This research aims to evaluate the underground CO2 storage site at Sleipner in the North Sea, Norway, by applying a newly proposed rock physics model to extract saturation from seismic data. Our focus is the CO2 storage reservoir unit of the Utsira Formation (a shallow saline formation where CO2 from produced natural gas is injected and stored) to validate the new rock physics model. The saturation information was used to determine the geometry of the CO2 plume within the different layers of the Utsira Formation. Variance attribute-a seismic attribute that measures the similarity of waveforms or adjacent traces over a given window was used to understand the vertical CO2 migration through thin shale layers.
To conduct the study, open-access data from the CO2 DataShare portal, specifically the Sleipner dataset, is utilized. This dataset includes wireline log data from two wells and seismic cubes from the baseline survey in 1994 to various time-lapse surveys until 2010. By performing the 2010 time-lapse (4D) survey, key parameters such as Acoustic Impedance (Zp) and the P-to-S-wave velocity ratio (Vp/Vs) were obtained in the form of cubes. Those inverted cubes were then analyzed using rock physics model calibrated with the seismic data within the CO2 plume area to determine saturation in different layers of the siliciclastic Utsira Formation.
The variance attribute shows numerous rounded to oval-shaped features which acted as paleo-natural gas conduits related to sediment compaction. The suggested rock physics model for monitoring CO2 storage demonstrates its capability to delineate the plume area and estimate saturation through the time-lapse (4D) seismic survey. The findings indicate that among the nine Utsira sand reservoir layers, high CO2 saturation is observed in the upper layers (4-9), while layers 1-3 show low CO2 saturation with a sparse distribution. This difference could be attributed to a preferred vertical migration of CO2 due to buoyancy. The other factors attributed to the reflectors dimming, such as transmission losses through low-velocity CO2 layers and inelastic attenuation, could have a minor effect on the imaging as CO2 is still in a dense phase, and lab studies indicate a patchy velocity-saturation relation.
This study uses a technique that utilizes an innovative rock physics model incorporating Acoustic Impedance (Zp) and the P-to-S-wave velocity ratio (Vp/Vs) derived from inverted seismic data. By employing this model, even with limited well-log control, a deterministic solution is offered for evaluating subsurface CO2 storage sites and monitoring CO2 plume evolution. This contribution is significant in supporting the success of carbon capture and storage initiatives.
Carbon capture and storage (CCS) has emerged as a crucial technology for mitigating greenhouse gas emissions and combating climate change. One of the critical challenges in CCS projects is the accurate evaluation and monitoring of underground CO2 storage sites. The Sleipner site has pioneered successful CO2 storage since the late 1990s. The Sleipner field is a natural gas condensate field in the North Sea, about 250 kilometres west of Stavanger, Norway (Fig. 1a). To ensure the effectiveness and long-term stability of any CO2 storage project like Sleipner, it is essential to accurately assess the saturation, distribution, and migration of the CO2 plume within the subsurface formations.
Injecting CO2 separated from the natural gas produced at the Sleipner field in the North Sea began in 1996. The injection has been ongoing at a constant rate of approximately 0.8 Mt/a, and by 2020, approximately 18.5 million tonnes of CO2 had been stored in the Utsira Formation (Williams and Chadwick, 2021). The CO2 is stored in the Utsira Formation, a regional saline aquifer consisting mainly of loose, porous sand with high porosity and permeability and typically exceeding 200 m in thickness.”. Logs from well 15/9-13 in the Sleipner area indicate the presence of several thin shale layers within the reservoir, approximately 1-2 meters thick (Fig. 1b). These shale layers play a crucial role in controlling the spread of the CO2 plume. The sandstone-shale layers are not obvious on the baseline 3D seismic survey (Fig 1c); however, the layers became prominent over time on the subsequent time-lapse (4D) surveys as the CO2 has accumulated in the reservoir as a multi-tiered plume with individual layers of CO2 trapped beneath the thin shale layer (Fig. 1d). The mudstones act as barriers, impeding the upward flow of CO2. Nevertheless, it is apparent that the CO2 migrated through the mudstones via high permeability vertical pathways, although the specific mechanisms by which the mudstones are bypassed are currently unknown (Chadwick et al., 2004; Williams and Chadwick, 2021).
In recent years, seismic data analysis techniques have become valuable tools for evaluating CO2 storage sites. Seismic data provide useful information about the subsurface structure and fluid saturation, which are crucial for understanding the behavior of the CO2 plume. Traditional methods rely on interpreting seismic data based on empirical relationships, often limited by uncertainties and assumptions. However, advancements in rock physics models offer a promising approach to extract quantitative information about CO2 saturation from seismic data.
