Glacial-interglacial history of the Skaftafell region, southeast Iceland, 0–5 Ma Johann Helgason* Ekra Geological Consulting, Thorsgata 24, 101 Reykjavik, Iceland
Robert A. Duncan* College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon 97331-5503, USA
ABSTRACT Volcanic strata in the Skaftafell region, southeast Iceland, record a sequence of at least 16 glacial and interglacial intervals since 5 Ma. Two composite sections of 2 to 2.8 km thickness have been constructed from multiple, overlapping, cliff profiles. The timing of alternating sequences of subaerial lava flows, pillow basalts, and hyaloclastite deposits is provided by magnetostratigraphic mapping and K-Ar radiometric dating. We find that the frequency and intensity of glaciations increased significantly at ca. 2.6 Ma, and particularly since 0.8 Ma, amplifying topographic relief in this area from ,100 m to ;2 km. These changes correlate with increases in global ice volume, ice-rafted debris, and development from local to regional glacial conditions in the North Atlantic. Keywords: subglacial volcanism, K-Ar age dating, paleomagnetism, relief amplification, identification of glacial and interglacial periods, lithology of subglacially erupted volcanic strata.
INTRODUCTION The Skaftafell region, located ;50 km east of the neovolcanic eastern rift zone in southeast Iceland, is almost totally surrounded by the Vatnajo¨kull glacier, except in the south, where the Skeiara´rsandur forms a gently sloping coastal plain (inset, Fig. 1). Outlet glaciers have carved the volcanic strata in this area into a spectacular series of ridges and valleys that have national park designation. Frequent alternations of subaerially erupted lavas, subglacially erupted volcanic strata, and glaciofluvial sedimentary rocks provide a lithologic basis for tracing the glacial and interglacial history of southeast Iceland and climate variability in the North Atlantic region over the past 5 m.y. Iceland owes its construction to the coincidence of a mantle hotspot with the Mid-Atlantic spreading ridge. Crustal sections that form in rift zones from fissure eruptions and central volcanoes subside as they move laterally and are normally deeply buried by later volcanism (Palmason, 1980, 1986; Helgason, 1984, 1985). Older crustal sections are then exposed in eastern and western Iceland through isostatic uplift in response to removal of overlying volcanic material by erosion. By comparison, the strata exposed in Skaftafell, located outside the axial rift but nearly directly above the hotspot, have subsided slightly. Therefore, the Skaftafell strata provide a nearly continuous history of the competing processes of constructive volcanism and destruc*E-mail: Helgason—
[email protected]; Duncan—
[email protected].
tive erosion. In this paper, we present lithologic and paleomagnetic data and K-Ar radiometric ages for two composite stratigraphic successions. Together these provide complementary information on the growing intensity of glaciations and increasing topographic relief in southeast Iceland since 4.7 Ma. FIELD AREA Lava formations in Skaftafell can typically be traced for only a few hundred meters, as opposed to much larger distances (up to 60 km) for the older Tertiary strata of eastern Iceland. Erosional unconformities are common, and in the lower part of the region, stratigraphic markers are rare. We adopted the basic field classification for rocks in Iceland (Walker, 1963), in which rock strata are divided into formations on the basis of lava cooling units; certain intercalated units of sedimentary origin, because of their climatic importance, are described as independent formations. The subaerially erupted basalt lavas, as well as the subglacially formed volcanic cooling units were divided (on the basis of composition) into (1) coarse-grained olivine tholeiites, (2) fine-grained aphyric tholeiites, (3) plagioclase porphyritic units, and (4) thicker, fine-grained, aphyric basaltic andesites. The subglacial strata were further divided into units on the basis of lithologic characteristics that graded upward from lobes, pillows, pillow breccia, hyaloclastite breccia, and primary hyaloclastite, to reworked hyaloclastite. We used compositional and paleomagnetic data and K-Ar radiometric ages to correlate 70
cliff profiles that were separated by erosional hiatuses and abrupt changes in lithology over relatively short distances, and to provide a precise time scale for alternating glacial and interglacial successions. We constructed two composite stratigraphic successions, referred to as the Skaftafell and Hafrafell sections (Fig. 1). The Skaftafell section, composed of 35 formations, is 2.8 km thick, and the 1.9-km-thick section from the Hafrafell is composed of 30 formations. We drilled over 1000 core samples for paleomagnetic directions from a total of 248 volcanic units. To obtain the freshest material for K-Ar dating, we collected block samples from the massive, jointed interiors of lava flows. We examined samples in thin section to select 23 of the best-crystallized and least-altered material for age determinations. Additional selection criteria were stratigraphic position, K content, and loss-on-ignition measurements.1 RESULTS Brunhes-age glacial erosion and subglacial volcanism in the Skaftafell region account for deep incisions of valleys as well as accumulation of subglacially formed strata that have been added intermittently to the valley walls at various levels. Underneath these discontinuous young strata is an older and more continuous stratigraphic succession made up of 1GSA Data Repository item 200120, Data on paleomagnetic properties, K-Ar age, and stratigraphic formations, is available on request from Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301-9140,
[email protected], or www. geosociety.org/pubs/ft2001.htm.
