geochemical and organic carbon isotope studies across the continental permo^triassic boundary of...

Geochemical and organic carbon isotope studies across thecontinental Permo^Triassic boundary of Raniganj Basin,

eastern India

A. Sarkar a;�, H. Yoshioka b;1, M. Ebihara b, H. Naraoka b

a Department of Geology and Geophysics, Indian Institute of Technology, Kharagpur 721302, Indiab Tokyo Metropolitan University, Tokyo 192-03, Japan

Received 28 January 2002; received in revised form 23 September 2002; accepted 25 October 2002

Abstract

Organic carbon isotope and geochemical changes across a continental interior Permo^Triassic boundary sectionfrom the Raniganj Basin, India, indicate a V9x drop in organic carbon N

13C in the Early Triassic and synchroneityof this event throughout the Pangea. The study demands a common causative mechanism for the perturbation of theglobal carbon reservoir and not a combination of multiple causes. A global sea-level fall and oxidation of marine gashydrates possibly increased the 12CO2 input in the ocean^atmosphere system which caused a climatic shift from humidto warm semi-arid type and consequent extinction of land plants. Simultaneous increase in erosion from near-barrenlands, change in the erosional base level and provenance deposited the boundary sandstone with positive europiumanomaly and debris flow type matrix rich conglomerate. No extraterrestrial source, therefore, is needed to explain thisEu anomaly. The response of the terrestrial plant community to this perturbation of the carbon reservoir was,however, sluggish and the N

13C drop took place slowly, being maximal in the Early Triassic only.6 2002 Elsevier Science B.V. All rights reserved.

Keywords: P/T boundary; India; carbon isotope; REE

1. Introduction

The Permo^Triassic (P/T) extinction event wasthe most catastrophic one in the entire Phanero-zoic as about 57% and 95% extinction of marinefossils took place at family and species levels, re-spectively (Erwin, 1994). On land the events wereno less dramatic, manifested by the extinction of

the Permian mega£ora (e.g. Glossopteris ; Retal-lack, 1999), the global disappearance of coal (Re-tallack et al., 1996) and changeover to a new rep-tilian fauna (MacLeod et al., 2000). A rapidnegative excursion of N

13C of maximum up toV10x has invariably been found to be associ-ated with the marine P/T sections and is now usedas a global correlation tool (Holser et al., 1991;Grossman, 1994; Bowring et al., 1998). Until atpresent a large number of mechanisms have beenproposed to explain the events across the P/Tboundary, namely oceanic anoxia (Isozaki, 1997)or overturn of deep oceanic CO2 (Knoll et al.,

0031-0182 / 02 / $ ^ see front matter 6 2002 Elsevier Science B.V. All rights reserved.PII: S 0 0 3 1 - 0 1 8 2 ( 0 2 ) 0 0 6 3 6 - 3

1 Present address: Japan National Oil Corporation, Chiba261 0025, Japan.* Corresponding author.E-mail address: [email protected] (A. Sarkar).

PALAEO 2984 3-1-03

Palaeogeography, Palaeoclimatology, Palaeoecology 191 (2003) 1^14

R

Available online at www.sciencedirect.com

www.elsevier.com/locate/palaeo

mailto:[email protected]

http://www.elsevier.com/locate/palaeo

1996), the environmental greenhouse e¡ect (Retal-lack, 1999), volcanism (Renne et al., 1995), largescale release of methane gas hydrates (Morante,1996; de Wit et al., 2002) and impact (Bhandari etal., 1992; Retallack et al., 1998). Many of thesemodels were advanced to explain the negative car-bon isotope excursion, observed in the marine rec-ords, which can accommodate a large input of 12Cinto the ocean. However, for providing a unifyingglobal model it is equally important to knowabout the events on land. Compared to the ma-rine ones very little high resolution work has beendone in terrestrial P/T sections and most of ithas been reported from coal bearing sections inAustralia, Antarctica and South Africa (Retallackand Krull, 1999; Krull and Retallack, 2000). ThePangean reconstruction during P/T time showsthat India was in the Southern Hemisphere sand-wiched between Australia and Africa (Fig. 1). Ex-cepting the Tethyan (marine) P/T sequence of Spi-ti (Bhandari et al., 1992), no high resolution data

are available from peninsular India which couldbe an important input to any of the models men-tioned above. De Wit et al. (2002) have providedN13C variation across the Permo^Triassic bound-ary from number of Indian basins. Their materialsmainly derive from bore hole samples withoutmuch control on either lithology or facies andhave large sampling intervals (more than severalmetres).Furthermore, in view of the close connection

between geochemical changes in the oceans andextinction events in the earth history, as observedin case of the Cretaceous^Tertiary (C/T) event(Asaro et al., 1982; Kyte et al., 1985), it is worth-while to ask if any such changes took place acrossthe P/T. In particular, the oxygen poor condition,developed in the P/T ocean (Wignall and Twitch-ett, 1996), can be investigated by the use of redoxsensitive elements like U, V, Co, Mo and a rareearth element (REE) like Ce (Sarkar et al., 1993;Wang et al., 1986; Holser, 1997a,b). Bhandari et

Fig. 1. Pangea during the Permo^Triassic boundary with the location of the Raniganj Basin; also shown are other continentalGondwana basins preserving the boundary.

