A New Geochemical-Sequence Stratigraphic Modell

Two major (1, 2) and two minor (3, 4) petroleum Array systems dominated by terrigenous type III organicmatter are recognized. The highstand, lowstand 1,lowstand 2, and transgressive systems tract oilsaccount for about 46, 31, 15, and 8%, respectively,of the 61 oil samples and about 45, 32, 11, and 12%,respectively, of the estimated ultimate recoverablereserves from the fields represented by these sam-ples. These fields account for about 13 of the 16BBOE (billion bbl of oil equivalent) estimated ulti-mate recoverable reserves in the entire Kutei basin.(1) Waxy highstand oils (e.g., Handil, Nilam)occur mainly onshore in middle Miocene–Pliocenereservoirs. These oils originated from middle–upperMiocene coal and shale source rocks deposited incoastal-plain highstand kitchens now near the peakof the oil window.(2) Less waxy lowstand 1 oils (e.g., Perintis, Sisi,Ragat) occur offshore in middle–upper Miocene reser-voirs. These oils originated from middle–upperMiocene coaly source rocks deposited in deep-waterlowstand kitchens now mostly in the early oil window.(3) Lowstand 2 oils (e.g., Semberah 037) are sim-ilar to the lowstand 1 oils but occur mainly onshorein lower–middle Miocene reservoirs. These oilsgenerally are more mature than lowstand 1 oils andoriginated from lower–middle Miocene coalysource rocks.(4) Nonwaxy transgressive oils (e.g., Badak)occur mainly onshore in middle–upper Miocenereservoirs. These oils were generated at low ther-mal maturity from middle Miocene suboxic marineshales deposited near maximum flooding surfaces.Our three-dimensional geochemical-stratigraphicmodels for highstand and lowstand source rocksindicate that less fractional conversion of the kero-gen occurred than had been predicted by the gener-ally accepted stratigraphic model and classic type IIIkerogen kinetics; furthermore, two-dimensionalfluid flow modeling supports independent geo-chemical evidence for commingling of oils in theTunu field from highstand and lowstand kitchens towest and east, respectively. Finally, our model suc-cessfully predicted that oil and gas, rather than gas12AAPG Bulletin, V. 84, No. 1 (January 2000), P. 12–44.only, would be discovered at the recently drilled deep-water Merah Besar and West Seno fields. Geochemical analyses of oils from the Merah Besar field confirm that they belong in the lowstand 1 oil group.INTRODUCTIONTerrigenous-dominated marine deltas contain important accumulations of petroleum in many parts of the world. For example, the Mahakam Delta was one of the first locations where terrigenous organic matter in coals was suggested to have gen-erated and expelled liquid hydrocarbons (Durand and Paratte, 1983; Huc et al., 1986). After many years of debate, a general consensus has developed that deltaic coals can generate large quantities of oil and gas (Law and Rice, 1993; Scott and Fleet, 1994). Major oil accumulations have been linked to hydrogen-rich coaly source rocks, especially in the Upper Jurassic–Paleogene of Australia and New Zealand and low-latitude (<20°) Tertiary deltaic coastal deposits throughout Southeast Asia (Scott and Fleet, 1994, and references therein).The prolific Kutei basin in Kalimantan, which contains the Tertiary Mahakam Delta, has estimated ultimate recoveries of between 11 (Paterson et al., 1997) and 16 billion bbl of oil equivalent (BBOE) (IHS/Petroconsultants S.A., 1998). The first oil in the Mahakam Delta was discovered about 100 yr ago (1898) in the Louise 1 well located on the western-most, inboard anticlinal trend (Figure 1). SeveralPeters et al.13 Figure 1—Map shows approximately north-south anticlinal trends (elliptical shapes) and locations of oil samples in the Mahakam-Makassar area, Kutei basin. Inset at upper left shows location of study area in eastern Borneo, Kalimantan, Indonesia. Heavy stippling indicates oil trends and light stippling indicates gas trends within the anticlines. Genetic groups with symbols defined in the inset at lower right are based on statistical analysis of multivariate geochemical data (Figure 7). Stars indicate location of the Louise 1 and Panca 1 wells.giant oil and gas fields subsequently were discov-ered in this and other approximately north-south–trending anticlines. The offshore Attaka field (Figure 1) was discovered in 1970 and was the first commercial field in Kalimantan with esti-mated ultimate recoverable reserves of 1173 mil-lion barrels of oil equivalent (MMBOE). Other important fields include Tunu (4774 MMBOE), Handil (1549 MMBOE), Badak (1393 MMBOE), Nilam (1069 MMBOE), Tambora (348 MMBOE), and Bekapai (323 MMBOE) (IHS/Petroconsultants S.A., 1998).Despite the economic importance of Tertiary deltas, understanding of petroleum systems (Magoon, 1997) and