Quick viewing(Text Mode)

Global Changes During Carboniferous–Permian Glaciation of Gondwana: Linking Polar and Equatorial Climate Evolution by Geochemical Proxies

Global Changes During Carboniferous–Permian Glaciation of Gondwana: Linking Polar and Equatorial Climate Evolution by Geochemical Proxies

Global changes during glaciation of : Linking polar and equatorial climate by geochemical proxies

K. Schef¯er   Mineralogical-Petrological Institute, Bonn University, Poppelsdorfer Schloss, 53115 Bonn, Germany S. Hoernes  L. Schwark Geological Institute, Cologne University, Zuelpicher Strasse 49a, 50674 Cologne, Germany

ABSTRACT terglacial sequence in the Basin of The most prevalent glaciation occurred during the Carboniferous±Permian South Africa. By comparing climate proxy on the Southern Hemisphere Gondwana . Sediments from the Pennsylva- signatures from polar and equatorial regions, nian Dwyka deposited in the Karoo Basin of South Africa provide a complete we gain insight into the synchronicity of cli- record of glaciation and deglaciation phases. The direct correlation of glaciation events in mate processes. Because geochemical proxies southern Gondwana basins with the well-studied climate evolution of equatorial regions are affected by diagenesis, provenance was previously hampered by lack of precise radiometric dating. As dating has now become change, or , we tried to extract re- available for the Karoo Basin, the Gondwana glaciation can be viewed in a global paleo- liable information on climatic and sedimen- climatic framework with high temporal resolution. Element geochemical proxies (CIA tary evolution by combining several indepen- [chemical index of alteration], Zr/Ti, Rb/K, V/Cr) record three con®ned shifts in climate dent proxies. and paleoenvironment of the Karoo Basin. These shifts were induced by changes in , weathering rate, provenance, and redox conditions. Because of the low availability RESULTS and diagenetic overprint of , ocean and atmosphere pCO2 variations had to be Geologic Overview 13 13 reconstructed from ␦ Corg values of marine organic matter. The ␦ Corg signatures are The of the Karoo Basin was sum- affected by variable proportions of marine versus terrestrially derived organic matter and marized by Visser (1997) and Visser and its state of preservation. Organic geochemical investigations (TOC [total organic ], Young (1990). The can be dif- C/N, lipid biomarkers) indicate the organic matter in the central Karoo Basin was pri- ferentiated into four deglaciation sequences 13 marily of algal origin. In agreement with element proxies, the varying ␦ Corg values (DS I±IV, Fig. 2A), each with a glacial and mirror shifts in pCO2, rather than variations of organic-matter type. A covariation trend an interstadial phase (Visser, 1997). Absolute 13 between carbon isotope signatures of equatorial carbonates and ␦ Corg values from the age determinations based on sensitive high- Karoo Basin argues against local forcing factors and instead implies a global climate- resolution ion microprobe (SHRIMP) analysis control mechanism. The 5±7 m.y. duration of a complete glacial cycle is not in tune with of single zircons de®ned the duration of each any known orbital frequency. Processes such as changes in equator-pole temperature gra- cycle to ϳ5±7 m.y. (Bangert et al., 1999; dients or newly developing atmosphere-ocean circulation pathways can be regarded as Stollhofen et al., 2000). Each deglaciation se- controlling factors. quence consists of a zone with massive diamictites overlain by a terminal mudrock Keywords: paleoclimate, , Carboniferous±Permian glaciation, Karoo Basin, Dwyka (Fig. 2A). Visser (1997) documented changes Group. in ice-¯ow direction over southern Africa during DS I±IV (Fig. 1). Interpretation of INTRODUCTION these global changes because of isotopic ex- geochemical proxy signatures must differen- The global climate comprises a complex in- change among atmospheric, organic, and in- tiate climatically induced from provenance- terplay between atmosphere, land, and ocean. organic carbon reservoirs (Hayes et al., 1999). controlled variations. For a better understanding of global interac- The in¯uence of the Gondwana glaciation on tion of climate-forcing factors, it is important global climate is documented by variations in Geochemical