PALAEO 2984 3-1-03

A. Sarkar et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 191 (2003) 1^142

al. (1992) reported a positive europium (Eu)anomaly (unlikely to be found in terrestrial rocks)in the boundary samples of the marine Spiti sec-tion and inferred a possible impact of eucriticmeteorite (like Moore County) at the P/T bound-ary. Negative N

13C spikes, recorded at extinctionboundaries, has often been explained by a globalcollapse of the primary productivity or ecosystemdue to a bolide impact or shock heating of marinegas hydrates (Bowring et al., 1998 and referencestherein). Such an inference, however, requires fur-ther high resolution search from other P/T sec-tions both marine and continental. Here we reporthigh resolution organic carbon isotope and geo-chemical data (including REEs) across a continen-tal P/T section from the Raniganj Basin, the coalmining heartland of eastern India and discusstheir implications to the P/T events.

2. Geology and sampling

Continental Permian and Triassic rocks are ex-posed in many coal basins in India, e.g. the Bo-karo coal¢eld, Son^Mahanadi Basin. The bestsections are, however, preserved in the RaniganjBasin, West Bengal. The basin is an intra-cratonicone where the Permian rocks of the DamudaGroup unconformably lie over older Precambrianbasement commencing with the Early Permianglacial tillite beds of the Talchir Formation oftencorrelated with the Dwyka and Buckeye tillites ofSouth Africa and Antarctica, respectively(Krishnan, 1982). The Talchir Formation is con-formably overlain by sandstones of the Barakarand Ironstone shale formations. In Raniganj Ba-sin the maximum thickness of the Talchir, Bara-kar and Ironstone shale formations are V300,650 and 350 m, respectively (Krishnan, 1982).The topmost unit of the Damuda Group is repre-sented by the Raniganj Formation which in turnlies below the Panchet Formation. The contactbetween the Raniganj Formation of Late Permianage and the Early Triassic Panchet Formation iseither gradational or characterised by local un-conformity. In particular, the Banspetali sectionin Raniganj Basin, exposing the latest Raniganjand earliest Panchet rocks, has recently been sug-

gested as the type continental P/T section of India(Ghosh, 1994). The earliest map of the RaniganjBasin in general and the Banspetali area in par-ticular was prepared by Gee (1932). We have re-mapped the area (Fig. 2) for critically examiningthe lithological and faunal changes across the P/Tboundary and sampling for geochemical and iso-topic studies.The Banspetali area is located about 160 km

WNW of Calcutta (23‡37PN, 86‡54PE). Theuppermost Permian and earliest Triassic rocksare exposed along a stream (nala) section £owingnorthward almost perpendicular to the E^Wstriking beds. The sedimentaries occur within thePrecambrian basement, composed of granitegneiss, amphibolite and metaquartzites. The basinboundaries are faulted where both the Raniganjand Panchet rocks abut against the basement inthe southern part. Evidences of syn- and post-sed-imentation faulting are also numerous (Fig. 2).The Permian rocks of the Raniganj Formationinclude an alternating sequence of sandstone,shale and coal. The white/gray sandstones are pla-gioclase rich and ¢ne to medium grained. Theshales are dark, rich in carbonaceous material.The Panchet sandstones are, on the contrary,rich in unaltered pink coloured potash feldspareasily identi¢able in hand specimens. The Panchetshales are gray to olive green coloured, much dif-ferent from Raniganj ones. In the upper part bothPanchet sandstone and shale become reddish in-dicating the presence of ferruginous cements pos-sibly due to severe oxidising conditions. The top-most litho-unit in the area is red coloured highlyimmature, poorly sorted, pebbly sandstone andconglomerate of the Supra-Panchet Formationof possible Late Triassic age.The litho^biostratigraphy across the P/T

boundary of the Banspetali section is shown inFig. 3. The P/T boundary is recognised in the ¢eldby the last Permian coal seam (30 cm thick) oc-curring within dark shale. The coal seam acts as amarker horizon to identify the terminal Raniganjsedimentation and can be traced in another nalasection at Tentulrakh, V3 km east of Banspetali.In the present study area at least 400 m of Rani-ganj Formation are present below the terminalPermian coal seam but a considerable part of it

PALAEO 2984 3-1-03

A. Sarkar et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 191 (2003) 1^14 3

has been down faulted along the southern contactwith the Precambrian basement (Fig. 2). Coalseams completely disappear in the overlying Pan-chet rocks. According to Ghosh (1994), the top-most bed of Raniganj is represented by a 5.5-m-thick single sandstone bed overlying the last coalseam (Plate I). However, lithologically this sand-stone is more akin to the Panchet sandstones con-taining pink potash feldspar. The sandstone isoverlain by a 9-m-thick matrix-supported con-glomerate (Plate II), the upper 4 m of which isnot well exposed. The conglomerate is polymictic

with pebbles of quartz, feldspar and basementrock fragments set in sandy and clayey matrix.There is a lack of any grading and the large an-gular £oating clasts with poor sorting of this unitare much like debris £ow products (Postma et al.,1988; Sohn, 2000). The sandstone is convolutelaminated. Both the sandstone and conglomerateare laterally persistent and indicate large scale in-stability of the basin and related syn-sedimentarydeformation. The conglomerate is succeeded bygray to olive green shale^siltstone of the PanchetFormation (Plate III). Floristically the Raniganj is

Fig. 2. Geological map of part of the Raniganj Basin around the Banspetali area showing the sampling location of the P/T sec-tion.