the regional distributions of source rock organic facies in these settings is limit-ed. Most deep-water marine source rocks are widespread and consist of thin intervals (tens to hundreds of meters) with fairly uniform organic matter. Conversely, most deltaic marine source rocks are thick (hundreds of meters or more), con-tain complex mixtures of terrigenous and marine organic matter in coals and shales, and show greater source rock spatial and temporal variability. The oil-generative character of coaly rocks is diffi-cult to characterize by Rock-Eval pyrolysis (Peters, 1986). Our understanding of the relative impor-tance of coals vs. shales as sources for oil and gas also is incomplete. In addition, the expulsion effi-ciencies of coaly rocks are difficult to assess due to their tendency to adsorb hydrocarbons (Saxby and Shibaoka, 1986; Powell and Boreham, 1991; Sandvik et al., 1992).The source input and diagenetic processes that control the oil potential of coaly rocks also are poorly understood. Coal maceral composition does not clearly indicate oil potential because not all bio-logical macromolecules believed to contribute to kerogen have microscopically identifiable struc-tures (Tegelaar et al., 1989; Tegelaar and Noble, 1994). Amorphous, hydrogen-rich material may explain the oil potential of vitrinite-rich coals, par-ticularly those rich in desmocollinite (Clayton et al., 1991) or suberinite (Khavari Khorasani and Michelsen, 1991). Other workers suggested that morphologically distinct hydrogen-rich liptinitic mac-erals (e.g., cutinite) or desmocollinite (hydrogen-rich vitrinite) are key precursors for liquid hydro-carbons (Khavari Khorasani, 1987; Snowdon, 1991; Garcia-Gonzalez et al., 1997; Horsfield et al., 1988). Evidence suggests that microbial reworking of humic material in deltaic settings may improve generative potential (Khavari Khorasani, 1987; Clayton, 1993).Despite these problems, the geochemistry of crude oils and source rock extracts can be used to identify and predict the regional distributions of deltaic source rock facies. Our use of high-resolution biomarker analyses allows us to identify system-atic differences among oils that are controlled mainly by source input rather than by thermal maturity, biodegradation, or other secondary processes. Prior to our work, compositional dif-ferences among oils from the Mahakam Delta generally were attributed to secondary process-es rather than genetic differences (e.g., Paterson et al., 1997). This underlying assumption of con-stant source composition extends to the kinetic parameters used to predict the timing of oil gen-eration for Mahakam Delta source rock. Previous models of generation in this area typically use the classic type III kinetics based on a single Mahakam Delta coal sample (Tissot et al., 1987), ignoring the possibility that oil types, organic facies, and source rock kinetics may vary both spatially and temporally.ObjectivesThe main objectives of this study of the Maha-kam shelf and Makassar slope were to define petroleum systems based on geochemical oil-to-oil and oil-to-source rock correlations. In addition, we measured kerogen kinetic parameters for hydrocar-bon generation in source rocks that showed affini-ties to the different oil groups and used these mea-surements in basin modeling to improve our understanding of the history of hydrocarbon gener-ation and migration. Finally, we integrated the geo-chemistry with geological-geophysical data to improve future exploration and evaluation of this region.SamplesTable 1 lists crude oil (N= 61) and key potential source rock (N= 14) samples collected from the study area (Figure 1). Nearly all of the oil samples analyzed are from middle–upper Miocene or Pliocene reservoirs in the Balikpapan or Kampung Baru groups (Figure 2). The prospective source rock samples range in age from early to late Miocene and originated from coastal-plain to deep-marine environments (Table 2). The Appendix describes the analytical methods used in the study.Samples may be described using abbreviated names. For example, oil from the Bekapai 4 well in DST 1 (2252–2256 m) is named B4 1. Field descrip-tions of samples as gas condensates or oils are com-monly unreliable. Thus all fluid petroleum samples in the text are called oils, even though some show high API gravities (∼50°API) that approach gas con-densate. Condensates generally show gravities higher than 55°API (Hunt, 1996).14Mahakam Delta and Makassar Slope, IndonesiaDISCUSSION Geological SettingThe study area extends from the onshore delta-plain in the Kutei basin to the slope adjacent to the Makassar Straits that separate Kalimantan on Borneo from Sulawesi (Figure 1). The Makassar area appears to be underlain by remnant oceanic crust trapped