Later coring on the Tahitian reef crest, however,
showed systematic changes in coral assemblages which
corresponded to the timing of Mwp-1b (Cabioch et al.,
1999). Following Mwp-1a, the base of the sequence
started in lower-reef-front slope assemblages (>10 m)
but shallowed into upper-slope assemblages (<10 m)
shortly before Mwp-1b. Following that rise event, assemblages reverted to lower slope as waters deepened, but
quickly shallowed thereafter, and showed little or no subsequent changes related to Mwp-1c (Figure 2). In other
words, the reef crest recovered following the first two
meltwater pulses but did not register the last.
Despite the lack of clear facies changes registered at
Tahiti during the 8 ka meltwater pulse, further evidence
of reef demise and back-stepping has since been reported
from several other areas around this time. Off the north
coast of St. Croix, Hubbard et al. (2005) recovered core
sequences from the interval of modern reef initiation,
and found that A. palmata reef-crest assemblages had
developed at 12 m by 7.7 ka (Figure 1). The comparison
of this initiation age, with the terminal age of an early
Holocene reef crest off the southwest coast (Adey et al.,
1978), shows that reef back-stepping started at $8 ka at
a depth of $21 m and was complete by 7.7 ka at 12 m.
Although the precise age and depth of the early Holocene
reef crest are uncertain due to the lack of core coverage,
heavy encrustation of the cored A. palmata by coralline
algae indicates that the reef-crest position could be no
shallower than 18 m (cf. Steneck and Adey 1976). In other
words, these data indicate St. Croix’s reef-crest backstepped 4–7 m in 300 years or less.
Similar evidence of rapid back-stepping has also
recently been confirmed from southeast Florida, where
Banks et al., (2007) reported coral ages from a single core
and submarine grounding site (USS Memphis in 1993).
These data indicate that a reef-crest facies, consisting of
in-place A. palmata, initiated on the inner shelf 7.4 ka
ago at 10 m below sea level (see inner reef–tract in
Figure 3). Comparing the initiation age of this shallow reef
with the demise of a deeper, early Holocene reef crest
exposed in a sewage-outfall trench a further 3 km offshore
(Lighty et al., 1978), shows that reef back-stepping had
started by $8 ka at a depth of $17 m and was complete
by 7.4 ka at 10 m (see outer reef–tract in Figure 3). These
data therefore indicate that southeast Florida reef crests
also back-stepped 4–9 m in $450 years.
In addition to the Holocene and deglacial events, backstepping has also been documented during the last interglacial highstand (MIS-5e) when sea level was as much as
6 m higher than present. Along the Red Sea coast of Eritrea,
for example, an uplifted and tilted LIG reef sequence near
Abdur clearly shows two superimposed stages of shallowreef development (Bruggemann et al., 2004). The lagoon
and patch-reef section of the lower reef unit is truncated
by an intermittent marine-erosion surface and directly overlain by a 3 m crest and reef-front section of the upper-reef
unit. This implies that reef-crest development back-stepped
over an existing reef lagoon. But given the neotectonic
setting of this site, the possibility that co-seismic uplift produced reef back-stepping cannot be discounted.
A clearer example of back-stepping has recently been
described from the northeast Yucatan by Blanchon et al.,
(2009) and Blanchon (2010). Two superimposed fossil
reef units were documented. A lower patch-reef complex
and adjacent crest unit at þ3 m is overlain by a second reef
unit with a crest at þ6 m. Reliable radiometric ages confirmed both units were of last interglacial age but could
not differentiate between them, mainly due to subtle diagenetic alteration of corals in the lower unit. However,
the relative-age relations between the two units were clear.