PALAEO 2984 3-1-03


characterised by the Glossopteris^Vertebrariamega£oral assemblage while the Panchet containsthe Dicroidium^Lepidopteris assemblage. Palyno-morph studies indicate a striate^disaccate Densi-pollenites rich micro£oral assemblage in terminalRaniganj while the beginning of Panchet ismarked by the appearance of Lunatisporites andcavate and zonate triletes, Densoisporites, Lund-bladispora assemblage of Early Triassic age (Ti-wari and Tripathi, 1992; Ghosh, 1994). The onsetof the Panchet is also marked by the widespreadoccurrence of the fossil conchostracans generallyknown as estheriids belonging to the Cyzicus bio-zone. The typical assemblage comprises forms likeCyzicus bengalensis Tasch, C. mangaliensis Tasch,Cornia panchetella Tasch, indicating a typicalEarly Triassic age. Ghosh (1994) has identi¢ed

at least three estheriid biozones within the Pan-chet Formation mostly occuring in ¢ne grainedclastics. Speci¢c morphological features of the es-theriids along with sedimentary structures likemud cracks, root burrows indicate a shallowfreshwater, desiccated high salinity environmentfor the estheriid bearing rocks of the Panchet.Together, the lithology and biota across the Ra-niganj^Panchet boundary indicate pronouncedclimatic and tectonic change from a humid coalbearing regime to warm semi-arid climatic regime.Under the microscope the typical Raniganj sand-stones occurring below the last coal seam displaya well rounded, low matrix, matured texture. Thesandstones of Raniganj above the coal seam andPanchet exhibit an angular, high matrix, imma-ture texture. A well-preserved vertically oriented

Fig. 3. Litho- and biostratigraphy across the P/T section along with the N13Corg pro¢le; note large N

13Corg drop in Early Triassic.

PALAEO 2984 3-1-03


Plate I. Late Permian boundary sandstone from the Banspetali area, Raniganj Basin. The last coal seam occurs just below thislayer within the underlying dark shale of the Raniganj Formation.

Plate II. Matrix supported polymictic conglomerate at the P/T boundary.

PALAEO 2984 3-1-03


stem of Schizoneura has been found to be pre-served in a normally graded (into mud) coarsesiltstone just above the last coal seam indicatingrapid sedimentation under a turbidity £ow. Itseems that the in£ux of ¢ner clastics into the ba-sin suddenly increased towards the terminal Ra-niganj (Permian) reaching its peak just at theboundary when a regional debris £ow produceda matrix supported conglomerate. A similarlythick laterally persistent conglomerate at the P/Tboundary has also been found at several sectionsin Australia and is interpreted as the evidence ofrapid sedimentation rather than a time gap (Re-tallack, 1999). The increased sediment input mustbe a result of the change in the local base leveleither due to the upliftment of basement Precam-brian rocks or a global marine regression (Holserand Magaritz, 1987). Alternatively, an enhancederosion of the river banks due to the extinction ofPermian mega£ora might be the cause of this in-creased sediment input, a mechanism suggested byWard et al. (2000) for the P/T boundary event inKaroo Basin. Following Ghosh (1994), we placethe P/T boundary in Banspetali section at the

junction between the conglomerate and the over-lying gray siltstone yielding de¢nitive Early Trias-sic £ora. The sampling for isotopic and geochem-ical studies of both sandstone and shale/siltstoneunits have been made over 30 m of the exposedsections (excepting the conglomerate unit) span-ning the P/T boundary.

3. Methods

For organic carbon isotope the relatively freshunweathered samples were analysed. The pow-dered sample was repeatedly treated with 6 NHCl both at room temperature and at 60‡C andwashed to remove all carbonate phase. The sam-ples were freeze-dried and combusted at 1000‡Cin a Carlo Erba EA-1108 unit connected online toa Finnigan MAT delta-S mass spectrometer. Thepuri¢ed CO2 was measured for its N13C composi-tion with an average analytical precision ofN 0.1x. All the carbon isotope data are reportedagainst the Pee Dee Belemnite (PDB) standard.The trace and REEs were determined by instru-

Plate III. Earliest Triassic gray siltstone.

PALAEO 2984 3-1-03


mental neutron activation analysis. About 50 mgof each powdered (6 200-mesh) sample was irra-diated at the F-ring irradiation site (nominal ther-mal neutron £ux: 1.5U1012 cm32 s31) of the TRI-GA-II reactor at the Institute for Atomic Energy,Rikkyo University, Japan. The Q-rays emitted bythe samples were measured by highly pure Ge-detectors. Elemental abundance values were ob-tained by comparing the Q-ray intensities of thesamples with those of reference standard rockssuch as JB-1 (basalt) and JG-1 (granodiorite) sup-plied by the Geological Survey of Japan. The Al-lende meteorite reference sample prepared by theSmithsonian Institute, Washington, USA, was si-multaneously irradiated. For the calculation theliterature values provided by Jarosewich et al.(1987) and Ando et al. (1989) were used for JB-1, JG-1 and Allende, respectively.

4. Results and discussion

The N13C data are plotted against depth in Fig.

