Transcript
Sub-salt modelling in 3D – Integration of seismic, well and gravity data validated by drilling A. Price*, M. Flouret, S. Rouyer and C. Ayong-Mba of Total E&P.
77th EAGE Conference & Exhibition 2015 Madrid IFEMA, Spain, 1-4 June 2015
Introduction Offshore sub-salt seismic imaging along the West African margin is a challenge in many areas, and with complex salt geometry, seismic depth imaging alone faces certain limitations. In an effort to further de-risk structures sub-salt, we detail here integration with the gravity data in 3D with methodology and examples. We hope to demonstrate that the incorporation of an independent geophysical parameter, such as density can effectively help to de-risk these difficult targets.
Methodology 1) 3D forward gravity modelling was performed to build confidence in the sub-salt structures seen, but not confidently imaged in seismic data. The modelling including all known seismic structures and using density information from wells logs largely confirmed the presence of density anomalies that corresponded to the locations of the seismic structures of interest. Alternative scenario testing by forward modelling and inversion confirmed that the density anomalies could not be justified by shallower variations or other post-salt structures. 2) Conversion of a series of seismic velocity cubes used to improve the sub-salt imaging to density cubes for 3D forward gravity calculation and comparison with the observed gravity data. This conversion and comparison was aimed at helping understand which of a number of seismic velocity models was more globally consistent with respect to the observed gravity data. The high density contrast between the salt and carbonates produced a large amplitude gravity signal that can be readily modelled – enabling better discrimination between these high-velocity rock-types than with seismic data alone. Difficulties were encountered with the small variations seen between the various cubes as a result of fine velocity updates, displaying small and perhaps not significant minimised error between calculated and observed gravity for two cubes in particular. This difficulty was thought to be caused by the domination of the gravity signal due to larger sub-salt structures demonstrating reduced sensitivity to the post-salt velocity updates. It is interesting to note that for these comparisons the larger amplitude gravity signal is sourced from the sub-salt structures, which is our target level.
Data Two types of gravity data were available over the studied area (Figure 1), GETECH regional data and high-resolution gravity and magnetic data acquired during a recent 3-D seismic survey. These data were merged in order to cover the area of interest and reduce edge effects at the block/model boundary.
Figure 1: High-resolution Bouguer gravity data acquired in conjunction with a 3D seismic survey (centre) merged with regional gravity data with area of sub-salt The surfaces were interpreted from seismic depth data and included a merger of 3D (detailed area) and 2D (regional) are listed below: -
Sea Bed Base Oligocene, Top Albian Carbonates, Top Allochtonous Salt, Base Allochtonous Salt,
-
Top Autochtonous Salt, Base Autochtonous Salt, Aptian Unconformity, Intra-preSalt series Unconformity, Top Basement.
77th EAGE Conference & Exhibition 2015 Madrid IFEMA, Spain, 1-4 June 2015
Depth Filtering The 1st vertical derivative of the merged Bouguer gravity seen in Figure 2 displays anomalies at the location of sub-salt structures seen in the seismic data, as defined by base salt closures. These correlations were suspicious at first since the 1st vertical derivative of the Bouguer gravity usually emphasises shallow structure, and here we were seeing what appeared to be deeper (sub-salt) structure – further investigation was necessary. Initially, depth specific (match or pseudo depth, Pilkington and Cowan 2006) filters were tried to isolate the gravity signal at a theoretical (point source) depth of 8km – well below the seismic base of salt. That filter appears in Figure 3 showing again the anomalies at the locations of interest.
Figure 2: First vertical derivative of the Bouguer gravity - with the indicated anomalies in the gravity coinciding with base salt structures.
Figure 3: Depth specific filter to approx. 8 km depth of the Bouguer gravity data – note again the anomalies at the same locations.
Figure 4: ‘Back-stripped’ residual with the anomalies after all structure above base salt removed.
These depth specific filter results can still be ambiguous in depth, so to further test the existence of these structures, a 3D density model representation of the subsurface for the target area and its surroundings was built using the seismic surfaces and well log density data with some regional 2D seismic interpretation to augment coverage (see Talwani and Ewing 1960 for 3D gravity modelling details).
