Middle East Well Evaluation Review

Ancient coral reefs, long buried insedimentary rocks, are an impor-tant source of oil in the MiddleEast. These carbonate rock formations,created from the skeletal remains ofdiverse marine species, have beenaround for at least 600 M years. Somereefs formed in shallow coastal waters,while others developed around vol-canic islands a long way from any largecontinent. What factors controlled theoccurrence of these coral reefs, andwhich have the best reservoir potential?

In common with modern coral reefs,ancient reefs (figure 1.1) thrived inwarm, shallow water with very littleassociated sediment. Over geologicaltime many different types of organismhave dominated the reef environment,but all have achieved their best growthrates in shallow equatorial waters. Con-sequently, the diversity and abundanceof coral species decrease with distanceaway from the equator.

The typical island reef, developedaround an oceanic volcano, has littlepotential for oil and gas accumulation.In these reef ecosystems almost all theorganic material present is in the formof living organisms. Nutrients from deadorganisms are recycled very rapidly. Asa result, reefs do not generally produceenough excess organic material to gen-erate oil and gas. However, some reefsdo contain hydrocarbons, and the bestoil and gas targets are reefs whichdeveloped close to ancient continentalmargins. Here, organic input from anearby continental shelf may have gen-erated hydrocarbons or been incorpo-rated in shales which can act as traprocks.

The Arabian Peninsula spans a rangeof latitudes, so that in the south - theGulf of Aden and the Red Sea - condi-tions are, currently, favourable for coralgrowth, while the Eastern Mediter-ranean and The Gulf are marginal areas,with few coral species. The Red Sea isparticularly rich in modern coral reefs,especially along the edges of faultblocks.

Fig. 1.1: HIGH AND DRY: Anancient (Triassic) reef,thrust up through the rocksequence, makes up thecliffs of Jebel Misht inOman. Oil-bearing reefs,ranging in age from Triassicto Cretaceous may lieburied east of the Arabianpeninsula, beneath thick,overpressured Tertiaryshales. This image is takenfrom the cover ofO m a n '

Over geological time the ArabianPeninsula has passed through the equa-torial belt a number of times (figure 1.2).Optimum conditions for reef growthoccurred during the Precambrian, Juras-sic, Cretaceous and Middle Tertiary.

During the Miocene, between 5 M and20 M years ago, abundant coral structuresformed at shallow depths on the highestpoints of rotated fault blocks in the Gulfof Suez. Oil is produced from these reefsbut the complex pattern of faulting pre-cludes the development of supergiant(over 1 billion bbls) reservoirs.

Exploration drilling in the eastern partof the Mediterranean, on similar faultblocks, may lead to more Miocene reefdiscoveries. For example, in the Red Sea,Miocene oil accumulation occurredwhen evaporites were deposited on thereef, forming an excellent seal, trappinghydrocarbons in the porous reef rock.However, in some cases the evaporitelayers developed too early, preventingoil and gas migration into the carbonatehighs.

Tertiary rocks contain many of themost productive reefs found in the Mid-dle East. The reefs of the Precambrianalso contain important oil accumulations,while the reservoir potential of the Juras-sic is still under investigation.

Number 15, 1994.

A number of Tertiary reef prospectsare being developed across the MiddleEast. There are Eocene-Oligocene (earlyTertiary) reefs in northern Syria and Iraqand part of the Asmari reefs facies in Iranproduces hydrocarbons. More Tertiaryreefs have been discovered in Oman andin the Gulf of Aden. These discoveriesmay develop into important reservoirs inthe future, although they are not, as yet,in commercial production.

The oldest reefal reservoirs in Arabiaare located in the Precambrian rocks ofsouthern Oman which contain algal stro-matolitic reefs. Stromatolites are largeaccumulations of carbonate sediment andskeletal material bound together by algae.They are often dome-shaped and have adistinctive, finely-laminated appearance.In addition, deep drilling projects in mar-ginal areas of the Ara and Hormuz saltsare expected to reveal additional oil andgas in stromatolitic reservoirs.

Recently, reef- forming stromatoliteshave been discovered in Belize and Hon-duras, in Central America. Studies ofthese stromatolites should clarify the rolewhich these organisms play in the devel-opment of modern reef communities.

During the Jurassic, environmentalconditions in Arabia meant that ooliticcarbonate sands were more commonthan reefs. Oolites are rocks composed of

Tertiary

Cretaceous

Jurassic

Triassic

Permian

Carboniferous

Devonian

Silurian

Ordovician

Cambrian

Cen

ozoi

cM

esoz

oic

Pal

aeoz

oic

Precambrian

Presentsea level

0 0.5 1.0

Relative changes of sea levelFalling Rising

1st &2nd - Order Cycles

600

500

400

300

200

100

10oN 10oS 30oS 50oS

Equ

ator

Position of Arabia withrespect to the Equator

Sahara glaciation

Gondwanan glaciation

Pre

dom

inan

ce o

f car

bona

tes

Pre

dom

inan

ce o

f cla

stic

s

M years

Reefs absent

Incipient reefsReefs fully developed

World reef potential through time

Can have oomouldic

Aragonite threshold

Can have oomouldic

(Leached oolite)moulds

Ocean chemistry

Holocene

Fig. 1.2: Reefs have come andgone throughout geologicalhistory. The presence of well-developed reef facies (a) is linkedto sea level fluctuations. The timeswhen reefs have been almostabsent from the geological recordshow a rough correlation withperiods of rapid sea level change(b). Another factor controlling thedevelopment of coral reefs is themovement of Arabia relative tothe equator through time (c).Although close to the equatorduring the Devonian, Arabia wascompletely above sea level at thattime, so no coral reefs developedon the Arabian peninsula. Thedevelopment of ooids (which arechemical rather than biologicalcarbonates) is controlled byseawater chemistry. As oceanchemistry fluctuated through time(d) the changes influenced therelative stability of calcite andaragonite cements and thedissolution of ooids. This diagramis based on concepts developed byDavid Raup, M.W. Hughes Clarke,Peter Vail and Bruce Wilkinson.

carbonate grains (ooids) which form intidal deltas and other shallow environ-ments or shoals. They are not pro-duced by biological processes, ratherthe carbonate in each ooid precipitateddirectly from calcium carbonate(CaCO3 ) saturated seawater.

However, some patch reefs didoccur in the Jurassic. These small car-bonate structures were scattered acrossthe shallowest parts of the continentalshelf. The patch reefs found to the westof Riyadh, in Saudi Arabia, providedgood reservoirs on salt domes and anti-clinal structures in The Gulf and alongits southern shore.

The hydrocarbon potential of the rel-atively large Jurassic marginal reef inthe Gotnia Basin, has not been exam-ined in detail. However, a recent discov-ery in the Kuwait/Saudi Neutral Zone(South Umm-Guadir DW-1) by theKuwaiti Oil Company (KOC) and SaudiArabia Texaco should renew interest inthis potential reef target. Good dolomiti-zation of the shelf-edge reef trend hasbeen found, with one of the best source-rock sequences down dip in the GotniaBasin. However, effective seals are pre-sent which may separate source fromreefal facies, and could have hinderedoil and gas migration.

(a) (b) (c) (d)

Middle East Well Evaluation Review

Darwins other discovery

Charles Darwin, the naturalist whosetheories revolutionized our understand-ing of biological science, made a signifi-cant contribution to the study of reefs. In1842, while engaged on a scientific voy-age on the research ship HMS Beagle,Darwin proposed his subsidence theoryfor the development of coral atolls. Hespeculated that reef atolls had, at sometime in their development, been mar-ginal or fringing reefs (figure 1.3a)around an island. Darwin believed thatas the island started to subside, the fring-ing reefs continued to grow upwards at arate equal to, or greater than subsi-dence. Consequently, as the area of thecentral island decreased, a shallowlagoon formed between the island andthe growing reef (figure 1.3b). When theisland finally dropped below sea level,all that remained was a large lagoonfringed with upbuilding reef material (fig-ure 1.3c). Seismic surveys and drilling car-ried out as part of the 1991 Ocean DrillingProgram (Leg 143) appear to have con-firmed Darwins theory on atolls.

Darwin also speculated on the char-acteristics of a specific group, the Mal-dive atolls. He suggested that subaerialerosion and subsidence were the mostimportant factors in their development.However, recent studies suggest thatDarwins sea level fluctuations alone cannot account for all the complexities ofcarbonate deposits in the Maldives.

Seismic surveys and drilling in boththe Maldives and the Seychelles haveconfirmed existing plate tectonic recon-structions of the Indian Ocean. In theMaldives, carbonate deposition began inthe early Eocene with shallow water sed-iments resting on hot-spot basalts (figure1.4) related to the slightly older DeccanTrap basalt rocks of India. Very little isknown about the early development ofthe islands, but the limited evidence sug-gests that early rifting coincided withdevelopment of a graben system alongthe transform fault on which the Mal-dives carbonate platform and atollsgrew. The overall structure of the plat-form was probably influenced by crustalcooling effects.

Island reefs and reef platforms arenot potential exploration targets, sincethey generally lack a source rock or seal-ing layer, or both. Reef environmentssuch as the modern Maldives platformcontain large volumes of porous andpermeable rock, but the Maldives aretoo far from sources of organic materialto generate hydrocarbons and too smallto generate the sealing layers necessaryfor oilfield development.