3. We have not been able to analyse any samplefrom the conglomerate horizon as these samplesare extremely weathered due to high matrix con-tent. Although minor £uctuations occur, the N

13Cvalue remains more or less constant at an average325x level during the Late Permian and earliestTriassic immediately following the boundary. Asigni¢cant drop in N

13C value is observed atV800 cm above the P/T boundary which slowlyreturns to the pre-depletion level in the overlyingsediments. The average drop in N

13C is V6xwhile the maximum depletion between the pre-and post-boundary level is more than 9x. Thehighest drop is observed at the base of the olivesiltstone of the Panchet Formation. A similarlylarge depletion in organic carbon N

13C has beenobserved from several coal bearing Permo^Trias-sic sections of Australia and South Africa (Faureet al., 1995; Morante, 1996; Retallack, 1999).Large N13C depletion (V10x) at the P/T bound-ary has also been reported in the reptilian tusksfrom Karoo Basin, South Africa (MacLeod et al.,2000). Our data along with the Australian andAfrican ones indicate the synchroneity in carbonisotope pattern throughout the Pangea and sug-

gest that the change in the carbon reservoir (asobserved in the marine realm) indeed a¡ected thecontinental domain as well. Such a remarkablesimilarity in carbon isotope pattern from widelydi¡erent geographical localities also indicates thisto be a genuine climatic signal and not an artifactof mixing of various plant components (Foster etal., 1999). This is further supported by the factthat the N

13C gradually goes back to the pre-boundary level in the later part of Triassic eventhough the Glossopteris dominated Permian £orabecame completely extinct at the P/T boundary.Hansen et al. (2000), on the basis of magneticsusceptibility stratigraphy of both Permian andTriassic segments, have proposed a sedimentationrate (for the shale/siltstone part only) of V165cm/100 kyr for this section. This is similar tothe rate obtained from study of the absolute chro-nology of Permian authigenic illites (Dutta andSuttner, 1986). If true, the drop in N

13C and itsrecovery during an interval of 900 cm indicatethat the event took place over a period of s 0.5Ma. This is indeed a large time interval and sup-ports the observation of Retallack (1999) that theP/T event (at least the instability in the continen-tal carbon reservoir) was gradual rather thanabrupt. The gradational change in N

13C in theBanspetali section also indicates a rather completestratigraphic record and not a disconformity. Theabrupt changes found in some continental records(Morante et al., 1994; Morante, 1996) are possi-bly due to the incompleteness of the record. Thehigh resolution U^Pb geochronology from theMeishan section of China, however, indicated acatastrophic change in the marine carbon reser-voir (Bowring et al., 1998). It is possible thatthe response of the continental biosphere wassluggish compared to the marine system.De Wit et al. (2002) reported a N

13Corg pro¢leacross several Indian P/T boundary sections in-cluding the Raniganj Basin. Large oscillations inN13C have been observed where the gradual neg-ative trend in the Upper Permian and a similarpositive recovery in the Lower Triassic, separatedby up to three large negative 13C spikes, havebeen observed. The average N

13C value(V324x) is similar to that obtained in thepresent study (V325x), however, the Banspe-

PALAEO 2984 3-1-03


tali section records only one large negative dropin the earliest Triassic. We observe some smalleroscillations (Fig. 3) but the positive swings havenever been as large as reported by de Wit et al.(2002). This is indeed surprising as the sampling/analyses resolution is much higher in the presentcase compared to that reported by these authors.Since all of their samples come from bore holes, itis not possible to know exactly what lithologicaland facies attributes these isotope maxima andminima have. Our N

13C pro¢le is more akin tothose from Australian basins like Sydney, Bowenor Bonaparte or even some of the continuous ma-rine P/T boundary sections.It is, however, di⁄cult to explain the causative

mechanism of the N13C drop. Absence of (or

weak) iridium anomaly (Bhandari et al., 1992;Retallack et al., 1998) and gradual change in con-tinental carbon isotope pro¢les do not favour theimpact hypothesis. Since the N

13C depletion isfound both in marine and continental sediments,a mechanism involving the global perturbation ofthe carbon reservoir must be invoked. Isozaki(1997) provided some geological evidence ofsuper-anoxia in the global ocean at and acrossthe P/T boundary and hence the increased oxida-tion of marine organic matter cannot explain thedepletion. An alternative mechanism of oxidationof continental organic matter, due to collapse ofthe terrestrial ecosystem and global tectonic up-lift, has been proposed by Faure et al. (1995). Aslightly di¡erent mechanism has been proposed byWard et al. (2000) where the collapse of the ter-restrial biosphere and the consequent increasederosion enhanced the terrestrial negative carboninput into the ocean. An increasingly popularmodel of oxidation of methane hydrates due tothe global sea-level fall has also been forwarded(Morante, 1996; de Wit et al., 2002). Choosing acandidate from any of these models is di⁄cult,although it is tempting to speculate that the 9-m-thick matrix supported conglomerate foundjust before the boundary in Banspetali (alongwith the similar ones observed elsewhere) was in-deed the result of massive debris £ow caused dueto the fall in sea-level and the consequent changein base level of erosion (and oxidation of gas hy-drates?). The sudden increase in matrix content

(composed of very ¢ne sand sized quartz grainsand clay) in the laterally persistent 5.5 m-thickconvolute laminated sandstone and the presenceof turbidite preserving a vertical plant stem nearthe P/T transition also indicate a change in theerosional pattern in the source and £uid dynam-ics. This is, however, not to undermine the localbasinal tectonics (evidenced by basin boundaryfaults) which could have acted together with thesea-level fall (Holser and Magaritz, 1987) to bringthe lithological changes across the boundary. Thecoupled ocean^atmosphere model calculation ofde Wit et al. (2002) indicates that the release of12CO2 could be episodic as indicated by multiplenegative N

