Carroll and Wartes, 2003.pdf

INTRODUCTION

Organic carbon burial in modern lakes may rival the mag-nitude of burial in many nearshore marine settings (Dean andGorham, 1998; Einsele et al., 2001), suggesting that thesedeposits should be considered an important carbon sink. Deanand Gorham (1998) concluded that Holocene carbon burial insmall lakes accounts for the majority of this carbon burial,based on estimates derived from post-glacial lakes in Min-nesota. A smaller, but still significant rate of carbon burial wasestimated to occur in large lakes (>5000 km2). Carbon burial by

lakes and inland seas was estimated to represent about one-thirdof the total global burial rate.

Collectively, nonmarine depositional environments, includingreservoirs, account for approximately 75% of present day organiccarbon burial. Einsele et al. (2001) pointed to rapid accumulationrates and high preservation factors in lakes as causal factors andnoted that burial of organic carbon increases with increaseddrainage basin area and change to wetter and warmer climate.

Ancient lacustrine stata deposited in tectonically subsidedbasins have long been recognized to include thick intervals ofhighly organic-rich mudstone, which is often economically

Geological Society of AmericaSpecial Paper 370

2003

Organic carbon burial by large Permian lakes, northwest China

Alan R. CarrollMarwan A. Wartes

Department of Geology and Geophysics, University of Wisconsin, 1215 W. Dayton Street, Madison, Wisconsin 53711, USA

ABSTRACT

Permian strata of the Junggar-Turpan-Hami Basins represent one of the thickestand most laterally extensive lacustrine deposits in the world, yet they are very poorlyknown outside of China. Deposition spanned approximately 30 m.y., from the Sakmar-ian through Changhsingian epochs. Continuous intervals of organic-rich lacustrinemudstone may exceed 1000 m, and the total thickness of lacustrine and associated non-marine strata locally exceeds 4000 m. Early Permian basin subsidence coincided withregional normal faulting and associated volcanism, interpreted to result from extensionor transtension of newly amalgamated accretionary crust. In contrast, relatively uni-form regional subsidence occurred during the Late Permian, most likely due to flexurecaused by renewed regional compression. The maximum expansion of Permian lakesduring the Wordian to Capitanian postdates any evidence for significant normal fault-ing or volcanism. Organic-rich mudstone facies cover an area at least 900 × 300 km,indicating that at their maximum, the Permian lakes were comparable in size to theCaspian Sea. An organic-rich profundal mudstone section in the south Junggar Basinhas been ranked as the thickest and richest petroleum source rock interval in the world,with total organic carbon content averaging 4% and commonly exceeding 20%. TotalLate Permian carbon burial is estimated at 1019 gC. Maximum organic carbon burialrates are estimated at 4 × 1012 gC/yr, equivalent to approximately 4–8% of estimatedglobal carbon burial rates during this time. Permian lakes in northwestern China werebroadly synchronous with other large lakes (or inland seas) in South America andAfrica that also formed due to the amalgamation of Pangea. Collectively, these basinsrepresent a large and heretofore unrecognized organic carbon sink that may have influ-enced atmospheric CO2 concentrations.

Keywords: Junggar, Turpan, Hami, lacustrine, climate, stratigraphy.

Carroll, A.R., and Wartes, M.A., 2003, Organic carbon burial by large Permian lakes, northwest China, in Chan, M.A., and Archer, A.W., eds., Extremedepositional environments: Mega end members in geologic time: Boulder, Colorado, Geological Society of America Special Paper 370, p. 91–104. ©2003Geological Society of America

91

exploited as oil shale. Three of the world’s five thickest and rich-est petroleum source rock intervals were deposited in lake basins(Demaison and Huizinga, 1991). Many of these deposits alsocover very large areas. For example, the Neocomian-Barremianorganic-rich lacustrine mudstone strata associated with the open-ing of the south Atlantic extend at least 3000 km along theAfrican and South American continental margins across numer-ous basins (cf., Ojeda, 1982; Mello and Maxwell, 1990; Katz andMello, 2000). In contrast to the deposits of relatively ephemeralHolocene post-glacial lakes, carbon buried by large tectonic lakebasins is typically removed from the atmospheric reservoir formillions to hundreds of millions of years. However, carbon buriedin tectonic lake basins has generally been ignored in global massbalance calculations. These deposits represent potentially impor-tant carbon sinks that may have helped to ameliorate pastepisodes of global warming.

The Permian Junggar-Turpan-Hami Basins in northwestChina collectively represent one of the largest known Phanero-zoic lake basins (Wartes et al., 2000, 2002) and include theworld’s thickest and richest petroleum source rock interval(Demaison and Huizinga, 1991). Organic-rich lacustrine mud-stone is distributed over an area roughly equivalent to that of themodern Caspian Sea. These strata also span an important tectonictransition, from late Paleozoic continental amalgamation to recur-rent intraplate collisional deformation (Allen and Windley, 1993;Allen et al., 1991, 1995; Carroll et al., 1990, 1992, 1995; Hen-drix et al. 1992; Graham et al., 1993). This paper summarizes

what is known about the origin of the Junggar-Turpan-Hamideposits and provides the first estimates of the magnitude andrate of carbon burial they represent.

Tectonic Setting of Northwest China

The Junggar, Turpan, and Hami Basins (Figs. 1 and 2) over-lie oceanic materials that have been interpreted as part of thePaleozoic Altaid accretionary complex (Sengör et al., 1993).Exposed basement lithologies consist chiefly of highly deformedOrdovician through Carboniferous volcanogenic turbidites, withlocal occurrences of mafic and ultramafic igneous rocks and chert(e.g., Coleman, 1989; Feng et al., 1989; Carroll et al., 1990,1995). Precambrian rocks are limited to the central Tian Shan andareas to the south. Isotopic (87Sr/86Sr and εNd) evidence from latePaleozoic granitic plutons that pierce basement lithologies con-firms their oceanic affinity and reflect an increased influence ofPrecambrian continental crust to the south (Hopson et al., 1989,1998; Wen, 1991). The nature of the substrate hidden beneath theJunggar Basin sedimentary strata is unknown, but it is inferred tobe similar to the basement lithologies that crop out on all sides ofthe basin. Total crustal thicknesses in this region have been esti-mated at approximately 40 km (Lee, 1985), which locallyincludes 15 km or more of relatively undeformed Carboniferousthrough Cenozoic sedimentary rocks (Hendrix et al., 1992; Gra-ham et al., 1993; Carroll et al., 1995). Uplift of the BogdashanRange (Fig. 2) first occurrred in Late Triassic to Early Jurassic

92 A.R. Carroll and M.A. Wartes

Figure 1. Location of the Junggar-Turpan-Hami Permian lake depositsand the Altaid orogenic complex (areaof Altaids modified from Sengör etal., 1993).