between relict subduction zones in northwestern Borneo and western Sulawesi (Malecek et al., 1993). The onshore portion of the Mahakam Delta overlies a series of tightly folded anticlines and broad synclines known col-lectively as the Samarinda anticlinorium, which resulted from inversion of the Paleogene basin (Chambers and Daley, 1995). Offshore Mahakam Delta areas show at least two phases of deforma-tion. Middle Miocene and older rocks exhibit compressional folding and thrusting, whereas the overlying upper Miocene–Pliocene strata are affected only by extensional faulting (Malecek et al., 1993).Regional cross sections demonstrate that Mio-cene and younger sediments fill a large depocenter offshore from the present Mahakam Delta and sug-gest that the current drainage system existed by the earliest Miocene (e.g., Paterson et al., 1997).Palynology indicates that the present-day tropical climate and dense rainforest vegetation have exist-ed since the Pliocene and perhaps as early as the Miocene (Caratini and Tissot, 1988).StratigraphyEarly efforts toward stratigraphic correlation of Miocene and younger rocks stemmed from hydro-carbon discoveries in the onshore Samarinda anti-clinorium. Limited biostratigraphic and low-quality seismic data necessitated the use of lithostrati-graphic correlation based mainly on well logs (e.g.,Huffington and Helmig, 1980, 1990). Zone names such as “Upper Main” and “Shallow” generally could not be extended beyond a field boundary (e.g.,DeMatharel et al., 1980; Verdier et al., 1980). WithPeters et al.15Figure 2—Generalized stratigraphic column (left) shows distribution of oil and rock samples (after Marks et al.,1982) and Neogene sequence lithostratigraphy, chronostratigraphy, and local zone names (right) of the Mahakam Delta and Makassar slope. Coastal onlap chart is from Haq et al. (1987).16Mahakam Delta and Makassar Slope, Indonesiaimproved seismic and biostratigraphic data, more regional marker names such as “Alpha” and “Beta”saw common usage in well reports and Indonesian Petroleum Association field papers (e.g., Sujatmiko and Irawan, 1984). These markers generally were compared to lithostratigraphic nomenclature (e.g.,“Balikpapan Beds,” later “Group”), suggesting that they also had limited chronostratigraphic signifi-cance (Figure 2).Total and the Institut Français du Pétrole (Total-IFP) produced the first significant chronostratigraph-ic correlation model for the Mahakam Delta after a 4-yr sequence stratigraphy study (Burrus et al., 1992;Duval et al., 1992a) (Figure 3, top). Integration of biostratigraphy, well logs, and seismic data from Total acreage identified a series of regional sequence-bounding unconformities and condensed sections, similar in age to those thought to have global significance (Vail et al., 1977; Haq et al.,1987). The early stratigraphic and geochemical work on the Mahakam Delta (e.g., Combaz and DeMatharel, 1978; Durand and Oudin, 1979;Verdier et al., 1980) provided the basis for a region-al hydrocarbon charging model (Duval et al.,1992b). These workers recognized the possibility of dramatic basinward shifts in sedimentation dur-ing lowstands and the deposition of thick carbon-ates at the shelf-margin break. The Total-IFPPeters et al.17Figure 3—Comparison of the Total-IFP model for the Mahakam Delta (top) (after Burrus et al., 1992; Duval et al.,1992a) with the new geochemical-stratigraphic model described in this paper (bottom). Location of XX ′is given in Figure 1. Highstand and lowstand kitchens are discussed in the text. Estimated ultimate recoveries (BBOE = billions of bbl of oil equivalent) refer to the corresponding fields (IHS/Petroconsultants S.A., 1998). Triangles indicate shelf break during specified intervals of deposition. White arrows indicate present-day direction of petroleum migration from local kitchens. Burial depth for the 10.5 Ma surface (dotted line) in the lowstand kitchen area of our model is about 3–4 km, depending on the two-way transit time vs. depth curve that is used (i.e., Perintis 1 vs. Sisi 1 wells), but is about 6 km based on the earlier model (top).sequence stratigraphic model was instrumental in successful efforts directed at reversing Total’s pro-duction decline, and initially included Tunu field reservoir characterization (Duval et al., 1992b), fol-lowed by later discoveries of the Northwest Peciko and Sisi fields.Although the Total-IFP sequence stratigraphic model represents a major improvement over previ-ous lithostratigraphic and log-defined correlation schemes, it has limited ties to the slope and basin, including the Makassar Straits. Complete calibra-tion of sequence-bounding unconformities requires downdip correlations. The Total-IFP model shows most of the shelf margins near or just seaward of