Arabian Sea

Indian Ocean

India

Africa Maldives

SaudiArabia

Reefs

Island records

The subsidence history of a small vol-canic island and subsequent growth ofan Eocene coral reef has been recon-structed using geochemical logging car-ried out as part of the Ocean DrillingProgram. Experts from Columbia Uni-versitys Borehole Research Group stud-ied a composite volcanic-carbonatesequence in the Indian Ocean, using aGeochemical Logging Tool (GLT*) tounravel the geological history of the area.

The volcanic rocks proved to bemainly vesicular olivine basalts whichshowed weathering effects. This sug-gested that they may have formed thesurface of a volcanic island. These rockswere overlain by plagioclase basaltswith high concentrations of titanium (Ti)and iron (Fe). Basalts can be assigned togeochemical families, with the chemi-cal composition of the rocks indicating

Fig. 1.3: GOING,GOING, GONE:Darwins theoryof volcanic islandsubsidence hasbeen confirmedby geochemicalanalysis carriedout during theOcean DrillingProject.According toDarwin, beforesubsidence thevolcanic islandwould have had afringing reef (a).When the islandstarted to subsidecoral growthcontinued at arate greater thanor equal tosubsidence.Eventually, alagoon appearedbehind the reef(b) and, assubsidencecontinued, theisland droppedbelow sea levelleaving an emptylagoon ringedwith coral (c).

Fig. 1.4: HOT SPOTS: The Maldives are oceanislands produced by movement of the IndianPlate over a mantle hot spot which pushesbasalt through the crust to form volcanicislands. As each island moves away from thehot spot the volcano loses its source of magmaand the island starts to subside.

(a)

(b)

(c)

Number 15, 1994.

probable tectonic setting. High concen-trations of iron, aluminium and silicon(figure 1.5) are typical of basalts foundin volcanic island settings. The alu-minium spikes in the lower part of thelog correspond to weathered soil lay-ers deposited between the basalt flows.

A thin calcarenite zone, interpretedfrom core as a beach deposit, is fol-lowed by a distinctive, titanium-richbasalt layer which marks the end of vol-canic activity on the island.

The reef proved more difficult tosample, with a core recovery rate ofonly 5%. However, this was enough toindicate a transition from high-energygrainstones to low-energy packstones.This, together with successive faunalchanges, confirmed that the waterabove the reef was getting deeper. Car-bonates, in stark contrast to basalts, arecharacterized by low concentrations ofaluminium, iron and silicon and, ofcourse, a high calcium content.

Fig. 1.5: WHERES THE BEACH?Geochemical logging is proving apowerful tool for stratigraphicinterpretation. The geochemicaldata retrieved from this boreholeshow a calcium-rich carbonate(reef) deposit overlying an iron-rich basalt (volcanic complex). Aseparate calcium-rich layer,identified from core material as abeach deposit, occurs in themiddle of the volcanic sequence.Columbia University, BoreholeResearch Group, 1990.

0 TiO2 5

100 SiO2 5 100 CaO 0

0 FeO 20 0 S 10

50 Al2O3 0

8000

7900

8100

8200

8300

Dep

th b

elow

the

rig fl

oor

Ti Si Fe Ca S Al

ft

Benthicforaminifera,gastropods,bryozoans,molluscs,solitary andcolonial corals

Grainstones

Subaerialvesicular olivinebasalt lava flows(1-5m thick)with lateriticweathering

Calcarenite(beach)

High Ti/Feplagioclasebasalt

Volcanic island

Fringing reef

Packstones

The sulphur curve in this log showszones of high sulphur concentrationwithin the reef. These sulphur peakshave been interpreted as sulphate evap-orites. Experts have speculated that thelog shows sea level changes with lowstands marked by relatively high sul-phur content. Sulphur peaks at wave-lengths of 25 ft and 50 ft could be relatedto the Eocene low sea level stands (at36 M, 40 M, 42 M, 49 M and 54 M yearsago) recorded in the Vail eustatic sealevel curve. This example illustrates thevalue of geochemical techniques indetermining the geological history of asequence from raw element datarecorded direct from logs.

The karst (subaerial weathering oflimestone) features found in manyancient reefs underline the importanceof fresh water diagenesis, which causedleaching of less stable carbonate miner-als. Fresh water zones fluctuate throughtime as sea level rises and falls in rela-

tion to the atoll islands and the platform.Geothermally-heated fluids, which risealong crustal fracture and fault systems,mix with cooler seawater and fresh waterlenses which are scattered within upperparts of the atolls and the platform.

Permo-Triassic exotics - large rockmasses which have been structurallyemplaced into a sedimentary sequence -were thrust onto the slope/shelf at theeastern edge of Arabia during the Creta-ceous. These may be fragments ofancient atolls and platforms similar tothose in the Maldives. Although theexotics seen at the surface contain nohydrocarbons, there may be oil and gas-filled reef blocks beneath Tertiary shalesoff the east coast of Arabia. If thesedeeper blocks could match the produc-tivity of similar reefs in the Gulf of Mex-ico, they would represent a major newexploration target in the Middle East.However, the overpressured Tertiaryseal, which may also be the source rock,is a serious obstacle to deep drillingoperations.

Sorting out the salt

The salt dome structures which underlieprolific reservoirs in The Gulf - includingthe Permo-Triassic Khuff carbonates, theJurassic grainstone reservoirs and Creta-ceous grainstone/reefal reservoirs -sometimes form islands, but have verylittle in common with the ocean atolls.Formations overlying the salt domes areclosely interlinked and continuous withformations which surround the dome.Hydrological studies suggest that thesereefs often form part of regional aquifers.The numerous evaporite layers found inPermo-Triassic and Jurassic formations,and the thick clay-rich shales depositedduring phases of low sea level, are wide-spread and form effective seals: ideal forthe development of giant oil and gasreservoirs.

Middle East Well Evaluation Review

The Egyptian experience

The Belayim carbonate facies, and theequivalent Gemsa Formation, developedalong Egypts Gulf of Suez as scatteredand separate carbonate deposits. In thenorthern part of the Darag Basin, theBelayim carbonates were deposited in avery shallow marine sabkha environment.In the west central part of the Gulf of Suez,in the Ras Gharib and Ras Fanar fields,the Belayim was deposited as a reef com-plex. In the southern Gulf of Suez thesecarbonates are represented by reefal lime-stones, such as those in Gemsa Field, andalso sabkha carbonates.

Structural factors control the distribu-tion of the Belayim carbonate facies,which were deposited on tilted anderoded pre-Miocene fault blocks (figure1.7). These fault blocks developed duringthe opening of the Gulf of Suez and areformed by NW-SE oriented Clysmic faults.Fractures associated with the structureshave enhanced secondary porosity, per-meability and hydrocarbon potential.

Fig. 1.6: ALL SHOOK UP: The dolomite replaces limestone layer-by-layer and itssmall-scale distribution through the reservoir rock sequence is difficult topredict in this Miocene reservoir sequence in the Gulf of Suez. Only detailedgeochemical logging techniques can provide a quantitative view of dolomitediagenesis in this complex sequence; a mixture of calcite, dolomite, quartz,pyrite, glauconite and two types of clay.

Diagenesis, dissolutionand dolomite

Porosity evolution in rocks is a com-plex, but vital part of reservoir develop-ment and a clear understanding of thisprocess is crucial in the search for oiland gas. Porosity varies within rock lay-ers. Where the porosity of a reservoirlayer falls below a threshold, or cut-off,it ceases to be a viable reservoir. Thiscut-off value varies from reservoir toreservoir.

A picture of the porosity distributionin each reservoir zone depends on aclear understanding of reservoir geo-chemistry (figure 1.6).

Other elements in the porosity equa-tion are sedimentary geochemistry,pressure and temperature of burial, fluc-tuating sea level and changing porefluid composition. Reservoir analystsmust build a composite picture ofporosity, extrapolating and interpolatingdata between wells from the start ofdrilling.

Dolomite makes a difference

Dolomite mineralization can play amajor part in influencing reservoir prop-erties such as porosity and permeabil-ity. The conversion of pure limestones(CaCO3 ) to dolomite (CaMg(CO3 ) 2 ) is a gradual processwhich can start almost as soon as thecarbonate sediments have beendeposited. Dolomite crystallization iscaused by seawater interacting withfresh water lenses or pore water in car-bonate rocks. This dolomitizationprocess can take place in hypersalineponds where there are freshwaterlenses in the sediment or in coastallakes which are subjected to intenseevaporation.

While dolomites can be produced ina number of ways, the chemicalchanges involved do not vary. Magne-sium from the seawater replaces someof the calcium present in the originallimestone. The concentration of magne-sium in dolomite is much higher than inthe seawater from which it was derived.

Dolomite crystallization and dissolu-tion processes often control porositydevelopment in carbonate reservoirs.Early dolomitization can preserveporosity which might be lost by com-paction effects and calcite cementation.