(Hendrix et al., 1992; Greene et al., 2001). Prior to that time, theJunggar, Turpan, and Hami Basins were united, although a partialdrainage divide may have existed between Junggar and Turpan-Hami Basins (Wartes et al., 2002).

The unified Junggar-Turpan-Hami Basin occupies the site of arelict Early to Middle Carboniferous sea that filled with marinedeep water through shelf facies following the extinction of arcmagmatism in the Late Carboniferous (Carroll et al., 1990, 1995).The tectonic setting of this area prior to the Permian is ambiguousand may be interpreted as either a remnant ocean basin or backarcbasin (Carroll et al., 1990; Windley et al., 1990; Allen et al., 1991)that became tectonically isolated from the sea. A relatively contin-uous shoaling-upward succession in excess of 2-km-thick recordsthe Late Carboniferous through Early Permian retreat of marinewaters from the southern Junggar Basin (Carroll et al., 1995).Overlying Permian nonmarine strata total over 4 km in thickness.A similar marine to nonmarine transition is recorded in southernBogdashan exposures of the Turpan-Hami Basin, although thesestrata are generally thinner and change more abruptly.

Stratigraphy and Depositional Environments

Absolute chronostatigraphic control for nonmarine Permianfacies in the Junggar-Turpan-Hami Basins is poor due to faunaland floral endemism. However, relative age assignments atapproximately the stage level are possible through comparison offaunal, floral, and palynolomorph assemblages reported from dif-ferent units (Wartes et al., 2000; Wartes et al., 2002). Radioiso-topic dating of volcanic units is limited to the Early Permian, butthe results are consistent with the established biostratigraphicframework (Wartes et al., 2000). These data permit regional cor-relations between Permian nonmarine stratigraphic units exposedin the southern Bogdashan, and between surface and subsurfacelithologies within and adjacent to the Junggar Basin.

Lacustrine facies assemblages range from fluvial-lacustrineto evaporative, based on physical features and biomarker geo-chemistry, and have been previously described by Carroll et al.(1992), Carroll (1998), Wartes et al. (2000), Bohacs et al. (2000),and Carroll and Bohacs (2001). Outcrop exposures of Lower Per-

Organic carbon burial by large Permian lakes 93

Figure 2. Location of Junggar and Turpan-Hami Basins, with maximum known extent of Upper Permian lake deposits (mod-ified from Wartes et al., 2002). Field localities: AR—Aiweiergou, HU—Huoshaoshan, NT—North Tian Shan, SH—Shisan-jifang, SJ—South Junngar Composite Section, TI—Tianchi, TS—Taoshuyuan, TX—Tian Shan Xiang, UR—Urumqi,XD—Xidagou, ZB—Zaobishan. A–A′ is line of section in Figure 3; B–B′ is line of section in Figure 4.

mian lacustrine facies are restricted to the southern Bogdashan,and are represented by portions of the Zaobishan and YierxituFormations (Fig. 3). These fluvial-lacustrine to fluctuating pro-fundal facies (cf., Carroll and Bohacs, 1999) are interbedded withmafic to intermediate volcanic rocks, and deposition was local-ized by Early Permian normal faulting. They typically lie uncon-formably on underlying Carboniferous marine facies. In contrast,Lower Permian facies of the Junggar Basin are dominantlymarine and generally conformable with underlying strata (Fig. 4).The thickness of Lower Permian strata in the subsurface wasdetermined using a combination of well logs and reflection seis-mic data, and facies similar to those in outcrop were verified bycore examination (Greene et al., 2001).

Upper Permian facies in both basins are exclusively nonma-rine and dominantly siliciclastic and record a wide range of lacus-trine depositional environments. Average carbonate contents inthe southern Junggar Basin range between 20% and 30%, withdolomite common in the most organic-rich facies (unpublishedX-ray diffraction and X-ray fluorescence spectrometry data).Upper Permian strata overlie a major regional unconformity insouthern Bogdashan exposures, but they appear to be con-formable with Lower Permian nonmarine strata in the northwest-ern Bogdashan (Figs. 3 and 4). The Tarlong Formation in thesouthern Bogdashan and northern Tian Shan consists dominantlyof fluctuating profundal facies, with mudstone, sandstone, andminor limestone cyclically interbedded on the scale of meters to


Figure 3. East-west stratigraphic cross section of Upper Carboniferous through Lower Triassic strata outcrops on southern flank of Bogdashan (mod-ified from Wartes et al., 2002; see Fig. 2 for location). AR—Aiweiergou, SH—Shisanjifang, TX—Tian Shan Xiang, ZB—Zaobishan, T1—LowerTriassic, T2—Middle Triassic.

tens of meters (Figs. 5A and 5B). Laminated mudstone and well-preserved fish fossils attest to deposition under deep, low oxygenconditions (cf., Olsen 1986), mostly likely in a stratified lake,whereas mudcracks indicate periodic dessication. The Tarlongcorrelates with the Lucaogou and Hongyanchi Formations of theJunggar Basin (Fig. 4). These units record a gradual progressionfrom evaporative (Fig. 6A) to fluvial-lacustrine facies (Fig. 6C) inexposures on the northwestern flank of the Bogdashan, based onphysical evidence for desiccation and synaeresis (mudcracks) andbiological marker compounds in mudrock extracts (Carroll,1998). Fluctuating profundal facies of the Lucaogou Formationreach a thickness of approximately 1300 m (Figs. 6B and 7). Thelower Lucaogou Formation is commonly dolomitic and containspossible synaeresis cracks (Fig. 5D), which suggests fluctuatingsalinity. The upper Lucaogou Formation includes at least 500continuous meters of laminated mudstone with no evidence of

subaerial exposure. Laminae thicknesses vary cyclically from~0.1 to 2.0 mm over a vertical scale of several meters. Variationsin laminae thickness and organic matter content were interpretedby Carroll (1998) to result from changes in the rate of clastic sed-iment supply in a deep, sediment-starved lake basin. Finally, theCangfanggou Group records a gradation to fluvial-lacustrinefacies in both basins (Figs. 3 and 4). Fluvial sandstone and lacus-trine mudstone are interbedded on the scale of meters with noapparent cyclicity, and plant fossils, freshwater molluscs, and evi-dence for paleosols are common.

