pulse was centered at 8,000 calendar years (Toscano
and Macintyre, 2003) and produced a rise of 6.5 m in
<140 years (Blanchon and Shaw, 1995).
Subsequent attempts to test the 3-step model of Caribbean sea-level rise, and corroborate the existence, timing,
and magnitude of melt-water pulses, have only been partially successful. The most successful sea-level reconstruction is from the Sunda Shelf between the
Indonesian archipelago and the Vietnam peninsula
(Hanebuth et al., 2000). Two core transects between the
126 m and 48 m isobaths recovered a transgressive
sequence of organic-rich tidal-flat and mangrove-swamp
deposits (Hanebuth and Stattegger, 2003). Radiocarbon
dating of in-situ organics above the Pleistocene basement
allowed the reconstruction of sea level (Æ 1 m) from
21 to 11 kyrs ago. This high-resolution reconstruction
showed that Mwp-1a was recorded as a 16 m rise in
300 years starting at 14.6 calendar kyrs ago, almost identical to the rate and magnitude calculated from the Caribbean drowned reefs (13.5 Æ 2.5 m in 290 years;
Blanchon and Shaw, 1995). The only difference was
the timing: on the Sunda Shelf Mwp-1a occurred
400 years before it did on Barbados (Figure 1). Although
there has been considerable argument about which of
these is correct (e.g., Weaver et al., 2003; Stanford
et al., 2006), the difference in age between these sites
may simply be the result of down-slope transport of
younger A. palmata clasts onto the oldest drowned reef
on Barbados seen at the top of core RGF-9 (Figure 1).
Such a possibility could easily be tested by dating the
lowest section of the up-slope reef (seen at the base of
core RGF-12). Down-slope transport would cause an
age overlap between the drowned and up-slope reefs.
Less precise reconstructions using corals with larger
depth ranges have had more limited success. For example,
a single core from the Huon Peninsula of Papua New
Guinea recovered an uplifted 52 m thick coral sequence
which recorded continuous reef accretion between 11,000
and 7,000 radiocarbon years BP (Chappell and Polach,
1991). All coral samples were later re-dated to obtain
more accurate
230
Th ages (Edwards et al., 1993) and
corrected for tectonic uplift (1.9 mm/yr). The sea-level
curve based on these new data showed a distinct period of
reduced melting between 12.3 kyrs and 11.0 kyrs, followed
by a clear Mwp-1b centered at 11 kyrs (Figure 1). It was
claimed that the period of reduced melting prior to the
meltwater pulse was shorter than at Barbados, but corals
used to identify sea-level position at Huon have larger depth
ranges than A. palmata and therefore are not as reliable in
pinpointing the precise elevation of sea level (Chappell
and Polach 1991). Also, the assumption that uplift was uniform in this highly-active neotectonic terrain cannot be
confirmed.
Cores recovered from the modern reef-crest around the
more tectonically-stable island of Tahiti found a thicker
90 m sequence of continuous reef accretion back to
$13,800
230
Th yrs BP (Bard et al., 1996; Montaggioni
et al., 1997). This sequence therefore started accreting
immediately following Mwp-1a implying that its effects
in Tahiti were the same as in Barbados, and that a rapid
sea-level rise had caused reef drowning and up-slope reef
back-stepping (Blanchon, this volume). As in the other
records, the coral age/depth data, when corrected for subsidence (0.25 mm/yr) showed possible evidence for
a period of reduced melting prior to Mwp-1b, but it was
significantly shorter (11.5–11 kyrs) than that in either
Barbados or PNG. Interestingly, however, the Tahitian
corals consistently plot deeper than those at Barbados,
indicating that they may not be as precise at indicating
sea level as claimed (Blanchon, 1998). This is supported
by two observations: the increase in offset of the Tahiti
corals through time, and the pattern of shallowing of the
cored reef sequence (Figure 1). Both observations indicate
that the reef prograded after Mwp-1a so that, in a vertical
core, progressively deeper reef-front corals were encountered through time. In other words, Tahitian cores represent deeper-reef accretion when compared to those in
Barbados. New cores collected during the IODP-310
Tahiti Leg, should help determine if this offset is indeed
the result of deeper coral growth or a function of differing
glacio-isostatic-adjustment histories. One preliminary
result so far indicates that Mwp-1a at Tahiti has a similar
timing to that identified from the Sunda Shelf at 14.6 ka
(Deschamps et al., 2009).