Proxies 13 to investigate paleoclimate records (Crowley and ␦ C measured on from equato- Multiple geochemical investigations reveal Baum, 1992; Berner, 1994; Frakes et al., 1992; rial regions (Bruckschen et al., 1999; Veizer climate variations during deposition of the Martini, 1997). In the view of today's discussion et al., 1999). Dwyka and Lower Permian on future climate change (IPCC, 2001), the The Karoo Basin of South Africa was part in the Karoo Basin. Zr/Ti ratios study of a icehouse-greenhouse transition of a major depocenter during the late Paleo- indicate provenance change (Fig. 2B) for the may prove valuable. The most extensive zoic (Fig. 1). Equivalent glacial sedimentary interstadial phases (DS II median Zr/Ti ϭ Phanerozoic glaciation lasted 90 m.y. (Crow- deposits exist in , South Africa, 0.076; DS III median Zr/Ti ϭ 0.044), whereas ell, 1978) and its termination occurred during Namibia, Tanzania, , India, and glacial phases exhibit constant Zr/Ti ratios the Carboniferous±Permian on the southern (Crowell, 1978; Caputo and Crow- (median 0.058). Higher Zr contents point to a hemispherical Gondwana supercontinent. ell, 1985; Veevers and Powell, 1987). Sedi- granitic rock composition, thus identifying the Glacial conditions in the Southern Hemi- mentological and mineralogical investigations northern highlands as sediment source region sphere and their contemporaneous effect on (BuÈhmann and BuÈhmann, 1990) have been during the phases of DS I and II. marine and terrestrial life in equatorial regions carried out for glacial sedimentary sequences Sequences DS III and IV received clastic de- in¯uenced the atmospheric CO2 content. The of the Karoo Basin, but only preliminary geo- bris from source regions located near the drop in pCO2 (Berner, 1994) at the end of the chemical analyses of the Dwyka Group sedi- southern magmatic arc associated with sub- Carboniferous coincides with times of con- mentary rocks exist (Visser and Young, 1990). duction along the paleo-Paci®c plate margin trasting climate evolution between polar and This study provides the ®rst detailed geo- (Visser, 1997). equatorial regions. Carbon isotopes record chemical record covering the entire glacial-in- Glacial deposits are interrupted by intersta-

᭧ 2003 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected]. Geology; July 2003; v. 31; no. 7; p. 605±608; 4 ®gures. 605 dial sedimentary units representing temperate climate conditions. Cyclic climate variations during DS I±IV induced severe sea-level change because the ice volume accumulated almost exclusively on land. These changes are recorded by various proxy signatures, includ- ing the chemical index of alteration (CIA; af- ter Nesbitt and Young, 1982), Rb/K ratio (Campbell and Williams, 1965), V/Cr ratio (Jones and Manning, 1994), and organic- matter content (Figs. 2C±2F). Variations in CIA indicate rapid transitions from glacial to interglacial phases (Fig. 2C). Low temperatures and rare vegetation favor physical weathering during glacial periods. The diamictites were deposited as subaqueous aprons and fans forming close to the ice- grounding line (Visser, 1997). Vegetation ex- pansion, soil formation, and warmer and more humid climate during interstadials increased chemical weathering rates, as indicated by el- evated CIAs. Differentiation between fully marine phases during the interstadials and limnic or brackish conditions induced by sea-level drawdown during glaciation is re¯ected by the Rb/K ratio (Fig. 2D). Stadial phases reveal median Rb/K ratios of 4.09, DS I and II interstadials yield Rb/K ratios of 4.78, and DS III and IV inter- Figure 1. Paleogeographic reconstruction of Karoo Basin dur- stadials give Rb/K ratios of 5.56 (Fig. 2D). ing late Paleozoic (after Visser and Praekelt, 1996). Major ice- This agrees with the most pronounced sea- ¯ow directions recorded in deglaciation sequences I, II, and III level rise, the Eurydesma-transgression of DS (after Visser, 1997) indicate clockwise change in provenance. III (Visser, 1997) inducing de®cient PÐParana Basin; KÐKaroo Basin; KHÐKalahari Basin. bottom water conditions associated with water column strati®cation upon sea-level highstand. Anoxia is indicated by higher V/Cr ratios cor- relating with elevated accumulation and pres- ervation of organic carbon (Fig. 2E and F).