Dolomitization often occurs as aresult of repeated sea level changes andthe mixing of hypersaline basinal brinesand normal seawater which accompa-nies these changes. At the same time,leaching of less stable skeletal compo-nents (aragonite) occurs, along the plat-form margins, increasing porosity. Dur-

ing long periods of rising sea level(marine transgressions) dolomite miner-alization may spread to carbonates atthe centre of the platform. The extent ofdolomitization, and its effect on reser-voir properties, depends on the volumeand salinity of the hypersaline brines.Large volumes of mouldic, vuggy andintercrystalline porosity can be createdby marine transgressions.

Changing sea level and water chem-istry also influence the composition ofcommon pore - filling cements. Calcitecementation is retarded because the cal-cium carbonate in solution is incorpo-rated into the precipitating dolomite.When calcite cementation is inhibited,the development of anhydrite cementsis the most important porosity-reducingmechanism. In some cases, both pri-mary porosity and early-generated sec-ondary porosity have been filled byanhydrite cements.

Calcite

Dolomite

MontmorilloniteQuartz

Oil

WaterKaolinite

Pyrite

x400ft

x500ft

Number 15, 1994.

Impermeable salt and anhydritewhich surround carbonate deposits inthe southern Gulf of Suez, for example atGebel al Fessayan, prevented hydrocar-bon migration into potential reservoirzones. Clearly, the position of salt andanhydrite layers is crucial in any evalua-tion of reservoir potential in carbonates.Miocene carbonate facies vary through-out the Gulf of Suez. The supratidalsabkha deposits at the northern end ofthe basin, around Ras Fanar Field and inthe Darag Basin, generally have poorreservoir potential. In contrast, the reefcomplexes of Nullipore facies found inthe Ras Gharib, Ras Fanar, Ras Bakr andGemsa fields are excellent reservoirs.

Fig. 1.7: The reef reservoirswhich developed in EgyptsGulf of Suez are found in avariety of positions along thetrend of Miocene fault blocks.

Fig. 1.8: Dolomite mineralizationdevelops gradually, spreadinggrain by grain through thereservoir rock. The progressivegrowth of dolomite crystalswithin a Miocene reservoir isshown in these microscopephotos (a-c). In the final stages ofdolomitization pore space can befilled by dolomite (d).

Abu Shaar el QibliZeit Bay Gemsa

Miocene Sea Level

Modern sea level

15-20mYounger reef

Older reef

Dolomite

Fig. 1.9: The interaction of seawater withfresh water (arrows) provided an idealenvironment for the replacement of calciteby dolomite after deposition of the olderreef. The upper limit of dolomitedevelopment coincides with maximum sealevel. The younger reef is presentlyundergoing dolomitization. Radioactivedating indicates the older reef cycle isbetween 350,000 and 270,000 years old.The younger reef cycle was depositedbetween 140,000 and 60,000 years ago(Strasser et al., 1992).

These images provided by Denise Stone, Amoco Prod. Res., Houston

(a) (b)

(c) (d)

Early dolomitization and subsequentdissolution of the dolomite crystals werevital steps in porosity development in Gulfof Suez carbonates. Other factors favouringthe development of high-porosity rocksincluded; skeletal aragonite dissolutioncoupled with late corrosion of anhydrite,and fine grained sediments. The porosity ofMiocene carbonates could reach valuesbetween 15% and 30% following deep bur-ial, and associated fracturing and late disso-lution of anhydrite cements, carbonategrains and even, in some cases, the rockmatrix. The corrosive fluids capable of alarge - scale, late - stage dissolution wereprobably associated with source rock mat-uration or basinal shale compaction.

Dolomite mineralization developsgradually (figure 1.8) and chemicalchanges can halt the process at anystage. Unfortunately, some of the majorfields (e.g. along the Shoab Ali Trendand the Kareem Formation of the ZeitBay Field) contain chalky microporosityand are only partly dolomitized. Marly-shaly units, which overlie potentialreservoir zones, probably kept thedolomitizing fluids out of the carbonate.

Modern dolomitization effects can beseen in the Gulf of Suez (figure 1.9)where freshwater from the surface andfrom the basement mix with seawater.

Middle East Well Evaluation Review

Dolomite close up

At any given depth, dolomitesequences seem to havegreater porosity than a lime-stone sequence. Most of thisporosity difference is due toporosity retention in thedolomite.

The factors which encourage dolomi-tization are reduced sulphate content inseawater (which typically occurs duringgypsum and anhydrite precipitation);dilution of seawater where the ionic con-centration is lowered while the molarMg:Ca ratio is maintained; raising of theMg:Ca ratio by evaporation; and temper-ature increases during burial. Climate isa factor, since dolomitization commonlyoccurs in arid depositional systems. Sealevel fluctuations also mix fresh waterand marine fluids in subsurface pore sys-tems - another cause of dolomitization.

Studies in the Khuff, Arab and Asmariformations and in Miocene carbonatesfrom the Gulf of Suez, indicate that all ofthe factors mentioned above played apart at some stage in the development ofdolomite in these major reservoirs.

In most Cretaceous reservoirs thedominant factors were groundwater andseawater mixing.

Pleistocene sea level fluctuations inthe Gulf of Suez and Red Sea are believedto be the main factors in stratigraphicvariations of reef dolomites, although theclimate and tectonics of any area willalways influence dolomite mineralization.

Age dating of dolomitized sequencesindicates that major cycles of dolomitedevelopment correlate well with the100,000 year cycle of eccentricity in theEarths orbit (Milankovitch cycles). Thiseccentricity affects the amount of solarradiation reaching the Earth and, therefore,has a profound effect on global climate. Cli-matic variations may, in turn, controldolomite development. The smallersequences which comprise the individualreef sequences are believed to be con-trolled by sea level fluctuations every21,000years, a cycle which relates to move-ment of the Earths axis. These smallerdepositional sequences have reefal andlagoonal facies which represent transgres-sive stages and coral rubble and siliciclas-tics associated with sea level highstands.

At the southern end of Sinai, aroundSharm el Sheikh, studies of Pleistoceneand younger reefs by Andre Strasser etal. (1992) underline the importance ofseawater and groundwater mixing in thedolomitization of the reefs and associ-ated sediments (figure 1.10).

The carbon and oxygen isotope val-ues in the older (Pleistocene) reefs (fig-ures 1.11 and 1.12) indicate a fresh waterinfluence on carbonate mineralization,whereas the younger reef samples showvalues typical of dolomite mineralizationin normal marine waters. Results fromthe Sinai reefs resemble findings fromPacific atolls and reefs in Latin Americawhere mixing zone dolomitization is themost important mechanism.

However, evidence for other mecha-nisms has emerged recently. Thermalpumping - hot water rising from depth tomix with seawater is the focus of currentresearch - while small-scale seawaterfluctuations, such as tides, may also pro-mote dolomite mineralization.

Thi

s im

age

and

thos

e op

posi

te w

ere

prov

ided

by

And

re S

tras

ser,

Inst

itut o

f Geo

logy

, Uni

vers

ity o

f Frib

ourg

, Sw

itzer

land

.

Fig. 1.10: Pointed aragonite crystals growing into pore space in the younger reef, Red Sea, southern end of the Sinai Peninsula.

Andre Strasser et al. (1992): Sequential Evolution and

Diagenesis of Pleistocene Coral Reefs (South Sinai, Egypt).

Sedimentary Geology, 78, pp. 59-79.

Number 15, 1994.

Fig. 1.11 (above): This Scanning Electron Microscope (SEM) image shows rhombs of dolomite growing over crystals of calcite which contain highconcentrations of magnesium. This rock is from the Pleistocene (older) reef.

Fig. 1.12 (below): Dolomite and high-magnesium calcite crystals are invaded by small needle-like crystals of aragonite. This change in the older reefis in response to changing water compositions as seawater and groundwater mix.

Middle East Well Evaluation Review16

2400

2425

2450

2475

2500

2525

2550

2575

2600

2625

Dep

th ft

0 25 50 75 100 0 25 50

Permeability(md)

Block calcite cement(%)

Fig. 1.13: NOW YOUSEE IT, NOW YOUDON'T: Permeabilitycomes and goes inKirkuk Field, Iraq. Thisplot of dolomite layersshows that the zones oflow permeability in thesequence correlate withlayers where late stagecalcite cements havedeveloped. It is difficultto reconcile thepresence of this calcitecement with burialdiagenesis. Fresh waterlenses preserved in thesediment seem themost likely cause ofthis patchy calcitecementation.

Developing dolomite

In the Cretaceous and Tertiary rocks ofthe Middle East there is patchy develop-ment of algal and foraminiferal lime-stone, with some coral and associateddetrital limestones. These were not con-nected with true fringing reefs, barrierreefs or reef banks. Some modern exam-ples can be seen in The Gulf today.

F.R.S. Henson, in his 1950 AAPGpaper, suggested the term reef-shoalsfor these small reefs which lack rigidfore-reef walls. He also recognized thatthere were massive rudist accumulationsmaking up banks which he observed out-cropping in north east Iraq (Upper Creta-ceous between Bekhme Gorge and Aqra,and Late Middle Cretaceous at Pir-i-Mugrun). Dolomite mineralization devel-oped in a variety of tectonic settings innorthern Iraq and in Syria (figure 1.12).

The Shuaiba reef of Bu Hasa Field inAbu Dhabi may be a rudist bank, butIbrahim Marzouk, Supervisor of Reser-voir Geology at the Abu Dhabi NationalOil Company (ADNOC), has indicatedthat wrench faulting may have affectedthe topography of the reef buildup.