In the context of the Sleipner CO2 storage project, previous studies have highlighted the importance of seismic data analysis in assessing CO2 storage sites and monitoring CO2 plume migration. Several researchers focussed on seismic attributes and quantitative seismic interpretation and 4D monitoring analysis (Chadwick et al., 2010; Bitrus et al., 2016; Furre et al., 2017). CO2 migration modeling and simulations (Boait et al., 2011; Karstens et al., 2017) have been carried out to match results with the 4D seismic. Full wave inversion (FWI) (Queißer and Singh, 2013; Raknes et al., 2015; Yan et al., 2019), poststack and prestack inversions (Clochard et al., 2009; Chadwick et al., 2010; Delepine et al., 2011; Ghosh et al., 2015) have been performed to extract the CO2 plume extent from seismic data. Their results demonstrated the potential of seismic data to provide valuable insights into CO2 migration patterns and storage efficiency. Rock Physics models from laboratory experiments and their usage to convert a reservoir model into a seismic velocity model of the CO2 plume have been carried out (Williams and Chadwick, 2021), which emphasized integrating rock physics models with seismic information for estimating CO2 saturation. However, the applicability and effectiveness of these models depend upon the P-wave velocity response.
Building upon the existing literature, this research aims to assess the saturation in the Sleipner CO2 reservoir using a newly proposed rock physics model (Fawad and Mondol, 2022a, 2022b) that utilizes the /S- wave velocity information. The study focuses on the Utsira Formation, a shallow saline formation where CO2 injection and storage occur. By using seismic data from the 2010 time-lapse survey and well-log data from the Sleipner dataset, the suggested rock physics model incorporating Acoustic Impedance (Zp) and the P-to-S-wave velocity ratio (Vp/Vs) is applied to estimate CO2 saturation in different layers of the Utsira Formation sandstone. The analysis also utilizes the variance attribute, a seismic attribute that measures waveform similarity, on the baseline survey (94p07) to investigate the vertical migration of CO2 through the shale layers. The findings shed light on the geometry and saturation distribution within the Utsira Formation and provide insight into the behavior of the CO2 plume over time. By employing this innovative rock physics model and seismic data analysis, this research aims to contribute to the effective evaluation and monitoring of subsurface CO2 storage sites. The findings from this study will not only enhance the understanding of CO2 plume behavior in the Sleipner area but also provide valuable insights for future CCS initiatives. The successful implementation of carbon capture and storage is crucial in achieving global net zero climate goals, making this research an important step towards sustainable and environmentally responsible energy practices.
Figure 1. a) Location of the CO2 storage site Sleipner, offshore Norway. The baseline seismic survey 94p07 with exploration well 15/9-3 and injection well 15/9-A-16 are shown in the map (NPD, 2021). b) Available wireline log data in well 15/9-13, the missing sonic data was generated from a digitized deep resistivity (Rt) log using the Faust equation (Faust, 1953). Since the bulk density (RHOB) log was erroneous due to an over-guaged hole, a synthetic density log (RHOB-Gardner) was created using the Gardner’s equation (Gardner et al., 1974). Thin shale layers on the various sandstone layers can be seen on the GR log. c) A random seismic line passing through well 15/9-13 of 3D survey 94p07 over the Sleipner area. The reflectors delineating the various layers are plotted on the seismic with both the injection well (15-9-A-16) and the earlier exploration well (15-9-13). The location of the line is shown on map a. d) The random seismic line of monitor survey 10p11 shows the effect of CO2 plume on seismic reflectors.
The variance attribute was extracted using the baseline survey (94p07), and saturation calculations were carried out on monitor seismic volume 10p11. The 10p11 volume was processed for imaging; therefore, there were no constraints to process it with a 4D protocol, leading to better results.
The vertical well 15/9-13 was an exploration well drilled in 1982 to appraise the Sleipner gas-condensate field. Due to a bad hole in the shallow Utsira Formation, the density log values were not reliable. Also, the sonic (DT) data points were missing within the Utsira level. The deep resistivity log (Rt) was not made public; however, a hard copy scan of the composite log is available on the NPD Factpages (NPD, 2021). We carefully digitized the deep resistivity log (Rt) from the hardcopy scan, and using Faust equation (Faust, 1953), a synthetic sonic was generated within the missing interval, ensuring the trend fits with the upper and lower available data points. Gardner equation (Gardner et al., 1974) was applied on this sonic (Dt-Faust) to extract a synthetic density log (RHOB-Gardner), making sure the values were comparable with the density log available in well 15/9-A-16 (Fig.1b). Since shear wave (Vs) logs were not acquired in both wells, we generated synthetic Vs using the Greenberg and Castagna (1992) method, incorporating the volume of shale (VshGR) from the gamma-ray log for mineral constraint.
