REVIEW OF LITERATUR.....

SPE 117735

Building a Deterministic 3D Model of Tar Mat deposits in a Carbonate Reservoir in a Geologically Consistent Manner: A Case Study from Offshore Abu Dhabi Philippe J. Ruelland, Christoph T. Lehmann, Khalil I. Al Hosany, David O. Cobb, ADMA-OPCO

Copyright 2008, Society of Petroleum Engineers This paper was prepared for presentation at the 2008 Abu Dhabi International Petroleum Exhibition and Conference held in Abu Dhabi, UAE, 3–6 November 2008. This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract The presence of a Tar Mat in carbonate reservoirs in the Gulf region is common. Tar Mats occur on different scales from thin seams along stylolites and porous intervals to thick successions within the reservoirs. In the latter case, it raises important issues regarding field development options and well placement optimization. The Tar Mat in Field A, offshore Abu Dhabi, was detected in the first exploration well in 1969. In the 1970s and 1980s, 7 additional wells were drilled; the most recent well was drilled in 2007. Tar Mat sections were identified in all the appraisal wells. The top of the Tar Mat can be clearly seen in the cored reservoir sections where it plugs the larger pore space. The bottom of the Tar Mat is not as simple to determine visually. Detection of Tar Mat is less reliable in cuttings and logs especially in tight zones and in the sections with lower reservoir quality. Thin section observations on the Tar Mat interval in the various wells appear to show that it is not present in heavily calcite cemented intervals related to paleo-water legs. This hypothesis on the Tar Mat generation is based on Gulf analogues, in which the Tar Mat reflects a fossilized paleo-oil-water contact. In Field A, the top and bottom Tar Mat surfaces are not flat and their deformation reflects the growth of the structure after Tar Mat generation ended. The bottom Tar Mat surface is considered as the latest paleo-oil-water contact preserved in the field. These observations help in building a 3D model of the Tar Mat away from well control and, therefore, decrease the uncertainty in predicting the distribution of Tar Mat in the field. Better predicting the distribution of Tar Mat will have a significant impact in the successful development of this field. INTRODUCTION

Field A (Fig. 1) is a four-way dip structure located offshore Abu Dhabi. Field A was discovered by well A-1 in 1969. 8 Appraisal wells have since been drilled, mainly in the 1970s and 1980s; the last well A-9 was drilled in 2007. The full-field development of Field A is planned in the near future.

2 SPE 117735

The Four-way dip structure formed, like many similar structures offshore Abu Dhabi, over a salt-related anti-form. The salt is regionally known to be Late Precambrian or Early Cambrian in age and was re-activated over different tectonic phases.

The Oil-bearing reservoirs are the Arab A (A2) to D formations, and the

Tar Mat is an essential feature in this field. This was recognized very early in the discovery and appraisal wells.

Tar Mat has always been considered as a crucial heterogeneity to take into account prior to developing the field. The water injection scheme would be significantly impaled by the location of the Tar Mat. However no satisfactory model for the Tar Mat geometry had been proposed so far. A simple surface interpolation of well tops between wells was the input to early static model definition; no geological input was applied to understand the distribution of Tar Mat away from the wells.

KNOWLEDGE ON TAR MAT DEPOSIT IN GULF AREA FIELDS

Tar is a common occurrence in the Gulf area Oil Fields. It is most often present as thin seams either in the matrix or filling in stylolites and fractures. In a few fields, Tar Mat occurs as thick units that act as very strong horizontal barriers. Knowing the Tar Mat thickness and extension is important in order to estimate the oil in place as well as to locate water injectors correctly. The risk of placing water injectiors below the Tar Mat would have a critical impact on field pressure maintenance.

Thick Tar Mats have already been described in fields in Kuweit (Minagish Field) by Al-Ajmi et al, 2001, in Abu Dhabi (Fields named S and Z) by Carpentier et al, 1998 and 2006, and in Qatar (Bul Hanine Field) by Jedaan et al, 2006.

The geometry of the Tar Mat filled interval varies, but in all cases seems either related to the position of a paleo-WOC or to the presence of extended low permeability layers in the reservoir. In the case of paleo-WOC related Tar Mats, the reservoir volume in which the asphaltenes deposited can take the geometry of a large and thick ring (ex: Minagish Field, Kuweit)

Where the Tar Mat is related to the presence of a low permeability layer (ex: Fields S and Z, described by Carpentier et al), its geometry is related to the structural deformation of the layer after the Tar Mat was deposited.