q 2001 Geological Society of America. For permission to copy, contact Copyright Clearance Center at www.copyright.com or (978) 750-8400. Geology; February 2001; v. 29; no. 2; p. 179–182; 2 figures; Data Repository item 200120.
179
Figure 1. Skaftafell and Hafrafell volcano-stratigraphic successions were compiled from rock compositions, lithologic characteristics, magnetic polarities, and KAr age determinations from 71 individual profiles. Inset shows location of Skaftafell study area.
subaerial lavas and subglacial volcanic strata and sedimentary rocks, dating back into the Tertiary. The lithologic diversity in the region provides a high degree of stratigraphic resolution despite erosional hiatuses of varying intensity. We present paleomagnetic correlations and K-Ar age determinations from the two volcanic sections in Figure 1. In general, measured ages conform with stratigraphic position, and closely spaced samples produced concordant ages. Abrupt age steps usually confirm erosional hiatuses recognized in the field. Rocks 180
from the Skaftafell region are somewhat more alkalic than the average Iceland axial-rift tholeiitic basalt (Prestvik, 1985), and higher K contents have produced relatively high proportions of radiogenic 40Ar and precise ages (2%–5% uncertainty for samples older than 1 Ma). For low-K or very young samples, analytical uncertainties are larger. Concentrations of 36Ar in all samples were low, however, indicating that addition of atmospheric Ar during alteration has not been a significant problem in the samples selected for dating. Although there is evidence from the lower-
most units for shallow burial (chabazite to mesolite zeolite zones), the ages determined here appear to provide reliable estimates of the time of crystallization—a conclusion reached for much older, but carefully selected samples from zeolite-grade Icelandic lava flows (McDougall et al., 1976). DISCUSSION Recognition of Glacial and Interglacial Conditions The stratigraphic formations (Fig. 1) are identifiable as either glacial or interglacial inGEOLOGY, February 2001
tervals. A glacial interval is defined as a period when the region was covered with thick ice under which volcanism built up discontinuous pillow-basalt ridges. Thus, subaqueously formed volcanic rocks, composed mainly of pillow basalt, hyaloclastic sedimentary rocks, and breccias, are the main indicators of glacial intervals. By contrast, interglacial intervals are periods in which more continuous, subaerial lava flows and thin volcaniclastic sedimentary beds alone make up the volcanic section. The two composite sections extend over a lateral distance of 20 km in a direction roughly perpendicular to strike and are correlatable for the period of ca. 4 to 0.8 Ma. Our correlations suggest that three glacial intervals in Hafrafell (G7, G8, and G9) are not seen in Skaftafell. Likewise, five glacial intervals in Skaftafell (G3, G4, G5, G10, and G11) are not found in Hafrafell. However, eight glacial intervals occur in both sections. Thus, 16 intervals are recorded in one or both sections. An additional interval in Skaftafell is possibly of glacial origin—i.e., G0, at the base of the section between units VG2 and VG1, dated at 4.6 to 4.7 Ma. Strata representing the interval of 3 to 2.4 Ma are missing, presumably through erosion, from the Hafrafell section. In Skaftafell this interval includes glacial intervals G3 to G5. Although the same number of glacial intervals occurs in both of the composite sections for certain periods, the exact correlation between individual glacial intervals has so far not been completed. If one or more of these intervals occurred in only one of these sections, then the total number of intervals recorded would be increased. An important advance of this study is direct radiometric dating of subglacial strata, in addition to subaerial lava flows that bracket glacial intervals. Examples are glacial intervals G3 at 2.82 (60.04) Ma and G5 at 2.52 (60.10) Ma. The more numerous age determinations of subaerial lavas in this study provide tighter limits on the timing of interglacial periods. With increasing stratigraphic height, individual formations are more limited in lateral extent, and greater variability exists in thickness between time horizons, determined from paleomagnetic boundaries and radiometric ages. Growth of Local Topographic Relief The regional topography of Iceland is a broad shield, rising from sea level to ;700 m along the central rift axis. Maximum elevation is displaced to the southeast, in the vicinity of the hotspot. From examination of the volcanic record preserved in the Skaftafell area, we observe that local topographic relief has been superimposed on the regional pattern largely as a result of glacial conditions developing over the past 5 m.y. As illustrated in Figure 2, we GEOLOGY, February 2001
Figure 2. Local topographic relief at Skaftafell has been amplified since 5 Ma as subglacially erupted ridges and deepening valleys developed under thickening ice.