The Sleipner 2019 Benchmark model within the available CO2 dataset contains various velocity surfaces and cubes. Using these data, we made a similar velocity model as suggested in the information. The given depth horizons were converted to time surfaces using the velocity model. These surfaces were snapped on the 94p07 peaks or troughs according to the nature of the interface. The troughs and peaks represent sandstone and shale tops, respectively. The snapped surfaces were gridded, subsequently obtaining zones from Utsira L-1 to the Caprock (Fig. 1c).
The deviated well 15/9-A-16 is an injection well with deep resistivity, density, neutron, and gamma ray logs available. A sonic, however, was either not present or was not made public. We used the 15/9-13 modified sonic data to display the deviated well (15/9-A-16 ) in the time domain (Fig. 1c&d).
The prestack seismic data was in the form of partial offset stacks. These stacks were converted to three angle stacks (10°- 20°- 30°) using the well 15/9-13 modified sonic after tying it with the 10p11 monitor survey. A prestack inversion of the 10p11 seismic was carried out using a commercial software (Fawad et al., 2020, 2021).
A New Rock Physics Model
In an earlier paper (Fawad and Mondol, 2022a), we introduced a new interactive rock physics model that directly related Zp with the Vp/Vs ratio for predicting fluid saturation (CO2 saturation in this case). The model can be calibrated with the well-log data interactively without using the Hertz-Mindlin model (Mindlin, 1949), Hashin–Shtrikman bounds (Hashin and Shtrikman, 1963), or Gassmann fluid substitution (Gassmann, 1951). The proposed model is nonlinear similar to the CPEI (Curved Pseudo-Elastic Impedance) attribute (Avseth et al., 2014; Avseth and Veggeland, 2015), however, with physical meanings and flexibility. To estimate the fluid saturation, the calibrated model can be applied directly to the seismic-derived Zp and Vp/Vs cubes. Following is the newly suggested rock physics model that was applied after calibration on the Zp and Vp/Vs ratio data cubes obtained by pre-stack inversion to estimate the target fluid saturation (SCO2) (Fawad and Mondol, 2022a, 2022b):
where VPma and VPw are the P-wave velocities of the mineral matrix, and brine, respectively, VPCO2 is the apparent P-wave velocity of CO2, ρma is the density of mineral grains, ρCO2 is the apparent density of CO2, ρw is the density of brine, Zp is acoustic impedance, G is the mineralogy/shaliness coefficient, α is Vs/Vp ratio of the mineral/rock matrix, and n is the stress/cementation coefficient.
Since no well has been drilled to assess the storage after the CO2 injection commenced, the rock physics model (Fawad and Mondol, 2022a, 2022b) was calibrated using the 10p11 inverted Zp and Vp/Vs data from Inline 1113 within the time window covering the Caprock to base Utsira Formation. First the mineral matrix value (α) was selected, i.e., 0.676. The mineralogy/shaliness coefficient ‘G’ was selected as 0.96. The stress/cementation coefficient ‘n’ that controls the slope of the iso-SCO2 contour lines was set to 0.73 to calibrate the line with SCO2 = 0 with the brine-saturated sandstone trend from the data. Finally, the apparent VpCO2 and ρCO2 were iterated so that the 90% fluid saturation line approximately coincided with the lowest Zp with corresponding Vp/Vs ratio data points (Fig. 2). The apparent VpCO2 and ρCO2 were 1354 m/s and 0.83 g/cm3, respectively.
Figure 2. Rock physics model calibration using acoustic impedance (Zp) and Vp/Vs ratiodata from seismic inversion of the monitor survey 10p11. A seismic lnline (IL 1113) passing through the CO2 plume specifying the data from the top Caprock to the base Utsira Formation was used. After adjusting the brine-saturated sandstone trend, the apparent CO2 velocity (Vpco2) and density (co2) were iterated so that the 90% CO2 saturation line coincides with the lowest sandstone Zp and the corresponding low Vp/Vs ratio.