The preferred precipitation of the asphaltenes occurs in the higher permeability layers just above a WOC or a barrier, when secondary light oil meets a heavier oil already in-place. This model was described in 2001 for the Minagish Field, and is also proposed in the study of the Bul Hanine Field (2007).

The geochemical model for the Tar Mat generation established for the Bul Hanine Field seems to be a good analogue to what could have occurred in Field A. The model (Fig. 2) can be summarized as follows:

• Oil (primary) was expelled from the Source Rocks and filled the reservoir.

SPE 117735 3

• Asphaltene Precursors were Gravity segregated in the oil column.

• A secondary light oil charge triggered the precipitation of the asphaltenes just above the paleo-WOC or permeability barriers.

It should be noted that changes in P & T conditions could also result in asphaltene precipitation. Biodegradation of the oils is not a likely process as this phenomenon usually occurs when the reservoir temperature is below 70°C (reservoir temperature in Field A is 116°C).

FIELD A DATA OBSERVATIONS

Tar Mat presence is clearly observable in Field A cores. On the logs (Fig. 3) Tar Mat is characterized by a very high Rt value (high LLD or saturated ILD). However, as Tar Mat –filled fractures may not give a resistivity response as sharp as that of a Tar Mat-rich layer. There are some uncertainties as to where its Top and Bottom really are in non-cored wells, especially on the well depths of the Bottom Tar Mat in wells A-1, A-4 and A-8 which were not cored.

The Tar Mat in Field A occurs mainly in the Arab D reservoir although there is some minor evidence for Tar Mat in the Arab C reservoir from the A-6 and A-2 wells. Interestingly no Tar Mat is observed in the Arab D in well A-6.

Thickness of the Tar Mat varies from one well to the other (Fig. 4).Top and Bottom subsea well depths of the Tar mat also vary from one well to another, giving evidence that the Tar Mat interval in Field A is not a flat (or sub-horizontal) ring-shaped volume below the oil such as that in the Minagish Field, Kuweit.

The location of the Tar Mat in the wells is neither related to nor conformable to the stratigraphic horizons (Fig. 5).

At the crest of the structure, the Tar Mat tends to reach the bottom of the Arab D reservoir, whereas on the flanks it does not entirely fill the Arab D reservoir.

The evidence from the varying depths and thicknesses of the Tar Mat suggest that the distribution of Tar Mat occurs rather as a “body” that has an irregular shape. The difficulty lies in understanding why it is not flat and how it will be possible to model the envelope of this body. These are the reasons why strong geological and geochemical rationales are needed in the process.

GEOCHEMICAL AND GEOLOGICAL REASONNINGS

The Geochemical model described in Bul Hanine Field (Qatar) is an analogue to what happened in Field A (Fig. 2).

The reservoir is initially filled by oil. The reservoir undergoes a secondary in-fill by lighter oil (or by gas) that modifies the geochemistry of the primary oil and causes the asphaltenes molecules to precipitate and deposit either at the WOC or above any barrier inside the

4 SPE 117735

reservoir. As time progresses, the asphaltene deposits also act as barriers and cause asphaltene to deposit just above them.

The geological reasoning must take into account the geochemical model and the Field observations. The geochemical model implies that the WOC plays a crucial role. In Field A, however, neither the Top nor the Bottom Tar Mat surfaces are flat. The fact that the Tar Mat records a paleo-WOC is a key hypothesis.In Field A, the surfaces are not sub-horizontal. This observation suggests that the original sub-horizontal Tar Mat has been deformed at a certain stage in time. The structure continued to grow while the Tar Mat was being deposited and after the end of asphaltene precipitation in the field.The combination of structural growth and asphaltene precipitation also explains the variations in thickness of the Tar Mat as observed in the different wells.

This reasoning also implies that the Bottom surface of the Tar Mat is the very last horizontal surface (WOC) to be fossilized by the asphaltene precipitation process.

This geological rationale based on the geochemical model and field observations represents a good foundation for the building of the Tar Mat Top and Bottom envelopes. The construction is based on the structural deformation of an originally flat surface. But, as observed in some fields (Bul Hanine in Qatar and Fields S and Z in Abu Dhabi), asphaltene will also deposit above barriers (or low Permeability layers) inside the reservoir. Any Tar Mat-filled originally high permeability layer will now fall in the category of a local barrier.