Figure 2. and facies evolution of glacial Dwyka Group sedimentary succession; shading indicates interstadial phases. EÐ Eurydesma transgression; ®lled starsÐsensitive high-resolution ion microprobe (SHRIMP) ages after Bangert et al. (1999); open starÐage by Visser (1997). A: Simpli®ed lithology and climate. B: Zr/Ti (a provenance proxy). C: CIA (chemical index of alteration; weathering proxy). D: Rb/K (salinity proxy). E and F: V/Cr and total organic carbon (TOC) (redox proxies).

606 GEOLOGY, July 2003 re¯ect the cyclicity shown by the element ra- tios. The transition from the upper Dwyka Group (mean ␦13CofϪ22.75½) into postgla- cial Ecca Group (mean ␦13CofϪ22.25½) is characterized by a conspicuous excursion to lighter ␦13C values of Ϫ26.5½. This excur- sion correlates with enhanced tectonic and volcanic activity, which is documented by de- creasing 87Sr/86Sr ratios ca. 290 Ma (Veizer et al., 1999). Climate information can be extracted from the ␦13C record of marine bicarbonate. Marine carbonates suitable for reliable carbon isotope determinations were not deposited during sed- imentation of the Dwyka Group. Therefore, we used ␦13C values of the organic carbon in the sedimentary rocks to infer climate chang- es. The isotope variations are driven by the CO2 content in the atmosphere, which affects the organic carbon reservoir. Paired and organic carbon isotopes have been used to decipher changes in atmospheric pCO2 over Figure 3. Carboniferous±Permian climate evolution within the Karoo Basin. Filled 's history (Hayes et al., 1999). Veizer et starsÐsensitive high-resolution ion microprobe ages after Bangert et al. (1999); open al. (1999) compiled an extensive set of ␦13C, starÐage by Visser (1997). A: Invariable mean C/N ratios for glacial and interstadial 18 87 86 phases indicate a constant marine organic-matter source. In B and C, median ␦13C ␦ O, and Sr/ Sr data from paleotropical 13 13 values for glacial and interstadial phases are shown by gray shading. B: ␦ Corg for carbonate . The ␦ Corg values from 13 13 Dwyka sequence. C: ␦ Ccarb for equatorial regions after Veizer et al. (1999); ages south Gondwana and ␦ Ccarb values from -corrected by ؊1 m.y. Synchronous evolution of proxies from polar and equatorial equatorial regions (Fig. 3B and C) show co regions indicates global control. evolutionary trends during the Pennsylvanian. Variability lies within the error range of the age determinations. The marked drop ca. 303 The cyclicity of the lower Dwyka Group, as analyses of aliphatic hydrocarbon fractions Ma and the subsequent increase in the ␦13C shown by different element ratios, can also be (Fig. 4A and 4B) show a predominance of al- carb values (Fig. 3C) indicate a transition from a seen in the carbon isotope composition of or- kanes with carbon numbers in the n-C to n- 16 warm to a cold climate phase (Bruckschen et ganic matter (Fig. 3B). More negative ␦13C C range. These compounds derive from al- org 18 al., 1999). The co-variations of ␦13C values mark the full glacial environments (for gal biomass and are in strong contrast to land org from the Karoo Basin in southern Gondwana stadials of DS II, median is Ϫ25.13½, and for ±lipids, which maximize around n-C in 29 and the equatorial ␦13C of Bruckschen et stadials of DS III, median is Ϫ23.95½), their n-alkane distribution (Meyers, 1997). As carb al. (1999) con®rm a global change in the whereas interstadial phases in general are for the C/N ratios, no obvious differences can carbon isotope composition during the characterized by less negative ␦13C values be measured between samples from glacial or org