Opportunities in Oman

Occidental has recently announced thediscovery of a Lower Cretaceous reser-voir in northwest Oman, near the flank ofthe Middle Cretaceous high situated westof Safah Field. This should lead to a re-evaluation of this reef-bearing region.

Reef facies were deposited aroundthe Kirkuk Field during both the earlyCretaceous and the Oligocene, alongwith Eocene nummulitic shoals. The min-eralogy of the reef facies was originallycalcite (limestone) but many zones werelater dolomitized. Recent studies indicatethat dolomitization of the Eocene bankwas related to falling sea level. Thisoccurred before development of themajor unconformity which precedesdeposition of the Fars evaporites.

Fig. 1.14: The porosity andpermeability of dolomite(green) and limestone(blue) samples from thecore indicate a closerelationship between theseproperties. The graphshows two dolomitetrends. One (a) follows thelimestone trend, showingthat dolomitization hadlittle effect on the poresystem. The second trend(b) indicates thatdolomitization led to ahigher permeability for agiven porosity. In this casethe leaching associatedwith dolomitemineralization hasimproved pore systemconnectivity. This coredata is taken from theCretaceous sequenceshown in figure 1.16.

0

Core porosity (pu)

10 20 30 400.01

.01

1

10

100

1000

10000

Cor

e pe

rmea

bilit

y (m

d)

(a)

(b)

Sea Level

Evaporites &associated deposits

Globigerinal chalks,marls & limestone

Reef, back-reef &fore-reef limestones

Sublittoral shales &marls

Sands,conglomerates, etc.

Tectonic sink A (Grabenat dep

th)}}}

BC Barrier or

fringing reefSubmerged highwith bank reef

Submerged highwith disconformity

A......Deposited horizontally before upliftB......Deposited during and after upliftC......Deposited during subsidence

Fig: 1.12: Dolomites developed in a variety of tectonic settings in Syria and northern Iraq.

Number 15, 1994. 17

Fig. 1.15: Dolomite - related porositytypically develops in one of twoways. In the early stages individualdolomite crystals appear in thelimestone matrix (a). Dolomitemineralization continues untilindividual crystals come intocontact (b), and a framework mayemerge (c). Alternatively, chemicalchanges may favour the dissolutionof the dolomite crystals and therock may develop leached mouldicporosity (d). If dolomitemineralization continues, thedolomite framework may provemore durable than the limestonehost. Dissolution of the limestoneleaves a dolomite framework withinter-crystalline porosity (e).

(a)

(b)

(c) (d)

(e)

Fig. 1.16: Mineral analysis of thisCretaceous carbonate reservoir inthe Emirates used geochemicaldata collected by a GeochemicalLogging tool (GLT). The bestporosity is found in the lower partof the sequence, but the highpermeability values correlate withdolomitization.

Core mineralogy, isotopic variationsand rock examinations suggest theremay have been seven stratigraphic dis-continuities caused by sea level fluctua-tions. Detailed mineral analysis indicatesthat cementation variations from one sealevel fluctuation to the next account forthe permeability variations found in thesequence. Changing sea level alsoresulted in the development of calcitecementation following the dolomitizationphase. The best permeability is found indolomite intervals where blocky mete-oric calcite cements have not developed.Data from the Kirkuk Field in northernIraq (figure 1.13) shows this clearly.

Dolomitization can have a very pro-found influence on permeability andporosity (figure 1.14). This makes a clearunderstanding of the process crucial toreservoir development. Dolomitization isa complex process and dolomite - relatedporosity can develop in two ways (figure1.15). Dolomite crystals appear in thelimestone matrix and, as dolomitizationcontinues, may coalesce to form a frame-work. At this stage chemical changes candissolve the dolomite leaving leachedporosity, or may dissolve the remaininglimestone, producing inter-granularporosity and high permeability in a puredolomite rock.

Studies from around the Middle Eastshow that dolomites retain their porositylonger than interbedded or associatedlimestones. There are a number of rea-sons for this, perhaps the most importantbeing less physical and chemical com-paction and reduced cementation associ-ated with dolomites. In mixed carbonatesequences dolomites often show thehighest permeability values (figure 1.16).However, shallow dolomite reservoirswith relatively high porosity values canhave lower permeabilities than grain-stones with similar porosities.

Susan Herron et al. (1992), Geochemical Logging of a

Middle East Carbonate Reservoir. Jour. Pet. Tech.

November 1992.

Oil

Moved HydrocarbonSecondary Porosity

Core permeability

Permeability(Core)

0.01 (md) 10000

Water

Calcite

Dolomite

SPI

0 (PU) 50

x200ft

x300 ft

Middle East Well Evaluation Review18

Subtle traps revealed inthe Middle East

Most of the giant anticlines and largereef reservoir bodies in the Middle Easthave been surveyed and drilled. Newreef and carbonate shoal reservoirs arelikely to be smaller than those in exist-ing fields, and will only be foundthrough careful processing andinformed interpretation of 3D seismicsurveys. In The Gulf region, many Creta-ceous reservoir zones are not dolomi-tized. Consequently, depositional char-acteristics are the most important factorin understanding oil and gas accumula-tions. Seismic surveys are thereforebeing evaluated for depositional charac-teristics as well as reservoir structure.

A team of seismic experts from Geco-Prakla/GeoQuest recently summarizedan integrated seismic processing strat-egy which can be applied in carbonate

exploration and reservoir characteriza-tion. Figure 1.17 (a to c) shows theirwork on a prograding carbonate plat-form and aggrading shoals similar tothose seen in northern Iraq and Syriaand northern and eastern Arabia. Thefirst step (figure 1.17a) is a preliminaryinterpretation of the structure, seismicsequence analysis and interpretation ofdepositional facies. The next stage (fig-ure 1.17b) produces a complete inter-pretation of depositional environment,using all available data from systemstracks and depositional sequences. Inthe third and final stage (figure 1.17c),synthetic modelling is carried out tocheck the interpretation and to give anindication of the geophysical risk factorsin the area.

Risk evaluation is a vital step in newexploration areas where the seismic,structural and depositional interpreta-tions are usually based on limiteddatasets.

Fig. 1.17: STEP BY STEP:After preliminaryinterpretation had beencarried out (a) theinterpreters brought togetherall existing data from systemtracks and sequences topresent an integrated pictureof the reservoir (b). This waschecked and the potentialrisks for developmentassessed (c) before anymajor productioncommitment was made.

(a)

(b)

(c)

Klaus Fischer et al. (1993) Remarks on Exploration Tools:

Integrated Exploration Strategy being applied to Carbonate

Environments. SPE Middle East Oil Show.

Basin

Shelf interior / marginal mounds

Shelf interior

Slope upper ramp

1.6 sec

1.8 sec

2.0 sec

2.2 sec

2.4 sec

2.0 sec

1.9 sec

5000 m 10000 m

Number 15, 1994. 19

Mar

k E

lliot

t, G

eoQ

uest

, Lon

don

Chr

isto

phe

M.

RUDISTS, REEFS AND RESERVOIRS

During the Cretaceous an aberrantgroup of large bivalves, the rudists,moved into the reef environment.These particular organisms filled thehigh-energy shoal so successfully thatmany geologists think of rudists asreef builders.

For 40M years rudists dominatedthe tops of shoaling highs and theedges of carbonate platforms. Theseunusual bivalves have one long cylin-drical valve hinged with a flat lid(figure 1.18). Rudists, like recentbivalves, filtered seawater for food.The elongate valve helped keep therudists feeding mechanism highabove the sediment-rich layer whichwould have clogged their food gath-ering system. This adaptationallowed them to feed almost continu-ously, stopping only when verystrong currents lifted muddy sedi-ment from the sea floor.

Rudists did not replace coralscompletely, they simply took overpart of the environmental nichewhich corals had exploited in thepast. As rudists moved into the envi-ronments where corals had been lesssuccessful, their shape evolved toovercome the soft mud problemswhich had faced the corals.

The hippuritids and radiotitidsformed the most striking of all in-place rudist congregations, with indi-viduals sometimes so denselypacked together that they resembled

colonial organisms (figure 1.19). Thesedense clusters were most common inquieter water. In high energy facies, thecaprinids were dominant.

The best reservoirs in the Creta-ceous are typically carbonate sandgrainstones or rudist shallow-marinecarbonate deposits. Of the latter, themost significant are the Middle Creta-ceous rudist facies which form banks,thickets and biostroms (fossil rich lay-ers). The rudists did not build reefs, nordid they form large bioherms, but theyare a vital component of many Creta-ceous reservoir rocks.

The best rudist reservoir facies arethose which contain a high proportionof skeletal aragonite (from Caprinid)shells. The leaching (dissolution andremoval) of aragonite, an unstable car-bonate, has produced important, sec-ondary porosity in the form of largevugs or cavities in the limestone. Thereservoir potential of a horizon is oftenenhanced if the aragonite intervals weresubaerially exposed after deposition.Increased porosity related to this typeof exposure can be seen in the Natihreservoir, in Oman. The leaching associ-ated with fresh water lenses during sub-aerial exposure is often high in the car-bonate reservoir, but not necessarily atthe very top of the sequence.

