Did the Antarctic Ice Sheets Expand During the Early Pliocene?

Did the Antarctic ice sheets expand during the early Pliocene?

P.J. Bart Department of Geology and Geophysics, Louisiana State University, Baton Rouge, Louisiana 70803, USA

ABSTRACT Seismic data show that glacial unconformities are located within lower Pliocene strata on the Antarctic continental shelves. The glacial unconformities are signi®cant because they provide direct evidence that the Antarctic ice sheets advanced despite the generally warmer climates and elevated sea levels that characterized most of the early Pliocene. The magnitudes of peak eustatic lowstands and 18O enrichments indicate that the ice volume on Antarctica may have exceeded today's ice volume by approximately 18%, which suggests that the ice-sheet grounding events on the shelves probably were associated with larger than present ice volumes on two to three occasions during the early Pliocene.

Keywords: Antarctica, ice sheet, global warming, Pliocene, seismic stratigraphy.

INTRODUCTION tion seismic studies have been previously con- Table 1 shows the depth and two-way trav- On the basis of marine data from the south- ducted (the eastern Ross Sea, Alonso et al., eltime ranges of the lower Pliocene strata from ern high latitudes, Kennett and Hodell (1993) 1992; the Antarctic Peninsula, Bart and An- the DSDP-ODP drill sites. At Sites 271 and argued that early Pliocene temperatures of derson, 1995; Prydz Bay, Cooper et al., 1991). 1097, I used a velocity of 2000 m/s to convert Antarctic surface waters increased by a max- These studies were selected because the re- the depths to two-way traveltime. At Site 739, imum of ϳ3 ЊC and that the volume of the sults are directly correlated to DSDP-ODP a geophysical survey indicates a velocity of Antarctic ice sheets ¯uctuated within narrow sites where lower Pliocene strata were sam- 2125 m/s between the sea¯oor and 130 m be- limits, resulting in a maximum sea-level rise pled and because the data-acquisition design low the sea¯oor (mbsf), whereas below 130 of ϳ25 m. Despite the evidence for ice-vol- for the seismic surveys was similar. In addi- mbsf, a velocity of 2625 m/s is indicated. ume reductions, the large magnitudes of eu- tion, these continental shelves received drain- Theoretically, a thick ice sheet could thin static lowstands (Haq et al., 1987) and 18O age from the three primary components of the and advance across the continental shelf with enrichments (Shackleton et al., 1995) suggest Antarctic cryosphere, the West Antarctic ice a minimal change in its overall ice volume. To that ice volumes may have also expanded to sheet, Antarctic Peninsula ice sheet, and the evaluate if ice-sheet advances were also as- considerably larger than present levels during East Antarctic ice sheet (Fig. 1). sociated with signi®cant ice-volume increases, the early Pliocene. If these lowstands and 18O enrichments were due to extreme expansions of the Antarctic ice sheets, direct evidence of such should exist on the Antarctic continental shelves in the form of glacial unconformities. Thus far, lower Pliocene strata have been sam- pled at Deep Sea Drilling Project±Ocean Dril- ling Program (DSDP-ODP) sites on the eastern Ross Sea (Hayes et al., 1975), Antarctic Peninsula (Barker et al., 1999), and Prydz Bay (Barron et al., 1989) continental shelves (Fig. 1). In this study, I evaluate the seismic stratigraphy of lower Pliocene strata on the Ant- arctic shelves to address these questions: Did the Antarctic ice sheets advance across the shelf during the early Pliocene, and if so, were the advances indicative of large increases in the overall ice volume? Given the current global warming and the potential effect of ice- sheet ¯uctuations on sea level, understanding the behavior of the Antarctic ice sheets in the early Pliocene, a time of warmer than present climatic conditions, is a fundamentally impor- tant issue. Figure 1. Location map showing East Antarctic ice sheet (EAIS), West Antarctic ice sheet METHODS (WAIS), and Antarctic Peninsula ice sheet (APIS). Deep Sea Drilling Project (DSDP) and Ocean Drilling Program (ODP) sites and seismic pro®les referred to in text are indicated, as This investigation was con®ned to three are elevation contours (white dashed lines), ice ¯ow lines (thin white solid lines) and ice- continental shelf areas for which high-resolu- drainage divides (thick white solid lines).