The topographic elevation (in TWT) of the top surfaces of the various layers were plotted side by side with the variance attribute at the surface and the maximum magnitude of SCO2 within a layer interval (Figs. 3-14). Following is the description of the results:
Base Utsira Formation
Base Utsira Formation shows a change in elevation from highs to lows where the Utsira Sandstones were deposited (Fig. 3). The variance attribute shows a high-density distribution of circular features on a high in the southwest. In other places, these circular features do not show any affinity with the high elevations; rather, these are distributed at the edges of the structures. These rounded features are considered to be high permeability channels to transport natural gas and other compaction-derived fluids vertically upwards (Williams and Chadwick, 2021).
Figure 3. Base Utsira Formation two-way-time (TWT) map, and Seismic Variance values draped on the Base Utsira surface. A yellow ellipse marked on the Base Utsira variance map shows a high density distribution of circular features on a topographic high in the southwest.
Utsira – Layer 1
This layer close to the injection point possibly does not represent the true elevation due to the pull-down effect of the CO2 plume on the seismic (Fig. 4). There are numerous high variance circular features close to the injection point. The CO2 saturation is low in the interval saturation map with few point-anomalies close to the injection area, which means high saturation of CO2 is present in some conduits through which the CO2 migrated to the upper layers . Overall, the density of high-variance points is higher on the top of Layer 1 compared to the Base Utsira Layer.
Figure 4. Top Utsira Layer 1 two-way-time (TWT) map, Seismic Variance values draped on the top Utsira Layer 1 surface, and the SCO2 maximum magnitude within the Layer 1 interval. A yellow circle highlights the variance density close to the injection point.
Utsira – Layer 2
Layer 2 shows a structural high just above the injection point (Fig. 5). The variance anomaly pattern is similar to Layer 1, except the variance density is lower close to the injection point compared to layer 1. The structural high above the injection point exhibits few high saturation anomalies and point-anomalies, possibly representing high permeability conduits still containing high concentrations of CO2.
Figure 5. Top Utsira Layer 2 two-way-time (TWT) map, Seismic Variance values draped on the top Utsira Layer 2 surface, and the SCO2 maximum magnitude within the Layer 2 interval. The yellow circle highlights the variance density close to the injection point. The variance density close to the injection point is lower on top Layer 2 compared to Layer 1.
Utsira – Layer 3
There is a structural high just above the injection point; however, Layer 3 shows low CO2 saturation, indicating the CO2 readily migrated to the upper layers (Fig. 6). Minor CO2 saturation point anomalies are evident, representing some CO2-bearing conduits.
Figure 6. Top Utsira Layer 3 two-way-time (TWT) map, Seismic Variance values draped on the top Utsira Layer 3 surface, and the SCO2 maximum magnitude within the Layer 3 interval. A CO2 saturation point-anomaly (highlighted by a yellow arrow) corresponds to a variance point anomaly indicating the CO2-bearing conduit continues to the top Layer 3 surface.
Utsira – Layer 4
There is a NE-SW trending structural high close to the injection point (Fig. 7). The variance map has a similar pattern as Layer 3 with a comparatively high density of variance point-anomalies. The CO2 accumulation on the saturation map shows the same trend as the structural strike.
Figure 7. Top Utsira Layer 4 two-way-time (TWT) map, Seismic Variance values draped on the top Utsira Layer 4 surface, and the SCO2 maximum magnitude within the Layer 4 interval. The Layer 4 variance map has a similar pattern as Layer 3 (in Fig. 6) with a comparatively high density of variance point-anomalies (indicated by a yellow ellipse).
Utsira – Layer 5
Layer 5 also has a NE-SW trending structural high above the injection point (Fig. 8). The variance point-anomaly density is similar to Layer 4. Layer 5 exhibits the largest CO2 storage area on the saturation map. The variance point-anomalies above this saturation distribution are generally of low magnitude.
Figure 8. Top Utsira Layer 5 two-way-time (TWT) map, Seismic Variance values draped on the top Utsira Layer 5 surface, and the SCO2 maximum magnitude within the Layer 5 interval. The Layer 5 variance point-anomalies are generally of low magnitude above the high CO2 saturation distribution (demarcated by a yellow ellipse).
Utsira – Layer 6
There is a N-S trending high northeast of the injection point on the elevation map (Fig. 9). A N-S trend in the variance point-anomalies is present in the middle of the area; however, the plume saturation seems to follow the structural high with a string of saturation point anomalies interpreted to be the conduits for CO2 migration to the overlying layers. These conduits are not obvious on the baseline seismic variance map possibly indicating a subsequent activation in response to the CO2 plume.