Fig. 1.18: IDENTITY CRISIS: It looks like acoral, but it's a bivalve. Rudists grew one longvalve to keep their filter feeding mechanismclear of muddy sediments.

Fig. 1.19: FAMILY TREE: The rudists, unlikecorals, were not colonial organisms. However,they normally crowded together to formmounds, with successive generations buildingon top of their parents.

Middle East Well Evaluation Review20

Fig. 1.20: This seismic line shows the facies changes which occur across the platform edge and into the basin. The sedimentary lobesdeveloped during the High Stand Systems Tract (HST) of the Natih e Member are remarkably clear on this seismic section. During thisperiod of high and stable sea level sediment was prograding from the NE towards the SW. This sedimentation was terminated by a sea leveldrop, creating a sequence boundary (SB).

Pictures of the prospect*Explorationists often have to deal withvery complex sedimentological andstructural problems in prospective areas.Their aim is to understand the detail ofreservoir variations, while drawing all ofthe information together into a compre-hensive picture of reservoir develop-ment and overall hydrocarbon potential.

Petroleum Development Oman (PDO)carried out an evaluation study of theSirat structure, making use of sequencestratigraphic techniques.

The Sirat Prospect, in the Natih For-mation of Oman, has been the focus ofintensive seismic and geological model-ling. This formation consists of stackedlimestone cycles separated by relativelythin shaly beds. The depositional envi-ronment of the Natih Formation has var-ied from deep water shales, with charac-teristic marine fossils such as ostracodsand planktonic foraminifera, to very shal-low marine packstones, grainstones andrudstones with abundant largerforaminifera and rudists.

In the upper part of the Natih e Mem-

ber a number of sedimentary lobesdeveloped (figure 1.20). These pro-graded from the shallowest parts of theshelf, building out to deeper water at theedge of the shelf. Maximum water depthduring this progradation was probablyno more than 100 m.

Each lobe contains a cycle of rocktypes changing from deep waterdeposits at the base to shallow sedi-ments at the top.

Sequence stratigraphy attempts toclassify sediments and sedimentarypackages by their relationships tochanging sea levels (rise, fall, rate ofchange) for local and worldwide (eusta-tic) changes. This allows us to define dif-ferent packages or sequences consist-ing of a Transgressive Systems Tract(TST), a Highstand Systems Tract (HST)and, in deeper areas, a Lowstand Sys-tems Tract (LST). Sequences are sepa-rated by sequence boundaries (SBs)created by sea-level fall. During times ofmaximum rate of sea-level rise, a Maxi-mum Flooding Surface (MFS) isdeposited. Sequences with their sys-

tems tracts and surfaces can all be recog-nized on seismic lines, giving vital cluesto the structure and likely compositionof sediments. Micropalaeontology pro-vides important, additional informationabout the sequences.

Sequences are ranked, according totheir importance and the type ofchanges which they represent. The 1st-order sequence boundary is moreimportant than a 2nd-order boundaryand so on.

The cyclic response from theGamma-Ray log has been used to definetwo 2nd-order sequences (Sequence Iand II) and a number of smaller 3rd-order sequences (figure 1.21). The top ofthe Natih e Member is identified as animportant sequence boundary. Theshorter period cyclicity defines the vari-ous members (a - g) which constitute theNatih Formation. In shallow areas onlythe TST and the HST are present. In thebasin at the southwestern end of theseismic section (figures 1.20 and 1.21) aLST developed.

*Taken from: - Sequence Stratigraphy and HydrocarbonHabitat of the Natih Formation in Oman. Presented by WytseSikkema (Petroleum Development Oman) at the 1993 AAPGInternational Conference, The Hague, The Netherlands.

SW NE

Number 15, 1994. 21

Fig.1.21: Sequence stratigraphic analysis of the area revealed two 2nd-order sequences (Sequence I and Sequence II) and a number of3rd-order sequences. Two maximum flooding surfaces have been identified and the top of the e Member is an important sequenceboundary. By correlating seismic lines with this analytical approach to sedimentary structures, experts can assess the structural history ofthe area and determine the risks associated with any given prospect.

The sequence stratigraphic recon-struction has a number of implicationsfor the prospectivity of the Sirat struc-ture. The prograding lobes of Sequence I,as seen on the seismic section, containexcellent reservoirs. Porosity and reser-voir permeability were enhanced by theexposure of the sediments whichoccurred during low sea level phases.This reservoir quality, coupled with theclear images available using seismictechnology, suggest that this would be anexcellent prospect. However the sedi-ments at the top of the Natih e levelwere deposited in a shallow environ-ment and are of poor sealing quality,thereby downgrading the prospect.

LST deposits are present but willprobably be low value reservoirs. Thesesediments normally contain a high pro-portion of fine clastic sediment whichwould reduce porosity. In addition, theabsence of rudist fragments suggests thatinitial porosity was low.

Three important reflectors relating tothe sequence stratigraphy can be seenon the seismic line: the maximum flooding surface ofSequence I (basal e Member) the sequence boundary between I andII (near top of e Member) the maximum flooding surface ofSequence II.

The seismic view

Explorationists integrated seismic lineswith well data in the sequence strati-graphic model, to reconstruct the depo-sitional environment of the SiratProspect. Deep water sediments occur aroundthe maximum flooding surface. The rela-tively deep limestone-shale alternationsare represented by a reflective seismicfacies containing a number of continu-ous, high-amplitude reflectors. The rudist accumulations are some-times visible as high-amplitude discon-tinuous reflectors. Thick deposits of shallow marine car-bonates appear as low amplitude trans-parent seismic facies.

The study concluded that despitethe excellent reservoir qualities of theHST lobes, the limited sealing capacityof overlying sediments made develop-ment of the Sirat Prospect a high-riskproject. A further conclusion from thestudy was that sequence stratigraphicmethods could be used to reconstructthe detailed sedimentary history of thearea and to predict the character of therocks in the sequence.

x300m

x500m

x500m

x300m

SB

mfs

mfs

Middle East Well Evaluation Review22

Cyclic sequences

The Middle Cretaceous is one of themain hydrocarbon producing horizonsin Oman and offshore Dubai. Sedimentssuch as the Middle Cretaceous NatihFormation were part of a Mesozoic plat-form carbonate succession, accumulat-ing around intrashelf depressions on theeastern edge of the Arabian peninsula.To the north west, in the Emirates, theequivalent reservoir rocks are known asthe Mishrif Formation (figure 1.22). Elfhas recently discovered oil in the sameformation offshore Qatar.

The Natih limestones are separatedfrom the deeper Shuaiba reservoir car-bonates by the Nahr Umr Shale. This,and the Fiqa Shale which overlies theNatih Formation, act as regional seals.

The Natih Formation is cyclic, com-prising a succession of coarsening-upward sequences. Each cycle consistsof deep marine shales and mudstonesgrading up to shallow marine rudistpackstones and grainstones. Emergencesurfaces occur at the top of each cycle.The cyclic sequence was caused byeustatic sea level changes, although itappears that deposition of the Natih For-mation was halted by tectonic uplift.

Away from the local highs, typifiedby shallow water deposits, the lime-stones interfinger with two deepermarine shales. These have significantorganic content and a rich fauna ofplanktonic foraminifera. The cycles haveformed the basis of a scheme of subdivi-sions (members labelled a to g) forthe Natih Formation.

Regional uplift during the Jurassiceffectively reduced average sea leveland led to the deposition of evaporitesover much of the Arabian carbonateplatform. In some areas uplift raised thesediment above sea level and there isevidence of subaerial erosion. Compres-sion, as the Arabian and Eurasian plateswere forced together, caused rapid sub-sidence along the plate boundaries. Thiswas followed by the spread of transgres-sive seas across the Arabian platform(figure 1.23).

On two more occasions during theCretaceous, uplift pushed topographichighs to a position where they wereeroded. Both phases were followed byrapid subsidence and shale deposition.The transgressive seas which developedafter these events became areas ofdeposition for the three main carbonatemegasequences which covernortheastern Arabia namely; theThamama/Kahmah, Wasia and Arumagroups. Each megasequence contains

NW SEDubai Oman

70

80

90

100

110

MillionYears

?

?Simsima

Shuaiba

Fiqa shale

Juweiza

MutiHalul-Ilam

Laffan

MishrifKhatiyahAhmadi

Mauddud

Nahr Umrshale

ac

b

efg

Aru

ma

grou

pW

asia

gro

upU

pper

Low

erC

reta

ceou

s

Nat

ih

Eroded

Eroded

??

?

?

?M

iddl

e

Fig. 1.22: TIMEZONES: The studyof rudistassemblages andthe discovery ofammonites withinthe Natih Formationhave provided aprecise correlationof time lines withinthe sequence.Correlationbetween outcropsections allowedexplorationists todevelop aconceptualsequencestratigraphic modelwhich includes thesubdivisions (a to g) used insubsurface studies.

Fig. 1.23: PRIME SITES: The best locationsfor rudist buildups were in the shelf settingas shown in this Middle Cretaceous map ofThe Gulf area.

numerous depositional cycles (3rd-order or parasequences) related tosmall-scale sea level fluctuations.

Understanding the cycles, and defin-ing which areas were most suitable forreef and shoal development, is an essen-tial part of the interpretation. Thesedepositional factors control the natureand location of the Cretaceous carbon-ate reservoirs.