Permian Basin Evolution

After retreat of marine waters, Early Permian extension ledto the creation of localized nonmarine depocenters that were con-trolled, in part, by normal faults. Normal faults have also been


Figure 4. North-south stratigraphic cross section of Upper Carboniferous through Triassic outcrops (NT, AR, and South Junggar) and wells (HU, XD)(modified from Wartes et al., 2002; see Fig. 2 for location).

postulated to exist beneath the Junggar Basin (Bally et al., 1986;Liu, 1986; Peng and Zhang, 1989), but the precise timing of thesestructures is unconstrained. Allen et al. (1995) and Sengör andNatal’in (1996) hypothesized that these structures and the originof the Alakol, Junggar, and Turpan Basins were all related to LatePermian-Early Triassic sinistral shear. However, field relation-ships near Taoshuyuan in the southern Bogdashan show that amajor north-south trending normal fault there is actually EarlyPermian (Wartes et al., 2002; Figs. 2 and 3). This fault and asso-ciated volcanic rocks are roughly coeveal with Early Permianbasaltic magmatism in the northwest Tarim Basin and to the east

in the Beishan, suggesting a widespread episode of extension fol-lowing the late Paleozoic amalgamation of this region. Carroll etal. (1995) and Wartes et al. (2002) interpreted this extension tohave been dominantly east-west oriented, occurring during ongo-ing north-south compression (present orientation). However,regional shear remains a viable alternative hypothesis.

Early Permian paleogeography was characterized by a com-plex series of fault-related sub-basins containing mostly nonma-rine fill in the Turpan-Hami area and a relict marine basin in theJunggar area (Fig. 8). The nature of the drainage divide betweenthese areas is unclear, but it may be topography inherited from


AB

CD

Figure 5. Outcrop photographs of Upper Permian lacustrine facies. A: Tarlong Formation mudstone with interbedded sandstone at Aierwaigou (see“AR” in Fig. 2 for location). Person circled for scale. B: Detail of Tarlong Formation at Aierwaigou (note slumped sandstone bed). C: LucaogouFormation at Tianchi aqueduct excavation (see “SJ” in Fig. 2 for location). Person circled for scale. D: Detail of possible synaeresis cracks in Lucao-gou Formation at Tianchi aqueduct excavation.

Figure 6. Representative detailed measured sections of Upper Permian lacustrine and associated alluvial facies exposed in western Bogdashan (mod-ified from Carroll, 1998). Arrows indicate deepening and shallowing trends. A: Jingjingzigou Formation at Tianchi aqueduct excavation. B: Lucao-gou Formation at Tianchi aqueduct excavation. C: Hongyanichi Formation at Urumqi. TOC—total organic content.

the late stages of a Carboniferous magmatic arc co-located withthe Bogdashan (c.f., Coleman, 1989; Carroll et al., 1990; Windleyet al., 1990; Allen et al., 1991). In contrast, late Permian subsid-ence was more uniformly distributed, resulting in widespreaddeposition of relatively continuous lacustrine facies and alluvialfacies (Fig. 9). Geohistory analysis of outcrop sections from thenorthwestern Bogdashan indicates that subsidence ratesincreased markedly during the Late Permian, most likely due to

lithospheric flexure in response to uplift of an ancestral Tian ShanRange (Watson et al., 1987; Graham et al., 1990; Carroll et al.,1990, 1995; Allen and Windley, 1993). Greene et al. (2001) andWartes et al. (2002) proposed that the Turpan-Hami Basin occu-pied a wedge-top position, which would account for the thinnerUpper Permian deposits preserved there and the development ofa basal Upper Permian unconformity. In contrast, the southernJunggar Basin is interpreted as a flexural foredeep, whichreceived over 4 km of Upper Permian nonmarine basin fill. Themaximum extent of lakes occurred during deposition of theLucaogou, Tarlong, and equivalent formations.

Organic Matter Burial

High enrichments of organic carbon in the Lucaogou Forma-tion were documented by Graham et al. (1990), who reported val-ues up to 34% total organic carbon and Type I kerogen. Vitrinitereflectance values for the same samples indicated that these rockshave reached the early stages of petroleum generation (Graham etal., 1993; Carroll et al., 1992; Carroll, 1998), so the original totalorganic carbon values may have been slightly higher prior to onsetof oil generation and migration. Oils on the northwestern, north-ern, and northeast flanks of the Junggar Basin have biomarker dis-tributions indicating derivation from Upper Permian lacustrinefacies (Clayton et al., 1997), and at least two fields in the Turpan-Hami Basin also produce similar oils (Greene, 2000). Outcrop-ping mudstone facies in the southern Bogdashan are relativelyweathered and appear to have been exposed to higher levels ofthermal maturation, but preliminary data suggest that they alsohad high original total organic carbon contents (Greene, 2000).

To determine the average total organic carbon of the Lucao-gou Formation mudstone facies, which is also referred to as “oilshale,” Carroll et al. (1992) conducted sampling at fixed, regularintervals from an aqueduct excavation in the northwestern Bog-dashan (Fig. 7). They reported that total organic carbon values of20% or more are common, and that total organic carbon valuesover a continuous 800-m interval average 4%. Our subsequentinvestigations have shown that similar but less well-exposedorganic rich mudstone facies continue stratigraphically above theaqueduct excavation for at least another 500 m. The total thick-ness of the Lucaogou Formation averaging 4% total organic car-bon at this locality is taken to be approximately 1300 m.