GEOMETRICAL OBSERVATIONS AND 2D MODEL

In order to validate the idea of Tar Mat deposition at a paleo-WOC from a geometrical point of view, it was necessary to investigate the relative localization of the Top and Bottom Tar Mat at the wells with regards to the Top of Arab D reservoir (Fig. 6).

Cross-plots 1 and 2 illustrate this first investigation: Top Arab well depths correlate better with the Bottom Tar Mat well depths than with the Top Tar Mat well depths. The regularity of the Bottom Tar Mat possibly illustrates that it is the last fossilized paleo-WOC. Thicker asphaltene deposits occurring above local permeability barriers may explain that the Top Tar Mat surface is locally irregular.

In Fig. 7, the Top Arab D well depths are plotted against the isochore between the Top Arab D and Bottom Tar Mat (An isochore thickness is the vertical depth difference between two surfaces, unlike the isopach thickness which is the True thickness when considering the structural dip of the surfaces). Fig. 7 very clearly shows that a linear relationship exists on the flanks of the structure between the Top Arab D well depths and the Isochore thickness between Top Arab D and Bottom Tar Mat. The crestal wells behave differently: the isochore gets thinner as the Tar Mat bottom is constrained by the basal horizon of the Arab D reservoir. The reason why the crestal area does not fit the regression is due to the Arab D unit being filled with Tar Mat down to its base, as observed in wells A-5 and A7 (Fig. 4).

Well A-6, as mentioned earlier, has not encountered the Tar Mat in the Arab D reservoir. The well was used to verify the consistency of the regression law, which predicts a thinning to zero of the Tar Mat thickess on the flanks of the structure. In Fig. 8, well A-6 is added and fits very well with the expected intersection of the bottom tar Mat and the Top Arab D.

SPE 117735 5

All this evidence seems to confirm the deformed paleo-WOC hypothesis.

Tar Mat in the crestal wells fills the Arab D reservoir down to the bottom (except perhaps for well A-1, which was not cored), therefore these thicknesses cannot be used to model the Tar Mat body. However this observation gives an indication that the Tarmat developed after the onset of structuration on Field A. The structuration was re-initiated after the primary oil was in- place and during secondary oil migration, which triggered asphaltene precipitation.

The bottom Tar Mat is the last fossilized paleo-WOC, once the asphaltene stops precipitating (ex: end of the secondary light hydrocarbon charge, or complete plugging in the Bottom-most part of the Tarmat). In the structural context of Field A, the Tar Mat “body” started off as a flat surface and then thickened preferentially in the central part of the Field where the deformation was the greatest. The Tar Mat in-fill was initially planar and progressively took on a deformed non planar 3D geometry.

PETROGRAPHIC EVIDENCE OF DIFFERENCES IN CEMENTATION ABOVE AND BELOW THE TAR MAT In field A, the cabonates, especially the grainstone-dominated facies, went through extensive diagenesis spanning from syndepositional processes to early diagenetic processes, to burial processes. Samples from the grainstone-dominated facies in most of the wells show early diagenetic rim cements around grains followed by equant calcite cements. Large pore space might be filled by blocky calcite cement which formed during burial and can be seen in some of the thin sections.

Hydrocarbon migration inhibits extensive burial diagenesis (cf. example from well A7 in Fig. 9). In the oil leg, the grainstones show syn-depostionnal to early diagenetic fabrics and to a lesser extent burial diagenetic fabric. The pore throat distribution of the grainstones is bi-modal; porosity ranges from 12 to 25% and permeability from 10 to 150mD, depending on the intensity of cementation.

In the aquifer, carbonates show a more pronounced late stage calcite cementation in the form of blocky cement (Fig. 9). This is the case in Well 6 (Northern tip of the Field A structure) where Tar Mat is not developed in the Arab D. This observation supports the previous assumption (brought forward from a geometrical point of view) that during Tar Mat generation in the Arab D reservoir, the reservoir section in the well 6 area was located in the water leg. In the grainstone facies similar to the ones observed in well 7, characterized by a similar bi-modal pore throat distribution, reservoir quality is suppressed by more extensive burial diagenesis; blocky spar cements are more pronounced, reservoir quality is reduced. In the more heavily cemented grainstones, only observed in Well A6, porosity ranges from 7 to 14% and permeability is less than 4mD.