The best reservoirs are generallyfound in the upper part of each megase-quence. This is due to upward shallow-ing, the abundance of coarse grain car-bonate particles, leaching caused bysubaerial exposure and the presence ofparticularly effective seals immediatelyabove the uppermost carbonate units ineach megasequence.

Moderncoastline

BasinShelf

Number 15, 1994. 23

Fig. 1.24: Typical mottled fabric of the UpperCretaceous Ilam carbonate reservoir faciesrevealed by FMI imagery and in core (inset) fromFateh Field.

Fig. 1.27: Karst surfaces seen in borehole imagesand core within the Mishrif reservoir provideclear evidence of repeated subaerial exposure.

Fig. 1.26: Electrical imagery and equivalent MishrifReservoir core sample (inset) of a rudist shoalingbuildup in Fateh Field. An arrow indicates the sealmark of the MDT pressure probe, the tool used todefine reservoir pressure and permeability. A styloliteseam can be seen just beneath the seal mark.

Fig. 1.25: FMI image and core interval (inset) ofthe unconformable contact at the top of theMiddle Cretaceous Mishrif reservoir. This unit isoverlain by Upper Cretaceous Laffan Shale.

Mishrif reservoirs

Rudist reefal-shoaling deposits com-prising the Cretaceous Mishrif Forma-tion, which is partially equivalent tothe Natih Formations of Oman, are themajor reservoirs in many fields inDubai and eastern Abu Dhabi. Thedomal Fateh Field is the largest off-shore Mishrif-age field in the Emirates.It was discovered in 1966, despite theabsence of Mishrif rocks from the dis-covery well - the result of pronouncedpost-Cenomanian erosion on the crestof the structure. Typical structures andfabrics from wells in Fateh Field areshown in figures 1.24 to 1.27.

Other fields in the region, includingthe Shah Field in Abu Dhabi and theAwali Field in Bahrain, are character-ized by erosion of Mishrif and equiva-lent rocks.

Fluid inclusion data, maturation cal-culations and burial history modellingindicate that cementation by blockycalcite crystals and oil migration hap-pened about the same time, between theLate Miocene and Early Eocene. TheKhatiyah Shale, which lies directlybeneath the Mishrif, is believed to be themajor source rock for these reservoirs.

A number of depositional cycles,locally bounded by erosional uncon-formities, have been identified by geol-ogists of the Dubai Petroleum Com-pany. These unconformities arebelieved to have been caused byglobal sea-level fluctuations and upliftof the deep, Eocambrian Hormuz Salt.

The combination of sea level falland uplift probably led to the develop-ment of new, tectonically-controlledislands and erosion of these structures.Anticlinal fold belts and deep-seatedsalt deposits were raised to the surfaceof the Cretaceous sea which coveredmuch of the Middle East. Havingreached the surface, they were sub-jected to the mixing of fresh water andmarine water. The results of this mix-ing process can be seen along themountain fronts from Turkey to Oman.

Subsequent transgression over thesubaerially exposed islands was asso-ciated with deposition of the LaffanShales which seals the Mishrif andother, slightly younger, Cretaceousreservoirs.

x043.0 ft

x047.0 ft

x317.0 ft

x324.0 ft

x343.0 ft

x340.0 ft

Middle East Well Evaluation Review24

Evaluating variation

Examination of drill cuttings, core andwell logs reveals the vertical variation incarbonate reservoir sequences. Inte-grated studies of reservoir behaviour,particularly when these involve majorwaterflood projects, highlight the lateralvariations present in all reservoirs. Insome giant and supergiant reservoirsunderstanding the lateral variations hasnot been a priority. These variationswere not considered a problem since theflow rates were outstandingly high andstandard porosity well logs suggested lat-eral variations were not significant.

In addition to the hidden complexitiesin some major reservoir zones, there aremany zones with lower reservoir poten-tial, whose development has beendelayed until now. These include chalky,high-porosity but low-permeability zonessuch as the Hanifa and certain Thamamareservoirs. These zones must beappraised carefully and new techniqueshave emerged to meet the challenge.Recent 3D seismic surveys have indi-cated many more faults than previouslyseen, and 3D borehole imagery in highlydeviated and horizontal wells is providinga wealth of fracture data.

The role of fractures, either helping orhindering oil production, has been exam-ined in detail (Middle East Well EvaluationReview, Number 14) and this knowledgecan be applied where fracturing affectsthe reservoir zones. Careful interpretationand integration of results indicates thatmany of the simple structures mappedover Gulf salt domes and in fold-belt anti-clinal reservoirs are actually more compli-cated than early models suggested. On amore positive note, better models of com-plex structures should reveal oil-filledreservoir compartments and reservoirfacies on the flanks of existing fields. Thiswill offer new exploration opportunities.

1600

1200

800

400

00.01 0.1 1 10 100

Pore area (cm2)

Fre

quen

cy

Fig. 1.28:UP THEWALL:Canadianreef wallcontainingmoulds ofleachedstromatopo-roids. Thesizeanalysissummary isshown infigure 1.29.

Fig. 1.29: VUGS AND MOULDS: Pore sizefrequency distribution measured from largearea photos of reef. Blue areas indicate thenumber of whole vugs while the greenshows the number of edge vugs. (From McNamara et al. 1991).

R. Ehrlich (1971) Relative permeability characteristics of

vugular cores - their measurement and significance.

SPE Annual Meeting Paper 3553.

Once a reservoir engineer has char-acterized total reserves or storage capac-ity in a reservoir, the emphasis switchesto production. Well logging and coreevaluation often provide sufficient datato determine the porosity of carbonatereservoirs. The next step is to analysevertical and lateral variations in perme-ability. This type of information, pre-sented as an integrated model of perme-ability distribution, is essential for effi-cient production. However, secondaryporosity in carbonate rocks complicatesthis modelling process.

First stop - secondary porosity

Secondary porosity is not of secondaryimportance in reservoirs. In fact this type ofporosity, created after the reservoir rockhas been buried, has often proved the mostdifficult to quantify and the most importantfor reservoir development.

Secondary porosity has caused prob-lems in the majority of carbonate reser-voirs. Even the biggest grainstone reser-voirs, with intergranular porosity similar tothat found in sandstone, can exhibit a sur-prising range of secondary porosity. This isoften developed in the form of intercrys-talline pores, vugs, moulds of leached shellmaterial and micropores which may be nomore than a few microns in diameter.

The abundance of micropores makescarbonates difficult to evaluate accurately.They are not visible to the naked eye, oreven under a standard microscope. Thevery high magnifications possible with ascanning electron microscope (SEM) areusually necessary for accurate estimates ofsecondary porosity. The size of the micro-pores means they are generally filled withnon-moveable water, while larger pores inthe same rock contain varying proportionsof water and oil. Consequently, it is possi-ble to produce oil, without water cut, froma carbonate reservoir interval which con-tains more than 50% water.

Many new techniques are available formicropore imaging. Dielectric measure-ments, nuclear magnetic resonance andStoneley wave sonic energy have beenintroduced in recent years. At the sametime, computer modelling of 3D boreholeelectrical imagery is improving the defini-tion of large vugs and moulds which char-acterize some reservoirs.

Going for the vugular

Carbonate research projects indicate thatevaluating porosity is difficult but deter-mining permeability is impossible in thepresence of large vugs and moulds (figure1.28) - even when whole diameter coresare used. Ehrlich, in his studies of carbon-ate permeability, concluded that no corewould be large enough to represent thefull extent of interconnection in the poresystem. Thus, whole diameter core oranalysis of 3D borehole images must beverified using down hole well testing tech-niques or drill stem tests. This approach isthe only way to improve our understand-ing of large scale interconnectivty and pro-ducibility in vuggy zones.

Recently, researchers (McNamara et al.1991) at the University of Calgary, Canadafound that porosity defined by core analy-sis alone could be 30 % lower than theactual value (figure 1.29). However, sucherrors in evaluating vuggy or mouldicporosity are unavoidable in cases wherethe size of the vugs is comparable withcore diameter.

Pr o g r e s s i v e s o l u t i o n

Initial particle

Mould

Solution-enlargedmould

Vug

Number 15, 1994. 25

Fig. 1.32: MAKING THEMOULD (AND THE VUG):When a particle, sedimentarygrain or organic fragmentdissolves from thesurrounding rock matrix itleaves a mould. If dissolutioncontinues the original shapeof the mould is lost,producing a vug.

Fig. 1.30: (above) Analysis of themould population in a short section ofcore can be carried out using theFormation MicroScanner (FMS)* tool.This example, from offshore Bombay,India, shows a plot of vug size andarea, giving an indication of vugdensity in the rock sample.

Fig. 1.31: (left) Concentrations ofcoral/algal moulds and vugs in thisTertiary coralgal boundstone fromoffshore India have been revealedand quantified using FMS imageryand core by N.R. Devrajan and R.S.Iyer of The Oil and Natural GasCommission of India.

Detecting vugs and moulds

The petroleum industry devotes a lot oftime to mould and vug evaluation (figures1.30 and 1.31). However, strict definitionsof moulds and vugs are often ignored andusing the two terms synonymously canlead to confusion.