Subsurface mapping of the Junggar Basin indicates severaldepocenters where the Upper Permian mudstone reaches thick-nesses of up to 2000 m (Fig. 10). The precise age of thesedeposits is uncertain, but their indicated thickness is broadly con-sistent with that of the measured outcrop exposures in the Bog-dashan. Furthermore, biomarker distributions in oils producedfrom the northwestern and eastern flanks of the Junggar Basincorrelate closely with bitumen extracts from the outcroppingLucaogou Formation (Carroll et al., 1992; Clayton et al., 1997;Carroll, 1998; Fig. 10). Oils with similar biomarker characteris-tics are also produced from two fields in the Turpan-Hami Basin(Greene, 2000), suggesting that correlative facies also occur


Figure 7. Master measured sections showing total organic carbon(%TOC) at Tianchi aqueduct excavation and at Urumqi (see Fig. 2 forlocation). Together these sections constitute south Junggar compositesection, based on approximate correlation shown.

south of the Bogdashan. Permian-sourced oils are generallyabsent in areas with thick accumulations of Mesozoic strata (thesouthern Junggar and much of the Turpan-Hami Basins), appar-ently due to overmaturation of Permian source facies during deepburial (Carroll et al., 1992).

The absolute limit for the distribution of the LucaogouFormation and its equivalents is uncertain, due to very incom-plete mapping of areas surrounding the Junggar-Turpan-HamiBasins. However, similar facies have been reported from an areathat greatly exceeds the boundary of these basins. For example,intervals of organic-rich, Upper Permian mudstone ~100 m thickhave been reported in the Yili Basin to the west of Junggar and inthe Santamu Basin to the east (Wang, 1992; Cheng Keming, per-sonal commun., 1998). The Yili Basin rocks are offset by Ceno-zoic dextral strike-slip faulting between central and northern TianShan (Tapponier and Molnar, 1979), suggesting that they mayhave originally been deposited adjacent to the Turpan-HamiBasin. The possibility that Permian lake deposits continuedacross the border into Mongolia is tantalizing but entirelyuntested. Finally, organic-rich Permian mudstone facies havebeen reported in the Irtysh-Zayshan area of Kazahkstan, adjacentto the northwest corner of the Junggar Basin (J. Degenstein, per-

sonal commun., 1993). In total, these known occurrences are dis-tributed over an area roughly 900 × 300 km.

We estimate an average thickness for Lucaogou Formationand equivalent strata of 400 m for the area outlined in Figure 2.This assumption most likely leads to an underestimate of totalmud rock volume due to the preferential exposure of these stratain areas with relatively less Permian subsidence and greater post-Permian uplift. It is likely that the thickest intervals of mud rockremain buried beneath thick Mesozoic and Cenozoic cover strata.The ultimate limit of mud rock deposition is unknown due toincomplete preservation and the rudimentary state of geologicalmapping, but it is inferred to exceed the outlined area. Based onthe above assumptions, we estimate total carbon burial within thisinterval of approximately 1019 grams (Table 1).

DISCUSSION

Based on the thickness and average richness of organic-richrocks in the south Junggar Basin composite section (Fig. 7),Demaison and Huizinga (1991) ranked the Junggar Basin as hav-ing the highest cumulative hydrocarbon potential of any petro-leum-producing basin in the world (Table 2). This ranking isbased on source potential index, a parameter that is based onaverage Rock Eval S1 (thermally distilled hydrocarbons) plus S2(hydrocarbons derived from pyrolytic breakdown of kerogen; seeEspitalié et al., 1977, and Peters, 1986, for further explanation ofRock Eval techniques). Source potential index values thus under-state the total magnitude of carbon burial, since they do not


Figure 8. Schematic block diagram showing Early Permian evolution ofJunggar and Turpan-Hami Basins (modified from Wartes et al., 2002).TI—Tianchi, TX—Tian Shan Xiang, ZB—Zaobishan. C2a—AoertuFormation, C2l—Liushugou Formation, C2q—Qijiagou Formation,P1s—Lower Permian Shirenzigou Formation, P1ts—Lower PermianTaoshuyuan Formation, P1tx—Taoxigou Group, P1z—ZaobishanFormation. See Figures 3 and 4 for stratigraphy.

Figure 9. Schematic block diagram showing Late Permian (Wordian)evolution of Junggar and Turpan-Hami Basins (modified from Wartes etal., 2002). AR—Aiweiergou, TI—Tianchi. ZB—Zaobishan. See Fig-ures 3 and 4 for stratigraphy. NTS—North Tian Shan, P2t—TarlongFormation, P2l—Lucaogou Formation, P2p—Pingdiquan Formation.

account for carbon contained in refractory compounds. Nonethe-less, Demaison and Huizinga’s (1991) summary provides a use-ful means of comparing the relative importance of variousintervals of organic-rich rock. It is interesting to note that three ofthe top five such intervals reported represent lacustrine basins thatpreserve Type I kerogen. A different ranking might result if the

areal extent could be integrated with the SPI values to obtain anestimate of total hydrocarbon potential in each of these basins. Inparticular, Lower Cretaceous lacustrine strata associated withvarious south Atlantic margin rift basins, for example, the Buco-mazi and Marnes Noires Formations in offshore west Africa andequivalent units in various Brazillian basins, could potentiallyexceed the total carbon burial of the Permian Junggar-Turpan-Hami Basin due to the large area of the Lower Cretaceous basins.

The lacustrine Green River Formation (Eocene of Colorado,Utah, and Wyoming) has long been recognized as the world’slargest oil shale deposit, representing an estimated resource of 1 × 1012 bbls of oil in Colorado alone (Pitman et al., 1989) and1.5 × 1012 bbls total (J. R. Dyni, personal commun., 2002). Theseestimates are determined primarily from Fischer assay measure-ments and thus cannot be directly compared with the total organiccarbon measurements from the Junggar-Turpan-Hami Basins.


