Recent refinement in the precision of U-series dating
has encouraged several attempts to date the early stages
of reef development exposed in these deep sections. Dates
from the 17 m thick reef-crest section from the southeast
coast reported by Blanchon and Eisenhauer (2001) were
affected by diagenesis, and even those with reliable isotopic signatures showed no stratigraphic consistency. In the
middle of the unit, a single, strictly reliable age of 129 ka
suggested a significant part of the reefs development
predated the insolation maximum. But absence of replicate ages from this sample means that true-age variation
and hence the samples reliability cannot be verified.
A further attempt was made by Gallup et al. (2002), who
dated samples from the deep west-coast road-cut using
both
230
Th and
231
Pa isotopes. Although these two chronometers have similar chemical affinities and the same
uranium parent isotope, they provide an independent
check on true-age variation of the sample (Edwards
et al., 1997). A single, isotopically pristine coral from
the base of the terrace gave concordant ages of 136 ka,
indicating that it grew during the penultimate deglaciation
at an equivalent depth of 18 m below present SL (Gallup
et al., 2002). If this age is accurate, it implies that SL had
completed 80% of its rise prior to the insolation maximum, indicating that other factors besides orbital forcing
may have been involved in this early deglaciation. Samples from a lower unit also gave concordant ages of
129 ka implying that the SL rise was interrupted by
a reversal before attaining its highstand. Subsequent work
by Speed and Cheng (2004), however, demonstrated that
corals in this lower unit were detrital and thus have no significance for SL. In addition, more precise ages from
speleothems indicate that glacial termination II started at
129 ka (Chen et al., 2009), casting doubt on claims of early
reef development and the
231
Pa method of age
verification.
Huon Peninsula, Papua New Guinea
The age and elevation of uplifted flights of reef terraces on
the Huon Peninsula have also been the subject of extensive study and used to corroborate the SL change and
ice-volume over the last glacial cycle (Chappell, 1974,
2002; Aharon and Chappell, 1986; Chappell and Shackleton, 1986; Stein et al., 1993; Esat et al., 1999). Although it
is clear that the age of successive terraces increase
upslope, and that they consist of shallow-water corals,
the fundamental basis of SL reconstructions from individual terraces is their comparison with the depth-zonation
and architecture of local modern and Holocene reefs.
Unfortunately, little is known about these reefs or their
development (Edinger et al., 2001). As summarized in
Figure 2, basic descriptions of local reefs have been
reported by Chappell (1974) and Nakamori et al. (1994)
who indicate that they consist of a fringing and barrier-reef
complex separated by a narrow sediment-dominated
lagoon. Although not present in the lagoon, a depthzonation based largely on the presence of coralline algal
crusts and coral morphology has also been reported from
the barrier crest down the seaward slope by Nakamori
et al. (1994) and Matsuda et al. (1994).
A quantitative analysis of coral assemblages on the
crests and slopes of the Pleistocene terraces found that,
although zonation of assemblages generally reflected the
modern zonation, it could only be reliably differentiated
using the presence or absence of thick crusts of coralline
algae (Pandolfi, 1996). In most sections, coral communities shared more than half of their common species, and
in places zonation was almost indistinguishable. The only
indicator was the greater occurrence of Fungia spp. and
Favid head corals in the lower-slope zones. Additional differences between the modern and fossil communities was
that fragile-branching corals were underrepresented in
slope sections and some exclusively upper-slope corals
on modern reefs, such as A. gemmifera, were also common in the deep slope assemblage of the fossil reefs,
indicating down-slope transport (cf. Van Woesik et al.,
1991).
For the LIG terrace in particular, measured sections at
several locations show that reefal zonation is similar to
the Holocene, consisting of a fringing reef, a lagoon, and
a barrier-reef complex (Chappell, 1974; Aharon and
Chappell, 1986; Pandolfi and Chappell, 1994). But these
sections also show that individual reef zones are much
thicker than in the modern-reef zonation, indicating that
the LIG reef complex likely responded to rising SL
(Figure 2). A similar response was identified from the
Holocene reef complex itself, where a 52 m overthickened reef unit was recovered from a bore-hole on
the barrier crest (Chappell and Polach, 1991). Dating of
corals in this core showed that reef development initiated
13 ka ago but ceased by 6 ka due to uplift (Chappell and
Polach, 1991; Edwards et al., 1993; Ota and Chappell,
1999). On the basis of these data, several workers have
proposed that reef development during the LIG and Holocene had a simple progradational architecture with the barrier crest keeping pace with SL rise and migrating over its
seaward slope (Chappell, 1974, 1980; Chappell and Veeh,
1978; Chappell and Polach, 1991; Ota and Chappell,
1999). Yet this keep-up model is not adequately supported
by the data, given that neither coral zonation nor facies
geometries have been fully established for either the LIG
or Holocene reef complexes. It could be, for example, that
reefs did not always maintain their position at SL, and may
have been outpaced by meltwater pulses.
In fact, investigation of the LIG terrace has found that
a simple keep-up reef architecture is difficult to reconcile
with the actual stratigraphy, which apparently shows significant complexity (Stein et al., 1993; Esat et al., 1999).
