explanation that
230
Th ages are difficult to replicate on
sub-millennial timescales, as shown by the fact that
50% of ages with pristine isotopic values can have discordant
231 Pa ages (Scholz and Mangini, 2007). In addition,
disorderly growth is clearly untenable in conformable reef
sequences, given that multiple passes of the shore-face
would leave significant stratigraphic evidence of marine
erosion and/or subaerial exposure.
The problem of true-age variability in the ages of
Yokoyama et al., (2001) was partially addressed by
Thompson and Goldstein (2005) who recalculated them
to correct for bias introduced by open-system diagenesis.
Although this reduced the spread of ages from some reef
units, it did not significantly alter outlying ages, which
remained stratigraphically inconsistent and again plotted
in separate parts of the sea-level curve (Figure 2). This
failure to address fundamental stratigraphic consistency
of these age data and to continue making erroneous agedependant uplift corrections on every dated coral seriously
compromises the Glacial sea-level reconstruction of
Thompson and Goldstein (2005) and means that the sealevel peaks are merely dating artifacts.
Glacial Termination II (MIS-6)
Reports of meltwater pulses during the penultimate deglaciation (MIS-6) have also recently been made. On the
Huon Peninsula, Esat et al., (1999) reported that a cave
exposing the base of the last interglacial terrace (VII) some
90 m below its crest, contained in-situ corals with 1-cmthick annual bands, implying growth in a shallow-water
reef. These corals returned seven isotopically acceptable
230
Th ages of between 125.6 ka and 133.7 ka (i.e.,
a 7.2 ka true-age variation) which were averaged to give an
age of 130 Æ2 ka (a process which excluded a reliable age
of 115 ka). This was compared with a 134 Æ2 ka age average from four corals recovered from 5 to 12 m below the
reef-crest reported earlier by Stein et al., (1993), but four
significantly younger ages at 2 and 16 m were ignored.
By comparing average ages and elevations of these two
coral groups, Esat et al., (1999) claimed that sea-level fell
rapidly at least 70 m from the level of the 134 ka crest
group, and then rose rapidly at least 85 m from that of the
130 ka cave group, with both the fall and the rise occurring
within the 4 ka difference between the age groups. Unfortunately, the comparison between these two groups is questionable given that it ignores younger ages from both the
down-slope cave site and the up-slope barrier site reported
earlier by Stein et al., (1993). It is much more likely that
the stratigraphic age inversions at both sites reflect trueage variability (Scholz and Manginni, 2007) and that neither may be accurate (Blanchon, this volume). In the
absence of reliable ages, the stratigraphy dictates that the
cave is simply the oldest part of terrace VII which accreted
vertically as sea-level rose. As a result, claims of a rapid
70 m fall and an equally rapid 85 m rise at the Huon
Peninsula during the penultimate deglaciation are extreme
and poorly supported by the data.
Similar claims of extreme sea-level variability have
also been made from IODP-310 cores recovered from
the fore-reef slopes of Tahiti (Thomas et al., 2009). In
these cores, two corals recovered from 115 to 118 m below
present sea level returned reliable
230
Th ages of 133 and
137 ka respectively. Both corals are typical of shallowwater reef environments but are separated by a 30 cm section of much deeper-reef framework. This led Thomas
et al., (2009) to claim that a rapid deepening event
drowned the older coral and produced a limited deep reef
before sea-level fell rapidly leading to a resumption of
shallow reef development. In other words, the return to
shallow reefal environments represented a reversal in the
rapidly rising sea level during the penultimate deglaciation. Unfortunately, such an interpretation ignores the
more plausible possibility that the shallow-water corals
were in fact clasts transported down-slope into deeper
water. Examination of the cores in their Figures S2 and 3
supports this shallow-clast interpretation, with the deeper
coral at 118 m consisting of a 5 cm diameter coral clast
enclosed by other basalt- and coral-clasts (which were
even identified as clasts in the log description). The other
coral, which sits directly on top of the deeper framework,
could also be a clast given its small size (20 cm); the fact
that it contains surfaces encrusted by coralline algae,
upward growth, and apparently horizontal cavity fills
could merely be coincidental, indicating it fortuitously
landed with an upright orientation. However, a clast interpretation is supported by the lack of coral microbialites
which are typical of deglacial reef-framework of Tahiti
(Heindel, 2008). Regardless of the interpretation chosen,
such extraordinary claims of extreme sea-level changes
are clearly not supported by adequate data. As such, the
existence of meltwater pulses during the penultimate
deglaciation, although likely, remains to be proven by
credible reef data.
Last Interglacial (MIS-5e)
A better supported claim of a more modest meltwater
pulse has recently been made from reef development on
the Yucatan Peninsula during the last Interglacial (MIS5e). There, Blanchon et al., (2009) identified an episode
of reef-crest demise at +3 m and back-stepping to +6 m.
In some parts of the reef, this back-stepping was accompanied by erosion, but in others the sequence between the
two reefs was continuous, implying that they grew consecutively on an ecological timescale with no intervening hiatus (Figure 4). In turn, this consecutive development
requires that reef-crest back-stepping occurred on an ecological timescale, implying that it was caused by a 2–3 m
sea-level jump (Blanchon, 2010). Although the dating
precision precluded testing the rise rate, a test of the stratigraphic relations was made by dating samples from both
reef-crests. Unfortunately, only the +6 m crest returned
strictly reliable
230 Th ages which fell between 119.5 ka
and 117 ka. However, these ages were younger than
well-dated +3 m reefs that had developed in other areas,
688
MELTWATER PULSES
230
Th ages are difficult to replicate on
sub-millennial timescales, as shown by the fact that
50% of ages with pristine isotopic values can have discordant
231 Pa ages (Scholz and Mangini, 2007). In addition,
disorderly growth is clearly untenable in conformable reef
sequences, given that multiple passes of the shore-face
would leave significant stratigraphic evidence of marine
erosion and/or subaerial exposure.
