The second potential problem with the 8 ka jump was
the discovery of a deposit of A. palmata off Sand Key,
Florida, that apparently filled the elevation gap between
drowned and back-stepped reefs (Toscano and Lundberg
1998; Figure 1). This deposit developed between 12.5 m
and 9 m below present sea level and contained 9–7 ka
old corals. In other words, it developed at the same time
as the 8 ka old drowned reef reported 4–7 m further
down-slope by Lighty et al., (1978). Given that Banks
et al., (2007) confirmed that the down-slope reef morphology is directly analogous to that of modern Florida
breakwater reefs, this means that Sand Key deposit further up-slope cannot also be a breakwater reef. This
objection is supported by the fact that the deposit not
only contains age reversals, but an 86 ka old extraformational clast reworked from the underlying foundation. The mismatch in elevation with coeval down-slope
reefs, together with the evidence of reworking, clearly
indicates that the Sand Key deposit is not of reefal origin
but is a coastal boulder-rampart deposited by storms and
hurricanes. As noted by Blanchon and Perry (2004), such
problems highlight the importance of identifying in-place
A. palmata reef framework using textural and taphonomic
criteria in order to differentiate it from either boulder
ramparts deposited up to 5 m above present sea-level, or
submerged accumulations which can be deposited in much
deeper water.
Until recently, there was scant support for reef drowning related to the 8 kyr Mwp-1c outside the Caribbean
(e.g., Bard et al., 1996). But in 2008, Harris et al., reported
early Holocene reef development in the Gulf of
Carpentaria, northern Australia. Multi-beam sonar and
short drill cores demonstrated widespread reef development between 23 m and 33 m water depth starting at
10.5 kys and ceasing by 8 kyrs. Unfortunately, the coral
sequences were poorly developed and it is unknown if
these reefs grew at sea level or were submerged below it.
Their surface elevations however are remarkably consistent over wide areas indicating they were likely breakwater reefs. Evidence for the 8 ka Mwp-1c has also been
reported from other sedimentary systems, notably deltas
and estuaries (Stanley and Warne 1994; Bratton et al.,
2003; Hori and Saito 2007) and isolation basins in
glacio-isostatic rebound terranes (Yu et al., 2007).
Last glaciation (MIS-3)
Reports of meltwater pulses prior to the last deglaciation
are more controversial. For example, there have been multiple claims of melt-water pulses during the last Glacial
between 19 ka and 65 ka. The first of these is a proposed
melt-water pulse at 19 ka which is based on a core transect
across the Bonaparte Gulf, in northern Australia between
depths of 147 m and 34 m (Yokoyama et al., 2000). These
cores recovered shelf and marginal-marine to brackishwater facies based on the foraminiferal and bivalve assemblages, which were used to radiocarbon date the sequence.
Using age/depth data of the transition from the low-stand
brackish-water facies into the succeeding marginal-marine
facies, it was argued that meter-scale elevation differences
between cores was the result of a melt-water pulse centered at 19 cal ka BP. The validity of this sea-level reconstruction, however, has been contested by Shennan and
Milne (2003), who highlighted the presence of hiatuses
reported in a later study (Yokoyama et al., 2001), inconsistencies between cores, and doubts concerning the depth
range of the facies units. As a consequence, the interpretation of a meltwater pulse at 19 kyr is suspect.
The second report of a 19 ka meltwater pulse is from the
north-east Irish coast where a presumed fluvial incision
was filled by transgressive marine foraminiferal mudstones along a coastal area undergoing active glacioisostatic uplift (Clark et al., 2004). The uniform age of
these deposits and the apparent transgression over
a fluvially-incised subaerial surface were interpreted to
result from a rapid 10 m rise in sea level at 19 cal ka BP.
Unfortunately, alternative interpretations of the origin of
the incised valley, such as marine ravinement, cannot be
ruled out. In addition, the correction of the ages for marine
reservoir has been disputed by Hanebuth et al., (2009).
Such problems once again render the interpretation of
a meltwater pulse at 19 ka as ambiguous.
Further reports of fully Glacial (MIS-3) stage meltwater
pulses have come from the Huon Peninsula. There, the
ages of uplifted reef terraces between 30 ka and 65 ka
apparently matched rapid climate changes and ice-sheet
events seen in Greenland ice and north Atlantic sediment
cores (Chappell, 2002). A sea-level reconstruction from
these terraces used the thickness and geomorphology of
reef-crest and intertidal units to model the relative rate
and magnitude of sea-level rise by making assumptions
about reef type and accretion rate. Then, these relative
rises were used to reconstruct absolute sea-level position
using
230 Th ages of the reefs and applying uplift corrections (Figure 2). This preliminary reconstruction showed
5 rapid sea-level-rise events of between 10 m and 20 m
within duration of 1–2 ka; the timing of some of these rises
coincided with rapid-ice loss events from the Laurentide
Ice Sheet, known as Heinrich events (Hemming 2004).