Moulds are pores formed by the selec-tive removal, normally by solution, of anexisting rock particle such as a shell frag-ment, crystal or grain. The resulting poros-ity is referred to as mouldic porosity and isdescribed according to the type of particleremoved; e.g. oomouldic for an oolitic rockwhere ooids have been dissolved.

If the leaching of the original particle goesbeyond the point at which it can be identi-fied the hole is a referred to as a vug (figure1.32). The condition of the hole, not its size,determines whether it is a mould or a vug.

The authors of the basic reference oncarbonate porosity, Philip Choquette andLloyd Pray, suggested that a vug which islarge enough to be examined from theinside should be referred to as a cave.They also defined micropores as thosewhich have a diameter or cross-sectionwhich averages less than 1/16 mmwhether the pores are equidimensional,platy or tabular.

The full capabilities of the ModularDynamic Tester (MDT)* tool include thedefinition of vuggy reservoir zones whichcannot be characterized by core or bore-hole imagery even when combined withother well logs. Even the RFT tool has lim-ited applications for vuggy intervals. Testsoften fail due to lack of seal or the pres-ence of a tight patch resulting in a drytest. Fractures in low porosity patches fur-ther complicate the situation. However,the MDT tool has inflatable packers whichcan be placed above and below the vuggyzone to isolate it. The zone can be definedby FMI/FMS tools or core data.

While testing vuggy zones the MDTtool can be configured to include a con-ventional probe and an inflatable packermodule. The tool can then provide probemeasurements, and allows the operator touse the inflatable packers when a seal isnot possible in the best fractured orvuggy interval.

The MDT tools pumpout module canbe used for the dual packer approachwhich often succeeds where RFT attemptsfail. Packer spacing can be set to matchthe small intervals defined in FMI/FMS (orUBI in oil-base muds) or core data, to aminimum of 3 ft. This minimum size actu-ally provides a surface area thousands oftimes greater than the standard RFT orMDT probe. In this respect it can bethought of as a small-scale DST- type testwhich provides a pressure buildup with aradius of investigation just under 100 ftinto the formation. This figure varies withthe pore system in the formation.

x752 ft

x753 ft

x751ft

x373.4 ft

x373.0 ft

Middle East Well Evaluation Review26

Evaluation of isolated zones is nor-mally achieved by pressure tests. How-ever fluid samples for evaluation can betaken from vuggy zones or even low per-meability or thin bed intervals. This ben-efit is derived from the large seal andsample area created by the dual pack-ers. Tough sampling situations requireuse of the MDT tools pumpout module,fluid analyzer and sample throttling; anapproach which relies on the tools mod-ular design.

Since pressure and fluid content read-out is done at the surface, the test needonly continue until the formation fluid isdetected. This appears after the flow ofdrilling fluid which invaded the forma-tion has been pumped out. This type ofarrangement can replace the moreexpensive drill stem test and offers ahigh degree of safety. The MDT tool hasbeen used for production testing forwells with high hydrogen sulphide (H2S)concentrations.

Revealing reservoir permeability

Measured slowness, derived from lowfrequency Stoneley waves, can be usedto evaluate the permeability of hydrocar-bon reservoirs. At low frequencies theStoneley wave produces fluid flow whichis related to the connectivity of porespace. By comparing observed slownesswith elastic slowness computed for a for-mation with no fluids we can calculatepermeability (figure 1.33). Elastic slow-ness is calculated using three factors whichhave a direct effect on Stoneley wave prop-agation- formation density, borehole fluiddensity and shear slowness.

Stoneley attenuation provides analternative to permeability estimatesbased on slowness. In permeable forma-tions Stoneley waves are attenuated byfluid moving in the pore space (figure1.34) to a degree proportional to fluidmobility in the formation. From thisvalue engineers can derive the qualityfactor, Q (inverse attenuation), which isdirectly related to reservoir permeabil-ity. The calculation used to derive per-meability from the quality factorinvolves values for pore fluid and bulkelastic moduli, and for porosity andborehole diameter.

The technique was tested on a datasetcollected from a high porosity, pure car-bonate reservoir in Saudi Arabia. Theslowness and attenuation techniqueswere applied to data gathered using aDipole Shear Sonic Imager (DSI*) tool.

The predicted permeability was mod-ified to simulate a synthetic flowmeterprofile. Agreement between slowness-derived permeability and the flowmeterprofile was very good.

Squirting flow

Pore throatto pore

Edge to centre

Crack lubricationfacilitatingfriction

Biotic fluid flowwith boundaryshear

Fig. 1.34: SLOWING THE FLOW: Thereare several mechanisms whichcontribute to the attenuation of shearwaves. The lubricating effects ofliquids in cracks absorb energy. Themechanisms involved are fluid flowwith boundary shear effects andsquirting flow (which occurs whenfluids are forced through narrow porethroats between grains). By measuringenergy absorption we can estimaterock permeability. From Johnston,Toksoz and Timur (1978).

Fig. 1.33: Stoneley permeability values can be calibrated by RFT tool testsand confirmed by core data where available. The MDT tool provides agreater range of permeability than the RFT tool used here to calibrate theStoneley energy-derived permeability. The lithology/porosity/fluid columnis an ELAN-computed result.

x950ft

x000ft

x050ft

M. Petricola and B. Frignet (1992) A Synergetic Approach to

Fracture and Permeability Evaluation from Logs. 5th Abu

Dhabi Petroleum Conference.

D. Johnston, M. Toksoz and A.Timur (1978) Attenuation of

Seismic Waves in Dry and Saturated Rocks. Geophysics 44.

Permeability(Core)

0.01 (md) 1000

Permeability(Stoneley)

0.01 (md) 1000

ShearSlowness170 (us/f) 90

StoneleySlowness

250 (us/f) 200

Permeability(RFT)

0.01 (md) 1000

Thick deposits of shallowmarine carbonates appear aslow-amplitude transparentseismic facies.

Number 15, 1994. 27

Imagery + Stoneley analysis + OH Logs Reservoir data for well testing strategy and optimum MDT tool configuration

(FMI) (DSI) (ELAN)(UBI) Lithology(ARI) Porosity

Saturation

Permeability progress

Permeability measurements in carbon-ate reservoirs present a major challengeto well logging analysts. A group ofexpert analysts and geophysicists inDubai, Abu Dhabi, Egypt, Saudi Arabiaand at the Oil and Natural Gas Commis-sion (ONGC) - Schlumberger JointResearch Council in India, have testedcarbonate reservoir permeability usingStoneley wave data. Present efforts areconcentrated on sample analysis andStoneley frequency using the DSI toolwhich samples at lower frequencies thanthe earlier Array Sonic tool (figure 1.35).

The permeabilities found in shoalingsequences, where coarse particles over-lie fine chalky facies with micropore sys-tems, have been characterized using theDSI tool. The tool has also found suc-cess in reservoirs where there are avariety of secondary porosity types.

RFT tool permeability data can beused to calibrate permeability profilesdefined by Stoneley wave data. In theexample, core permeability data fromone inch diameter plugs, taken at onefoot intervals, compare favourably withthe DSI tool and RFT tool profiles. How-ever, in carbonate reservoirs where thepore system is heterogeneous, thematch is often poor, despite accuratepermeability measurements.

This situation typically arises wheneach measurement relates to a differentrock volume. Whole core analysis isrecommended for permeability charac-terization in the heterogeneous poresystems found in many carbonatereservoirs.

The MDT tool has already succeededin defining pressure, permeability andfluid content within complex carbonatereservoirs in the Middle East. The tools

30

20

10

0

Sto

nele

y ob

serv

ed -

ela

stic

(s

/ft)

0 1000 2000 3000 4000 5000

Frequency (Hz)

3000md1000md

300md

30md

10md

100md

Fig. 1.35: At low frequencies the Stoneley wave produces fluid flowwhich is related to the connectivity of pore space (permeability). Thisplot shows the sensitivity of Stoneley slowness to frequency - in therange measured by the DSI tool - for a water-saturated sandstone.From Cheung and Liu (1988).

Electric power

Pump-out

Sample

Optical analyzer

Hydraulic

Single probe

Packer

Fig. 1.36: The MDTtool can definepressure,permeability andfluid content withincomplex carbonatereservoirs. The toolsmodular designallows the operatorto select the optimumconfiguration foreach task. The MDTtool is reliable inreservoirs, wherepermeability rangesfrom hundreds ofmillidarcies tohundredths of amillidarcy.

Thin layered porosity

Interwoven porosity

Isolated non-porous

Fractured porosity

Shale barriers & baffles

Isolated porosity

Porosity

Non-porous (or low-porosity)rock

Uniform high porosity

Layered porosity

Thin porous layers

Uniformly non-porous

Thin non-porous layers

modular design allows the operator toselect the optimum configuration foreach task (figure 1.36). The MDT tool isreliable in challenging reservoirs, suchas those where permeability ranges fromhundreds of millidarcies to hundredthsof a millidarcy. To devise a high-qualityMDT tool test we require informationfrom several sources (e.g. electricalimagery and Stoneley). Only by combin-ing data from several sources can we besure of maximizing test efficiency.

Fig. 1.37: The main types of carbonateporosity heterogeneity revealed byborehole imagery.

