Map of the Caribbean Basin, western Atlantic Ocean, adjacent continents and islands, localities referred to in the text, tectonic features (from Prentice et al., 2010), and southern limits of the Laurentide ice sheet during the past two glacial periods (Flint, 1971; Dyke et al., 2002). 

Contexts in source publication

Context 1

... prospect of a warmer Earth in the near future has increased concern about sea-level rise resulting from melting of large ice sheets in Greenland and Antarctica. One of the most effective means of understanding the magnitude of possible sea-level rise is study of high sea levels during past warm interglacial periods. High sea stands of the past are often preserved in the geologic record as emergent reef terraces with a “ staircase ” type of geomorphology on tectonically rising coasts, such as those on Barbados or Haiti ( Fig. 1a). In contrast, on tectonically stable coasts, or those that are subsiding slowly, such as the Florida Keys or the Cayman Islands, high sea stands are recorded as limestones composed of stacked reef tracts, separated by subaerial exposure surfaces with weathered zones or paleosols (Fig. 1b). In the deep-sea oxygen isotope record, the most recent previous interglacial periods are represen- ted by Marine Isotope Stages (MIS) 5.5 (~125 ka), 7 (~200 ka), 9 (~300 ka), and 11 (~400 ka) (Imbrie et al., 1984; Rohling et al., 2009, 2010). As pointed out in the 2007 IPCC Report, these four interglacial periods differ from one another in their degree of warmth, duration, and timing relative to the Earth's orbital geometry (Jansen et al., 2007). Of the past warm periods, the last interglacial or MIS 5.5 has received the most attention (see reviews in Kopp et al., 2009 and Muhs et al., 2011). There have been fewer studies of MIS 7 and MIS 9. There has been an increasing interest in MIS 11 (~400 ka) because the Earth's orbital con fi guration at that time was similar to that of today (Berger and Loutre, 1991, 2002). Indeed, Berger and Loutre (2002) speculate that MIS 11 could have been an exceptionally long and warm interglacial period and a suitable analog for a future climate on Earth. To investigate past interglacial high sea stands, we studied coral reef terraces on the island of Curaçao, Leeward Antilles islands, in the southern Caribbean Sea (Fig. 2). The Leeward Antilles islands of Curaçao, Bonaire and Aruba are situat- ed in the southernmost part of the Caribbean Sea (Fig. 2). Nearby, the Caribbean plate underthrusts the South American plate at a shallow angle north of the Leeward Antilles ridge, a major topographic/bathymetric high that includes Aruba, Bonaire and Curaçao (Hippolyte and Mann, 2011). This subduction process has probably been in progress since the Paleogene. However, recent data indicate that the main sense of motion of the Caribbean plate, relative to the South American plate, is east-west, at a rate of ~20 mm/yr (Hippolyte and Mann, 2011). The geology of Curaçao consists of a core of Cretaceous diabase fl anked by Paleogene sedimentary rocks, in turn overlain by the Seroe Domi Formation, a thick sequence of Neogene marine limestones and siliciclastic sandstone (De Buisonjé, 1974). Along the coastal margins of Aruba, Curaçao and Bonaire there are spectacular fl ights of Quaternary marine terraces (Figs. 3 and 4). Numerous workers in the early part of the 20th century recognized the geologic signi fi cance of the marine terraces on Curaçao and De Buisonjé (1974) summarizes these early studies. Alexander (1961) mapped and measured the terraces and was perhaps the fi rst investigator to recognize that the terraces on the islands of the Leeward Antilles were the products of Quaternary interglacial/glacial sea-level fl uctuations superimposed on steady tectonic uplift. This fundamental con- cept is key to our present understanding of marine terrace origins on many coastlines and islands worldwide. Nevertheless, Alexander (1961) interpreted the terraces on Curaçao as erosional landforms. Although De Buisonjé (1974) thought that this interpretation applied to some of the terraces on the leeward (southern) side of the island (where they are cut on the Seroe Domi Formation), he recognized that most of the terraces, particularly on the windward (northern) side, are constructional coral