3D Gravity Modelling Sedimentary densities were assigned in the 3D gravity modelling by a hyperbolic fit of the density vs. depth of near-by well logs - representing compaction driven density increase. Fixed densities were chosen for the high-density, high-velocity carbonate layers and salt. Those densities remained fixed through all modelling iterations.
77th EAGE Conference & Exhibition 2015 Madrid IFEMA, Spain, 1-4 June 2015
The 3D gravity model described above was then used to more quantitatively assess the sub-salt structure by ‘back-stripping’ the gravity signal generated by all structures down to base salt (Figure 4). Here again, we clearly see the sub-salt gravity anomalies associated with the target locations.
Scenario Testing Additionally, the question was asked as to what alternate density configuration would be required to fit the observed gravity data for various scenarios – this of course to asses any ambiguity or uncertainty in the 3D gravity model representation. One scenario is a sub-salt density inversion restricted to a range of densities from 2.50 to 2.65 g/cm3, in the layer defined between Aptian unconformity and the intra-pre-salt-series unconformity that would fit the observed data with results seen in Figure 5. In places, the model indicates high densities required to fit the data in addition to basement topography mapped in the seismic data: such anomalies may be adjusted by a higher basement interpretation in the seismic or by higher density sediments above basement. Another scenario tested was a basement topography of 2.70 g/cm3 that would be required to match the observed gravity with the resulting surface appearing in Figure 6 (in this scenario, the intra-pre-Salt series Unconformity is adjusted in term of depth). In both these scenarios, the target structures are still clearly visible.
Figure 5: Sub-salt density inversion – again the same structures are indicated.
Figure 6: 3D Basement topography inversion of 2.70 g/cm3. Same structures indicated.
Other scenarios tested with an effort to fit the observed data using denser salt layers of 2.20 g/cm3, denser Madiela of 2.70 g/cm3, and lighter Basement of 2.65 g/cm3 all resulted in degraded fits with the observed data compared to the base model (not shown here) and were thus discarded.
Velocity/Density Cube Testing One of the key concerns for the modelling performed in the previous sections was that a constant density of 2.62 g/cm3 was used for the Madiela carbonate layers. The Madiela is known to vary laterally in velocity/density and it was though necessary to account for this variation in some way. The seismic velocity cube contains information on this lateral variation. These variations were included by conversion of the 3D velocity cube to density by well calibration (calibrated Gardner’s relation – see Gardner et. al. 1974) for all sediments with the exception of the salt layers held at a constant 2.15 g/cm3. We then compared the various seismic velocity cubes for improved sub-salt imaging. The comparison was performed by 3D forward calculation of the cubes and comparison with the observed gravity data. Of course, there are residual differences between calculated and observed gravity data for all cubes, however we expected that these differences will minimise for the more ‘realistic’ cubes with better depth imaging results. Those differences were quantified as the standard deviation (SD) of the gravity difference for each cube – an example appearing in Figure 7. Due to the limited size of the target area (box indicated in Figures 1-6 and 7), a merge with the larger, static cube was necessary to account for edge effects in the rapidly laterally varying salt and carbonate structure. 77th EAGE Conference & Exhibition 2015 Madrid IFEMA, Spain, 1-4 June 2015
Figure 8: Standard Deviation ‘error’ between Figure 7: Calculated and observed gravity calculated and observed gravity data for 6, 3D difference map from an early seismic velocity gravity/velocity models cube. In general, these comparisons have a positive gravity difference (red/white) implying relative density/velocity deficit, and negative difference (blue/purple) the opposite. All differences are displayed as a 50 km high-pass filter suppressing regional/deeper differences that are out of scope. Note that for Figure 7, the location of the target structures are near zero difference between a low and a high (green) with possible denser/faster material required to the north-east and the opposite to the south-west, which is thought to be a residual feature deeper in the model (crustal level). Although these remaining differences suggest that the velocity/density cube modelling still does not account for some deeper structures, they can still be used to compare and contrast additional cubes as the unaccounted-for structures will presumably not change. Such a comparison appears in Figure 8 for 6 different velocity updates, suggesting that cube # 2 minimises the SD gravity difference errors. Velocity cube # 6 was used as an end-member test of the Madiela carbonate velocities required to flatten the base salt and was very inconsistent with the 3D gravity data modelling and the observed data. Consequently, a correlation was seen between velocity cubes 2 and 4 producing improved subsalt imaging, thus validating the idea that 3D gravity modelling can be used to globally ‘choose’ a more correct velocity model to better constrain seismic imaging – particularly in sub-salt settings.