᭧ 2001 Geological Society of America. For permission to copy, contact Copyright Clearance Center at www.copyright.com or (978) 750-8400. Geology; January 2001; v. 29; no. 1; p. 67±70; 3 ®gures; 1 table. 67 TABLE 1. DEPTHS AND TWO-WAY TRAVELTIMES OF LOWER PLIOCENE STRATA suring the vertical-elevation change of coastal Ross Sea Site 271* Antarctic Peninsula Site 1097² Prydz Bay Site 739§ onlap from the end of the highstand to the depth time depth time depth time beginning of the lowstand. The Haq et al. (mbsf) (msbsf) (mbsf) (msbsf) (mbsf) (msbsf) (1987) global-cycle chart indicates three eu- 204 204 150 150 100 94 static lowstands in the early Pliocene (Fig. 251 251 436 436 150 137 3A). The sea-level elevations are approxi- *Deep Sea Drilling Program, Eastern Basin Ross Sea shelf (Hayes and Frakes, 1975) mately Ϫ15 m, ϩ60 m, and Ϫ20 m relative ²Ocean Drilling Program, Antarctic Peninsula Paci®c shelf (Barker et al. 1999) to present-day sea level. Greenlee and Moore §Deep Sea Drilling Program, Prydz Bay shelf (Barron et al., 1989) (1988) recognized four eustatic lowstands (Fig. 3B). In their most conservative estimate (i.e., 100 m paleo±water depth model), the the seismic evidence from the Antarctic ocene strata. On the Antarctic Peninsula con- lowstand sea-level elevations are Ϫ38 m, shelves was interpreted in light of two eustatic tinental shelf, Bart and Anderson (1995) Ϫ35 m, Ϫ4m,andϪ1 m relative to present- records (Haq et al., 1987; Greenlee and found seismic evidence for at least 31 glacial day sea level. Moore, 1988) and two ␦18O records (Hodell unconformities. At least six glacial unconfor- and Venz, 1992; Shackleton et al., 1995). mities, numbers 23, 24, 25, 26, 27, and 28, 18 are located within the lower Pliocene strata O ENRICHMENTS DEPOSITION AND EROSION ON THE (Fig. 2B). On the Prydz Bay shelf, an area Shackleton et al. (1995) constructed a high- 18 ANTARCTIC SHELVES receiving drainage from the East Antarctic ice resolution ␦ O record from deep-sea benthic In the current interglacial, the Antarctic ice sheet, Cooper et al. (1991) subdivided the foraminifers at ODP Site 846 (Fig. 3C), which sheets terminate near the coast or on the inner stratigraphic section into four stratigraphic is located in the eastern equatorial Paci®c shelf, and the outer continental shelves are es- units (PS.1±PS.4). At least one topset surface Ocean. The sampling interval was 10 cm, sentially sediment starved (Anderson, 1999). is located in the lower Pliocene strata. Thus which corresponds to a temporal resolution of During glacial periods, ice sheets advanced far, no detailed mapping of this topset surface ϳ2500 yr. At this site, there are three pro- 18 18 well into the marine realm. At these times, the has been conducted to determine whether it is nounced O enrichments (Fig. 3C). The ␦ O 18 inner shelves became a zone of net erosion a glacial unconformity (Fig. 2C). values of these O enrichments are ϳ3.3½, (ten Brink et al., 1995), and sediments eroded ϳ3.55½, and ϳ3.29½, respectively. The raw from the continent and inner shelves were in- EUSTATIC LOWSTANDS ␦18O data from Site 846 show that each 18O corporated into basal debris zones (Alley, The magnitudes of eustatic lowstands were enrichment is supported by more than one 1989). On the shelves, the most rapid erosion estimated primarily from seismic data by mea- data point. For example, a tight cluster of ®ve and sediment transport occurred beneath ice streams, which are wide (several tens of kilometers) zones of fast-¯owing ice contained within slow-moving ice. At the mouths of ice streams, subglacial sediments were released as sediment gravity ¯ows that were deposited as low-angle prograding foresets. If the ice sheet advanced to the shelf edge, poorly sorted ter- rigenous sediments were supplied to upper slope depocenters. Within these depocenters, strata typically have an overall topset and foreset geometry. On seismic pro®les, topset surfaces that exhibit regional extent (several tens of kilometers) and broad glacial-trough topography are interpreted as glacial unconformities (Anderson, 1999).