Pennsylvanian. (interstadials of DS I have a median of interglacial phases. This similarity points to- Ϫ24.83½, those of DS II have a median of ward a consistent marine origin for the organic Ϫ23.19½, and those of DS III have a median matter deposited in the basin center. Deltaic DISCUSSION AND CONCLUSION of Ϫ22.86½). Organic matter derived from sedimentation systems with strong terrigenous Independent geochemical proxies re¯ect marine primary production incorporates the input were established along the northeastern climate variations during deposition of stadial 13 and interstadial sediments of the glacial Dwy- ␦ C signature of atmospheric CO2 and thus basin margin (Visser, 1997). High C/N ratios reveals the changes between glacial and inter- (Ͼ30), pristane/phytane ratios greater than 8, ka Group within the central Karoo Basin. Dur- ing stadial phases, sea-level lowstand caused glacial pCO2 values (Jasper and Hayes, 1990). and the presence of long-chained aliphatic hy- Variation in proportions of marine versus ter- drocarbons indicate a strong terrestrial input sediment deposition under nonmarine condi- rigenous organic matter sources, of deg- into the basin margins (Figs. 4C and 4D). By tions, as indicated by sedimentological and radation, and burial history can alter the car- comparing the geochemical ®ngerprints of or- paleoecological evidence (Visser, 1997). Or- bon isotopes of bulk organic matter (Meyers, ganic matter from central and marginal basin ganic matter deposited in distal regions, how- 1997). Origin and degree of diagenetic over- positions, we demonstrate that the distal basin ever, reveals no terrestrial in¯uence. During print of organic matter were assessed by de- received no signi®cant amounts of terrestrial interstadial phases, sea level rose signi®cantly, termination of C/N ratios and biomarker com- organic matter during deposition of the entire establishing full-marine conditions that per- position. C/N ratios Ͻ10 indicate an origin of Dwyka Group. Because the organic carbon sisted throughout deposition of the postglacial organic matter from marine (Meyers, was derived exclusively from marine organic Ecca Group. 1997), whereas terrigenous organic matter matter, variations in its carbon isotopes must The global nature and synchronous onset of yields C/N ratios of Ͼ20. Samples from the originate from CO2 variations between glacial regionally observed climate variations are Dwyka sequence give C/N ratios of Ͻ8, in- and interglacial phases. Admixture of terrige- documented by the simultaneous shifts in the dicative of marine-derived organic matter. nous organic matter from vegetation cover of ␦13C values of equatorial carbonates and the Glacial and interglacial phases show no sig- the southern Gondwana during ice- ␦13C values of organic carbon from the south- ni®cant variation in C/N ratios (Fig. 3A). Gas free phases can be excluded. ern-latitude Karoo Basin. The ϳ5 m.y. peri- chromatography±mass spectrometry (GC-MS) In DS IV carbon isotopic signatures do not odicity in the duration of stadial-interstadial

GEOLOGY, July 2003 607 Berner, R.A., 1994, Geocarb II: A revised model of atmo- spheric CO2 over Phanerozoic time: American Journal of Science, v. 294, p. 56±91. Bruckschen, P., Oesmann, S., and Veizer, J., 1999, Isotope stratigraphy of the European Carboniferous: Proxy sig- natures for ocean chemistry, climate, and tectonics: Chemical Geology, v. 161, p. 127±163. BuÈhmann, C., and BuÈhmann, D., 1990, Clay minerals as pa- laeoenvironment indicators exempli®ed on a Karoo Se- quence from the Bothaville area, South Africa: South African Journal