conclude that local topographic relief was less than ;100 m until 3 Ma, was probably less than 600 m during the period from 3 until 0.8 Ma, but is now up to 2000 m. Accumulation of snow from year to year to form glaciers depends on many factors, but most important are regional temperature, precipitation, and local elevation. It is unlikely that the low local relief in the Skaftafell area prior to ca. 0.8 Ma could have sustained glaciers through interglacial intervals such as the present, given that local glaciers today survive only at elevations above ;1000 m. The recent and drastic relief changes in Skaftafell have thus elevated the region to an altitude where temperature and precipitation favor glacier formation. There remains some uncertainty as to when the deep valleys in the region formed. We see no evidence for valleys exceeding 1 km depth in the Skaftafell region until quite recently
(since 0.8 Ma). This finding suggests that local glaciers, comparable to the present glaciers in Iceland, may not have survived through interglacial intervals until Brunhes time. We link the growing local topographic relief to intensification of glaciations (i.e., increasing ice thickness) over the past 5 m.y. as follows. Preglacial Period (Before 4 Ma, ,10 m Local Relief). No evidence for bedrock erosion is observed in the lowest sedimentary horizon in Skaftafell (stage I-1), dated at ca. 4.6 Ma. Erosion associated with the oldest glacial horizon (G-1), dated both in Skaftafell and Hafrafell at ca. 4 Ma, is expressed by striated surfaces on the underlying lava flow. During this stage, lavas accumulated with rare sedimentary interbeds, and we see no example of lavas thinning out because of erosion. Early Landscape-Forming Period (4 to 3 Ma, 100 m Local Relief). In the 3.2 to 3.5 Ma part of the Hafrafell section, for example, we observe, in cliff profiles only 500 m apart, that a 40-m-thick formation of interglacial lavas has been removed by erosion from the first but not the second and a 90-m-thick lava formation is present in the second but not the first. There is evidence in some outcrops for minor subglacial volcanic activity. Main Landscape-Forming Period (3 to 0.8 Ma, 400 to 600 m Local Relief). The oldest subglacial volcanic ridge of significant positive relief (200 m) found in the area is dated at 2.8 Ma (G3), but younger ridges from this period are up to 400 m in height. Such topographic highs have probably led to channeling of ice since ca. 3 Ma and thus to deepening of the valleys through glacial erosion of the underlying lava flows. An example of this is in Jokulfell (5 km west of Skaftafell) where a 140 m thickness of volcanic section has been removed. From stratigraphic relationships, it is clear that this erosion took place prior to 2.5 Ma but after 3.35 Ma. The combination of positive relief produced by subglacially formed volcanic ridges (up to ;400 m) and negative relief caused by deepening valleys (100 to 200 m) led to a total local relief of as much as 600 m during this period. Present Period (Since 0.8 Ma, 2000 m Relief). At some point, the main ice flow from the highland ice sheet eroded deep and wide valleys that could no longer be refilled by volcanic activity during interglacial stages. These valleys thus became permanent channels for ice flow during successive glacial intervals. They became progressively deeper and wider, as indicated by a cap of Brunhes-age volcanic strata extending sporadically from valley floors to ridge crests of the Skaftafell region. Hence, the record of glacial and interglacial intervals is far less complete than for earlier times. Precise dating of the onset of this period is not possible with the present data, but it cannot be older 181
than about 1 Ma. It is most likely that there was a progressive increase in relief throughout the Brunhes chron. The present difference in height from the top of the volcanic section to the valley floor in Skaftafell and Hafrafell is 1800 m and 1200 m, respectively. The corre¨ ræfajo¨kull area, 15 km sponding relief for the O to the east and the highest elevation in Iceland, is at least 2000 m. Skaftafell in Relation to the Ice Sheet and Volcanic Centers In Borgarfjordur of western Iceland, McDougall et al. (1977) studied volcanic strata of similar age to that of the Skaftafell sections, and observed a complete magnetic polarity record for this period (i.e., no major erosional hiatuses). They did not, however, find subglacially formed volcanic material comparable to the accumulations we observe in the Skaftafell region. They reported only seven glacial intervals and made no note of subglacially formed volcanic ridges. The explanation