Figure 9. Top Utsira Layer 6 two-way-time (TWT) map, Seismic Variance values draped on the top Utsira Layer 6 surface, and the SCO2 maximum magnitude within the Layer 6 interval. A yellow ellipse on Layer 6 variance map highlights N-S trending point-anomalies in the middle of the area.
Utsira – Layer 7
Similar to Layer 6, a N-S trending high north of the injection point is present on the elevation map (Fig. 10). The variance map is also similar to that of Layer 6, except the N-S trending variance point-anomalies are not present in the middle of the area. A plume with high saturation of CO2 is present just south of the injection point with a NNE-SSW trend. Further NNE, there is a sparse saturation trend indicating most of the CO2 moved vertically upwards in the overlying Layer 8.
Figure 10. Top Utsira Layer 7 two-way-time (TWT) map, Seismic Variance values draped on the top Utsira Layer 7 surface, and the SCO2 maximum magnitude within the Layer 7 interval. A plume with a high saturation of CO2 is present just south of the injection point with a NNE-SSW trend.
Utsira – Layer 8
A small structural high is present slightly north of the injection point (Fig. 11). On the variance map, the point-anomalies are sparse, depicting the low density of the vertical conduits. A NNE-SSW trending high CO2 saturation anomaly indicates a better storage potential for this layer.
Figure 11. Top Utsira Layer 8 two-way-time (TWT) map, Seismic Variance values draped on the top Utsira Layer 8 surface, and the SCO2 maximum magnitude within the Layer 8 interval. The NNE-SSW trending high CO2 saturation anomaly above the injection point indicates a better storage potential for this layer.
Utsira – Thick Shale
Similar to Layer 8, there is a small structural high above the injection point, and the variance anomalies are sparse with an absence of high magnitude anomalies in the vicinity of the injection point (Fig. 12). There are minor CO2 saturation anomalies above the structural high close to the injection point, showing the possibility of CO2 upward migration through re-activated vertical conduits within the shale.
Figure 12. Top Thick Shale two-way-time (TWT) map, Seismic Variance values draped on the top Thick Shale surface, and the SCO2 maximum magnitude within the Thick Shale interval. On the saturation map there are minor CO2 saturation anomalies above the structural high close to the injection point; however, the point-anomalies density is very low on the variance map (yellow ellipse). Either the in-situ conduits are closed near the top Thick Shale surface or the conduits not obvious on the base survey were later re-activated to allow an upward CO2 migration through the shale.
Utsira – Sand Wedge
On the Sand Wedge (or Layer 9) elevation map, there is a north- trending structural high from the injection point with two parallel channels merging in a lobe further northwards (Fig. 13). The variance point-anomaly density is low close to the injection point. The saturation map shows a NNE-SSW trending plume with high CO2 saturation. The two northwards trending channels are visible, with the left seeming to provide a horizontal conduit for CO2 to fill the sand lobe.
Figure 13. Top Sand Wedge two-way-time (TWT) map, Seismic Variance values draped on the top Sand Wedge surface, and the SCO2 maximum magnitude within the Sand Wedge interval. The Sand Wedge point-anomaly density is low close to the injection point, highlighted by a yellow circle on the variance map. The saturation map shows a NNE-SSW trending plume with high CO2 saturation.
The Caprock above the Utsira Formation sandstones shows a similar elevation as that of Sand Wedge (or Layer 9) (Fig. 14). The variance distribution and density are higher than the underlying three layers (Figs. 11-13); however, on the structural high close to the injection point, the density of point-anomalies is comparatively low on the variance map. On the CO2 saturation map, there are a few minor saturation anomalies in the NNE of the injection point and two high saturation anomalies in the south and southwest. The southwestern anomaly is a little off from the main CO2 plume area.
Figure 14. Top Caprock two-way-time (TWT) map, Seismic Variance values draped on the top Caprock surface, and the SCO2 maximum magnitude within the Caprock interval. On the Caprock variance map, the density of point-anomalies (demarcated by a yellow ellipse) is comparatively low on the structural high close to the injection point. On the Caprock saturation map, two high saturation anomalies in the south and southwest are highlighted with yellow arrows, corresponding to low values on the Cap Rock variance map. The southwestern anomaly is a little off from the main CO2 plume area.