of Geology, v. 93, p. 505±513. Campbell, F.A., and Williams, G.D., 1965, Chemical com- position of of (Lower Creta- ceous) of central Alberta, : American Associ- ation of Petroleum Geologists Bulletin, v. 49, p. 81±87. Caputo, M.V., and Crowell, J.C., 1985, Migration of glacial centers across Gondwana during Paleozoic : Geo- logical Society of America Bulletin, v. 96, p. 1020±1036. Crowell, J.C., 1978, Gondwana glaciation, cyclothems, con- tinental positioning, and climate change: American Journal of Science, v. 278, p. 1345±1372. Crowley, T.J., and Baum, S.K., 1992, Modeling late Paleo- zoic glaciation: Geology, v. 20, p. 507±510. Frakes, L.A., Francis, J.E., and Syktus, J.I., 1992, Climate modes for the Phanerozoic: Cambridge, Cambridge University Press, 274 p. Hayes, J.M., Strauss, H., and Kaufman, A.J., 1999, The abun- dance of 13C in marine organic matter and isotopic fractionation in the global biogeochemical cycle of carbon during the past 800 Ma: Chemical Geology, v. 161, p. 103±125. IPCC (Intergovernmental Panel on Climate Change), 2001, Figure 4. Dwyka sedimentary lipid distribution; n-alkanes indicated by carbon number, and Third assessment report: Climate change 2001: http:// pristane and phytane denoted as Pr and Ph, respectively. A and B: Aliphatic hydrocarbon www.ipcc.ch (February 2003). compositions of samples from (A) interstadial and (B) glacial phases document a fully ma- Jasper, J.P., and Hayes, J.M., 1990, A carbon isotope record of CO2 levels during the late : Nature, rine origin of organic matter. C and D: Distribution of n-alkanes from basin margins reveals v. 347, p. 462±464. 13 a strong input of land-plant material. Variation of ␦ Corg values in central basin thus results Jones, B., and Manning, D.A.C., 1994, Comparison of geo- from change in pCO2. chemical indices used for the interpretation of palaeo- redox conditions in ancient mudstones: Chemical Ge- ology, v. 111, p. 111±129. Kutzbach, J.E., Guetter, P.J., and Washington, W.M., 1990, Simulated circulation of an idealized ocean for Pan- cycles does not ®t any known orbital frequen- (Saltzman, 2003) decoupled the evolution of gean time: Paleoceanography, v. 5, p. 299±317. cy and requires further investigation. the Panthalassian and Paleotethys Oceans. Martini, I.P., 1997, Late glacial and postglacial environmental The clockwise drift of the southern Gond- As a result of reduced equatorial to polar changes, Quaternary, Carboniferous±Permian, and Pa- leozoic: Oxford, Oxford University Press, 343 p. wana continent shifted the Karoo Basin into temperature gradients, retreated and Meyers, P.A., 1997, Organic geochemical proxies of palae- polar cold regions and can be regarded as the sea level rose. The processes of CO release oceanographic, paleolimnologic and paleoclimatic pro- 2 cesses: Organic Geochemistry, v. 27, p. 213±250. primary trigger for glaciation processes. This and temperature increase, linked in a positive Nesbitt, H.W., and Young, G.M., 1982, Early cli- mechanism is, however, insuf®cient to explain feedback, caused global warming. The ®nal mates and plate motions inferred from major element chemistry of lutites: Nature, v. 299, p. 715±717. the cyclic climate variations as well as the ter- termination of the Carboniferous±Permian Saltzman, M.R., 2003, Late Paleozoic : Oceanic gate- minal and abrupt breakdown of the ice shields. glaciation phase in southern Africa must be way or pCO2?: Geology, v. 31, p. 151±154. Shaviv, N.J., 2002, Cosmic ray diffusion from galactic spiral Warm currents along the eastern coast of Pan- viewed to have resulted from the interaction arms, iron , and a possible climatic connec- (Kutzbach et al., 1990) transported moist of different climate-controlling factors, includ- tion: Physical Review Letters, v. 89, 051102. Stollhofen, H., Stanistreet, I.G., Bangert, B., and Grill, H., air to polar regions. In combination with high ing variations in pCO2 values, the continent- 2000, Tuffs, tectonism and glacially related sea level albedo rates and low pCO2, icehouse condi- ocean con®guration, and continental topogra- changes, Carboniferous±Permian, southern Namibia: tions were established and glaciers started to phy, as well as changing atmospheric and Palaeogeography, Palaeoclimatology, Palaeoecology, v. 161, p. 127±150. accumulate in high latitudes. High tempera- oceanic circulation pathways. As the solar sys- Veevers, J.J., and Powell, C.A., 1987, Late Paleozoic glacial ture gradients between equatorial and polar re- tem passes through one of the four galactic episodes in Gondwanaland re¯ected in transgressive- regressive depositional sequences in Euramerica: Geo- gions favored the development of a psychro- spiral arms, the cosmic ray ¯ux increases, logical Society of America Bulletin, v. 98, p. 475±487. sphere where deep ocean cold-water currents which may trigger global glaciation (Shaviv, Veizer, J., Ala, D., Azmy, K., Bruckschen, P., Buhl, D., Bruhn, F., Carden, G.A.F., and Diener, A., 1999, 87Sr/ reached equatorial regions. Feedback mecha- 2002). The poor correlation between the cos- 86Sr, ␦13C and ␦18O evolution of Phanerozoic seawa- nisms intensi®ed the cooling rates and favored mic ray ¯ux maximum and the timing of the ter: Chemical Geology, v. 161, p. 59±88. Visser, J.N.J., 1997, Deglaciation sequences in the Permo± the ®xation of water as glaciers on southern Pennsylvanian glaciation excludes this exter- Carboniferous Karoo and Kalahari basins of southern . This in turn caused a global fall in nal forcing mechanism and underlines the Africa: A tool in the analysis of cyclic glaciomarine sea level. During regression phases, CO ®x- need for further investigation. basin ®lls: Sedimentology, v. 44, p. 507±521. 2 Visser, J.N.J., and Praekelt, H.E., 1996, Subduction, mega- ing carbonate-complexes were exposed above shear systems and late Paleozoic basin development in sea level and carbonate weathering led to in- ACKNOWLDEGMENTS the African segment of Gondwana: Geologische Rund- schau, v. 805, p. 632±646. creasing pCO . In concert with phases of en- P. Meyers and an anonymous reviewer are gratefully ac- 2 knowledged for their constructive reviews. We thank Deut- Visser, J.N.J., and Young, G.M., 1990, Major element geo- hanced tectonic activity (Stollhofen et al., sche Forschungsgemeinschaft for ®nancial support of this chemistry and of the Permo±Carbon- study under contract Ho868/21. iferous glacigene Dwyka Formation and postglacial 2000), elevated atmospheric CO2 release ini- mudrocks in southern Africa: Palaeogeography, Pa- tiated greenhouse conditions. Elevated tem- laeoclimatology, Palaeoecology, v. 81, p. 49±57. REFERENCES CITED peratures established new air-ocean circulation Bangert, B., Stollhofen, H., Lorenz, V., and Armstrong, R., Manuscript received 9 December 2002 pathways and possibly a retreat of the CO2- 1999, The geochronology and signi®cance of ash-fall Revised manuscript received 7 March 2003 tuffs in the glaciogenic Carboniferous±Permian Dwyka Manuscript accepted 16 March 2003 ®xing psychrosphere. Most importantly, the Group of Namibia and South Africa: Journal of Afri- Pangean closure of the equatorial seaway can Earth Science, v. 29, p. 33±49. Printed in USA

608 GEOLOGY, July 2003