for the lithologic differences between these regions appears to be that the Skaftafell area has, since 2.8 Ma, been more elevated and much closer to the center of the ice sheet that formed during glacial intervals and thus has developed a much greater local topographic relief. Another consideration is that the strata in western Iceland probably subsided at a faster rate within an accreting axial rift zone, so that any subglacially formed volcanic ridges might now be buried below sea level. Our results conform with the initiation of continental glacial conditions in other parts of the Northern Hemisphere by 2.8 Ma, followed by a second amplification at ca. 0.8 Ma (e.g., Pisias and Moore, 1979; Clark et al., 1999). In response to periods of thick ice cover, local topographic relief increased dramatically after 0.8 Ma, from less than 600 m to at least 2000 m, but how fast this increase happened is still unresolved. We suggest that up until ca. 2.8 Ma the Iceland shield reached maximum elevations of ;700 m, the current elevation of the rift axis nearest the hotspot, and glacier ice did not survive through interglacial intervals except perhaps on local stratovolcanoes. Cooler temperatures then allowed the buildup of sufficiently thick ice that subglacial volcanic ridges began to form and increased the local topographic relief enough to sustain small glaciers through interglacial intervals. These higher elevations survived in Skaftafell because the region does not subside as fast as crust created at the axial rifts. The formation of subglacial volcanic ridges and valley deepening continued to amplify local topographic relief through Brunhes time and caused southeast Iceland to be partly covered with a thick ice sheet (Vatnajo¨kull) during recent interglacial intervals (i.e., since 0.8 Ma). 182
Skaftafell in a Regional and Global Context Eiriksson et al. (1990), Eiriksson and Geirsdottir (1991), and Geirsdottir and Eiriksson (1994) have investigated the temporal and regional distribution of evidence for glaciations in Iceland back to ca. 9 Ma. They concluded that well over 20 glacial intervals have occurred in this period. Because these events have been dated by interpolation between magnetic polarity boundaries only, their timing is uncertain, as is the correlation between studied sections and the significance of possible erosional hiatuses. Hence, an increase in glacial activity since ca. 3 Ma is apparent, but quantification of discrete events (such as timing and distribution) was not possible from their data. Geirsdottir and Eiriksson (1994) proposed that the Skaftafell area has been covered with ice since late Miocene time, but our evidence shows permanent ice only in Brunhes time. Evidence for the onset and intensification of Northern Hemisphere glaciations since Late Pliocene time has come from variations in d18O recorded in benthic foraminifers (e.g., Mix et al., 1995), which represent continental ice-volume changes. The gradual increase to more positive d18O after ca. 3.5 Ma reflects slow global ice buildup, whereas the largeramplitude changes beginning ca. 1 Ma indicate swings from glacial to interglacial conditions. Studies of ice-rafted debris found in deep-sea drilling cores from the North Atlantic (Jansen et al., 1990; Jansen and Skoholm, 1991) provide additional evidence that intermittent sea ice existed from ca. 6–7 Ma along the Norwegian and East Greenland margins (Larsen et al., 1994). However, significant increases in ice-rafted debris at 2.6 Ma probably indicate the onset of true continental glaciations. The growth of topographic relief in southeast Iceland beginning at 2.5 to 3 Ma and intensification at about 0.8 Ma is consistent with this glacial history for the North Atlantic. ACKNOWLEDGMENTS Helgason was supported by the Icelandic Science Fund and the Alexander von Humboldt Foundation (Germany). Duncan was supported by the National Science Foundation (USA) and a travel grant from the Oregon State University Research Foundation. We thank Leo Kristjansson for generous use of his paleomagnetic laboratory; the late Ragnar Stefansson and his family at Skaftafell and later Freysnes for their invaluable assistance and hospitality; and John T. Andrews and Nicholas Eyles for constructive reviews. REFERENCES CITED Clark, P.U., Alley, R.B., and Pollard, D., 1999, Northern Hemisphere ice-sheet influences on global climate change: Science, v. 286, p. 1104–1111. Eiriksson, J., and Geirsdottir, A., 1991, A record of Pliocene and Pleistocene glaciations and climatic change in the North Atlantic based on variations in volcanic sedimentary facies in Iceland: Marine Geology, v. 101, p. 147–159.