Several earlier studies highlighted the presence of point-discontinuities, vertical conduits, or gas chimneys in the Sleipner area (e.g., Bitrus et al., 2016; Williams and Chadwick, 2021). These chimneys are associated with bright spots in the baseline seismic (Fig.15a). The boundary between Base Utsira Formation and the underlying Hordaland Shale shows pronounced relief (Fig. 3) created by soft sediment deformation while the Utsira Sand deposition took place (Boait et al., 2011). The Base Utsira surface shows a dense distribution of circular features or point-discontinuities on the variance attribute on a structural high in the southwest (Fig. 3). In other places, these are distributed at the edges of the structures, possibly related to the faults bounding these structures. These features possibly provided vertical conduits for earlier natural gas/fluid flow. The variance attribute shows a high density of these point discontinuities on Utsira top layer 1 to top layer 7 (Figs. 4-10). This could be due to the upward branching of gas chimneys starting from the Base Utsira surface that itself exhibits a comparatively low density of variance anomalies (Fig. 15b). On top of Layer 8, the top Thick Shale and top Sand Wedge, the density of discontinuities is apparently low (Figs. 11-13). This can be interpreted as high permeability sands, which facilitated the vertical gas flow without a need for vertical conduits; similarly, the Thick Shale is possibly not continuous and may contain high permeability sand facies. The top surface of the Caprock again shows a high density of variance anomalies (Fig. 14), indicating that the paleo gas flow took place through gas chimneys as the Caprock was impermeable.
Figure 15. a) A baseline survey (94p07) seismic inline showing gas chimneys and associated bright spots in the overburden above the Utsira CO2 storage reservoir. b) Same inline with variance attribute
The caprock interval shows some high anomalies on the SCO2 attribute (Figs. 14&16a). Initially, it seemed a CO2 intrusion in the caprock; however, these anomalies correspond to bright spots in the baseline seismic survey representing in-situ shallow gas accumulations (Fig. 16b&d) related to gas chimneys (Fig. 15a&16a).
Earlier studies showed concerns about CO2 saturation monitoring that Vp decreases significantly as the low percentage of CO2 is introduced, yet unable to precisely measure subsequent increases in CO2 saturation. (Furre et al., 2017). Also, Zp is only strongly sensitive to CO2 saturation of <30% or so (Boait et al., 2011). However, the CO2 plume in the Utsira reservoir is still in a dense phase (Boait et al., 2011; Falcon-Suarez et al., 2014); therefore, we can rule out the gas effect on seismic velocity. Furthermore, lab experiments show a rock physics relation with a combined patchy fluid distribution and squirt flow effects, which provides a practical Vp-Saturation relationship (Falcon-Suarez et al., 2014; Williams and Chadwick, 2021).
It has already been demonstrated that a fluid saturation of up to 100% can be quantified on a Zp-Vp/Vs plane, even in the case of gas (Avseth et al., 2005; Fawad and Mondol, 2022a). Using the seismic-derived Zp and Vp/Vs data for saturation estimation, most of the uncertainties are associated with the inversion procedure itself (Avseth et al., 2016; Fawad and Mondol, 2022a). This method can be applied only in siliciclastic reservoirs, assuming no CO2 – rock chemical reaction took place. Due to a difference in resolution between the wireline log data and seismic, calibrating the model using wireline logs often results in an up-scaled profile in seismic. The other uncertainties include lateral changes in mineralogy or shale volume within the reservoir, causing a slight change in the reference brine saturated trend compared to the original calibration (Fawad and Mondol, 2022a).
There was an uncertainty delineating Layer 1 on the baseline survey due to the pull-down effect. Therefore, except for Utsira Layer 1, the CO2 plume followed a shape controlled by the topography of the top surface, signifying the buoyancy effect. Our SCO2 estimation showing high saturation in the upper layers confirms the earlier flow simulations in the Sleipner storage area (Boait et al., 2011; Williams and Chadwick, 2021) that the dimming of the deeper layers in the time-lapse seismic is not purely a velocity attenuation effect, but at least in part a real change in saturation due to upward CO2 migration (Fig. 16c). The saturation data can be used to extract a 3D geometry of the CO2 plume in a reservoir (Fig. 16d).
Figure 16. 3D view of Sleipner CO2 storage area. a) The Top Caprock shows SCO2 anomalies within the Caprock layer interval:, however, the presence of earlier gas chimneys and bright spots within the Caprock interval exhibited on the baseline seismic 94p07 in b) confirms that the anomalies represent in situ shallow natural gas. c) A random line from the SCO2 cube passing through well 15/9-13 shows that the CO2 saturation is high in the upper layers. In the lower layers, the CO2 saturation is low, and the distribution is sparse. The white arrows indicate the perforated zone representing the CO2 injection point. d) A 3D plume geometry extracted using a threshold SCO2 > 50%, all the gas accumulation above the Sand Wedge within the Caprock represent in situ shallow natural gas.