13C excursions. At Banspetali too atleast two such (smaller) excursions are observedbefore the ¢nal large one at the earliest Triassic.Interestingly, the ¢rst excursion also coincideswith a major geochemical (REE) anomaly (Fig.3) as discussed below.Ward et al. (2000), on the basis of facies and

£uvial architecture analyses of Karoo Basin, pro-vided strong evidence of the dramatic changeoverin river morphology from meandering to braidedsystems across the P/T boundary. The changeoverwas the consequence of basin wide landscape de-stabilisation and increased erosion of both soiland river banks due to terrestrial plant die-o¡during this transition. The increase in sedimentyield characteristically takes place in the earliestTriassic similar to the one found in the RaniganjBasin. It would be interesting to see if such a rivermorphology induced facies change is also pre-served across the Indian P/T boundary and reallyforms a global phenomenon. Nevertheless, thematrix rich boundary conglomerate and abruptlyoverlying olive siltstone^shale package (Fig. 3) in-dicate that it could well be the product of a rapid£ash £ood, as proposed by Ward et al. (2000),followed by a quick deceleration of energy. Arapid increase in marine 87Sr/86Sr ratio (Faure etal., 1995; Morante, 1996) at the P/T boundaryalso indicates a dramatic increase in the continen-tal input due to the enhanced erosion of the riv-ers.Geochemically, both the Raniganj and Panchet

shale/siltstone formations have a higher Fe con-tent (4^7%) compared to the boundary sandstone

PALAEO 2984 3-1-03


(2^3.5%); so are U (maximum 8 ppm against1 ppm), Ni, Co, and Cr which are relatively moreconcentrated in organic rich ¢ner clastics of pos-sible over bank origin. The U concentration in thePanchet rocks is less (2^5 ppm) compared to thatof Raniganj (6^8 ppm) and is consistent with amore oxidising climatic condition during theEarly Triassic. The Na concentration in the sand-stone is higher than that in the shale/siltstone dueto the potash and alkali feldspar rich modal min-eralogy. The chondrite normalised REE patternsof coal and dark shale of Raniganj, the sandstonenear the boundary and gray shale, olive shale/silt-stone of the Panchet Formation are shown in Fig.4. The 4REE of the coal/shale/siltstone of eitherRaniganj or Panchet is much higher (203 to 348ppm) compared to the boundary sandstone (59^124 ppm) which is consistent with the fact that the

clays preferentially concentrate about 1.5 timesmore REE than sandstones (Henderson, 1984;Wang et al., 1986). It is interesting to note thatthe REE composition of the P/T boundary sand-stone is near-chondritic. The most notable featurein Fig. 4 is the presence of a positive europiumanomaly (Eu/Eu*, i.e. the ratio of the measuredEuN value to the interpolated Eu value betweenneighbouring SmN and GdN) in all the samples ofthe boundary sandstone (1.2^1.59) compared toeither Raniganj shale (0.33^0.35) or Panchetshale/siltstone (0.04 to 0.63) which exhibit normalnegative europium anomalies. A similarly positiveeuropium anomaly has earlier been reported in alimonitic layer at the P/T boundary from the ma-rine carbonate sequence of Spiti Himalayas(Bhandari et al., 1992). The Banspetali section isthe second location where the positive europium

Fig. 4. Chondrite normalised REE patterns of sediments across the P/T boundary; note positive Eu anomaly in all the samplesof boundary sandstones.

PALAEO 2984 3-1-03


anomaly has been observed. Compilation of sedi-ment REE data shows that the positive europiumanomaly in the post-Archaean rocks is very raredue the more evolved nature of the crust generat-ing the sediments (Taylor and McLennan, 1985).The only exceptions are sedimentary barite (whereEu2þ substitutes Ba2þ) and some iron formationsand sul¢des (via interaction of hydrothermal sul-¢des with felsic volcanic rocks; Cullers and Graf,1984). In the present case neither any correlationis observed between Eu/Eu* and the Ba content ofthese rocks nor any evidence of hydrothermal ac-tivity in this clastic sandstone. Therefore, none ofthese mechanisms can explain the positive Euanomaly. Change in redox condition, causingchange in the Eu2þ/Eu3þ ratio (Henderson,1984), is not responsible either as we do not ¢ndany correlation between the (Ce/La)N ratio (Cebeing redox sensitive) and Eu/Eu*. Positive Euanomalies in carbonate rich river sediments havealso been reported where the Eu2þ replacement ofSr2þ in carbonate minerals has been held respon-sible for such an e¡ect (Leleyter et al., 1999).However, the boundary sandstone does not con-tain carbonate either in framework or in matrix/cement. Bhandari et al. (1992) indicated the pos-sibility of the impact of an eucritic meteorite atthe P/T boundary of Spiti which exhibits positiveeuropium anomaly. However, the europiumanomaly at Spiti is con¢ned within an only thin(V5-cm) limonitic layer at the boundary whereasat Banspetali all three samples from the bottom,top and middle of the 5.5 m-thick sandstone ex-hibit the anomaly. In no case such a thick sand-stone bed can be the product of an instantaneousdeposition as is required for a meteorite impact.Therefore, we consider that the positive Eu/Eu*anomaly is not produced due to any syn-sedimen-tary or extraterrestrial process rather that it re-£ects the REE geochemistry of the provenancerock from which the sandstone minerals are gen-erated.Lithologically the sandstone is composed of