These observations suggest that hydrocarbon migration and Tar Mat formation suppressed extensive burial diagenesis, while carbonates within the waterleg or below a paleo-WOC were subjected to a more extensive and longer burial diagenesis reducing porosity and permeability. No Tar Mat developed in Well A6 because at the time of Tar Mat generation the interval was below the paleo-WOC.

6 SPE 117735

3D DETERMINISTIC MODEL OF THE TAR MAT ENVELOPPE

The proposed deterministic 3D model of the Tar Mat envelope in the Arab D formation is based on the observations described earlier. A bottom Tar mat surface has been constructed from the Top Arab D surface on the basis of the regression line shown in Fig. 8.

Fig. 10 illustrates the construction steps that enabled to build the Bottom Tar Mat surface. The Bottom Tar Mat geometrical surface intersects the Bottom Arab D surface; this does not compromise the method, it simply is a consequence of the observations in the crestal wells. The Bottom Tar Mat surface from the regression was subsequently cut where it lay beneath the Bottom Arab D reservoir.

To construct the Top Tarmat surface, it was necessary to relate it to the extrapolated Bottom Tar Mat surface. The depths, at which the Bottom Tar Mat would have been found in the crestal wells, had the Arab D reservoir been thick enough, were computed. Fig. 11 shows the cross-plot of these depths versus the Top Tar Mat well depths. There is a very clear linear regression between them. The variables being independent, the regression between Top and Bottom Tar Mat depths is valid. Fig. 12 illustrates the construction of the Top Tar Mat surface.

Once the two surfaces were constructed then the Bottom Tar Mat surface was merged with the Bottom Arab D (Fig. 10b) where the former is deeper than the latter, and the surfaces were adjusted at and around the well according to well observations.

The methodology was successfully applied in 3D. The resulting Tar Mat envelope, constructed deterministically, has been explicitly integrated in the cellular model (Fig. 13); a Net to Gross (NTG) parameter equal to ZERO has then been attributed to the Tar Mat-bearing cells.

A palynspastic reconstruction (Fig. 14) was also attempted in order to illustrate the growth of the structure, the oil in-fill and the gradual Tar Mat deposit. The real timing of the growth of the Field A structure is unknown today.

CONCLUSION

In Field A, the Tar Mat is significant both because of its thickness and its extension. It impacts greatly on the development scheme of the field.

This study is the first attempt to bring geochemical, geological and geometrical reasoning together to generate a consistent 3D model of the envelope of the expected Tar Mat-filled reservoir volume. The geometrical approach is based on the data available to date (Tops and Bottoms observed in cores).

The approach is based on observed geometrical relationships between the Top Arab D and the Bottom Tar Mat wells depths.

Observations on the differences in cement types above and beneath the Tar Mat are consistent with the predicted localisation of the Bottom Tar Mat surface; cementation was continuing below the Tar Mat while it was less prominent in the Oil leg.

SPE 117735 7

The Top and Bottom Surfaces of the Tar Mat body represent the envelope of the zone where a Tar Mat is expected. Inside the envelope, the asphaltene precipitation took place initially in higher permeability layers and then continued in lower permeability layers or zones; there may also remain patches where Tar Mat did not deposit.

The proposed approach for modeling the Tar Mat has lead to the construction of a deterministic 3D Tar Mat body. The method is a quick-look approach based on geometry only; nevertheless it does follow a strong geologically-consistent reasoning.

The depths of Top and Bottom Tar Mat are not always very obvious to pick visually, as the dark grey to black colour of the Tar Mat may not be discrimiminant enough (especially in dark grey limestones). Image resistivity logs do spot the thickest intervals but not necessarily the “disseminated” Tar Mat. Thin sections and specific geochemical analyses are the best available techniques today to observe and characterize the Tar Mat.

As the Bottom Surface of the Tar Mat is deformed, Oil is expected to have filled the space between the Tar Mat and the water. The evidence of Oil staining on the cores in well A-6, located structurally low on the northern flank of the structure, tends to confirm this hypothesis. The 3D model of the Tar Mat envelope has enabled to estimate the potential volume of oil underneath it.

This approach for constructing a 3D model of a Tar Mat envelope seems applicable to any field, regarding that the geochemical and geological hypothesis are solid. The method can also be used in order to test any paleo-WOC hypothesis.