Further dating of the Glacial reef terraces on the Huon
Peninsula by Yokoyama et al., (2001) attempted to provide an independent sea-level reconstruction and model
glacio-isostatic adjustment (Figure 2). Unfortunately, this
new reconstruction conflicted in terms of timing, number,
and magnitude of rapid sea-level rise events with the
reconstruction of Chappell (2002). This is largely because
Yokoyama et al., (2001) ignored the problem of true-age
variability (Scholtz and Manginni, 2007) and considered
all isotopically-reliable ages to be valid sea-level indicators, despite their lack of stratigraphic consistency. These
errors were compounded by making an age-dependant
uplift correction for every dated coral, rather than making
this correction for a core-group of ages that best
represented the time-span of individual reef-crest development. Together these errors led to single outlying ages
forming the peaks of rapid sea-level-rise events, and
686
MELTWATER PULSES
the discovery of a deposit of A. palmata off Sand Key,
Florida, that apparently filled the elevation gap between
drowned and back-stepped reefs (Toscano and Lundberg
1998; Figure 1). This deposit developed between 12.5 m
and 9 m below present sea level and contained 9–7 ka
old corals. In other words, it developed at the same time
as the 8 ka old drowned reef reported 4–7 m further
down-slope by Lighty et al., (1978). Given that Banks
et al., (2007) confirmed that the down-slope reef morphology is directly analogous to that of modern Florida
breakwater reefs, this means that Sand Key deposit further up-slope cannot also be a breakwater reef. This
objection is supported by the fact that the deposit not
only contains age reversals, but an 86 ka old extraformational clast reworked from the underlying foundation. The mismatch in elevation with coeval down-slope
reefs, together with the evidence of reworking, clearly
indicates that the Sand Key deposit is not of reefal origin
but is a coastal boulder-rampart deposited by storms and
hurricanes. As noted by Blanchon and Perry (2004), such
problems highlight the importance of identifying in-place
A. palmata reef framework using textural and taphonomic
criteria in order to differentiate it from either boulder
ramparts deposited up to 5 m above present sea-level, or
submerged accumulations which can be deposited in much
deeper water.
Until recently, there was scant support for reef drowning related to the 8 kyr Mwp-1c outside the Caribbean
(e.g., Bard et al., 1996). But in 2008, Harris et al., reported
early Holocene reef development in the Gulf of
Carpentaria, northern Australia. Multi-beam sonar and
short drill cores demonstrated widespread reef development between 23 m and 33 m water depth starting at
10.5 kys and ceasing by 8 kyrs. Unfortunately, the coral
sequences were poorly developed and it is unknown if
these reefs grew at sea level or were submerged below it.
Their surface elevations however are remarkably consistent over wide areas indicating they were likely breakwater reefs. Evidence for the 8 ka Mwp-1c has also been
reported from other sedimentary systems, notably deltas
and estuaries (Stanley and Warne 1994; Bratton et al.,
2003; Hori and Saito 2007) and isolation basins in
glacio-isostatic rebound terranes (Yu et al., 2007).
Last glaciation (MIS-3)
Reports of meltwater pulses prior to the last deglaciation
are more controversial. For example, there have been multiple claims of melt-water pulses during the last Glacial
between 19 ka and 65 ka. The first of these is a proposed
melt-water pulse at 19 ka which is based on a core transect
across the Bonaparte Gulf, in northern Australia between
depths of 147 m and 34 m (Yokoyama et al., 2000). These
cores recovered shelf and marginal-marine to brackishwater facies based on the foraminiferal and bivalve assemblages, which were used to radiocarbon date the sequence.
Using age/depth data of the transition from the low-stand
brackish-water facies into the succeeding marginal-marine
facies, it was argued that meter-scale elevation differences
between cores was the result of a melt-water pulse centered at 19 cal ka BP. The validity of this sea-level reconstruction, however, has been contested by Shennan and
Milne (2003), who highlighted the presence of hiatuses
reported in a later study (Yokoyama et al., 2001), inconsistencies between cores, and doubts concerning the depth
range of the facies units. As a consequence, the interpretation of a meltwater pulse at 19 kyr is suspect.
The second report of a 19 ka meltwater pulse is from the
north-east Irish coast where a presumed fluvial incision
was filled by transgressive marine foraminiferal mudstones along a coastal area undergoing active glacioisostatic uplift (Clark et al., 2004). The uniform age of
these deposits and the apparent transgression over
a fluvially-incised subaerial surface were interpreted to
result from a rapid 10 m rise in sea level at 19 cal ka BP.
Unfortunately, alternative interpretations of the origin of
the incised valley, such as marine ravinement, cannot be
ruled out. In addition, the correction of the ages for marine
reservoir has been disputed by Hanebuth et al., (2009).
Such problems once again render the interpretation of
a meltwater pulse at 19 ka as ambiguous.
Further reports of fully Glacial (MIS-3) stage meltwater
pulses have come from the Huon Peninsula. There, the
ages of uplifted reef terraces between 30 ka and 65 ka
apparently matched rapid climate changes and ice-sheet
events seen in Greenland ice and north Atlantic sediment
cores (Chappell, 2002). A sea-level reconstruction from
these terraces used the thickness and geomorphology of
reef-crest and intertidal units to model the relative rate
and magnitude of sea-level rise by making assumptions
about reef type and accretion rate. Then, these relative
rises were used to reconstruct absolute sea-level position
using
230 Th ages of the reefs and applying uplift corrections (Figure 2). This preliminary reconstruction showed
5 rapid sea-level-rise events of between 10 m and 20 m
within duration of 1–2 ka; the timing of some of these rises
coincided with rapid-ice loss events from the Laurentide
Ice Sheet, known as Heinrich events (Hemming 2004).
Further dating of the Glacial reef terraces on the Huon
Peninsula by Yokoyama et al., (2001) attempted to provide an independent sea-level reconstruction and model
glacio-isostatic adjustment (Figure 2). Unfortunately, this
new reconstruction conflicted in terms of timing, number,
and magnitude of rapid sea-level rise events with the
reconstruction of Chappell (2002). This is largely because
Yokoyama et al., (2001) ignored the problem of true-age
variability (Scholtz and Manginni, 2007) and considered
all isotopically-reliable ages to be valid sea-level indicators, despite their lack of stratigraphic consistency. These
errors were compounded by making an age-dependant
uplift correction for every dated coral, rather than making
this correction for a core-group of ages that best
represented the time-span of individual reef-crest development. Together these errors led to single outlying ages
forming the peaks of rapid sea-level-rise events, and
686
MELTWATER PULSES