A 300 000-yr coral reef record of sea level changes, Mururoa atoll (Tuamotu archipelago, French Polynesia)

Abstract

The accurate dating of fossil coral reefs is of prime importance in determining the timing of deglaciation events and thus understanding the mechanisms driving glacial–interglacial cycles. So far, the most useful coral reef records of past sea level changes are those related to the last deglaciation and the Last Interglacial period. U/Th ages for older isotopic stages are more limited, due to the scarcity of datable material, reflecting diagenetic alteration. Most data are from emergent parts of reefs and reef terraces in active subduction zones where relative sea level records may be biased by variations in rates of tectonic uplift. New constraints on sea level changes over the past 300 000 yr are based on high-precision U-series age measurements of successive reef units on Mururoa. These have been cored in four continuous 300-m-long drill holes with seaward inclinations of 30 to 45° on the northeastern rim of the atoll. Past sea level positions have been calculated from the radiometric ages of corals by correcting the present depth of subsurface horizons both for thermal subsidence and for depositional palaeodepth. The location of this atoll at a considerable distance from former ice sheets (‘far field’) minimises the influence of glacio–isostatic rebound. Prominent units formed during four periods of relative sea level highstands, including the Holocene and stages 5 (∼125 ka), 7 (∼212 ka) and 9 (∼332 ka). These are primarily composed of coralgal frameworks that grew in very shallow water. Three periods of relative low stand correspond to stages 2 (∼17–23 ka), 4 (∼60 ka) and 8 (∼270 ka) during which small reefs developed in association with large bioclastic accumulations. Good agreement with the timing of sea level changes based on oxygen isotope measurements in deep-sea cores is noted for most of the dated reef units. We report here the first accurate coral record of the Last Glacial Maximum in the Pacific, 135–143 m below the present sea surface, suggesting that sea level may have been lower than expected during this period.

Introduction

Coral reefs are sensitive recorders of past sea level and environmental changes. Their accurate dating by mass spectrometry is of prime importance for the determination of the timing of deglaciation events and thus for the understanding of the mechanisms driving glacial–interglacial cycles (Edwards et al., 1987, Chen et al., 1991). The most useful coral reef records of past sea level changes are related to the last deglaciation (Fairbanks, 1989, Bard et al., 1990, Bard et al., 1996a) and the Last Interglacial period approximately 125 ka (Edwards et al., 1987, Hamelin et al., 1991, Chen et al., 1991, Stein et al., 1993, Gallup et al., 1994, Stirling et al., 1995, Stirling et al., 1998, Eisenhauer et al., 1993). Unfortunately, U/Th ages for older isotopic stages are limited due to the scarcity of datable material arising from diagenetic alteration and post-depositional migration of U and Th isotopes (Hamelin et al., 1991). After Broecker et al. (1968), the most complete records of sea levels for the last glacial cycle and/or stage 7 (reefs of the penultimate Interglacial) have come from coral reef terraces exposed on the Huon Peninsula, New Guinea (Chappell, 1974, Bloom et al., 1974, Chappell and Veeh, 1978, Stein et al., 1993, Chappell et al., 1996), Barbados (Mesolella et al., 1969, Bard et al., 1990, Gallup et al., 1994) and Sumba (Pirazzoli et al., 1993, Bard et al., 1996b). These generally coincide with major sea level highs predicted by the astronomical theory of climate change (Milankovitch, 1941) and, to a lesser extent, to reef units formed during glacial periods. However, they concern uplifted and presently emerged parts of reefs and reef terraces in active subduction zones. Vertical tectonic movements in such zones may be large and are often discontinuous, implying that apparent sea level records may be biased by variations in the rates of uplift. Furthermore, reef terraces in uplifted areas are commonly subject to diagenetic alteration. Hence, there is a clear need to study sea level changes in tectonically stable regions or in areas where vertical movements are slow and/or regular, restricting the possible alteration of deposits in subaerial environments.

Limestone columns beneath mid-oceanic atolls contain excellent records of past sea level changes, including both highstands and lowstands, and have frequently been referred to as ‘dipsticks’ (Wheeler and Aharon, 1991). Thermal subsidence due to gradual cooling of the oceanic lithosphere is the primary driving component of atoll subsidence and tends towards linearity with time (Parsons and Sclater, 1977, Detrick and Crough, 1978). Furthermore, because Pacific atolls at a considerable distance from former ice sheets (‘far field’) the influence of glacio–isostatic rebound is minimised (Peltier, 1991, Mitrovica and Peltier, 1991). The eustatic function is best estimated from sea level data collected far from the former ice margins where the glacio–isostatic effects are smaller (see Lambeck et al., 2000).

The present study concerns four continuous drill holes 300 m in length, with seaward inclinations of 30 to 45°, on two sites, pK 5 (cores 5/30 and 5/45) and pK 8 (cores 8/30 and 8/40) on the NE rim of Mururoa atoll by the French Commissariat à l’Energie Atomique (Fig. 1). Additional data are derived from vertical cores drilled through the modern reef rim (core ‘Françoise’) and in the lagoon (cores ‘FR Lagoon 2’ and ‘Janie 17/1’). Here we document the compositions of successive reef units and constrain sea level changes over the last 300 000 yr. Detailed sedimentological data, and the reconstruction of the geometry and development through time of the atoll will be the subject of a separate paper.