For example, between the Kwangam Gorge and the slopes
above the villages of Sialum and Kwambu, there are several well-studied sections through the fringing, lagoon,
and barrier-reef complex (Figure 2). These sections show
that the outer barrier (terrace VIIb) forms a narrow ridge
with local reef-crest remnants (Stein et al., 1993). Its seaward face exposes coral and Halimeda gravels with
LAST INTERGLACIAL AND REEF DEVELOPMENT
625
has encouraged several attempts to date the early stages
of reef development exposed in these deep sections. Dates
from the 17 m thick reef-crest section from the southeast
coast reported by Blanchon and Eisenhauer (2001) were
affected by diagenesis, and even those with reliable isotopic signatures showed no stratigraphic consistency. In the
middle of the unit, a single, strictly reliable age of 129 ka
suggested a significant part of the reefs development
predated the insolation maximum. But absence of replicate ages from this sample means that true-age variation
and hence the samples reliability cannot be verified.
A further attempt was made by Gallup et al. (2002), who
dated samples from the deep west-coast road-cut using
both
230
Th and
231
Pa isotopes. Although these two chronometers have similar chemical affinities and the same
uranium parent isotope, they provide an independent
check on true-age variation of the sample (Edwards
et al., 1997). A single, isotopically pristine coral from
the base of the terrace gave concordant ages of 136 ka,
indicating that it grew during the penultimate deglaciation
at an equivalent depth of 18 m below present SL (Gallup
et al., 2002). If this age is accurate, it implies that SL had
completed 80% of its rise prior to the insolation maximum, indicating that other factors besides orbital forcing
may have been involved in this early deglaciation. Samples from a lower unit also gave concordant ages of
129 ka implying that the SL rise was interrupted by
a reversal before attaining its highstand. Subsequent work
by Speed and Cheng (2004), however, demonstrated that
corals in this lower unit were detrital and thus have no significance for SL. In addition, more precise ages from
speleothems indicate that glacial termination II started at
129 ka (Chen et al., 2009), casting doubt on claims of early
reef development and the
231
Pa method of age
verification.
Huon Peninsula, Papua New Guinea
The age and elevation of uplifted flights of reef terraces on
the Huon Peninsula have also been the subject of extensive study and used to corroborate the SL change and
ice-volume over the last glacial cycle (Chappell, 1974,
2002; Aharon and Chappell, 1986; Chappell and Shackleton, 1986; Stein et al., 1993; Esat et al., 1999). Although it
is clear that the age of successive terraces increase
upslope, and that they consist of shallow-water corals,
the fundamental basis of SL reconstructions from individual terraces is their comparison with the depth-zonation
and architecture of local modern and Holocene reefs.
Unfortunately, little is known about these reefs or their
development (Edinger et al., 2001). As summarized in
Figure 2, basic descriptions of local reefs have been
reported by Chappell (1974) and Nakamori et al. (1994)
who indicate that they consist of a fringing and barrier-reef
complex separated by a narrow sediment-dominated
lagoon. Although not present in the lagoon, a depthzonation based largely on the presence of coralline algal
crusts and coral morphology has also been reported from
the barrier crest down the seaward slope by Nakamori
et al. (1994) and Matsuda et al. (1994).
A quantitative analysis of coral assemblages on the
crests and slopes of the Pleistocene terraces found that,
although zonation of assemblages generally reflected the
modern zonation, it could only be reliably differentiated
using the presence or absence of thick crusts of coralline
algae (Pandolfi, 1996). In most sections, coral communities shared more than half of their common species, and
in places zonation was almost indistinguishable. The only
indicator was the greater occurrence of Fungia spp. and
Favid head corals in the lower-slope zones. Additional differences between the modern and fossil communities was
that fragile-branching corals were underrepresented in
slope sections and some exclusively upper-slope corals
on modern reefs, such as A. gemmifera, were also common in the deep slope assemblage of the fossil reefs,
indicating down-slope transport (cf. Van Woesik et al.,
1991).
For the LIG terrace in particular, measured sections at
several locations show that reefal zonation is similar to
the Holocene, consisting of a fringing reef, a lagoon, and
a barrier-reef complex (Chappell, 1974; Aharon and
Chappell, 1986; Pandolfi and Chappell, 1994). But these
sections also show that individual reef zones are much
thicker than in the modern-reef zonation, indicating that
the LIG reef complex likely responded to rising SL
(Figure 2). A similar response was identified from the
Holocene reef complex itself, where a 52 m overthickened reef unit was recovered from a bore-hole on
the barrier crest (Chappell and Polach, 1991). Dating of
corals in this core showed that reef development initiated
13 ka ago but ceased by 6 ka due to uplift (Chappell and
Polach, 1991; Edwards et al., 1993; Ota and Chappell,
1999). On the basis of these data, several workers have
proposed that reef development during the LIG and Holocene had a simple progradational architecture with the barrier crest keeping pace with SL rise and migrating over its
seaward slope (Chappell, 1974, 1980; Chappell and Veeh,
1978; Chappell and Polach, 1991; Ota and Chappell,
1999). Yet this keep-up model is not adequately supported
by the data, given that neither coral zonation nor facies
geometries have been fully established for either the LIG
or Holocene reef complexes. It could be, for example, that
reefs did not always maintain their position at SL, and may
have been outpaced by meltwater pulses.
In fact, investigation of the LIG terrace has found that
a simple keep-up reef architecture is difficult to reconcile
with the actual stratigraphy, which apparently shows significant complexity (Stein et al., 1993; Esat et al., 1999).
For example, between the Kwangam Gorge and the slopes
above the villages of Sialum and Kwambu, there are several well-studied sections through the fringing, lagoon,
and barrier-reef complex (Figure 2). These sections show
that the outer barrier (terrace VIIb) forms a narrow ridge
with local reef-crest remnants (Stein et al., 1993). Its seaward face exposes coral and Halimeda gravels with
LAST INTERGLACIAL AND REEF DEVELOPMENT
625