The framework of the upper-reef unit was infiltrated by
shelly beach-gravel as sea level fell at end of the interglacial, but infiltration of lower reef was prevented by cap of
crustose coralline algae. This infiltration pattern shows
that the upper reef was younger and must have been alive
shortly before sea level fell, and that the lower-reef was
older and was dead when sea level fell. Areas of continuous accretion between the lagoonal patch reefs of the
lower unit and reef crest and back-reef of the upper unit,
however, require that the demise of the lower-reef was
ecologically synchronous with initiation of the upper-reef
tract. In other words, back-stepping took place on an ecological timescale. The relative differences between the elevations of the reef crests and flats in the two reef units, and
the presence of 1.5-m tall colonies at the base of the uppercrest unit, indicate that this back-stepping was a result of
a 2–3 m sea-level jump at the end of the last interglacial
(Blanchon et al., 2009).
Reef back-stepping in subsiding terranes
As suggested by Schlager (1981), the combination of subsidence and pulsed glacio-eustatic sea-level rise provides
an ideal mechanism to trigger reef back-stepping.
A good example has been described from the Huon Gulf,
Papua New Guinea, where rapid and oblique convergence
of the Australian and West Pacific plates, and intervening
microplates, has produced a foreland basin with high rates
of vertical displacement across its collisional axis (e.g.,
12 mm/year; Abers and McCaffrey 1994). Flexure of the
basin’s cratonic margin has caused submergence and
onlap during the last $450 kyr and produced as many as
14 back-stepping platforms, with 7 confirmed as reefal
in origin but with only 1 returning a reliable radiometric
age (Galewsky et al., 1996; Webster et al., 2004a, b). Coral
fragments from the base of platform PII recorded an age of
$60 ka, but there was no direct evidence for it being reefal
in origin (Webster et al., 2004a). A single age from the
talus of platform PXII, that is clearly reefal in origin, indicates that it was likely drowned by a major deglacial sealevel rise associated with the transition from MIS-10 to 9
(Galewsky et al., 1996; Webster et al., 2004b). However,
the number of confirmed reefal platforms clearly exceeds
the number of major deglaciations, and indicates that
interstadial sea-level-rise events may also be responsible
for platform drowning (Webster et al., 2009).
80
BACK-STEPPING
showed systematic changes in coral assemblages which
corresponded to the timing of Mwp-1b (Cabioch et al.,
1999). Following Mwp-1a, the base of the sequence
started in lower-reef-front slope assemblages (>10 m)
but shallowed into upper-slope assemblages (<10 m)
shortly before Mwp-1b. Following that rise event, assemblages reverted to lower slope as waters deepened, but
quickly shallowed thereafter, and showed little or no subsequent changes related to Mwp-1c (Figure 2). In other
words, the reef crest recovered following the first two
meltwater pulses but did not register the last.
Despite the lack of clear facies changes registered at
Tahiti during the 8 ka meltwater pulse, further evidence
of reef demise and back-stepping has since been reported
from several other areas around this time. Off the north
coast of St. Croix, Hubbard et al. (2005) recovered core
sequences from the interval of modern reef initiation,
and found that A. palmata reef-crest assemblages had
developed at 12 m by 7.7 ka (Figure 1). The comparison
of this initiation age, with the terminal age of an early
Holocene reef crest off the southwest coast (Adey et al.,
1978), shows that reef back-stepping started at $8 ka at
a depth of $21 m and was complete by 7.7 ka at 12 m.
Although the precise age and depth of the early Holocene
reef crest are uncertain due to the lack of core coverage,
heavy encrustation of the cored A. palmata by coralline
algae indicates that the reef-crest position could be no
shallower than 18 m (cf. Steneck and Adey 1976). In other
words, these data indicate St. Croix’s reef-crest backstepped 4–7 m in 300 years or less.
Similar evidence of rapid back-stepping has also
recently been confirmed from southeast Florida, where
Banks et al., (2007) reported coral ages from a single core
and submarine grounding site (USS Memphis in 1993).