Figure 10. Distribution of Upper Permian lacustrine petroleum source facies in the Junggar Basin (modified from Carroll et al., 1992) and biomarker“fingerprints” of Permian-source oils. Inset diagrams show m/z 218 mass fragmentograms for alkane fractions of Junggar oils and for one rockextract from outcropping organic-rich mudstone (“Tianchi”). Each peak on fragmentograms corresponds to a different steroidal compound. Blackpeaks indicate specific groups of sterane isomers and highlight relative distribution of compounds with 27, 28, or 29 carbon atoms in each sample.Note that Permian-sourced oils from rock extract and from oils produced in western, central, and eastern Junggar Basin all have similar sterane dis-tributions, characterized by low C27 and subequal C28 and C29 homologues. In contrast, Jurassic-sourced oils from southern Junggar Basin have lowC27 and C28 but high C29, typical of coaly source facies. Potential Upper Permian source facies in southern Junggar Basin are buried by up to 11 kmof post-Permian strata and are thus overmature with respect to oil generation.

However, 1.5 × 1012 barrels equates to approximately 1.6 × 1017 gof oil, which is two orders of magnitude less than the estimatedtotal carbon buried in the Junggar-Turpan-Hami Basins. Thiscomparison fails to account for refractory organic carbon com-pounds in the Green River Formation, but the higher level of ther-mal maturity in the Junggar-Turpan-Hami deposits at leastpartially offsets the deficit. Source potential index values offer analternative means of comparison. The Laney Member of theGreen River Formation in Wyoming, which is the stratigraphicequivalent of the Parachute Creek Member in Colorado, has anSPI value that is about one third that of the Upper Permian in theJunggar-Turpan-Hami Basin (Table 2). Total source potentialindex for the Green River Formation may locally equal or exceedthat of the south Junggar Basin, but the Green River FormationLakes at their maximum extent covered only about one-third ofthe area of the Junggar-Turpan-Hami Lakes.

Although it is clear that Permian lacustrine mudstone in west-ern China basins contains a very large mass of organic carbon, cal-culation of accurate carbon burial rates is difficult due to poor agecontrol. One typical approach to this problem is to assume thatmudstone laminations are annual in nature and therefore representvarves. If this is the case and sedimentation rates were reasonablyconstant, then the Lucaogou Formation and its equivalents weredeposited over a period of approximately 3 m.y., based on countsof its finest-scale laminations. This duration is broadly consistentwith other constraints on the age of the Lucaogou interval and isreasonable when compared with mudstone accumulation ratesmeasured in other nonmarine basins (Table 3). However, we

observed considerable variation in lamination thickness in outcrop,ranging between ~0.1 and 2.0 mm. Thicker laminae appear to besystematically associated with lower % total organic carbon, sug-gesting that considerable variation occurred in the flux of inorganicsediment (Carroll, 1998). If so, then the actual duration of theLucaogou interval may be less than 3 m.y.

Based on the above assumptions, the calculated organiccarbon burial rate per unit area for the Junggar-Turpan-Hami


Lake is comparable to the average rate for modern lakes andmuch lower than the maximum seen in modern eutrophic lakes(Dean and Gorham, 1998; Table 3). If valid, this calculated ratesuggests that the preservation of large amounts of organic mat-ter corresponded to a period of relatively modest primaryaquatic productivity. Carroll (1998) argued on the basis of tex-tural and geochemical evidence that high %total organic carbonvalues in the Lucaogou Formation resulted, in part, from lowrates of inorganic sedimentation in a deep, chemically stratifiedlake with oxygen-depleted bottom water. This conclusion isalso consistent with Dean and Gorham (1998) and others’observation that large modern lakes and inland seas typicallybury organic carbon at much lower rates than do smalleutrophic lakes, due to relatively lower nutrient concentrationsin the surface waters of the larger bodies.

Despite the relatively modest rate of organic carbon burialper unit area beneath the Junggar-Turpan-Hami Lake, its extremesize resulted in a very large magnitude of total organic carbonburial per year. This mass has not been included in previous phys-ical assessments of organic carbon burial. For example, Bernerand Canfield (1989) stated that “The organic carbon content ofnon-coal-containing continental clastics is essentially zero.” Theyestimated that global organic carbon burial rates decreased fromapproximately 100 × 1018 to 50 × 1018 gC/yr during the Permian.Depending on the actual timing of the Junggar-Turpan-HamiLake, its organic carbon burial therefore represented 4–8% ofglobal organic carbon burial. This amount is clearly not zero, sug-gesting that large lakes may be more important geological reser-voirs of organic carbon than previously thought.

The occurrence of other large, inland water bodies in thesouthern hemisphere at about the same time helps to reinforce thehypothesis that such features were globally significant sinks fororganic carbon. An inland water body that was apparently severaltimes the size of the Junggar-Turpan-Hami Lake covered parts ofBrazil and southern Africa during the latest Early Permian orearly Late Permian, and its deposits were preserved in the Paranáand Great Karoo Basins (Oelofsen, 1987; França et al., 1995;Ziegler et al., 1996; Fig. 11). The precise size of this water bodyis uncertain because these deposits are incompletely exposed, butit may have been large enough to ameliorate the otherwise harshclimatic conditions that would have prevailed in the interior ofGondwana (Yemane, 1993; Kutzbach and Ziegler, 1993). TheIrati and Whitehill Formations (South America and southernAfrica) contain organic-rich mudstone facies similar to theLucaogou Formation, commonly with greater than 10% totalorganic content (e.g., Correa da Silva and Cornford, 1985; Oelof-sen, 1987). These somewhat controversial deposits have been var-iously interpreted to record either a fresh to brackish-water lake(e.g., Correa da Silva and Cornford, 1985; Faure and Cole, 1999)or else an inland sea (e.g., Teichert, 1974; Mello et al., 1993;Visser, 1994; Goldberg, 2001). The precise age relationship ofthese units to the Lucaogou Formation is unknown. In contrast tothe Lucaogou Formation, the oil shale facies of the Irati andWhitehill Formations are relatively thin (meters to tens of meters).