In addition to the deglacial events, evidence for
a meltwater pulse at 8 kyrs (Mwp-1c) has also been further
supported by additional discoveries of drowned reefs from
Grand Cayman, southeast Florida and the Gulf of
Carpentaria (Blanchon et al., 2002, Banks et al., 2007,
Harris et al., 2008). On the east coast of Grand Cayman,
a transect of short cores along the reef-front in 20 m of water
recovered a submerged A. palmata reef-crest whose surface
corals returned
230
Th ages of between 8.9 kyrs and 8.1 kyrs
BP (Blanchon et al., 2002). The depth of this drowned reef
also corresponded with an intertidal notch on the western
shelf at a depth of 18 m and indicated the establishment of
an 18-m shoreline between 8.1 cal and 7.6 cal ka. Comparison of the age/depth data between Grand Cayman and
other Caribbean islands, indicated a near synchronous
demise and back-stepping of A. palmata reefs across the
Caribbean around 8 kyrs ago in response to a rapid 6-m
jump in sea level (Blanchon et al., 2002). But two potential
problems with this jump hypothesis have been raised. The
first problem was that back-stepping could only be demonstrated by comparing the age of reef drowning in one location
with the age of up-slope reef initiation in another. In other
words, there were few sites where clear evidence of both
drowning and back-stepping had been reported together.
That changed recently with the reporting of new age/depth
data from an up-slope reef off the southeast Florida shelf
(Banks et al., 2007). Combining the calibrated radiocarbonage data of this new reef with those from an earlier report
of an 8 ka old drowned reef further down-slope (Lighty et al.,
1978), confirmed that at this site Mwp-1c at 8 kyrs caused the
back-stepping of A. palmata reef crests 6–7 m up-slope in
less than 580 years (Blanchon, this volume).
MELTWATER PULSES
685
and Macintyre, 2003) and produced a rise of 6.5 m in
<140 years (Blanchon and Shaw, 1995).
Subsequent attempts to test the 3-step model of Caribbean sea-level rise, and corroborate the existence, timing,
and magnitude of melt-water pulses, have only been partially successful. The most successful sea-level reconstruction is from the Sunda Shelf between the
Indonesian archipelago and the Vietnam peninsula
(Hanebuth et al., 2000). Two core transects between the
126 m and 48 m isobaths recovered a transgressive
sequence of organic-rich tidal-flat and mangrove-swamp
deposits (Hanebuth and Stattegger, 2003). Radiocarbon
dating of in-situ organics above the Pleistocene basement
allowed the reconstruction of sea level (Æ 1 m) from
21 to 11 kyrs ago. This high-resolution reconstruction
showed that Mwp-1a was recorded as a 16 m rise in
300 years starting at 14.6 calendar kyrs ago, almost identical to the rate and magnitude calculated from the Caribbean drowned reefs (13.5 Æ 2.5 m in 290 years;
Blanchon and Shaw, 1995). The only difference was
the timing: on the Sunda Shelf Mwp-1a occurred
400 years before it did on Barbados (Figure 1). Although
there has been considerable argument about which of
these is correct (e.g., Weaver et al., 2003; Stanford
et al., 2006), the difference in age between these sites
may simply be the result of down-slope transport of
younger A. palmata clasts onto the oldest drowned reef
on Barbados seen at the top of core RGF-9 (Figure 1).
Such a possibility could easily be tested by dating the
lowest section of the up-slope reef (seen at the base of
core RGF-12). Down-slope transport would cause an
age overlap between the drowned and up-slope reefs.