This study has improved greatly the understanding of the generation of Tar Mat in Field A. It has enabled to give a picture of the geometry of the expected Tar Mat filled envelope within the reservoir, and therefore has brought some key elements in the subsequent reservoir simulation and Field development scheme.

8 SPE 117735

References Jedaan N.M., Al Abdulmalik A., Dessort D., de Groen V.L.N., Fraisse C.J., Pluchery E., 2007, Characterisation, Origin and Repartition of

Tar Mat in the Bul Hanine Field in Qatar, ITPC11812 Al-Ajmi H, Brayshaw A.C, Barwise A.G, Gaur R.S, 2001, The Minagish Field Tar Mat, Kuwait: Its Formation, Distribution and Impact on

Water Flood, Gulf Arabia Vol.6, No.1, Gulf PetroLink, Bahrain Carpentier B., Arab H., Pluchery E.,Chautru J.-M., 2006, Tar Mats and Residual Oil Distribution in a Giant oil Field Offshore Abu Dhabi,

2006, Journal of Petroleum Science and Engineering 58 (2007), p472-490 Carpentier B., Huc A.-Y, Marquis F., Distribution and Origin of a Tar Mat in the S. Field (Abu Dhabi, UAE), SPE-49472

SPE 117735 9

FIGURES Figure 1: Field A Map at Top Arab D reservoir

Figure 2: Proposed mechanism for the Ta Mat formation in Bul Hanine Field (D. Dessort in Jedaan N.M et al., 2001)

TSR: Thermochemical Sulfate Reduction

10 SPE 117735

Figure 3: Well A-9, example of Tar Mat signature on logs (RT, FMI and on cores)

Figure 4: Tar Mat thickness variations in the cored wells (thicknesses in ft)

SPE 117735 11

Figure 5: Tar Mat deposit is not related to stratigraphy

12 SPE 117735

Figure 6: Cross-plots showing the relationship between Top Arab D and Top and Bottom Tar Mat at the wells 6a) Cross-plot 1: Top Arab D vs Top Tar Mat

6b) Cross-plot 2: Top Arab D vs Bottom Tar Mat

SPE 117735 13

Figure 7: Cross-plot showing the relationship of Top Arab D with the isochore calculated between Top Arab D and Bottom Tar Mat at the wells

Figure 8: In well A-6, no Tar Mat is observed in the Arab D reservoir. The extrapolation of the regression line defined for the flank wells tends to a zero-thickness of the Tar Mat down-flank. It is consistent with the observation in well A-6.

14 SPE 117735

Figure 9: Petrographic evidence for differences in cementation above and below the Tar Mat.

Figure 10: Bottom Tar Mat surface construction steps. 10a) The Bottom Tar Mat surface is built from the Top Arab D surface. This surface intersects the Bottom Arab D horizon.

SPE 117735 15

10b) The Bottom Tar Mat surface is merged with Bottom Arab D where the former is underneath the latter

10c) The surface is adjusted to locally the well markers

16 SPE 117735

Figure 11: Cross-plot of Top Tar Mat well depths and “calculated” Bottom Tar Mat well depths. The “calculated” depths are those at which the Bottom Tar Mat would be expected had the Arab D reservoir been thicker.

Figure 12: Top Tar Mat surface construction steps 12a) The Top Tar Mat surface is built from the calculated (un-merged) Bottom Tar Mat surface

SPE 117735 17

12b) The surface is adjusted locally to the well markers

Figure 13: Representation of the Tar Mat-filled reservoir volume in the 3D Grid, as a Net to Gross parameter equal to zero.

18 SPE 117735

Figure 14: Palynspastic reconstruction of the evolution of the structure and the Tar Mat infill of the reservoir. The exact timing of the events is unknown today. Dotted arrows represent the secondary lighter oil infill. 14a) Time T0: Initial Tar Mat generation, just above the WOC. The Tar Mat interval is horizontal.

14b) Time T1: The structure has grown; the Tar Mat filled reservoir volume thickens towards the center of Field A. In the crestal area, the Arab D unit is already filled down.

SPE 117735 19

14c) Time T2: The structural growth continues together with the asphaltene precipitation, local over-thicknesses are expected to form just above the initial Tar Mat Top.

14d) Time T3: Present-day; there is space between the likely current WOC and the Bottom Tar Mat surface for oil to be present.