Tunu field (triangles in Figure 3, top), which ex-plains why many workers believe that middle Miocene coastal-plain coal and shale source rocks occur mainly landward of Tunu field (Burrus et al., 1992). In the Total-IFP model, outer shelf and slope areas are distant from the source rocks and thus carry significant exploration risk; furthermore, the Total-IFP model does not explain deep-water hydro-carbon accumulations, such as in the Sisi-Nubi field trend on the outer shelf (Burrus et al., 1992).New Stratigraphic ModelIn our stratigraphic model (Figure 3, bottom), the interpreted shelf margins are located farther seaward than they are in the Total-IFP model, allow-ing for source rock deposition in less explored outer shelf and slope regions, and the source rocks show more favorable oil-window maturity. The new interpretation allows development of low-stand kitchens, i.e., thick depocenters seaward of the coastal plain, that are closer to the slope margin of the Mahakam-Makassar area than allowed by the Total-IFP model. Because the lowstand kitchens are downdip of the coeval shelf margin, the terrige-nous organic matter is believed to consist of trans-ported rather than in situ coals or coaly shales. Our model infers the 10.5 Ma horizon to be buried to only about 3–4 km depth, considerably shallower than the Total-IFP model (6 km, Figure 3). One rea-son for the difference is that the 10.5 Ma surface is a significant angular unconformity, which must be traced across dipping reflections rather than fol-lowing clinoforms downward as in the Total-IFP model. The inferred depth differences greatly impact both the expected thermal maturity of source rocks and the sandstone reservoirs. On the outer shelf, the middle Miocene is postmature based on the Total-IFP model, but much of it is within the oil window based on our new model.Ten depositional sequences were identified and mapped in the stratigraphic section of the Mahakam shelf and adjacent Makassar slope (Snedden et al., 1996). Standard sequence strati-graphic analysis allowed identification of sequence-bounding unconformities and maximum flooding surfaces. Seismic facies within sequences were cali-brated with available well (lithology) control. The sequences were dated using paleontological picks in well ties and compared to available coastal onlap charts (e.g., Haq et al., 1987). Several of these sequences lie below the distinctive 10.5 Ma sequence boundary (Figure 3, bottom). A major change is interpreted in tectonic style with compressional thrusting below and extensional normal faulting above this boundary.Deltaic depocenters identified by isochore map-ping of individual sequences show thick potential source kitchens near or at the shelf margin for sequences below the 10.5 Ma unconformity (Figure 4) (Snedden et al., 1996). Above the 10.5 Ma unconformity, the depocenters shift landward of the coeval shelf margin for each sequence. This shifting could be the result of tectonic differences before and after erosion, with the thick lowstand kitchens located seaward of the thrust fault zone. Alternately, this shifting could be the result of the global late Miocene–early Pliocene second-order sea level rise and transgression. These outer shelf and slope kitchens appear to have been filled by sediments rich in terrigenous organic matter dur-ing major falls in relative sea level and are referred to as lowstand kitchens. In contrast, highstand kitchens contain coals and organic-rich shales that accumulated landward of the shelf break during relative highs in sea level. The key difference between these two kitchen types from a sequence-stratigraphic standpoint is how the organic matter accumulated. Organic matter in highstand kitchens is largely in situ, whereas much of the organic matter in lowstand kitchens represents eroded coastal-plain detritus that has undergone signifi-cant transport.Source Rock IntervalsHighstand Systems Tract Coastal-Plain Coals The generally accepted source rocks for the Mahakam Delta crude oils are coals and shales from the Miocene Balikpapan Group (Schoell et al., 1983; Robinson, 1987) that have been buried deep-er than 2600 m (Perrodon, 1983). The Paleogene section is overpressured and believed to be too deep to contribute to petroleum accumulations. The post-Miocene section is not buried sufficiently to have entered the oil window. According to the Total-IFP model, the best source rocks consist of lower delta-plain coals (Duval et al., 1992a) that contain mainly huminite with minor exinite macer-als (Burrus et al., 1992). These rocks correspond to18Mahakam Delta and Makassar Slope, Indonesiachronostratigraphic sequences ranging in age from 15.5 to 5.5 Ma (Figure 2). Most workers believe that the middle Miocene coastal-plain