4D Seismic Pilot Successfully Interprets Carbonate Reservoir
You have access to this full article to experience the outstanding content available to SPE members and JPT subscribers.
This paper describes interpretation results of a 4D seismic-monitoring program in a challenging Middle East carbonate reservoir. The program consists of a 4D pilot [oceanbottom cable (OBC)] over a giant field divided into two phases. The authors discuss the difficulties faced by both phases of the pilot, and prove that a reliable 4D signal can be extracted over a Middle East carbonate reservoir.
The time-lapse, or 4D, seismic method consists of repeating seismic survey acquisitions over the production time of a field. Newer vintages are compared with previous ones, and the signal differences are interpreted to characterize effects of the production spatially. Though this technique is mature and has been widely used in clastic reservoir environments, it is still used rarely, in operational terms, to monitor carbonate fields.
An opportunity for a 4D proof-of-concept program in a Middle East marine carbonate environment was identified during a new OBC seismic acquisition in 2013. While the survey was intended for 3D imaging purposes, the operator decided to acquire a 4D pilot over a test area of approximately 25 km2 (Phase 1) while repeating as best as possible the 1994 acquisition design (baseline for this 4D study). The goal was, first, to process this test, and then, if a reliable 4D signal (above noise level) could be established, to propose interpretation in accordance with the field-production mechanisms.
Field-Production and 4D Monitoring Objectives
The principal oil-producing reservoir belongs to the Upper Jurassic and is a mixture of limestones, dolomite, and anhydrite. It contains the primary 4D monitoring target for this pilot. First oil was produced in the 1960s, followed by peripheral water injection in the 1970s and crestal gas injection in the 1990s. Thus, the 1994 4D baseline survey is an OBC seismic acquired after a rich and complex production history.
Four-dimensional monitoring is justified by the fact that many expensive wells are drilled every year and their positioning is not straightforward. In particular, the crucial fluid-movement prediction (gas and water) is challenging and, thus, is considered the main monitoring objective.
Phase 1 Acquisition Status
Repeating baseline acquisition parameters during the monitoring survey is crucial. Any deviation from this principle is likely to generate a 4D noise that will be difficult to remove.
In this pilot, taking advantage of a seismic crew mobilization for a modern wide-azimuth OBC acquisition, the field operator decided to try to repeat, in a small patch, an older OBC survey as a 4D pilot. In fact, this legacy vintage, which has a very different narrow-azimuth design, only was repeated over three swaths. The geometrical repeatability was good, except in the obstruction areas.
Phase 1 4D Processing Challenges
The seismic baseline and monitor data sets were reformatted from archived cartridge data. All shot points and receiver positions were checked, as well as the geometry of acquisition using the analysis of refraction first breaks. Indeed, some subtle errors in geometry that are hardly noticeable in 3D can have a significant effect in 4D.
Overall, 4D processing was complicated because of the high noise level recorded on the data, particularly the incoherent noise related to seabed currents. The water-bottom depth and near-surface characteristics explain the important Scholte waves and guided waves contaminating the data in the area. These were reduced through filtering methods, avoiding any modern adaptative techniques which, by definition, are less 4D-compliant.
The very-shallow-water context (approximately 20 m) characterized by a hard water bottom and strong shallow reflections led to aggressive data pollution by nonrepeatable multiples. These were mitigated not only by PZ summation but also with a specifically designed shallow-water demultiple flow and high-resolution radon filtering.
Because of a relatively small petroelastic response to reservoir modification, 4D in carbonate was believed to have a small probability of success. Indeed, the 4D response is challenging because of the rock type, but several examples have emerged that prove it is highly context-dependent and that there is no a priori killing factor. Therefore, a useful tool to estimate and understand potential 4D response to production is a rock-physics model. Despite issues in the quality of some sonic log measurements, a 4D rock-physics model was built on the basis of key wells available in the 4D pilot area.
4D Seismic-Attribute Generation and Interpretation Strategy
At the end of the 4D processing, after an initial 4D quality-control step and 4D noise analysis, meaningful 4D attributes were extracted through a fully integrated inversion-process chain, including three steps with back loops. The work flow consisted of the following stages:
- Implementation of warping, which is a purely data-driven inversion process recovering time shifts between baseline and monitor data while generating 4D relative velocity-change attributes.
- 4D calibration delivering both relative velocity and density changes for local calibration of anomalies.
- 4D inversion is performed after calibration and from aligned baseline and monitor data to estimate relative impedance changes.