The problem of true-age variability in the ages of
Yokoyama et al., (2001) was partially addressed by
Thompson and Goldstein (2005) who recalculated them
to correct for bias introduced by open-system diagenesis.
Although this reduced the spread of ages from some reef
units, it did not significantly alter outlying ages, which
remained stratigraphically inconsistent and again plotted
in separate parts of the sea-level curve (Figure 2). This
failure to address fundamental stratigraphic consistency
of these age data and to continue making erroneous agedependant uplift corrections on every dated coral seriously
compromises the Glacial sea-level reconstruction of
Thompson and Goldstein (2005) and means that the sealevel peaks are merely dating artifacts.
Glacial Termination II (MIS-6)
Reports of meltwater pulses during the penultimate deglaciation (MIS-6) have also recently been made. On the
Huon Peninsula, Esat et al., (1999) reported that a cave
exposing the base of the last interglacial terrace (VII) some
90 m below its crest, contained in-situ corals with 1-cmthick annual bands, implying growth in a shallow-water
reef. These corals returned seven isotopically acceptable
230
Th ages of between 125.6 ka and 133.7 ka (i.e.,
a 7.2 ka true-age variation) which were averaged to give an
age of 130 Æ2 ka (a process which excluded a reliable age
of 115 ka). This was compared with a 134 Æ2 ka age average from four corals recovered from 5 to 12 m below the
reef-crest reported earlier by Stein et al., (1993), but four
significantly younger ages at 2 and 16 m were ignored.
By comparing average ages and elevations of these two
coral groups, Esat et al., (1999) claimed that sea-level fell
rapidly at least 70 m from the level of the 134 ka crest
group, and then rose rapidly at least 85 m from that of the
130 ka cave group, with both the fall and the rise occurring
within the 4 ka difference between the age groups. Unfortunately, the comparison between these two groups is questionable given that it ignores younger ages from both the
down-slope cave site and the up-slope barrier site reported
earlier by Stein et al., (1993). It is much more likely that
the stratigraphic age inversions at both sites reflect trueage variability (Scholz and Manginni, 2007) and that neither may be accurate (Blanchon, this volume). In the
absence of reliable ages, the stratigraphy dictates that the
cave is simply the oldest part of terrace VII which accreted
vertically as sea-level rose. As a result, claims of a rapid
70 m fall and an equally rapid 85 m rise at the Huon
Peninsula during the penultimate deglaciation are extreme
and poorly supported by the data.
Similar claims of extreme sea-level variability have
also been made from IODP-310 cores recovered from
the fore-reef slopes of Tahiti (Thomas et al., 2009). In
these cores, two corals recovered from 115 to 118 m below
present sea level returned reliable
230
Th ages of 133 and
137 ka respectively. Both corals are typical of shallowwater reef environments but are separated by a 30 cm section of much deeper-reef framework. This led Thomas
et al., (2009) to claim that a rapid deepening event
drowned the older coral and produced a limited deep reef
before sea-level fell rapidly leading to a resumption of
shallow reef development. In other words, the return to
shallow reefal environments represented a reversal in the
rapidly rising sea level during the penultimate deglaciation. Unfortunately, such an interpretation ignores the
more plausible possibility that the shallow-water corals
were in fact clasts transported down-slope into deeper
water. Examination of the cores in their Figures S2 and 3
supports this shallow-clast interpretation, with the deeper
coral at 118 m consisting of a 5 cm diameter coral clast
enclosed by other basalt- and coral-clasts (which were
even identified as clasts in the log description). The other
coral, which sits directly on top of the deeper framework,
could also be a clast given its small size (20 cm); the fact
that it contains surfaces encrusted by coralline algae,
upward growth, and apparently horizontal cavity fills
could merely be coincidental, indicating it fortuitously
landed with an upright orientation. However, a clast interpretation is supported by the lack of coral microbialites
which are typical of deglacial reef-framework of Tahiti
(Heindel, 2008). Regardless of the interpretation chosen,
such extraordinary claims of extreme sea-level changes
are clearly not supported by adequate data. As such, the
existence of meltwater pulses during the penultimate
deglaciation, although likely, remains to be proven by
credible reef data.
Last Interglacial (MIS-5e)
A better supported claim of a more modest meltwater
pulse has recently been made from reef development on
the Yucatan Peninsula during the last Interglacial (MIS5e). There, Blanchon et al., (2009) identified an episode
of reef-crest demise at +3 m and back-stepping to +6 m.
In some parts of the reef, this back-stepping was accompanied by erosion, but in others the sequence between the
two reefs was continuous, implying that they grew consecutively on an ecological timescale with no intervening hiatus (Figure 4). In turn, this consecutive development
requires that reef-crest back-stepping occurred on an ecological timescale, implying that it was caused by a 2–3 m
sea-level jump (Blanchon, 2010). Although the dating
precision precluded testing the rise rate, a test of the stratigraphic relations was made by dating samples from both
reef-crests. Unfortunately, only the +6 m crest returned
strictly reliable
230 Th ages which fell between 119.5 ka
and 117 ka. However, these ages were younger than
well-dated +3 m reefs that had developed in other areas,
688
MELTWATER PULSES