coals and coaly shales (15.5–10.5 Ma) are the primary source inter-val in the Tunu kitchen (e.g., Burrus et al., 1992).The Miocene coastal plain was linked to the east-flowing Mahakam River system, which deposited sediments in a subsiding basin oriented north-south parallel to and near the delta front. The Balikpapan Group contains up to 175 m of cumulative coal and 1750 m of cumulative shale (Thompson et al., 1985). This progradational phase of the delta has continued, with transgressive interruptions, to the present day (Allen et al., 1976; Oudin and Picard, 1982). Various upper delta-plain to delta-front facies and occasional transgressive limestones were deposited during delta progradation.The distribution of coastal-plain sediments and organic matter is strongly influenced by tidal and river flow. Domed (ombrogenous) peats fed by rainwater form on the upper coastal plain and extend basinward as far as the intertidal zone, where they abut against beach ridges or mangrove swamps (Anderson, 1964). Reworking of coastal-plain peats in tidal flat or lagoonal environments results in selective separation of soluble and insolu-ble organic matter and accumulation of liptinite-rich allochthonous drift peats and carbonaceous shales (Allen et al., 1979; Thompson et al., 1985). These liptinite-rich peats can be further enriched by fresh liptinitic material, such as leaf cuticles and damar resin from resin-producing trees growing behind the coastal plain.Although most Balikpapan Group coals contain abundant vitrinite group macerals, the hydrogen indices of the coals are commonly over 300 mg hydrocarbon/g TOC (total organic carbon) and the oxygen indices are usually below 15 mg CO2/g TOC (Thompson et al., 1985). Hydrogen indices over 200 mg hydrocarbon/g TOC are necessary for significantPeters et al.19 Figure 4—Schematic of the new geochemical-stratigraphic model and predicted distribution of Mahakam-Makassar area source rocks. Faults, which provide inferred migration pathways from source rocks to reservoirs, are not shown. HST = highstand systems tract, TST = transgressive systems tract, LST = lowstand systems tract, MFS = maximum flooding surface.oil-generative potential from coals (Hunt, 1996). Geochemical oil-to-source rock correlation indi-cates that a large genetic group of our samples, des-ignated highstand oils, originated from highstand systems tract middle–upper Miocene coals and coaly shales (discussed in following sections).Lowstand Systems Tract Coaly Shales Paleogeographic mapping based on our new stratigraphic model and the different geochemical compositions of oil samples in the study (discussed in following paragraphs) suggest that source rock facies other than the well-documented middle Miocene coastal-plain coals and coaly shales exist in the Mahakam-Makassar area. In our model, ter-rigenous organic matter accumulates in depocen-ters seaward of the coeval coastal plain in outer shelf to slope settings during lowstand systems tracts. Enhanced erosion of coastal-plain sediments during rapid falls in relative sea level could provide large quantities of terrigenous organic matter to the lowstand kitchens (Figure 3 bottom, Figure 4). These downdip depocenters might receive terrige-nous organic matter by a process similar to that responsible for gravity-flow sandstones on the outer shelf and slope (Figure 5). In that process, sand is eroded from highstand deltas, transported through incised valleys, and deposited as deep-water gravity-flow sandstones.Few downdip wells penetrate marine sections with high terrigenous organic matter in the study area. For example, the best shale source rocks in the deep-water Perintis 1 and Sisi 1 wells (Figure 1) contain only 1 to 3 wt. % TOC; however, sampling bias is possible because most exploration wells are drilled on structural highs where less terrigenous organic matter may have been deposited. Low TOC in many prolific offshore basins (e.g., Sabah Basin) (Anuar and Muhamad, 1997) suggests that sampling bias may occur elsewhere. The best evidence for terrigenous-rich deep-water source rocks offshore20Mahakam Delta and Makassar Slope, IndonesiaFigure 5—Block diagram of the lowstand valley system in the paleo-Mahakam Delta that is inferred to have allowed the transport and preservation of significant terrigenous organic matter in deep-water settings.Mahakam Delta comes from observations of updip erosion and transport of terrigenous organic mat-ter. For example, cores of middle Miocene channel and incised-valley fills show that TOC and hydro-gen index can be high (Snedden et al., 1996). Seismic evidence for valley fills elsewhere in the delta supports this view.These terrigenous-rich lowstand depocenters generally lie above the 15.5 and 12.5 Ma sequence boundaries