The occurrence of large lakes as discussed above is directlyrelated to processes of continental convergence and orogenesisby the entrapment of thinned continental or oceanic crust withincollisional zones and by the development of internal drainagesbehind rising mountain ranges. Late Cenozoic examples of thisprocess may be found in the southern Caspian and PannonianBasins; both are former marine realms that were trapped withinthe developing Alpine orogenic zone (Zonenshain and LePichon, 1986; Mattick et al., 1988; Geary et al., 1989). TheBlack Sea and parts of the Mediterranean will likely suffer thesame fate in the future. Significant organic matter burial is com-mon in such basins, but the mechanisms of organic matterpreservation appear to be complex. In some cases, the burialrates of organic carbon may reach a peak during transitionbetween marine and lacustrine conditions. For example, Arthurand Dean (1998) noted that organic carbon-rich sapropel in theBlack Sea was deposited during the initial incursion of marinewaters through the Bosporus during the Early Holocene. Theseauthors suggested that marine spillover into the previous fresh-water lake resulted in vertical mixing of nutrients and that sub-sequent salinity stratification helped to maintain bottom-wateranoxia. However, in the case of the Junggar-Turpan-Hamideposits, no evidence for marine incursions has been reported,either because such incursions did not occur or else becausethese vast deposits have been inadequately investigated. Regard-less of the causes of organic matter preservation, lakes and


Lakes and inland seas

WordianJunggar-

Turpan-

Hami

Irati

Whitehill

Figure 11. Plate tectonic reconstruction of Pangea during Late Permian(Wordian), indicating approximate position of Junggar-Turpan-HamiLakes and of the lake or inland sea that deposited the Irati-Whitehill For-mations in Brazil and southern Africa (modified from Ziegler et al.,1996).

inland seas within convergent zones are capable of burying largetotal quantities of carbon due to frequent, high rates of tectonicsubsidence. In contrast to the much smaller eutrophic lakes thatare often geomorphic in origin, large tectonic lakes can perma-nently isolate organic carbon from the atmosphere over timeperiods lasting up to hundreds of millions of years.

ACKNOWLEDGMENTS

We thank K. Cheng, J. Degenstein, T. Hu, S. A. Graham,T. J. Green, and Y. Liang for helpful discussions of parts ofthis study. Financial support has been provided by the Donorsof the Petroleum Research Fund of the American ChemicalSociety, Conoco, Texaco, the Graduate School of the Univer-sity of Wisconsin, and the Stanford-China Industrial Affiliates.We are grateful to Charles Oviatt and Walter Dean for con-structive reviews.

REFERENCES CITED

Allen, M.B., and Windley, B.F., 1993, Evolution of the Turfan Basin, Chinesecentral Asia: Tectonics, v. 12, p. 889–896.

Allen, M.B., Windley, B.F., Zhang, C., Zhao, Z.Y., and Wang, G.R., 1991, Basinevolution within and adjacent to the Tien Shan range, NW China: Journal ofthe Geological Society, London, v. 148, p. 369–378.

Allen, M.B., Sengör, A.M.C., and Natal’in, B.A., 1995, Junggar, Turfan, andAlakol basins as Late Permian to Early Triassic extensional structures in asinistral shear zone in the Altaid orogenic collage, Central Asia: Journal ofthe Geological Society, London, v. 152, p. 327–338.

Arthur, M.A., and Dean, W.E., 1998, Organic-matter production and preservationand evolution of anoxia in the Holocene Black Sea: Paleoceanography,v. 13, p. 395–411.

Bally, A.W., Chou, I.M., Clayton, R., Eugster, H.P., Kidwell, S., Meckel, L.D.,Ryder, R.T., Watts, A.B., and Wilson, A.A., 1986, Notes on sedimentarybasins in China—Report of the American Sedimentary Basins Delegation tothe People’s Republic of China: United States Geological Survey Open-FileReport 86-327, 108 p.

Bohacs, K.M., Carroll, A.R., Neal, J.E., and Mankiewicz, P.J., 2000, Lake-basintype, source potential, and hydrocarbon character: An integrated sequence-stratigraphic geochemical framework, in Gierlowski-Kordesch, E.H., andKelts, K., eds., Lake basins through space and time: American Associationof Petroleum Geologists Studies in Geology 46, p. 3–33.

Berner, R.A., and Canfield, D.E., 1989, A new model for atmospheric oxygenover Phanerozoic time: American Journal of Science, v. 289, p. 333–361.

Carroll, A.R., and Bohacs, K.M., 2001, Lake type control on hydrocarbon sourcepotential in nonmarine basins: American Association of Petroleum Geolo-gists Bulletin, v. 85, p. 1033–1053.

Carroll, A.R., 1998, Upper Permian lacustrine organic facies evolution, southernJunggar Basin, NW China: Organic Geochemistry, v. 28, p. 649–667.

Carroll, A.R., and Bohacs, K.M., 1999, Stratigraphic classification of ancientlakes: Balancing tectonic and climatic controls: Geology, v. 27, p. 99–102.

Carroll, A.R., Liang, Y., Graham, S.A., Xiao, X., Hendrix, M.S., Chu, J., andMcKnight, C.L., 1990, Junggar Basin, northwest China: Trapped late Paleo-zoic ocean: Tectonophysics, v. 181, p. 1–14.

Carroll, A.R., Brassell, S.C., and Graham, S.A., 1992, Upper Permian lacustrineoil shale of the southern Junggar Basin, northwest China: American Associ-ation of Petroleum Geologists Bulletin, v. 76, p. 1874–1902.

Carroll, A.R., Graham, S.A., and Hendrix, M.S., 1995, Late Paleozoic tectonicamalgamation of northwestern China: Sedimentary record of the northernTarim, northwestern Turpan, and southern Junggar Basins: Geological Soci-ety of America Bulletin, v. 107, p. 571–594.

Clayton J.L., Yang, J., King, J.D., Lillis, P.G., and Warden, A., 1997, Geochem-istry of oils from the Junggar Basin, northwest China. American Associationof Petroleum Geologists Bulletin, v. 81, p. 1926–1944.

Coleman, R.G., 1989, Continental growth of northwest China: Tectonics, v. 8,p. 621–635.

Correa da Silva, Z.C., and Cornford, C., 1985, The kerogen type, depositionalenvironment and maturity, of the Irati Shale, Upper Permian of ParanáBasin, southern Brazil: Organic Geochemistry, v. 8, p. 399–411.

Dean, W.E., and Gorham, E., 1998, Magnitude and significance of carbon burialin lakes, reservoirs, and peatlands: Geology, v. 26, p. 535–538.