ANTARCTIC STRATIGRAPHY Interpretations of seismic pro®les from three Antarctic shelves are shown in Figure 2. The seismic-stratigraphic interpretations are from previous studies of the eastern Ross Sea (Alonso et al., 1992), Prydz Bay (Cooper et al., 1991), and the Antarctic Peninsula (Bart and Anderson, 1995). The gray-shaded areas on the pro®les are my correlations of the lower Pliocene strata (Fig. 2). On the eastern Ross Sea outer continental shelf, an area receiving drainage from the West Antarctic ice sheet, Alonso et al. (1992) identi®ed eight glacial unconformities (labeled Figure 2. Interpreted line drawings of seismic pro®les from (A) eastern Ross Sea (after 1±8 in Fig. 2A). At least one glacial uncon- Alonso et al., 1992), (B) Antarctic Peninsula (after Bart and Anderson, 1995), and (C) Prydz formity, number 5, is located within lower Pli- Bay (after Cooper et al., 1991). Gray shading indicates lower Pliocene strata.

68 GEOLOGY, January 2001 sphere ice sheets did not develop until the late Pliocene (Berggren, 1972), the solid vertical line at ϩ7 m in Figure 3 (A and B) corresponds to the expected sea-level elevation if the Greenland ice sheet did not exist and the Antarctic ice sheets existed in their present con®guration. If during the early Pliocene, the Antarctic ice sheets attained the ice volume reached at the last glacial maximum (LGM) as described by Hughes et al. (1981), then the Antarctic ice sheets would have been larger than at present by 24 m SLE. However, because 7 m SLE of the additional ice volume was grounded below sea level (and hence would not affect sea level), eustatic sea level would have fallen by only 17 m SLE. There- fore, the dash-dot vertical line at Ϫ10 m in Figure 3 (A and B) corresponds to the expected sea-level elevation if the Antarctic ice sheets expanded to the volume attained at the LGM. That the peak eustatic lowstands on the eustatic records far exceed the ϩ7 m and Ϫ10 Figure 3. A: Eustatic curve after Haq et al. (1987). B: Eustatic curve after Greenlee and Moore m values suggests either that the volume of (1988). C: ␦18O curve after Shackleton et al. (1995). D: ␦18O curve after Hodell and Venz the Antarctic ice sheets exceeded the ice vol- (1992). See text for discussion. ume attained at the LGM or that the sequence- stratigraphy estimates of maximum eustatic falls are too large. Taken at face value, the raw samples (3.46½, 3.66½, 3.55½, 3.38½, recovered from Site 739 were composed en- eustatic evidence suggests that larger than pre- and 3.44½) contributed to the smoothed peak tirely of massive diamictite. They interpreted sent land-based ice volumes may have existed of 3.55½ at 4.0 Ma. These 18O enrichments the diamictite as waterlaid till deposited in on a few occasions during the early Pliocene. far exceed the mean ␦18O value for the early proximity to the ice-sheet grounding line (the Pliocene (ϳ2.85½) and also exceed today's zone where the ice sheet lifts off the sea¯oor expected median ␦18O value for the site to form an ice shelf). The location of proximal EARLY PLIOCENE ICE VOLUME: (dashed line at 3.2½). till at Site 739, ϳ35 km from the paleo±shelf IMPLICATIONS FROM ␦18O RECORDS Hodell and Venz (1992) constructed a ␦18O edge, indicates that the East Antarctic ice The present-day expected median ␦18O val- record from deep-sea benthic foraminifers at sheet advanced at least to the outer shelf ues (dashed line in Fig. 3, C and D) are pri- ODP Site 704 (Fig. 3D), which is located in sometime during the early Pliocene. The seis- marily a result of the present-day con®gura- the sub-Antarctic south Atlantic Ocean. The mic evidence is signi®cant because it dem- tions of the Antarctic ice sheets and Greenland lower frequency of ¯uctuations is due to a onstrates that the warmer than present cli- ice sheet, and the current temperature of the coarser sampling interval. At Site 704, there mates that characterized most of the early deep seas in which modern benthic foramini- are two peak 18O enrichments. The ␦18O val- Pliocene did not preclude advance of the Ant- fers secreted their shells. The limiting value ues are ϳ3.07½ and ϳ3.02½, respectively. arctic ice sheet. on the maximum cooling of the deep-sea wa- These 18O enrichments exceed the mean ␦18O ter temperature is probably the temperature of value (ϳ2.8½) for the early Pliocene and EARLY PLIOCENE ICE VOLUME: the Antarctic Bottom Water at the site where closely match today's expected median ␦18O IMPLICATIONS FROM EUSTATIC it is produced on the Antarctic continental value for the site (dashed line at 3.0½). RECORDS shelves. Weyl (1968) argued that this bottom- Given the relatively high frequency and water production was reduced during the gla- GROUNDING EVENTS OF THE large magnitude of the early Pliocene sea-lev- cial periods when ice cover on the shelves was ANTARCTIC ICE SHEET el falls, volume-change mechanisms not in- more permanent. Given that the seismic evi- The seismic evidence illustrates that the volving ice (e.g., reduction in the volume of dence suggests that the Antarctic ice sheets West Antarctic ice sheet grounded at the pa- the mid-oceanic ridges and hotspots; thermal grounded on the continental shelves in the ear- leo±shelf edge at least once during the early contraction of the ocean waters; charging of ly Pliocene, the extent of the production sites Pliocene, whereas the Antarctic Peninsula ice groundwater reservoirs) probably did not con- must have been reduced during these glacial sheet advanced to the shelf edge at least six tribute signi®cantly to the peak eustatic low- episodes. If the supply of cold Antarctic Bot- times (Fig. 2, A and B). The seismic evidence stands. Hence, the early Pliocene lowstands tom Water to the deep sea was reduced during on the Prydz Bay shelf is not suf®cient to di- must contain a signi®cant glacio-eustatic com- these glacial periods, then the temperature of rectly determine whether the East Antarctic ponent. In terms of ice volume, the present- the deep-sea waters probably would not have ice sheet advanced during the early Pliocene, day sea-level elevation (dashed line in Fig. 3, cooled further beyond the present-day deep- but the topset surface from Prydz Bay (Fig. A and B) is primarily controlled by the Green- sea temperature of ϳ1 ЊC. If this reasoning is 2C) suggests that that ice sheet may have ad- land ice sheet, which contains an ϳ7 m sea- accepted, then the early Pliocene peak 18O en- vanced across the shelf. In addition, Barron et level equivalent (SLE), and the Antarctic ice richments beyond today's median expected al. (1989) found that the lower Pliocene units sheet. However, because the Northern Hemi- ␦18O values probably contain a signi®cant ice-