The variance attribute shows numerous point-discontinuities on top of Utsira Layers 1 to 7. Layer 8, the overlying Thick Shale, and Sand Wedge/ Layer 9 show a low density of these discontinuities, indicating high vertical permeability sands and sand facies within the Thick Shale. Estimating SCO2 using the newly conceived rock physics model predicts high saturation in the upper layers (4-9) compared to the lower layers (1-3), signifying an upward CO2 migration. These results support the earlier CO2 plume simulation studies on the Sleipner storage area.
We are thankful for the facilities, support, and funding (for project SF23001) provided by the Center for Integrative Petroleum Research (CIPR), King Fahd University of Petroleum and Minerals (KFUPM). We are obliged for the data provided by the CO2 Storage Data Consortium, Equinor AS with Sleipner partners (Vår Energi ASA, PGNiG Upstream Norway AS, KUFPEC Norway AS), and Norwegian Petroleum Directorate (NPD). We appreciate GeoSoftware for providing HampsonRussell academic software licenses, dGB Earth Sciences for OpendTect, LR for Interactive Petrophysics, and SLB for Petrel.
Declaration of AI-assisted Technologies
During the preparation of this work, the authors used ChatGPT 3.5 for structuring the Abstract and the Introduction part. Subsequently, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
Manzar Fawad and Nazmul Haque Mondol have been granted a patent in Norway (NO346572B1 – Rock physics model for fluid identification and saturation estimation in subsurface reservoirs), and a patent is pending in the United States (US63/140,891).
Avseth, P., and T. Veggeland, 2015, Seismic screening of rock stiffness and fluid softening using rock-physics attributes: Interpretation, 3, SAE85–SAE93.
Avseth, P., T. Mukerji, and G. Mavko, 2005, Quantitative Seismic Interpretation: Applying Rock Physics Tools to Reduce Interpretation Risk: Cambridge university press.
Avseth, P., T. Veggeland, and F. Horn, 2014, Seismic screening for hydrocarbon prospects using rock-physics attributes: The Leading Edge, 33, 266–274.
Avseth, P., A. Janke, and F. Horn, 2016, AVO inversion in exploration—Key learnings from a Norwegian Sea prospect: The Leading Edge, 35, 405–414.
Bitrus, P. R., D. Iacopini, and C. E. Bond, 2016, Defining the 3D geometry of thin shale units in the Sleipner reservoir using seismic attributes: Marine and Petroleum Geology, 78, 405–425.
Boait, F., N. White, A. Chadwick, D. Noy, and M. Bickle, 2011, Layer spreading and dimming within the CO2 plume at the Sleipner Field in the North Sea: Energy Procedia, 4, 3254–3261.
Chadwick, A., G. Williams, N. Delepine, V. Clochard, K. Labat, S. Sturton, M.-L. Buddensiek, M. Dillen, M. Nickel, and A. L. Lima, 2010, Quantitative analysis of time-lapse seismic monitoring data at the Sleipner CO 2 storage operation: The Leading Edge, 29, 170–177.
Chadwick, R. A., P. Zweigel, U. Gregersen, G. A. Kirby, S. Holloway, and P. N. Johannessen, 2004, Geological reservoir characterization of a CO2 storage site: The Utsira Sand, Sleipner, northern North Sea: Energy, 29, 1371–1381.
Clochard, V., N. Delépine, K. Labat, and P. Ricarte, 2009, Post-stack versus pre-stack stratigraphic inversion for CO 2 monitoring purposes: a case study for the saline aquifer of the Sleipner field, in SEG Technical Program Expanded Abstracts 2009, Society of Exploration Geophysicists, 2417–2421.
Delepine, N., V. Clochard, K. Labat, and P. Ricarte, 2011, Post‐stack stratigraphic inversion workflow applied to carbon dioxide storage: application to the saline aquifer of Sleipner field: Geophysical Prospecting, 59, 132–144.
Falcon-Suarez, I., L. North, and A. Best, 2014, Experimental rig to improve the geophysical and geomechanical understanding of CO2 reservoirs: Energy Procedia, 59, 75–81.
Faust, L. Y., 1953, A velocity function including lithologic variation: Geophysics, 18, 271–288.
Fawad, and N. H. Mondol, 2022a, Monitoring geological storage of CO2 using a new rock physics model: Scientific Reports, 12, 1–11.
Fawad, M., and N. H. Mondol, 2022b, Rock physics model for fluid identification and saturation estimation in subsurface reservoirs, Patent NO346572B1: , 1–15.