reef terraces, similar to those on the island of Barbados. De Buisonjé (1974) identi fi ed three distinct facies in the terraces, seaward to landward: (1) a barrier reef zone, dominated by Acropora palmata ; (2) a lagoonal or back-reef zone dominated by Montastraea annularis ; and (3) an innermost lagoonal or back- reef zone dominated by Siderastrea . In remapping the Quaternary deposits of Aruba, Curaçao and Bonaire, De Buisonjé (1974) recognized several emergent terraces on Curaçao named, from oldest to youngest: Highest Terrace (~ 90 up to ~150 m), Higher Terrace (~50 to 85 m), Middle Terrace I (~15 to ~ 25 m), Middle Terrace II (~ 25 to ~ 45 m), and Lower Terrace (~6 to 12 m). De Buisonjé (1974) states that Middle Terrace II deposits overlie Middle Terrace I deposits in places and thus are stratigraphically younger even though they are topographically higher. Herweijer and Focke (1978) report that the Lower Terrace on Curaçao consists of two constructional reef tracts (the lower called the Cortalein, the higher called the Hato) separated by a prominent discontinuity. Pandol fi et al. (1999), Pandol fi (2001), Pandol fi and Jackson (2001) and Meyer et al. (2003) provide details of the coral fauna from the Hato unit. Most previous age determinations of reef terraces on Curaçao have been on the Lower Terrace. Schubert and Szabo (1978) report alpha- spectrometric U-series ages of ~125 ka, correlating the Hato unit of the Lower Terrace to the peak of the last interglacial period, or MIS 5.5. Hamelin et al. (1991) analyzed two corals from the Hato unit of the Lower Terrace on the leeward side of the island using thermal ionization mass spectrometric (TIMS) U-series dating and report ages similar to those of Schubert and Szabo (1978). Although Schubert and Szabo (1978) concluded that the upper part of the Lower Terrace represents MIS 5.5, they speculated that terraces representing MIS 7 (~250 – 200 ka) and MIS 9 (~350 – 300 ka) were missing on Curaçao. In contrast, Herweijer and Focke (1978) hypothesized that the lower, Cortalein unit of the Lower Terrace could date to MIS 7. Schellmann et al. (2004) conducted electron spin resonance (ESR) analyses of corals from both units of the Lower Terrace. They report ages of 140 ka to 101 ka for the Hato unit and 226 ka to 189 ka for the Cortalein unit, in support of Herweijer and Focke's (1978) hypothesis. There have been a few attempts to date one of the older terraces on Curaçao, all of them on Middle Terrace I and all apparently very close to Boca San Pedro on the windward coast of the island (Fig. 3). McFarlane and Lundberg (2002) report a TIMS U-series age on an unrecrystallized coral from this terrace of ~405 ka. Their sample has a calculated initial 234 U/ 238 U value that is higher than seawater, indicating that it is probably biased old, but following Gallup et al.'s (1994) model, the coral would fall within the age range of MIS 11. Schellmann et al. (2004) report four ESR ages on coral from Middle ...

Context 2

... prospect of a warmer Earth in the near future has increased concern about sea-level rise resulting from melting of large ice sheets in Greenland and Antarctica. One of the most effective means of understanding the magnitude of possible sea-level rise is study of high sea levels during past warm interglacial periods. High sea stands of the past are often preserved in the geologic record as emergent reef terraces with a “ staircase ” type of geomorphology on tectonically rising coasts, such as those on Barbados or Haiti ( Fig. 1a). In contrast, on tectonically stable coasts, or those that are subsiding slowly, such as the Florida Keys or the Cayman Islands, high sea stands are recorded as limestones composed of stacked reef tracts, separated by subaerial exposure surfaces with weathered zones or paleosols (Fig. 1b). In the deep-sea oxygen isotope record, the most recent previous interglacial periods are represen- ted by Marine Isotope Stages (MIS) 5.5 (~125 ka), 7 (~200 ka), 9 (~300 ka), and 11 (~400 ka) (Imbrie et al., 1984; Rohling et al., 2009, 2010). As pointed out in the 2007 IPCC Report, these four interglacial periods differ from one another in their degree of warmth, duration, and timing relative to the Earth's orbital geometry (Jansen et al., 2007). Of the past warm