GEOLOGY, January 2001 69 volume component. The solid lines in Figure shelf in Prydz Bay. The magnitudes of peak Antarctica: Implications from Leg 119 dril- 3 (C and D) correspond to the expected ␦18O eustatic lowstands and 18O enrichments in the ling, in Barron, J., et al., Proceedings of the Ocean Drilling Program, Scienti®c results, value if the Antarctic ice sheet existed in its early Pliocene suggest that global ice volumes Volume 119: College Station, Texas, Ocean present con®guration and the Greenland ice exceeded that attained at the LGM on a few Drilling Program, p. 5±25. sheet did not exist (calculated by using a con- occasions. Because there was no signi®cant Greenlee, S.M., and Moore, T.C., 1988, Recognition version factor of 0.1½ ϭ 10 m SLE). If dur- ice in the Northern Hemisphere until the late and interpretation of depositional sequences and calculation of sea-level changes from ing the early Pliocene, the Antarctic ice sheets Pliocene and because nonice volume effects stratigraphic dataÐOffshore New Jersey and attained an ice volume equal to that reached probably did not contribute signi®cantly to the Alabama, Tertiary, in Wilgus, C.K., et al., by those ice sheets during the LGM (i.e., ϩ17 peak lowstands and 18O enrichments, I sur- eds., Sea-level changes: An integrated ap- m above sea level and ϩ7 m below sea level mise that early Pliocene Antarctic ice-sheet proach: Society of Economic Paleontologists and Mineralogists Special Publication 42, ϭϩ24 m SLE; Hughes et al., 1981), then the grounding events probably were associated p. 329±353. ␦18O values would be enriched by ϳϩ0.34½. with larger than present ice volumes despite Haq, B.U., Hardenbol, J., and Vail, P.R., 1987, The conversion factor used to estimate the the generally warmer climates and higher sea Chronology of ¯uctuating sea levels since the ␦18O value for the additional 24 m SLE on levels. Triassic: Science, v. 235, p. 1156±1167. Hayes, D.E., Frakes, L.A., et al., 1975, Initial re- Antarctica is the same as that used by Kennett ports of the Deep Sea Drilling Project, Volume and Hodell (1993) to compute the maximum ACKNOWLEDGMENTS 28: Washington, D.C., U.S. Government Print- reduction of the Antarctic ice sheets during the I thank George Denton and Eugene Domack for ing Of®ce, 1017 p. early Pliocene (i.e., 0.1½ ഠ 7 m SLE). constructive comments and suggestions that im- Hodell, D.A., and Venz, K., 1992, Toward a high- proved the manuscript. At ODP Site 846, the three peak 18O en- resolution stable isotopic record of the South- ern Ocean during the Pliocene-Pleistocene 18 richments far exceed the ␦ O value expected REFERENCES CITED (4.8 to 0.8 Ma), in Kennett, J.P., and Warnke, 18 for today, and the largest O enrichment ex- Alley, R.B., 1989, Sedimentation beneath ice D.A., eds., The Antarctic paleoenvironment: A ceeds the ␦18O value expected for an LGM- shelvesÐThe view from ice stream B: Marine perspective on global change (Part 1): Antarc- type ice-volume expansion of the Antarctic ice Geology, v. 85, p. 101±120. tic Research Series, v. 56, p. 265±310. Hughes, T.J., Denton, G.H., Andersen, B.G., Schil- sheets (dash-dot line in Fig. 3C). At ODP Site Alonso, B., Anderson, J.B., Diaz, J.T., and Bartek, L.R., 1992, Plio-Pleistocene seismic stratigra- ling, D.H., Fastook, J.L., and Lingle, C.S., 