Conclusions Coincident gravity anomalies and seismic base salt closures were observed in this locality. Various efforts were made to justify the gravity anomalies seen as structures other than sub-salt with no success. The gravity anomalies are larger in amplitude and not equivalent to any possible post-salt structure, independent of density. In short, the scenario gravity modelling suggests that these are indeed sub-salt structures, most likely localised basement or some other high density closures - which is consistent with the seismic imaging. Additional modelling of the various velocity cubes used for sub-salt seismic imaging in 3D using gravity can be used to globally select the ‘best’ velocity model to better constrain deep seismic imaging – particularly in sub-salt settings where imaging is difficult. Note that the north-west target has since been drilled, intersecting gas/condensate in sandstone reservoirs validating the sub-salt interpretation and the petroleum system in this area. Additional sub-salt structures identified using this methodology are currently being evaluated by new seismic, gravity and magnetic acquisition.
Acknowledgements We wish to thank Total and our partners, Marathon for their support and their permission to publish, also to remember Christian Ayong-Mba for his contributions, who sadly is no longer with us.
References Gardner, G. H. F., Gardner, L. W. and Gregory, A. R. [1974], Formation Velocity and Density – the Diagnostic basics for Stratigraphic Traps, Geophysics 39, No. 6 p 770. Pilkington, M. and Cowan, R. [2006], Model-based separation filtering of magnetic data, Geophysics 71 No. 2. p L71. Talwani, M. and Ewing, M. [1960], Rapid Computation of Gravitational Attraction of Three-Dimensional Bodies of Arbitrary Shape, Geophysics 25 p203.
77th EAGE Conference & Exhibition 2015 Madrid IFEMA, Spain, 1-4 June 2015
SUB-SALT MODELLING IN 3D – INTEGRATION OF SEISMIC, WELL AND GRAVITY DATA VALIDATED BY DRILLING A. Price*, M. Flouret, S. Rouyer and C. Ayong-Mba of Total E&P.
Summary
● Introduction ● Data Used ● Depth Filtering ● Methodology ● 3D Gravity Modelling ● Velocity/Density Cube Modelling ● Scenario Testing ● Conclusions 3
Introduction – The Problem
KPSDM with AGC – M06
Due to the high-velocity, complex geometry salt/carbonates seismic imaging is challenging Sophisticated velocity modelling for depth migration can help – but which model? 4
Data – Bathymetry and model extent
North Target
South Target
Polygons are seismic mapped base salt closures – but what’s underneath? What’s causing them?
5
Data – Marine gravity merged with regional data
Bouguer Gravity 2.20 g/cm3 – coincident anomalies? Salt or carbonate structures ?
6
Data – 1st vertical derivative gravity
Bouguer Gravity, first vertical derivative – correlation not perfect ? Salt or carbonate structures convoluting the signal ?
7
Data – 1st vertical derivative with deep structure
Bouguer Gravity, first vertical derivative – correlation with deep structure
8
Methodology – how to cope with salt and carbonates
Initial seismic velocity cube
Conversion Velocity -> Density
Density Cube loaded into 3D gravity modelling tool
Calculate gravity differences
Density Cube
Invert basement structure
Water – constant 2.20 g/cc Post Salt Velocity/Density Cube Intra--Salt Velocity/Density Cube Intra Salt 2.15 g/cc const. Sub--salt 2.550 g/cc const. Sub Unconformity 2.65 g/cc const. Basement 2.70 g/cc const.