Fawad, M., J. A. Hansen, and N. H. Mondol, 2020, Seismic-fluid detection-a review: Earth-Science Reviews, 103347.
Fawad, M., M. J. Rahman, and N. H. Mondol, 2021, Seismic reservoir characterization of potential CO2 storage reservoir sandstones in Smeaheia area, Northern North Sea: Journal of Petroleum Science and Engineering, 205, 108812.
Furre, A.-K., O. Eiken, H. Alnes, J. N. Vevatne, and A. F. Kiær, 2017, 20 years of monitoring CO2-injection at Sleipner: Energy Procedia, 114, 3916–3926.
Gardner, G. H. F., L. W. Gardner, and A. R. Gregory, 1974, Formation velocity and density—The diagnostic basics for stratigraphic traps: Geophysics, 39, 770–780.
Gassmann, F., 1951, Über die elastizität poröser medien: Vierteljahrss-chrift der Naturforschenden Gesellschaft in Zurich 96, 1-23: Paper Translation at http://Sepwww. Stanford. Edu/Sep/Berryman/PS/Gassmann. Pdf.
Ghosh, R., M. K. Sen, and N. Vedanti, 2015, Quantitative interpretation of CO2 plume from Sleipner (North Sea), using post-stack inversion and rock physics modeling: International Journal of Greenhouse Gas Control, 32, 147–158.
Hashin, Z., and S. Shtrikman, 1963, A variational approach to the theory of the elastic behaviour of multiphase materials: Journal of the Mechanics and Physics of Solids, 11, 127–140.
Karstens, J., W. Ahmed, C. Berndt, and H. Class, 2017, Focused fluid flow and the sub-seabed storage of CO2: Evaluating the leakage potential of seismic chimney structures for the Sleipner CO2 storage operation: Marine and Petroleum Geology, 88, 81–93.
Mindlin, R. D., 1949, Compliance of elastic bodies in contact: J. Appl. Mech., ASME, 16, 259–268. NPD, 2021, Norwegian Petroleum Directorate Fact-pages. Available at https://factpages.npd.no/. Accessed February 1, 2021.
Queißer, M., and S. C. Singh, 2013, Full waveform inversion in the time lapse mode applied to CO2 storage at Sleipner: Geophysical Prospecting, 61, 537–555.
Raknes, E. B., B. Arntsen, and W. Weibull, 2015, Three-dimensional elastic full waveform inversion using seismic data from the Sleipner area: Geophysical Journal International, 202, 1877–1894.
Williams, G. A., and R. A. Chadwick, 2021, Influence of reservoir-scale heterogeneities on the growth, evolution and migration of a CO2 plume at the Sleipner Field, Norwegian North Sea: International Journal of Greenhouse Gas Control, 106, 103260.
Yan, H., B. Dupuy, A. Romdhane, and B. Arntsen, 2019, CO2 saturation estimates at Sleipner (North Sea) from seismic tomography and rock physics inversion: Geophysical Prospecting, 67, 1055–1071.
About the Authors
Manzar Fawad received his M.Sc. in Applied Geology from the University of Azad Jammu & Kashmir, Pakistan, an M.Sc. in Petroleum Geosciences from the Norwegian University of Science & Technology [NTNU], and a Ph.D. in Experimental Rock Physics from the University of Oslo, Norway. He is a multidisciplinary geoscientist with more than 25 years of petroleum geoscience, consulting, and research experience. His research interests include rock physics, petrophysics, geomechanics, and quantitative seismic to characterize source, reservoir, cap, and overburden rocks for the exploration of conventional and unconventional hydrocarbon, and geologic storage of CO2. Currently, Manzar Fawad is a Research Scientist at KFUPM, Saudi Arabia.
Nazmul Haque Mondol received a B.S. in geology from the University of Dhaka, Bangladesh, an M.S. in geology from the University of Dhaka, Bangladesh, an M.S. in petroleum geosciences from the Norwegian University of Science & Technology, Norway, and a Ph.D. in experimental rock physics from the University of Oslo, Norway. He is a professor at the University of Oslo and an advisor (adjunct position) at NGI, Oslo, Norway. He was a postdoctoral fellow at the University of Oslo, Norway (with grants from the Research Council of Norway under the PETROMAKS program), before permanently joining the University of Oslo. His research interests include rock physics, petrophysics, geophysics, geomechanics, geology, and seismic to characterize the source, reservoir, cap, and overburden rocks for exploration and exploitation of conventional and unconventional hydrocarbon resources and geological storage of CO2. He can be reached at firstname.lastname@example.org and email@example.com.