Heterogeneities defined by imagery

Middle East Well Evaluation Review28

0

100

200

300

400T

ime

(Mill

ion

year

s)

N

P

UCrI

J

Tr

P

C

D

S

O

0.7070 0.7080 0.7100

(after BP 1992)

E. C

ret

Late

Jur

assi

cM

iddl

e Ju

rass

icE

. Jur

assi

c

Valanginian

Barriasian

Tithonian

Kimmeridgian

Oxfordian

Callovian

Bathonian

Bajocian

Aalenian

Toarcian

Pliensbachian

145.6

152.1154.7

157.1

161.3

166.1

173.5178.0187.0

194.5

Sr seawater curve for the Jurassic(modified after Smalley et al, 1989)

87Sr/86Sr0.7065 0.7070 0.7075Ma Age

87Sr/86Sr0.7090

Ratio

Global seawater strontium curve

Fig. 1.38: NAME THE DATE: This simplified curve for global Sr isotope ratios in seawater illustratesthe principle of the isotopic dating methods. For a known 87Sr/86Sr ratio a vertical line can bedrawn. Wherever this line crosses the curve, the sample ratio matches the seawater ratio for thatparticular time. However, some isotopic ratio values occur at two or more places in the curve.When this happens age must be defined by alternative dating methods.

Fig. 1.39: This series of logs (a) from the UpperJurassic shows the Asab Oolite at well A.Strontium isotope dating indicates that rocks ofthe same age are also found in well B, but havebeen lost from the sequence at well C to thenorth east. The unconformable contact betweenthe Jurassic and Cretaceous beds in the third wellmarks a period of erosion or non-deposition.Core taken from this level (b) confirms theunconformity.

Chemical timing

Geochemistry is finding new applica-tions as a tool for explorationists andreservoir analysts. A few years ago,most geochemical surveys weredirected at identifying source rock. Thisled to new applications in maturationand migration studies. Today, laborato-ries are using petroleum geochemistryto tackle reservoir problems such asassessing heterogeneity.

Geochemical methods include deter-mining biomarkers in a sequence andthen using isotopes to fingerprint dif-ferent oils present in a reservoir. Recentstudies have investigated hydrocarbonvariation within reservoirs and clarifiedthe extent of compartmentalizationcaused by tar mats, shale barriers andsealing faults. This data is vital in estab-lishing models for development andproduction phases.

Rock geochemistry is especially use-ful where the reservoir rocks are notcomposed of the usual quartz, lime-stone or dolomite lithologies on whichlog interpretations are based. Problemscan even arise where mixtures of thesebasic lithologies are being investigatedfor basic formation evaluation. This hasencouraged the spread of geochemicalwell logging and core studies.

Well-to-well correlation can beenhanced by applying geochemical tech-niques to core or well log data. In thisway, we can identify geochemical varia-tions in major lithologies or the presenceof minor minerals in adjacent wells.

Log analysts who routinely use thelithology indications from density/neu-tron variations in simple lithology mix-tures are defining lithology by compar-ing a single element, hydrogen, to thebulk density of the formation. Interpre-tations from geochemical logs are basedon much more information.

At present, the gamma ray log is mostwidely used for correlation in carbonatesequences: the elements identified areuranium (U) thorium (Th) and carbon(C). Geochemical logging analyses use afurther nine elements for correlation.

Ocean chemistry

Ocean chemistry influences the compo-sition of minerals being deposited on thesea floor. However, the chemical com-position of the oceans varies throughtime and these variations control min-eral stability. Ocean chemistry is crucialin determining the proportions of arago-nite or calcite present on the sea bedand, consequently, in the accumulatedsediment which reservoirs contain. Thisproportion influences ultimate reservoirporosity and permeability.

Upp

er J

uras

sic

Low

erC

ret.

SW A B C70km105km

HabshanFm

HithEquiv

AsabOolite

LowerAsab

Ray

daF

mA

sab

For

mat

ion

NE(a)

Weighing the evidence

Isotopes are atoms of the same elementhaving different numbers of neutrons inthe nucleus and, therefore, differentatomic weights. The weight difference isimportant, and useful, because naturalprocesses such as evaporation, conden-sation and photosynthesis cause signifi-cant variations in the distribution of iso-topes within the various geochemicalcycles.

For example, the light oxygen iso-tope, 1 6 O, is concentrated inwater vapour when seawater evapo-rates. The 1 6O-enriched vapourtravels through the atmosphere towardsthe poles where it condenses and isincorporated in the polar ice sheets. Thedifferential evaporation of oxygen atomswhich occurs at the equator means thatthe 1 8 O / 1 6 O ratio inpolar ice caps is much lower than in sea-

(b)

Number 15, 1994. 29

1330

Core dataafter demagnetization

-50 0 50

Reverse Normal

Log data

NMRT SUMT

Normal

1335

1340

1345

Dep

th, m

Magneticcolourcode Normal

Reverse

0 1 2-1-2

Normal

water at the equator.These stable isotope ratios have varied

systematically over time and can, there-fore, be used to date rock samples andcorrelate sequences. The stable isotopicratios of strontium (Sr) and sulphur (S)are used in chronostratigraphic studies,confirming time gaps at unconformitiesand determining sedimentation rates.They can even be used to date diageneticevents such as dolomitization.

The principle of strontium dating relieson changes in 8 7 Sr / 8 6 Srthrough time and the assumption that theratio within seawater is uniform worldwideat any given time. Seawater curves forstrontium ratios have been plotted and cal-ibrated against the geological time scale(figure 1.38). This was done by analyzingthe Sr isotope ratio in carbonate and phos-phate from fossils of known ages.

Strontium dating, and correlations basedon strontium ratios, can be used when thereare few fossils and when biostratigraphiczonation is poor. Independent of facies andfossil occurrence, this technique can evenbe used to date evaporite sequences.

Very small samples are required (as lit-tle as 0.1 mg) to provide a reliable age,with uncertainty normally being +/-1 mil-lion years, or less. The technique can beapplied worldwide and, unlike fossil cor-relations, is completely objective.

While the benefits of this technique areobvious, there are some limitations.Weathering affects all isotopic systems,and Sr isotope ratios can be modified bycontamination from meteoric / mixingzone diagenesis and burial cements.Depending on the modifying mechanism,the 8 7Sr/8 6Sr ratios can beshifted towards values typical of younger,or older, rocks. If unaltered carbonatesamples are not available for Sr isotopestudies then adjustments must be made toaccount for sample impurities.

Brachiopods and belemnites, withtheir low - magnesium calcite skeletons,are little affected by diagenesis and thebest samples come from these and fromthe phosphates which make up fish andconodont fossils. Whole rock samples,with the exception of anhydrites, usuallygive less accurate dates since their8 7 Sr/8 6 Sr values have fre-quently been altered by diageneticprocesses.

The final problem occurs when one8 7Sr/8 6Sr value correspondswith two or more ages in the seawatercurve. Ratios recorded from rocks in theKimmeridgian have the same8 7Sr/8 6Sr ratio as found inBajocian rocks which are 10 M yearsolder. This problem occurs on both largeand small scales throughout the geologi-cal record. Isotope values recorded in theUpper Permian can be identical to thosein Cretaceous rocks, although there is lesschance of confusion between these units.Matching the isotope ratio to a position on

Fig. 1.40: POLE POSITION: Over geological time scales, the Earth's magnetic field switches polarity.These changes are recorded in rock sequences. This forms the basis of a new logging tool which cancorrelate polarity changes between wells, offering a high-resolution magnetic log. The magneticreversals shown here were revealed by combining data from the Nuclear Magnetic Resonance welllogging tool (NMRT*) with the induced field as measured by the Susceptibility Measurement tool(SUMT*). Modified from Arnaud Etchecopar et al., Oilfield Review, October 1991.

the curve is normally a problem only ifthe age of the sample layer is verypoorly constrained.

In 1991, the Abu Dhabi Company forOnshore Oil Operations (ADCO) carriedout an isotopic pilot study to resolvesome of the uncertainties in Jurassicstratigraphy. Early results were encourag-ing and the study expanded. Today, thedatabase consists of Sr isotope analysesfrom ooid grainstones, belemnites, limemudstones, anhydrites and bivalves.

Sr dating has provided evidence of adirect stratigraphic correlation betweenthe pelagic transgressive belemnite lagdeposit and the unconformity (a type IIsequence boundary) of the Jurassic-Cre-taceous contact and the intraclasticbelemnite horizon in the Asab Oolite (fig-ure 1.39a and b). This type of geochemi-cal correlation has helped to refine

regional stratigraphy. For example,anhydrites which had been included inthe lowermost Cretaceous Habshan For-mation, were re-assigned to the UpperJurassic while the Manifa Member(150.5 M years) was correlated with theAsab Formation (151.2 M years) and withthe lateral equivalent Qatar Formation(150.5 M years) using this technique.

A new logging technique (figure 1.40)which relies on reversals of the Earth'smagnetic field through geological time,has proved very successful for cross-well correlation. Rocks can retain themagnetization from previous magneticfields, a phenomenon called naturalremnant magnetism (NRM). The loggingtechniques which record this magneticmemory are very accurate and can beused worldwide. Absolute age correla-tions derived from reversals have beenmade between three wells drilled byTotal in the Jurassic sediments of the