Inclusion of the velocity model converted to density accounts for variable density laterally in the sediments/carbonates
9
3D Gravity Modelling – Initial model
Profile 5 2.15 g/cm3
Profile 4 Approx Target North
Profile 3 Approx Target South
Profile 2
DD 2.55 g/cm3
2.62 g/cm3
Main Unconformity
2.66 g/cm3 2.70 g/cm3
Basement
2.90 g/cm3
Profile 1
Initial fit with the gravity data is good – indicates close to the ‘true’ model.
10
3D Gravity Modelling - Back stripped, sub sub--salt response
Gravity effect of structure from seabed to base salt removed – better correlation.
11
3D Gravity Modelling - ‘Final’ Model
Profile 5
Profile 4 Approx Target North
Profile 3 Approx Target South
Profile 2
DD
2.62 g/cm3
2.55 g/cm3 2.66 g/cm3 2.70 g/cm3
Basement
2.90 g/cm3
Profile 1
Inverting for basement structure... structure
12
3D Gravity Modelling – Basement Inversion
Inverting for basement structure, better correlation with seismic closures – can we do more ?
13
Velocity/Density Cube Modelling - SEGY Velocity Cubes
Initial seismic velocity cube
Conversion Velocity -> Density
Density Cube loaded into 3D gravity modelling tool
Calculate gravity differences
Density Cube
Compare gravity differences
Water – constant 2.20 g/cc Post Salt Velocity/Density Cube Intra--Salt Velocity/Density Cube Intra Salt 2.15 g/cc const. Sub--salt 2.550 g/cc const. Sub Unconformity 2.65 g/cc const. Basement 2.70 g/cc const.
Do this for all depth migration velocity cubes – then compare. Gravity can give a ‘global’ view of the density/velocity field
14
Velocity/Density Cube Modelling - difference from Density Voxet Density Cross Sections Profile 5
Calculated Field Observed Field
Profile 4 Difference Gravity
Profile 3
Profile 2
Profile 1
Contractor velocity/density cube difference
15
Velocity/Density Cube Modelling - A post salt
SD difference 0.9389 0 9389 Negative difference means excess density (e.g. More salt or less carbonate needed)
16
Velocity/Density Cube Modelling - B man update post salt
SD difference 0.9660 0 9660 Negative difference means excess density (e.g. More salt or less carbonate needed)
17
Velocity/Density Cube Modelling - C 2nd update post salt
SD difference 0.9507 0 9507 Negative difference means excess density (e.g. More salt or less carbonate needed)
18
Velocity/Density Cube Modelling - D update post salt
SD difference 0.9517 0 9517 Negative difference means excess density (e.g. More salt or less carbonate needed)
19
Velocity/Density Cube Modelling - Comparison
A
SD diff 0.9389
B man
SD diff 0.9660
C 2nd
SD diff 0.9507
D
SD diff 0.9517
Improved sub-salt coherency correlating with better gravity fit
20
Example of Early Model
SD diff 0.9875
21
Example of 2nd Manual Update Regional Improvement
SD diff 0.9517
Less Noise in Pre-Salt
Better BoS 22
Velocity/Density Cube Modelling - Comparison
1 193 1.193 0 9517 0.9517 0 9507 0.9507 0 9660 0.9660 0 9389 0.9389 0 9875 0.9875 SD
2 707 2.707 (mGal)
Flat Base B Salt – Extreme Case e
Manu ual update witth regional ad djustment
2nd Manual M update e
1st Ma anual update
Manu ual update
Contrractor – Maskked to AOI
Contrractor – whole e Cube
23
Conclusions
● It seems the ggravityy modellingg supports pp the existence of sub-salt density anomalies that are mostly justified by the inclusion of structure between base salt and basement.
● Velocity cube to density conversions demonstrate a data difference minimum with the C second velocity update model suggesting this is the better model model, which also produces improved sub-salt sub salt depth imaging in the 3D seismic
● One O off th the closures l h has b been d drilled, ill d encountering t i sands d and d gas/condensate validating the interpretation
● A new acquisition of these data (seismic, gravity and magnetics) are being assessed to illuminate additional prospectivity
24
Acknowledgements
● The authors would like to thank the partners and Total for support and permission to present this work
25