pink potash feldpar and quartz with high matrixcontent. Although both plagioclase and potashfeldspars concentrate more Eu relative to otherminerals, even the post-Archaean arkosic rocksare characterised by a typically negative Eu

anomaly only (Taylor and McLennan, 1985).This is consistent with their derivation from neg-ative anomaly bearing typically upper continentalpotash rich granitic rocks. Hence the boundarysandstone must have had a very speci¢c litholog-ical provenance which gave rise to this anomaly.The Precambrian rocks with positive Eu anomalyare tonalite^tronjhemite gneiss (TTG), granodio-rite and quartz diorite. The TTG are mostly con-¢ned to the Archaean and their positive anomalyarises not because of feldspar accumulation but isdue to hornblende-melt equilibria (Cullers andGraf, 1984). The basement rocks of the easternIndian coal basin including the Banspetali belongto the Chotanagpur Gneissic Complex (CGC)which has been dated as Middle to Late Protero-zoic (Majumder, 1996; Roy et al., 2002). Also, thehornblende content in the sandstone, as an acces-sory mineral, is extremely low and hence therecannot be any TTG source responsible for thisanomaly. Diorite^granodiorites, produced by par-tial melting of amphibolite or eclogite, often dis-play a positive Eu anomaly (Cullers and Graf,1984). Although not present within the mappedarea of the Banspetali region, in general theCGC has a vast tract of diorite^granodioriterocks some showing rapakivi like textures of analkali feldspar core rimmed by plagioclases (Ma-jumder, 1996). It is possible that some of theseareas were acting as the provenance of the bound-ary sandstone. Either basement uplift or sea-levelfall and change in the erosion pattern might haveinduced a switch over in provenance from Eu de-pleted to Eu rich rocks near the boundary. Only adetail REE analysis of the various components ofthe CGC can identify the exact source of the pos-itive Eu anomaly in these rocks. In any case thepositive Eu anomaly in the Banspetali P/T sectionis caused by a terrestrial process only and there isno need to invoke an extraterrestrial one. A noteof caution, however, must be added at this point.Together, the near-chondritic REE compositionand the positive Eu anomaly demands that fur-ther search must be made to locate and identifyimpact components, if any, in this sandstone.Whatever may be the cause, the positive Eu

anomaly preceded the carbon cycle perturbation(N13C depletion) by a relatively long period of

PALAEO 2984 3-1-03


time unless both the sandstone and conglomeraterepresent instantaneous event beds of impact ori-gin. The anomaly also occurs immediately afterthe last very thin (V30 cm) Late Permian coalseam, coinciding with the ¢rst minor negativeN13C excursion, meaning the changes in sedimen-tation pattern and £oral extinction were almostcoeval. Oxygen isotope (N18O) analyses of earlyauthigenic clays in sandstones across the Rani-ganj/Panchet boundary showed a large increasefrom 6.6x to 10.4x and have been interpretedas a climatic changeover from a humid to warmsemi-arid type (Dutta and Suttner, 1986). Anoverall increase in the abundance of smectite,chlorite, and pink potash feldspar (Basu, 1976;Potter, 1978), speci¢c dry-regime sedimentarystructures (mentioned above) and the completedisappearance of coal in Panchet compared tokaolinite and the coal-measure rich Raniganj For-mation support such an interpretation. As men-tioned above, the extreme ferrugination in the ma-trix and clays in some of the sandstone and shaleof the Panchet Formation also indicates an in-creased aridity. We envisage that towards LatePermian time large scale release of 12C enrichedCO2 from oxidation of gas hydrates (via sea-levelfall) possibly introduced a climatic change acrossthe P/T boundary causing extinction of landplants. The ensuing base level readjustment alongwith increased erosion from near-barren landsand change in provenance deposited the boundarysandstone with positive Eu anomaly followed by adebris £ow type matrix rich conglomerate. Theresponse of the terrestrial plant community tothis perturbation of the carbon reservoir was,however, sluggish and manifested by a V9xN13C drop much later in the Early Triassic only.

5. Conclusions

Lithological, biotic, organic carbon isotope andgeochemical studies indicate pronounced climaticand/or tectonic change across a continental Per-mo^Triassic boundary section from the Banspetaliarea, Raniganj Basin, India. The large drop ofmaximum up to V9x in organic carbon N

13Cin the Early Triassic is similar to those observed

from Australian and African P/T sections and in-dicates synchroneity of this event throughout thePangaea. The N

13C depletion at the P/T boundaryin both the oceanic and continental domains in-dicate common causative mechanisms of the per-turbation of the global carbon reservoir and not acombination of multiple causes. A positive euro-pium anomaly, observed in the boundary sand-stone, can be explained by the change in sedimentprovenance without invoking the contributionfrom any extraterrestrial source. The sandstoneoccurs immediately after the last Permian coalseam which indicates that the changes in sedimen-tation pattern and £oral extinction were almostcoeval. The observed geological changes favourmodels of release of 12C enriched methane hy-drates due to the global end-Permian regression(via sea-level fall) which introduced a climaticchange (humid to warm semi-arid) across theP/T boundary causing extinction of land plants.The ensuing base level readjustment of the riversalong with increased erosion from near-barrenlands and change in provenance deposited bound-ary sandstone with positive Eu anomaly followedby a debris £ow type matrix rich conglomerate.The response of the terrestrial plant communityto this perturbation of the carbon reservoir was,however, sluggish (at Banspetali the event is spreadover s 0.5 Ma) compared to the marine ones andthe V9x N

13C drop took place much later in theEarly Triassic only. Any future modelling of theperturbation of the carbon reservoir at the P/Tboundary must take this aspect into account.