704, the fact that the magnitudes of the early phy of the Ross Sea: Evidence for multiple ice 1981, The last great ice sheets: A global view, Pliocene peak 18O enrichments exceed today's sheet grounding episodes, in Elliot, D.H., ed., in Denton, G.H., and Hughes, T.J., eds., The expected median ␦18O value suggests that the Contributions to Antarctic research III: Ant- last great ice sheets: New York, John Wiley & Sons, p. 221±261. Antarctic ice-sheet grounding events were as- arctic Research Series, v. 57, p. 93±103. Anderson, J.B., 1999, Antarctic marine geology: Kennett, J.P., and Hodell, D.A., 1993, Evidence for sociated with a larger than present ice volume. Cambridge, UK, Cambridge University Press, relative climatic stability of Antarctica during Judging from high-resolution data at Site 846, 289 p. the early Pliocene: A marine perspective: Geo- peak 18O enrichments are brief events. Hence, Barker, P.F., Camerlenghi, A., et al., 1999, Proceed- gra®ska Annaler, v. 75A, p. 204±220. ings of the Ocean Drilling Program, Initial re- Shackleton, N.J., Hall, M.A., and Pate, D., 1995, because of the coarser sampling at Site 704, Pliocene stable isotope stratigraphy of Site the peak 18O enrichments probably were un- ports, Volume 178: College Station, Texas, Ocean Drilling Program, 60 p. 846, in Pisias, N.G., et al., Proceedings of the 18 dersampled and aliased on the Site 704 ␦ O Barron, J., Larsen, B., et al., 1989, Proceedings of Ocean Drilling Program, Volume 138, Scien- record. Taken at face value, the Site 846 ␦18O the Ocean Drilling Program, Initial reports, ti®c results: College Station, Texas, Ocean Volume 119: College Station, Texas, Ocean Drilling Program, p. 337±355. record suggests that larger than present ice ten Brink, U.S., Schneider, C., and Johnson, A.H., volumes may have existed on three brief oc- Drilling Program, 942 p. Bart, P.J., and Anderson, J.B., 1995, Seismic record 1995, Morphology and stratal geometry of the casions during the early Pliocene. of glacial events affecting the Paci®c margin Antarctic continental shelf: Insights from of the northwestern Antarctic Peninsula, in models, in Cooper, A.K., et al., eds., Geology and seismic stratigraphy of the Antarctic mar- CONCLUSIONS Cooper, A.K., et al., eds., Geology and seismic stratigraphy of the Antarctic margin: Antarctic gin: Antarctic Research Series, v. 68, p. 1±24. Weyl, P.K., 1968, The role of the oceans in climatic The seismic and lithologic data from lower Research Series, v. 68, p. 75±95. change: A theory of the ice ages: Meteorology Pliocene strata from the Antarctic shelves Berggren, W.A., 1972, Late Pliocene±Pleistocene Monograph, v. 8, p. 37±62. demonstrate that the West Antarctic ice sheet glaciation, in Laughton, A.S., et al., Initial re- and Antarctic Peninsula ice sheet grounded at ports of the Deep Sea Drilling Project, Volume Manuscript received May 24, 2000 12: Washington, D.C., U.S. Government Print- the paleo±shelf edge in the eastern Ross Sea Revised manuscript received September 27, 2000 ing Of®ce, p. 953±963. Manuscript accepted October 4, 2000 and Antarctic Peninsula and that the East Ant- Cooper, A.K., Stagg, H., and Geist, E., 1991, Seis- arctic ice sheet grounded at least to the outer mic stratigraphy and structure of Prydz Bay, Printed in USA

70 GEOLOGY, January 2001