Eiriksson, J., Gudmundsson, A.I., Kristjansson, L., and Gunnarsson, K., 1990, Palaeomagnetism of Pliocene-Pleistocene sediments and lava flows on Tjornes and Flatey, North Iceland: Boreas, v. 19, p. 39–55. Geirsdottir, A., and Eiriksson, J., 1994, Growth of an intermittent ice sheet in Iceland during the late Pliocene and early Pleistocene: Quaternary Research, v. 42, p. 115–130. Helgason, J., 1984, Frequent shifts of the volcanic zone in Iceland: Geology, v. 12, p. 212–216. Helgason, J., 1985, Shifts of the plate boundary in Iceland: Some aspects of Tertiary volcanism: Journal of Geophysical Research, v. 90, p. 10,084–10,092. Jansen, E., and Skoholm, J., 1991, Reconstruction of glaciation over the past 6 Myr from iceborne deposits in the Norwegian Sea: Nature, v. 349, p. 600–603. Jansen, E., Skoholm, J., Bleil, U., and Erichsen, A., 1990, Neogene and Pleistocene glaciations in the Northern Hemisphere and late MiocenePliocene global ice volume fluctuations: Evidence from the Norwegian Sea, in Bleil, U., and Thiede, J., eds., Geological history of the polar oceans: Arctic versus Antarctic: London, Kluwer, p. 677–705. Larsen, H.C., Saunders, A.D., Clift, P., Beget, J., Wei, W., Spezzaferri, S., and Scientific and Shipboard Party, Ocean Drilling Project Leg 152, 1994, Seven million years of glaciation in Greenland: Science, v. 264, p. 952–955. McDougall, I., Watkins, N.D., Walker, G.P.L., and Kristjansson, L., 1976, Potassium-argon and paleomagnetic analysis of Icelandic lava flows: Limits on the age of Anomaly 5: Journal of Geophysical Research, v. 81, p. 1505–1512. McDougall, I., Saemundsson, K., Johannesson, H., Watkins, N.D., and Kristjansson, L., 1977, Extension of the geomagnetic polarity time scale to 6.5 m.y.: K-Ar dating, geological and paleomagnetic study of a 3,500-m lava succession in western Iceland: Geological Society of America Bulletin, v. 88, p. 1–15. Mix, A.C., Pisias, N.G., Rugh, W., Wilson, J., Morey, A., and Hagelberg, T.K., 1995, Benthic foraminifer stable isotope record from Site 849 (0–5 Ma): Local and global climate changes, in Pisias, N.G., et al., Proceedings of the Ocean Drilling Project, Scientific Results, Volume 138: College Station, Texas, Ocean Drilling Program, p. 371–412. Palmason, G., 1980, A continuum model of crustal generation in Iceland: Kinematic aspects: Journal of Geophysics, v. 47, p. 7–18. Palmason, G., 1986, Model of crustal formation in Iceland, and application to submarine midocean ridges, in Vogt, P.R., and Tucholke, B.E., eds., The western Atlantic region: Boulder, Colorado, Geological Society of America, Geology of North America, v. M, p. 87–97. Pisias, N.G., and Moore, T.C., Jr., 1979, The evolution of Pleistocene climate: A time series approach: Earth and Planetary Science Letters, v. 52, p. 450–458. Prestvik, T., 1985, Petrology of Quaternary volcanic rocks from Oraefi, southeast Iceland: Trondheim, Norway, Geologisk Institutt, Norges Tekniske Hojskole, Report 21, 81 p. Walker, G.P.L., 1963, The Breiddalur central volcano, eastern Iceland: Quaternary Journal of the Geological Society of London, v. 119, p. 29–63. Manuscript received August 2, 2000 Revised manuscript received November 6, 2000 Manuscript accepted November 7, 2000 Printed in USA GEOLOGY, February 2001