Acknowledgements

Part of this work was carried out during a visit-ing fellowship programme at Tokyo MetropolitanUniversity (TMU) granted to A.S. by the IndianNational Science Academy, New Delhi, and theJapan Society for Promotion of Science, Tokyo.We thank the TMU for providing experimentalfacilities. Financial grant from ISIRD, SRIC,IIT, Kharagpur towards ¢eld work is gratefullyacknowledged. A.S. thanks Dr. P.P. Chakrabartyand S.C. Ghosh for stimulating discussions. Thispaper considerably bene¢ted from the critical

PALAEO 2984 3-1-03


reviews by Drs. M.R. Saltzman and G.J. Retal-lack.

References

Ando, A., Kamioka, H., Terashima, S., 1989. 1988 values forGSJ rock reference samples, ‘Igneous rock series’. Geochem.J. 23, 143^148.

Asaro, F., Alvarez, L.W., Alvarez, W., Michel, H.V., 1982.Geochemical anomalies near the Eocene^Oligocene and Per-mian^Triassic boundaries. Geol. Soc. Am. Spec. Pap. 190,517^528.

Basu, A., 1976. Petrology of the Holocene sand derived fromplutonic source rocks: implications to paleoclimatic inter-pretations. J. Sediment. Petrol. 46, 694^709.

Bhandari, N., Shukla, P.N., Azmi, R.Z., 1992. Positive euro-pium anomaly at the Permo^Triassic boundary, Spiti, India.Geophys. Res. Lett. 19, 1531^1534.

Bowring, S.A., Erwin, D.H., Jin, Y.G., Martin, M.W., Davi-dek, K., Wang, W., 1998. U/Pb zircon geochronology andtempo of the end-Permian mass extinction. Science 280,1039^1045.

Cullers, R.L., Graf, J.L., 1984. Rare earth elements in igneousrocks of the continental crust: intermediate and silicic rocks-ore petrogenesis. In: Henderson, P. (Ed.), Rare Earth Ele-ment Geochemistry. Elsevier, pp. 275^316.

de Wit, M.J., Ghosh, S.C., Joy, G., de Villiers, S., Rakotoso-lofo, N., Alexander, J., Tripathi, A., Looy, C., 2002. Multi-ple Organic Carbon Isotope Reversals across the Permo^Triassic Boundary of Terrestrial Gondwana Sequences:Clues to Extinction Patterns and Delayed Ecosystem recov-ery. J. Geol. 110, 227^240.

Dutta, P.K., Suttner, L.J., 1986. Alluvial sandstone composi-tion and paleoclimate. II Authigenic mineralogy. J. Sedi-ment. Petrol. 56, 346^358.

Erwin, D.H., 1994. The Permo^Triassic extinction. Nature367, 231^236.

Faure, K., de Witt, M.J., Willis, J.P., 1995. Late Permianglobal coal hiatus linked to C-13 depleted CO2 £ux intothe atmosphere during the ¢nal consolidation of Pangea.Geology 2, 507^510.

Foster, C.B., Logan, G.A., Summons, R.E., 1999. The Per-mian^Triassic boundary in Australia: Organic carbonisotope anomalies relate to organofacies, not a biogeochem-ical ‘event’. Ninth Annual V.M. Goldschmidt conference.Lunar and Planetary Institute Contribution 971, pp. 87^88.

Gee, E.R., 1932. The Geology and Coal Resources of Rani-ganj Coal¢eld. Record Geological Survey of India, 61.

Ghosh, S.C., 1994. Study of Permo^Triassic boundary inGondwana sequence of Raniganj Basin, India. Ninth Inter-national Gondwana Symposium, vol. 1, pp. 179^193.

Grossman, E.L., 1994. The carbon and oxygen isotope recordduring the evolution of Pangea: Carboniferous to Triassic.Geol. Soc. Am. Spec. Pap. 288, pp. 207^228.

Hansen, H.J., Lojen, S., Toft, P., Dolenec, T., Jinnan, T, Mi-chaelsen, P., Sarkar, A., 2000. Magnetic susceptibility andorganic carbon isotopes of sediments across some marineand terrestrial Permo^Triassic boundaries. In: Yin, H.,Dickins, J.M., Shi, J.R., Tong, J. (Eds.), Permian^TriassicEvolution of Teyhys and Western Circum-Paci¢c. Elsevier,pp. 271^289.

Henderson, P. (Ed.), 1984. Rare Earth Element Geochemistry.Elsevier, 510 pp.

Holser, W.T., Magaritz, M., 1987. Events near the Permian^Triassic boundary. Mod. Geol. 11, 155^180.

Holser, W.T., Schonlaub, H.-P., Boeckelman, K., Magaritz,M., Orth, C., 1991. The Permian^Triassic of the Gartfun-kel-1 core (Carnic Alps, Austria): synthesis and conclusions.Abh. geol. Bundesanst. Wien 45, 213^232.

Holser, W.T., 1997a. Geochemical events documented in inor-ganic carbon isotopes. Palaeogeogr. Palaeoelimatol. Palaeo-ecol. 132, 173^182.

Holser, W.T., 1997b. Evaluation of the application of rareearth elements to paleoceanography. Palaeogeogr. Palaeocli-matol. Palaeoecol. 132, 309^323.