Steps 1 and 3 deliver full-field 4D attributes that can be mapped or interpreted in 3D through the delineation of 3D geobodies. Step 2 allows the precise, quantitative assessment of the uncertainty in Steps 1 and 3 by incorporating strong stratigraphic layering in the inversion process at the well location.
Throughout this work flow, 4D anomalies are compared with the noise level for validation. Some 4D anomalies can be discarded to avoid misinterpretation. Next, 4D composite maps are created and 4D anomalies are interpreted in terms of production phenomena and calibrated with well data. Once calibrated, unexpected phenomena (i.e., heterogeneities not predicted by the reservoir model) are highlighted as potentially corresponding to an added value.
Results and Interpretation
Fig. 1 illustrates impedance-change 4D attributes on a north/south sectional view. The figure shows clear different negative (impedance-decrease) and positive (impedance-increase) 4D anomalies exhibiting different magnitudes, described as follows:
- Light positive (blue) impedance changes in R3, R4, and R5 in the northern part of the section
- Strong positive (blue) impedance changes in R4 and R5 in the southern part of the section
- Light negative (orange) impedance changes in R4 and R5 in the northern part below the light-blue anomaly
- Medium to strong negative (orange/red) impedance changes on the upper part of the section
All anomalies are above the 4D noise level and, with confidence, are not linked to residual multiple issues. Thus, the question is whether this image can be interpreted in terms of production phenomena.
Demonstration, in this paper, focuses on the water-front-movement interpretation because it is a well-understood mechanism and is critical information for the field. Other anomalies can be interpreted with an associated level of confidence as described in the figure.
The first validated anomaly mentioned previously (light positive-impedance changes) is interpreted as the water-saturation change during the 4D time lapse. This is consistent petrophysically; if water replaces oil in the reservoir, then impedance increases. However, 4D interpretation is ambiguous and other production phenomena can explain impedance increase, so the interpretation must be assessed carefully.
The second step is to calibrate this interpretation to well data. A clear correlation can be seen between the 4D anomaly and water cut: All wells that began producing water during the 4D time lapse are superimposed to the 4D anomaly, and the opposite is also true. The interpretation is therefore robust and consistent with production data.
The 4D pilot was successful. One reason is because the monitor acquisition repeated the baseline vintage design as best as possible. However, the new full-field reservoir seismic acquisition is a totally different design (though both are OBC). While the 1994 data set was acquired with a parallel shooting (and, thus, narrow-azimuth), the new one is a source-spread acquisition for a wide-azimuth objective.
For verification, 4D processing was performed over a larger pilot area (Phase 2) to compare it with the Phase 1 4D results (the Phase 2 surface overlaps that of Phase 1). The key purpose of the pilot was to set up the optimal 4D processing work flow to handle these differences successfully.
As expected, the image of water movements is not as accurate as in Phase 1 (with a dedicated repeated 4D acquisition), but a smooth water front was interpreted and critical unexpected heterogeneities were highlighted.
This 4D pilot presents a successful case study. A validated 4D signal is established at reservoir levels using a thorough interpretation work flow and integration of production information. The interpretation of the water-front movement during the 4D time lapse is calibrated to well data and can be judged as robust, proving that a reliable 4D signal can be extracted over such carbonate reservoirs. The Phase 2 pilot has demonstrated that, although results will be less accurate than with a dedicated repeat of the base survey, they will still contain critical information about the water-front movement.
For a limited time, the complete paper SPE 193063 is free to SPE members.
4D Seismic Pilot Successfully Interprets Carbonate Reservoir
01 March 2019
Is the Cloud Mature Enough for High-Performance Computing?
Data volumes are growing at an exponential rate. How can high-performance computing solutions help operators manage these volumes? Will faster, stronger processors and cloud computing solutions be the answer?
Machine-Learning Approach Identifies Wolfcamp Reservoirs
This paper discusses a project with the objective of leveraging prestack and poststack seismic data in order to reconstruct 3D images of thin, discontinuous, oil-filled packstone pay facies of the Upper and Lower Wolfcamp formation.
Imaging Shows What is at Fault in Fracturing Job
Electromagnetic images can show where water flows during a hydraulic fracture. A test in the Anadarko Basin showed a fault there was a bigger hazard than expected.
Don't miss out on the latest technology delivered to your email weekly. Sign up for the JPT newsletter. If you are not logged in, you will receive a confirmation email that you will need to click on to confirm you want to receive the newsletter.
25 March 2019
14 March 2019
20 March 2019