Demaison, G., and B.J. Huizinga, 1991, Genetic classification of petroleum sys-tems: American Association of Petroleum Geologists Bulletin, v. 75,p. 1626–1643.

Einsele, G.,Yan, J., and Hinderer, M., 2001, Atmospheric carbon burial in modernlake basins and its significance for the global carbon budget: Global andPlanetary Change, v. 30, p. 167–195.

Espitalié, J., Laporte, J.L., Madec, M., Marquis, F., Lepat, P., Paulet, J., andBoutefeu, A., 1977, Méthode rapide de caractérisation des roches mères, deleur potentiel petrolier et leur degré d’évolution: Review Institut Françaisdu Pétrole, v. 32, p. 23–42.

Faure, K., and Cole, D., 1999, Geochemical evidence for lacustrine microbialblooms in the vast Permian Main Karoo, Paraná, Falkland Islands and HuabBasins of southwestern Gondwana: Palaeogeography, Palaeoclimatology,Palaeoecology, v. 152, p. 189–213.

Feng,Y., Coleman, R.G., Tilton, G., and Xiao, X., 1989, Tectonic evolution of thewest Junggar region, Xinjiang, China: Tectonics, v. 8, p. 729–752.

França, A.B., Milani, E.J., Schneider, R.L., Lopez-Paulsen, M., Suárez-Soruco,R., Santa Ana, H., Wiens, F., Ferreiro, O., Rossello, E.A., Bianucci, H.H.,Flores, R.F.A., Vistalli, M.S., Fernandez-Seveso, F., Fuenzalida, R.P.,Munoz, N., 1995, Phanerozoic correlation in southern South America, inTankard, A.J., Suarez-Soruco, R., and Welsink, H.J. eds., Petroleum basinsof South America: Tulsa, Oklahoma, American Association of PetroleumGeologists Memoir 62, p. 129–161.

Geary, D.H., Rich, J., Valley, J.W., and Baker, K., 1989, Stable isotopic evidenceof salinity change: Influence on the evolution of melanopsid gastropods inthe late Miocene Pannonian basin: Geology, v. 17, p. 981–985.

Goldberg, K., 2001, The paleoclimatic evolution of the Permian in the ParanáBasin in southern Brazil [Ph.D. dissertation]: Chicago, University ofChicago, 267 p.

Graham, S.A., Brassell, S., Carroll, A.R., Xiao, X., Demaison, G., McKnight,C.L., Liang,Y., Chu, J., and Hendrix, M.S., 1990, Characteristics of selectedpetroleum source rocks, Xinjiang Uygur Autonomous Region, China:American Association of Petroleum Geologists Bulletin, v. 74, p. 493–512.

Graham, S.A., Hendrix, M.S., Wang, L.B., and Carroll, A.R., 1993, Collisionalsuccessor basins of western China: Impact of tectonics inheritance of sandcomposition: Geological Society of America Bulletin, v. 105, p. 323–344.

Greene, T.J., 2000, Tectonics, basin analysis and organic geochemical attributesof Permian through Mesozoic deposits and their derivative oils of the Tur-pan-Hami Basin, northwestern China: [Ph.D. Dissertation], Stanford, Stan-ford University 392 p.

Greene, T.J., Carroll, A.R., Hendrix, M.S., and Graham, S.A., 2001, Sedimentaryrecord of intra-Mesozoic deformation and Middle to Late Triassic inceptionof the Turpan-Hami Basin, northwest China, in Hendrix, M.S., andDavis, G., eds., Paleozoic and Mesozoic tectonic evolution of central andeastern Asia—from continental assembly to intracontinental deformation:Geological Society of America Memoir 194, p. 317–340.

Hendrix, M.S., Graham, S.A., Carroll, A.R., Sobel, E.R., McKnight, C.L.,Schulein, B.J., and Wang, Z., 1992, Sedimentary record and climatic impli-cations of recurrent deformation in the Tian Shan: Evidence from Mesozoicstrata of the north Tarim, south Junggar, and Turpan Basins, northwestChina: Geological Society of America Bulletin, v. 104, p. 53–79.

Hopson, C., Wen, J., Tilton, G., Tang, Y., Zhu, B., and Zhao, M., 1989, Paleozoicplutonism in east Junggar, Bogdashan, and eastern Tianshan, northwestChina: Eos, v. 70, p. 1403–1404.


Hopson, C.A., Wen, J., and Tilton, G.R., 1998, Isotopic variation of Nd, Sr, andPb in Paleozoic granitoid plutons along an east Junggar-Bogdashan-Tian-shan transect, NW China [abs.]: Geological Society of America Abstractswith Programs, v. 30, p. 20.

Horváth, F., and Pogácsás, Gy., Contribution of seismic reflection data to chron-ostratigraphy of the Pannonian basin, in Royden, L.H., and Horváth, F.,eds., The Pannonian Basin; a study in basin evolution :American Associa-tion of Petroleum Geologists Memoir 45, p. 97–105.

Katz, B.J., and Mello, M.R., 2000, Petroleum systems of South Atlantic marginalbasins—An overview, in Katz, B.J., and Mello, M.R., eds., Petroleum sys-tems of south Atlantic margins: American Association of Petroleum Geolo-gists Memoir 73, p. 1–13.

Kutzbach, J.E., and Ziegler, A.M., 1993, Simulation of Late Permian climate andbiomes with an atmosphere-ocean model: comparison with observations:Philosophical Transactions of the Royal Society of London, B, v. 341,p. 327–340.

Lee, K.Y., 1985, Geology of the petroleum and coal deposits in the Junggar Basin,Xinjiang Uygur Zizhiqu, northwest China: U.S. Geological Survey Open-File Report 85-230, 53 p.

Liu, H., 1986, Geodynamic scenario and structural styles of Mesozoic and Ceno-zoic basins in China: American Association of Petroleum Geologists Bul-letin, v. 70, p. 377–395.

Mattick, R.E., Phillips, R.L., and Rumpler, J., 1988, Seismic stratigraphy and dep-ositional framework of sedimentary rocks in the Pannonian Basin in south-eastern Hungary, in Royden, L.H., and Horvath, F., eds., The PannonianBasin; a study in basin evolution: American Association of Petroleum Geol-ogists Memoir 45, p. 117–145.