Isozaki, Y., 1997. Permo^Triassic boundary super-anoxia andstrati¢ed superocean: records from lost deep sea. Science276, 235^238.

Jarosewich, E., Clarke, R.S. Jr., Barrows, J.N., 1987. The Al-lende meteorite reference sample. Smithsonian Contribu-tions in Earth Science 27, pp.1^12.

Knoll, A.H., Bambach, R.K., Can¢eld, D.E., Grotzinger, J.P.,1996. Comparative earth history and late Permian mass ex-tinction. Science 273, 452^457.

Krishnan, M.S., 1982. Geology of India and Burma, 6th ed.CBS Publ., 536 pp.

Krull, E.S., Retallack, G.J., 2000. 13C depth pro¢les from pa-leosols across the Permian^Triassic boundary: Evidence formethane release. Bull. Geol. Soc. Am. 112, 1459^1472.

Kyte, F.T., Smit, J., Wasson, J.T., 1985. Siderophile interele-ment variations in the Cretaceous^Tertiary boundary sedi-ments from Caravaca, Spain. Earth Planet. Sci. Lett. 73,183^195.

Leleyter, L., Probst, J.L., Depetris, P. et al., 1999. Rare earthelement speciation in river sediments. Journal of ConferenceAbstracts, Symposium on Aqueous Geochemistry, vol. 4,p. 1.

MacLeod, K.G., Smith, R.M.H., Koch, P.L., Ward, P.D.,2000. Timing of mammal-like reptile extinctions across thePermian^Triassic boundary in South Africa. Geology 28,227^230.

Majumder, S.K., 1996. Precambrian geology of peninsular In-dia. Indian Miner. 50, 139^174.

Morante, R., Veevers, J.J., Andrew, A.S., Hamilton, P.J.,1994. Determining the Permian^Triassic boundary in Aus-tralia using C-isotope chemostratigraphy. Aust. Pet. Explor.Assoc. J. 34, 330^336.

Morante, R., 1996. Permian and early Triassic isotopic recordsof carbon and strontium in Australia and a scenario ofevents about the Permian^Triassic boundary. Hist. Biol.11, 289^310.

PALAEO 2984 3-1-03


Postma, G., Nemec, W., Kleinspehn, K.L., 1988. Large £oat-ing clasts in turbidites: a mechanism for their emplacement.Sediment. Geol. 58, 47^61.

Potter, P.E., 1978. Signi¢cance and origin of big rivers. J. Geol.86, 13^33.

Renne, P.R., Zichao, J., Richards, M.A., Black, M.T., Basu,A.R., 1995. Synchrony and casual relations between Per-mian^Triassic boundaries and Siberian £ood volcanism. Sci-ence 269, 1413^1416.

Retallack, G.J., Veevers, J.J., Morante, R., 1996. Global earlyTriassic coal gap between Late Permian extinction and Mid-dle Triassic recovery of peat-forming plants. Bull. Geol. Soc.Am. 108, 195^207.

Retallack, G.J., Seyodolali, A., Krull, E.S., Holser, W.T., Am-bers, C.P., Kyte, F.T., 1998. Search for evidence of impactat the Permian^Triassic boundary in Antarctica and Austra-lia. Geology 26, 979^982.

Retallack, G.J., 1999. Postapocalyptic greenhouse palaeocli-mate revealed by earliest Triassic paleosols in the Sydneybasin, Australia. Geol. Soc. Am. Bull. 111, 52^70.

Retallack, G.J., Krull, E.S., 1999. Landscape ecological shift atthe Permian^Triassic boundary in Antarctica. Aust. J. EarthSci. 46, 786^812.

Roy, A., Sarkar, A., Jeyakumar, S., Ebihara, M., 2002. Mid-Proterozoic plume related thermal event in eastern Indian

craton evidence from trace elements, REE geochemistry andSr^Nd isotope systematics of ma¢c^ultrama¢c intrusivesfrom Dalma volcanic belt. Gondwana Res. 5, 133^146.

Sarkar, A., Bhattacharya, S.K., Sarin, M.M., 1993. Geochem-ical evidence for anoxic deep water in the Arabian Sea dur-ing the last glaciation. Geochim. Cosmochim. Acta 57,1009^1016.

Sohn, W.K., 2000. Depositional process of submarine debris£ow in the Miocene fan deltas, Pohans basin, SE Korea withspecial reference to £ow transformation. J. Sediment. Res.70, 491^503.

Taylor, S.R., McLennan, S.M., 1985. The Continental Crust:Its Composition and Evolution. Blackwell, Oxford, 312 pp.

Tiwari, R.S., Tripathi, A., 1992. Marker assemblage zone ofpore and spollen species through Gondwana Paleozoic andMesozoic sequence in India. Paleobotanist 40, 194^236.

Wang, Y.L., Liu, Y.G., Schmitt, R.A., 1986. Rare earth geo-chemistry of South Atlantic deep sea sediments: Ce anomalychange at V54 Ma. Geochim. Cosmochim. Acta 50, 1337^1355.

Ward, P.D., Montgomery, D.R., Smith, R., 2000. Altered rivermorphology in South Africa related to the Permian^Triassicextinction. Science 289, 1740^1743.

Wignall, P.B., Twitchett, R.J., 1996. Oceanic anoxia and theend-Permian mass extinction. Science 272, 1155^1158.

PALAEO 2984 3-1-03


geochemical and organic carbon isotope studies across the continental permo^triassic boundary of...

Documents