periods, the last interglacial or MIS 5.5 has received the most attention (see reviews in Kopp et al., 2009 and Muhs et al., 2011). There have been fewer studies of MIS 7 and MIS 9. There has been an increasing interest in MIS 11 (~400 ka) because the Earth's orbital con fi guration at that time was similar to that of today (Berger and Loutre, 1991, 2002). Indeed, Berger and Loutre (2002) speculate that MIS 11 could have been an exceptionally long and warm interglacial period and a suitable analog for a future climate on Earth. To investigate past interglacial high sea stands, we studied coral reef terraces on the island of Curaçao, Leeward Antilles islands, in the southern Caribbean Sea (Fig. 2). The Leeward Antilles islands of Curaçao, Bonaire and Aruba are situat- ed in the southernmost part of the Caribbean Sea (Fig. 2). Nearby, the Caribbean plate underthrusts the South American plate at a shallow angle north of the Leeward Antilles ridge, a major topographic/bathymetric high that includes Aruba, Bonaire and Curaçao (Hippolyte and Mann, 2011). This subduction process has probably been in progress since the Paleogene. However, recent data indicate that the main sense of motion of the Caribbean plate, relative to the South American plate, is east-west, at a rate of ~20 mm/yr (Hippolyte and Mann, 2011). The geology of Curaçao consists of a core of Cretaceous diabase fl anked by Paleogene sedimentary rocks, in turn overlain by the Seroe Domi Formation, a thick sequence of Neogene marine limestones and siliciclastic sandstone (De Buisonjé, 1974). Along the coastal margins of Aruba, Curaçao and Bonaire there are spectacular fl ights of Quaternary marine terraces (Figs. 3 and 4). Numerous workers in the early part of the 20th century recognized the geologic signi fi cance of the marine terraces on Curaçao and De Buisonjé (1974) summarizes these early studies. Alexander (1961) mapped and measured the terraces and was perhaps the fi rst investigator to recognize that the terraces on the islands of the Leeward Antilles were the products of Quaternary interglacial/glacial sea-level fl uctuations superimposed on steady tectonic uplift. This fundamental con- cept is key to our present understanding of marine terrace origins on many coastlines and islands worldwide. Nevertheless, Alexander (1961) interpreted the terraces on Curaçao as erosional landforms. Although De Buisonjé (1974) thought that this interpretation applied to some of the terraces on the leeward (southern) side of the island (where they are cut on the Seroe Domi Formation), he recognized that most of the terraces, particularly on the windward (northern) side, are constructional coral reef terraces, similar to those on the island of Barbados. De Buisonjé (1974) identi fi ed three distinct facies in the terraces, seaward to landward: (1) a barrier reef zone, dominated by Acropora palmata ; (2) a lagoonal or back-reef zone dominated by Montastraea annularis ; and (3) an innermost lagoonal or back- reef zone dominated by Siderastrea . In remapping the Quaternary deposits of Aruba, Curaçao and Bonaire, De Buisonjé (1974) recognized several emergent terraces on Curaçao named, from oldest to youngest: Highest Terrace (~ 90 up to ~150 m), Higher Terrace (~50 to 85 m), Middle Terrace I (~15 to ~ 25 m), Middle Terrace II (~ 25 to ~ 45 m), and Lower Terrace (~6 to 12 m). De Buisonjé (1974) states that Middle Terrace II deposits overlie Middle Terrace I deposits in places and thus are stratigraphically younger even though they are topographically higher. Herweijer and Focke (1978) report that the Lower Terrace on Curaçao consists of two constructional reef tracts (the lower called the Cortalein, the higher called the Hato) separated by a prominent discontinuity. Pandol fi et al. (1999), Pandol fi (2001), Pandol fi and Jackson (2001) and Meyer et al. (2003) provide details of the coral fauna from the Hato unit. Most previous age determinations of reef terraces on Curaçao have been on the Lower Terrace. Schubert and Szabo (1978) report alpha- spectrometric U-series ages of ~125 ka, correlating the Hato unit of the Lower Terrace to the peak of the last interglacial period, or MIS 5.5. Hamelin et al. (1991) analyzed two corals from the Hato unit of the Lower Terrace on the leeward side of the island