between surveys as well as due to feathering variations.
Data-dependent processes may introduce effects that are caused by nonrepeatable noise or changes in acquisition configuration. Since changes in reflectivity are important for 4-D interpretation, amplitude-preserving processing must be applied throughout the sequence. Travel time differences between base and monitor surveys may be used to map changes in the reservoir. Travel time differences can be very small, on the order of a few milliseconds; hence, it is vital that all surveys refer to the same zero-time.
Thus, a crucial part of processing is quality control. One or several horizons that are not affected by production changes can be used as reference horizons during the processing to identify acquisition- or processing-induced effects in the data.
Amplitude and phase must be monitored throughout the whole workflow, from loading field data to matching final datasets. Powerful visualization tools can be used to ensure that processing-related differences or other unwanted artifacts are not generated.
The use of 4-D data
New visualization and interpretation technology has made it possible to handle the large datasets that frequently are the products of 4-D projects. Two or three 3-D sets split in different offset ranges (for example, near, mid and far) normally are processed simultaneously, often in slightly different versions. Hence, there are cases where more than 50 3-D datasets have been produced in a 4-D project.
Since 4-D data must be available in due time so the results can be used to manipulate existing wells or place new wells, this sets strong requirements to planning and turning around a 4-D project. A normal marine 3-D survey typically takes 3 months to plan, 2 months to acquire and 4 months to process. Hence, the picture of the reservoir is at least 10 months old. To decrease the production cycle, several steps can be taken. For instance, a fixed processing stream with workflows and velocity fields stored on disk may improve the processing time.
A challenge for the use of 4-D data is to detect changes in partial liquid saturation because there is little sensitivity in compressional-wave velocity when the liquid saturation is in the range of 0% to 80%. Sea-bottom seismic allows recording of shear waves (S-waves). S-wave velocity generally is insensitive to changes in saturation, but S-waves can help differentiate between pressure and fluid changes in the reservoir.
An area where 4-D has shown its value is in mapping where injected gas goes. An early example is Gaz de France's use of shallow onshore reservoirs to store gas during the summer for production in winter and monitor the stock level by seismic. Similarly, Statoil uses seismic monitoring to see where unwanted CO2 moves when it is injected into a shallow sand (Utsira Fm) above the Sleipner field.
The most obvious use of 4-D is to map undrained areas or areas that are not drained as shown by the reservoir simulator. Some successful North Sea examples are the Troll, Gullfaks, Draugen, Magnus and Schiehallion fields.
When a 4-D survey is planned, the strength of the expected 4-D signal should be evaluated against the required repeatability. Many successful cases have proved that existing technology for acquisition of surface seismic gives good results even when different acquisition geometry has been applied on baseline and monitor surveys.
There clearly is a potential to improve seismic technology. Whether permanent sensors will be the best solution or the current technology will sustain remains to be seen. Interesting development work in both areas is ongoing.
New visualization technology is fundamental to use the information in 4-D data.
The cost of a monitor survey is normally small compared to potential benefits. The few numbers that oil companies have published show that 4-D seismic can increase the net revenues from a field by tens of millions of dollars.

Comments