Mello, M.R., and J.R. Maxwell, 1990, Organic geochemical and biologicalmarker characterization of source rocks and oils derived from lacustrineenvironments in the Brazilian continental margin, in Katz, B.J., ed., Lacus-trine basin exploraion: American Association of Petroleum GeologistsMemoir 50, p. 77–97.

Mello, M.R., Koutsoukos, E.A.M., Santos Neto, E.V., and Silva Tellas, J.A.C.,1993, Geochemical and micropaleontological characterization of lacustrineand marine hypersaline environments from Brazilian sedimentary basins:American Association of Petroleum Geologists Studies in Geology 37,p. 17–34.

Oelofsen, B.W., 1987, The biostratigraphy and fossils of the Whitehill and IratiShale Formations of the Karoo and Paraná Basins, in McKenzie, G.D., ed.,Gondwana six: Stratigraphy, sedimentology, and paleontology: AmericanGeophysical Union Monograph 41, p. 131–138.

Ojeda, H.A.O., 1982, Structural framework, stratigraphy, and evolution of Brazil-ian marginal basins: American Association of Petroleum Geologists Bul-letin, v. 66, p. 732–749.

Olsen, P.E., 1986, A 40-million year lake record of Early Mesozoic orbital cli-matic forcing: Science, v. 234, p. 842–848.

Peng, X., and Zhang, G., 1989, Tectonic features of the Junggar Basin and theirrelationshiop with with oil and gas distribution, in Zhu, X., ed., Chinese sed-imentary basins: Amsterdam, Elsevier, p. 17–31.

Peters, K.E., 1986, Guidelines for evaluating petroleum source rock using pro-grammed pyrolysis: American Association of Petroleum Geologists Bul-letin, v. 70, p. 318–329

Pitman, J.K., Pierce, F.W., and Grundy, W.D., 1989, Thickness, oil-yield, andkriged resource estimates for the Eocene Green River Formation, PiceanceCreek Basin, Colorado: U.S. Geological Survey Oil and Gas InvestigationsChart OC-0132, 6 sheets, 4 p.

Sengör, A.M.C., and Natal’in, B.A., 1996, Paleotectonics of Asia: Fragments of asynthesis, in Yin, A., and Harrison, M.T., eds., The tectonic evolution ofAsia: New York, Cambridge, p. 486–640.

Sengör, A.M.C., Natal’in, B.A., and Burtman, V.S., 1993, Evolution of the Altaidtectonic collage and Paleozoic crustal growth in Eurasia: Nature, v. 364,p. 299–307.

Smith, M.E., Pietras, J.T., Carroll, A.R., and Singer, B , 2001, Laser fusion40Ar/39Ar geochronology and cyclostratigraphy of the Green River Forma-tion: Testing micro- and meso-scale climate forcing [abs.]: Eos (Transac-tion, American Geophysical Union), v. 82, p. F1141.

Tapponnier, P., and Molnar, P., 1979, Active faulting and Cenozoic tectonics ofthe Tien Shan, Mongolia, and Baykal regions: Journal of GeophysicalResearch, v. 84, p. 3425–3459.

Teichert, C., 1974, Marine sedimentary environments and their faunas in Gond-wana area: American Association of Petroleum Geologists Memoir 23,p. 43–145.

Visser, J.N. J., 1994, A Permian argillaceous syn- to post-glacial forelandsequence in the Karoo Basin, South Africa, in Deynoux, M., Miller, J.M.G.,Domack, E.W., Eyles, N., Fairchild, I.J., and Young, G.M., eds., Earth’s gla-cial record: Cambridge, Cambridge University Press, p. 193–203.

Wang, X., 1992, Petroleum geological properties of the Permo-Carboniferous inthe Ili Basin: Oil and Gas Geology, v. 13, p. 219–228 (in Chinese with Eng-lish abstract).

Wartes, M.A., Carroll, A.R., and Greene, T.J., 2002, Permian stratigraphic evolu-tion of the Turpan-Hami Basin and adjacent regions, northwest China: Con-straints on post-amalgamation tectonic evolution: Geological Society ofAmerica Bulletin, v. 114, p. 131–152.

Wartes, M.A., Carroll, A.R., Greene, T.J., Cheng, K., and Hu, T., 2000, Permianlacustrine deposits of northwest China, in Gierlowski-Kordesch, E.H., andKelts, K., eds., Lake basins through space and time: American Associationfor Petroleum Geologists Studies in Geology 46, p. 123–132.

Watson, M.P., Hayward, A.B., Parkinson, D.N., and Zhang, Z.M., 1987, Plate tec-tonic history, basin development and petroleum source rock deposition inonshore China: Marine and Petroleum Geology, v. 4, p. 205–225.

Wen, J., 1991, Geochronological and isotopic tracer studies of some granitoidrocks from Xinjiang, China: Constraints on Paleozoic crustal evolution andgranitoid petrogenesis [Ph. D. thesis]: Santa Barbara, University of Califor-nia, 163 p.

Windley, B.F., Allen, M.B., Zhang, C., Zhao, Z.Y., and Wang, G.R., 1990, Paleo-zoic accretion and Cenozoic redeformation of the Chinese Tien Shan range,central Asia: Geology, v. 18, p. 128–131.

Yemane, K., 1993, Contributions of Late Permian paleogeography in maintaininga temperate climate in Gondwana: Nature, v. 361, p. 51–54.

Ziegler, A.M., Hulver, M.L., and Rowley, D.B., 1996, Permian world topographyand climate, in Martini, I.P. ed., Late glacial and postglacial environmentalchanges-Quaternary, Carboniferous-Permian and Proterozoic: New York,Oxford University Press, p. 111–142.

Zonenshain, L.P., and Le Pichon, X., 1986, Deep basins of the Black Sea andCaspian Sea as remnants of Mesozoic back-arc basins: Tectonophysics,v. 123, p. 181–211.

MANUSCRIPT ACCEPTED BY THE SOCIETY JANUARY 22, 2003


Printed in the USA

Documents

Carroll and Wartes, 2003.pdf