and include periods of extensive updip erosion, downdip subsidence, and deposition (Figure 4). Basin modeling (discussed in a following section) shows that two lowstand depocenters are within the window for hydrocarbon generation and thus are potential kitchens. The presence of two lowstand kitchens is supported by our geochemical data, which identify two distinct subgroups of low-stand oils. The lowstand 1 and lowstand 2 oils dif-fer in both source- and maturity-related geochemi-cal parameters, implying origins from source rocks of different age and thermal exposure. Our oil-to-source rock correlation indicates that the lowstand 1 and lowstand 2 oils originated from middle–upper and lower–middle Miocene source rocks, respectively.Transgressive Systems Tract Source Rocks Thick and laterally extensive coals and organic-rich shales can form during both highstand and lowstand systems tracts, given the proper balance of peat production and accommodation (Bohacs and Suter, 1997). Little algal organic matter is con-tributed to these sediments under conditions of high water turbidity, terrigenous organic matter input, and clastic deposition typical of the Maha-kam Delta; however, coals are rare during a trans-gressive systems tract because swamp sediments commonly are drowned before substantial organic matter can accumulate. Reduced water turbidity and clastic deposition can occur during the peak transgressions marked by maximum flooding sur-faces (MFS, Figure 4). Seismic evidence indicates several maximum flooding surfaces during the mid-dle Miocene, especially within the 15.5 and 13.8 Ma sequences (Figure 4).The geochemistry (discussed in a following sec-tion) of a small, but distinct, group of oils in our study suggests a transgressive marine-influenced source rock with significant algal input mixed with the terrigenous organic matter; furthermore, these oils geochemically correlate with extract from a source rock deposited in an inner shelf environ-ment as part of a transgressive systems tract (13.8 Ma). Although our highstand and lowstand oil sam-ples show genetic affinities for both coals and shales, the transgressive systems tract oils show affinities only for shales.Oil GeochemistryBulk PropertiesThe 61 oil samples show moderate to high API gravity (26–50°) and low sulfur (<0.1 wt. %, Table 1), typical of oils from Tertiary deltaic marine set-tings. The oils show high pristane/phytane ratios (Pr/Ph = 3.8–10.9) and stable carbon isotope ratios for saturated and aromatic hydrocarbons near –29 and –27‰, respectively. Many of the oils show a marked odd-to-even predominance of n-paraffins in the dominant nC27to nC33range. These character-istics typify oils generated from paralic or deltaic marine source rocks enriched in terrigenous organ-ic matter (e.g., Chung et al., 1992).The analyzed oils are broadly similar when com-pared using relatively insensitive source-related parameters. For example, pristane/nC17and phy-tane/nC18ratios (Pr/nC17and Ph/nC18, Figure 6) indicate that most of the oils originated from simi-lar source rocks containing mainly terrigenous, type III organic matter deposited under oxic condi-tions. These ratios decrease with thermal matura-tion and increase with biodegradation. The anoma-lous composition of Peciko oil (Figure 6) results from mixing of low-maturity oil with a later charge of nonbiodegraded condensate (discussed in a fol-lowing section). The compositions of Mutiara 48 and Bekapai 4 DST 5 oils (Figure 6) suggest that their source rock was deposited under less oxic conditions than the other samples.Oil-to-Oil CorrelationsPrevious geochemical studies suggest that Mahakam Delta oils originated from similar type III kerogen (Schoell et al., 1983) and that no chemical difference exists between the organic matter in nearby coals and shales acting as source rocks (Vandenbroucke et al., 1983); however, Huc et al. (1986) observed that Mahakam Delta coals gener-ate more high-molecular-weight n-alkanes than do nearby shales. These early observations do not jus-tify the conclusion that all differences in oil compo-sition in the Mahakam Delta area result from sec-ondary alteration processes. On the contrary, our data show clear source and maturity differences among the sampled oils.All of the oil samples in this study originated from deltaic-paralic source rocks that received vary-ing proportions of terrigenous organic matter. For example, higher plant biomarkers in the samples include C29steranes (Moldowan et al., 1985), bicadinanes (Alam and Pearson, 1990; Woolhouse et al., 1992; Armanios et al., 1995), oleanane (Moldowan et al., 1994), and tricyclic diterpanes (Noble et al., 1986; Alexander et al., 1987). The bicadinanes are derived from damar resins exudedPeters et al.21。