using thermal ionization mass spectrometric (TIMS) U-series dating and report ages similar to those of Schubert and Szabo (1978). Although Schubert and Szabo (1978) concluded that the upper part of the Lower Terrace represents MIS 5.5, they speculated that terraces representing MIS 7 (~250 – 200 ka) and MIS 9 (~350 – 300 ka) were missing on Curaçao. In contrast, Herweijer and Focke (1978) hypothesized that the lower, Cortalein unit of the Lower Terrace could date to MIS 7. Schellmann et al. (2004) conducted electron spin resonance (ESR) analyses of corals from both units of the Lower Terrace. They report ages of 140 ka to 101 ka for the Hato unit and 226 ka to 189 ka for the Cortalein unit, in support of Herweijer and Focke's (1978) hypothesis. There have been a few attempts to date one of the older terraces on Curaçao, all of them on Middle Terrace I and all apparently very close to Boca San Pedro on the windward coast of the island (Fig. 3). McFarlane and Lundberg (2002) report a TIMS U-series age on an unrecrystallized coral from this terrace of ~405 ka. Their sample has a calculated initial 234 U/ 238 U value that is higher than seawater, indicating that it is probably biased old, but following Gallup et al.'s (1994) model, the coral would fall within the age range of MIS 11. Schellmann et al. (2004) report four ESR ages on coral from Middle Terrace I, with estimates ranging from ~363 ka to ~544 ka. Thus, their data would also permit a correlation of Middle Terrace I with ...

Context 3

... rate back to ~200 ka, yields a range of possible paleo-sea levels on Curaçao during late MIS 7 from − 3.3 m to +2.3 m, relative to present. Using the same late Quaternary uplift rate, we also estimate paleo- sea level at the time of formation of Middle Terrace I (Fig. 7), but this requires careful interpretation of existing data. Lundberg and McFarlane (2002), McFarlane and Lundberg (2002) and Schellmann et al. (2004) suggest that Middle Terrace I could represent a high sea stand during MIS 11 at ~400 ka. De Buisonjé (1974) infers that deposits of Middle Terrace I underlie deposits of Middle Terrace II. Above Middle Terrace II, there is a notch cut in the cliff face of the outer edge of the Higher Terrace that Lundberg and McFarlane (2002) and McFarlane and Lundberg (2002) infer to have formed during MIS 11. Because the notch is undated, it could be older than MIS 11, as pointed out by Bowen (2010). There is no direct evidence that Middle Terrace II is of MIS 11 age either, although Lundberg and McFarlane (2002) and McFarlane and Lundberg (2002) infer that it dates to this time period. We take a conservative view that the maximum elevation recorded by a high stand in MIS 11 at ~400 ka is the highest elevation (~30 m) of the inner edge of Middle Terrace I that we measured at San Pedro (Fig. 7a). For San Pedro, we use the same late Quaternary uplift rate that we calculated for the Boca Cortalein/Boca Mansaliña area. Our GPS elevations of the notch at the inner edge of the Lower Terrace at San Pedro were hindered by poor satellite geometry because of the high (~27 m) cliff of the outer edge of Middle Terrace I. Nevertheless, we measured notch elevations here of 11.1 – 13.1 m, which bracket the inner edge elevation of 12.4 m at Boca Cortalein. Thus, we use the same uplift rate for San Pedro as we use for Boca Cortalein (Table 2). Results indicate that with a conservative estimate of last interglacial paleo-sea level (+6 m), MIS 11 sea level on Curaçao could have been +8.4 to +10 m, relative to present. Using a higher last interglacial sea level of +9 m yields much higher MIS 11 sea level estimates, from +17.4 m to +19.6 m (Table 2). Corals showing mostly closed-system histories from the Hato unit of the Lower Terrace have U-series ages ranging from ~ 126 ka (Boca Mansaliña, Boca Cortalein, Punta Halvedag) to ~ 118 ka (Un Boca and Boca Santu Pretu). If the corrected age of ~129 ka at Un Boca is accepted, then the high stand could have started even earlier. These ages indicate that the last interglacial period on Curaçao had a duration of at least 8,000 yr, and possibly ~ 11,000 yr, similar to what has been reported for tectonically stable coasts elsewhere (Chen et al., 1991; Stirling et al., 1995, 1998; Muhs et al., 2002, 2011). The Curaçao fossil