Less precise reconstructions using corals with larger
depth ranges have had more limited success. For example,
a single core from the Huon Peninsula of Papua New
Guinea recovered an uplifted 52 m thick coral sequence
which recorded continuous reef accretion between 11,000
and 7,000 radiocarbon years BP (Chappell and Polach,
1991). All coral samples were later re-dated to obtain
more accurate
230
Th ages (Edwards et al., 1993) and
corrected for tectonic uplift (1.9 mm/yr). The sea-level
curve based on these new data showed a distinct period of
reduced melting between 12.3 kyrs and 11.0 kyrs, followed
by a clear Mwp-1b centered at 11 kyrs (Figure 1). It was
claimed that the period of reduced melting prior to the
meltwater pulse was shorter than at Barbados, but corals
used to identify sea-level position at Huon have larger depth
ranges than A. palmata and therefore are not as reliable in
pinpointing the precise elevation of sea level (Chappell
and Polach 1991). Also, the assumption that uplift was uniform in this highly-active neotectonic terrain cannot be
confirmed.
Cores recovered from the modern reef-crest around the
more tectonically-stable island of Tahiti found a thicker
90 m sequence of continuous reef accretion back to
$13,800
230
Th yrs BP (Bard et al., 1996; Montaggioni
et al., 1997). This sequence therefore started accreting
immediately following Mwp-1a implying that its effects
in Tahiti were the same as in Barbados, and that a rapid
sea-level rise had caused reef drowning and up-slope reef
back-stepping (Blanchon, this volume). As in the other
records, the coral age/depth data, when corrected for subsidence (0.25 mm/yr) showed possible evidence for
a period of reduced melting prior to Mwp-1b, but it was
significantly shorter (11.5–11 kyrs) than that in either
Barbados or PNG. Interestingly, however, the Tahitian
corals consistently plot deeper than those at Barbados,
indicating that they may not be as precise at indicating
sea level as claimed (Blanchon, 1998). This is supported
by two observations: the increase in offset of the Tahiti
corals through time, and the pattern of shallowing of the
cored reef sequence (Figure 1). Both observations indicate
that the reef prograded after Mwp-1a so that, in a vertical
core, progressively deeper reef-front corals were encountered through time. In other words, Tahitian cores represent deeper-reef accretion when compared to those in
Barbados. New cores collected during the IODP-310
Tahiti Leg, should help determine if this offset is indeed
the result of deeper coral growth or a function of differing
glacio-isostatic-adjustment histories. One preliminary
result so far indicates that Mwp-1a at Tahiti has a similar
timing to that identified from the Sunda Shelf at 14.6 ka
(Deschamps et al., 2009).
In addition to the deglacial events, evidence for
a meltwater pulse at 8 kyrs (Mwp-1c) has also been further
supported by additional discoveries of drowned reefs from
Grand Cayman, southeast Florida and the Gulf of
Carpentaria (Blanchon et al., 2002, Banks et al., 2007,
Harris et al., 2008). On the east coast of Grand Cayman,
a transect of short cores along the reef-front in 20 m of water
recovered a submerged A. palmata reef-crest whose surface
corals returned
230
Th ages of between 8.9 kyrs and 8.1 kyrs
BP (Blanchon et al., 2002). The depth of this drowned reef
also corresponded with an intertidal notch on the western
shelf at a depth of 18 m and indicated the establishment of
an 18-m shoreline between 8.1 cal and 7.6 cal ka. Comparison of the age/depth data between Grand Cayman and
other Caribbean islands, indicated a near synchronous
demise and back-stepping of A. palmata reefs across the
Caribbean around 8 kyrs ago in response to a rapid 6-m
jump in sea level (Blanchon et al., 2002). But two potential
problems with this jump hypothesis have been raised. The
first problem was that back-stepping could only be demonstrated by comparing the age of reef drowning in one location
with the age of up-slope reef initiation in another. In other
words, there were few sites where clear evidence of both
drowning and back-stepping had been reported together.
That changed recently with the reporting of new age/depth
data from an up-slope reef off the southeast Florida shelf
(Banks et al., 2007). Combining the calibrated radiocarbonage data of this new reef with those from an earlier report
of an 8 ka old drowned reef further down-slope (Lighty et al.,
1978), confirmed that at this site Mwp-1c at 8 kyrs caused the
back-stepping of A. palmata reef crests 6–7 m up-slope in
less than 580 years (Blanchon, this volume).
MELTWATER PULSES
685