These data indicate that a reef-crest facies, consisting of
in-place A. palmata, initiated on the inner shelf 7.4 ka
ago at 10 m below sea level (see inner reef–tract in
Figure 3). Comparing the initiation age of this shallow reef
with the demise of a deeper, early Holocene reef crest
exposed in a sewage-outfall trench a further 3 km offshore
(Lighty et al., 1978), shows that reef back-stepping had
started by $8 ka at a depth of $17 m and was complete
by 7.4 ka at 10 m (see outer reef–tract in Figure 3). These
data therefore indicate that southeast Florida reef crests
also back-stepped 4–9 m in $450 years.
In addition to the Holocene and deglacial events, backstepping has also been documented during the last interglacial highstand (MIS-5e) when sea level was as much as
6 m higher than present. Along the Red Sea coast of Eritrea,
for example, an uplifted and tilted LIG reef sequence near
Abdur clearly shows two superimposed stages of shallowreef development (Bruggemann et al., 2004). The lagoon
and patch-reef section of the lower reef unit is truncated
by an intermittent marine-erosion surface and directly overlain by a 3 m crest and reef-front section of the upper-reef
unit. This implies that reef-crest development back-stepped
over an existing reef lagoon. But given the neotectonic
setting of this site, the possibility that co-seismic uplift produced reef back-stepping cannot be discounted.
A clearer example of back-stepping has recently been
described from the northeast Yucatan by Blanchon et al.,
(2009) and Blanchon (2010). Two superimposed fossil
reef units were documented. A lower patch-reef complex
and adjacent crest unit at þ3 m is overlain by a second reef
unit with a crest at þ6 m. Reliable radiometric ages confirmed both units were of last interglacial age but could
not differentiate between them, mainly due to subtle diagenetic alteration of corals in the lower unit. However,
the relative-age relations between the two units were clear.
The framework of the upper-reef unit was infiltrated by
shelly beach-gravel as sea level fell at end of the interglacial, but infiltration of lower reef was prevented by cap of
crustose coralline algae. This infiltration pattern shows
that the upper reef was younger and must have been alive
shortly before sea level fell, and that the lower-reef was
older and was dead when sea level fell. Areas of continuous accretion between the lagoonal patch reefs of the
lower unit and reef crest and back-reef of the upper unit,
however, require that the demise of the lower-reef was
ecologically synchronous with initiation of the upper-reef
tract. In other words, back-stepping took place on an ecological timescale. The relative differences between the elevations of the reef crests and flats in the two reef units, and
the presence of 1.5-m tall colonies at the base of the uppercrest unit, indicate that this back-stepping was a result of
a 2–3 m sea-level jump at the end of the last interglacial
(Blanchon et al., 2009).
Reef back-stepping in subsiding terranes
As suggested by Schlager (1981), the combination of subsidence and pulsed glacio-eustatic sea-level rise provides
an ideal mechanism to trigger reef back-stepping.
A good example has been described from the Huon Gulf,
Papua New Guinea, where rapid and oblique convergence
of the Australian and West Pacific plates, and intervening
microplates, has produced a foreland basin with high rates
of vertical displacement across its collisional axis (e.g.,
12 mm/year; Abers and McCaffrey 1994). Flexure of the
basin’s cratonic margin has caused submergence and
onlap during the last $450 kyr and produced as many as
14 back-stepping platforms, with 7 confirmed as reefal
in origin but with only 1 returning a reliable radiometric
age (Galewsky et al., 1996; Webster et al., 2004a, b). Coral
fragments from the base of platform PII recorded an age of
$60 ka, but there was no direct evidence for it being reefal
in origin (Webster et al., 2004a). A single age from the
talus of platform PXII, that is clearly reefal in origin, indicates that it was likely drowned by a major deglacial sealevel rise associated with the transition from MIS-10 to 9
(Galewsky et al., 1996; Webster et al., 2004b). However,
the number of confirmed reefal platforms clearly exceeds
the number of major deglaciations, and indicates that
interstadial sea-level-rise events may also be responsible
for platform drowning (Webster et al., 2009).
80
BACK-STEPPING