reefs are dominated by the rapidly growing Acropora palmata , and thus may be examples of "keep-up" reefs that respond quickly to a rising sea level (Neumann and MacIntyre, 1985). Sea level was probably already high, at or above modern sea level, by ~ 129 – 126 ka, when summer insolation at high latitudes in the Northern Hemisphere was peaking (Berger and Loutre, 1991). This contrasts with the present interglacial, wherein the peak of sea level lagged the Northern Hemisphere summer insolation high by several thousand years. The penultimate interglacial period, MIS 7, was a complex consisting of three high-sea stands, based on the Red Sea oxygen isotope record (Fig. 10). In the Mediterranean region, speleothem records also show evidence of three relatively high sea stands, with the oldest (about 249 – 231 ka) and youngest (about 201 – 190 ka) stands being higher than the intermediate-aged (about 217 – 206 ka) stand (Dutton et al., 2009), consistent with the Red Sea record. At least three MIS 7 terraces may be present on Barbados (Bender et al., 1979; Schellmann and Radtke, 2004). The youngest of these (~200 ka) is estimated to have been a high stand above present (Gallup et al., 1994, 2002; Schellmann and Radtke, 2004). On tectonically stable Bermuda, the Florida Keys, and Grand Cayman Island, an MIS 7 sea level dated to ~200 ka is estimated to be near present (Harmon et al., 1983; Vénzina et al., 1999; Muhs et al., 2002, 2011). Chappell (1974) estimated paleo-sea levels for emergent MIS 7 reefs very close to present on the rapidly rising coast of New Guinea. In contrast, on the very slowly uplifting island of Oahu, Hawaii and on the slowly subsiding island of Mururoa in the South Paci fi c, MIS 7 sea levels are estimated to be well below present (Sherman et al., 1999; Camoin et al., 2001; Fletcher et al., 2008). Some of the differences in estimates of MIS 7 sea levels could be due to glacial isostatic adjustment (GIA) processes (Nakada and Lambeck, 1989; Mitrovica and Peltier, 1991; Milne and Mitrovica, 2008; Raymo and Mitrovica, 2012). Potter and Lambeck (2003) and Lambeck et al. (2012) show that because of the proximity of Bermuda to the Laurentide ice sheet (Fig. 2), GIA processes have likely affected this island's sea-level record, at least during the last interglacial complex. In contrast, Curaçao, like Barbados, is distant from the Laurentide ice sheets of the past two glacial periods (Fig. 1) and thus can be considered a far fi eld location with respect to GIA effects. If so, apparent paleo-sea levels on Curaçao may be close to those of eustatic sea levels. Nevertheless, our estimates of sea level close to present for Curaçao are similar to those estimated for Bermuda, which should have been affected by GIA processes. Studies of the magnitude of the MIS 11 high sea stand at ~400 ka have generated highly divergent estimates, from ~22 m above present (Hearty et al., 1999; Lundberg and McFarlane, 2002) to a high sea stand near present (Bowen, 2010). The high MIS 11 sea level proposed by Hearty et al. (1999) is based on data from both Bermuda and the Bahamas, and there have been challenges to this interpretation of the record for both islands (McMurtry et al., 2007; Mylroie, 2008). As discussed above, GIA effects could have had a strong imprint on the sea-level record for Bermuda. Recent modeling by Raymo and Mitrovica (2012) indicate that at least 10 m of the alleged MIS 11 high stand on Bermuda can be accounted for by GIA effects. These workers estimate the eustatic component of sea level rise during MIS 11 to be on the order of +6 m to +13 m. Our most conservative estimates of the MIS 11 high stand on Curaçao indicate it could have been as high as +10 m or as low as +8.4 m, in broad agreement with Raymo and Mitrovica (2012) These conservative estimates are still higher than those from the Red Sea oxygen isotope record (Fig. 10). Even with the most cautious approach, our paleo-sea level estimate for MIS 11 at ~400 ka would still imply a very warm interglacial, requiring loss of most or all of the present Greenland and West Antarctic ice sheets (see discussion in Muhs et al., 2011). We stress, however, that better geochronology is needed for Middle Terrace I before fi rm conclusions can be drawn about sea-level history during this ...