Page 22 - 23

ARTICLE IN PRESS
1588 K. Lambeck et al. / Quaternary Science Reviews 23 (2004) 1567–1598
where the first two terms on the right-hand side
represent the eustatic and isostatic model-dependent
contributions, respectively. The third term dze(t) repre-
sents a corrective term to the nominal eustatic term
Dze(t) used to compute the isostatic term DzI(j, t) and
the last term is a tectonic contribution. Eq. (4) may be
reorganized as
DzeðtÞ þ dzeðtÞ ¼ DzÃe ðtÞ
¼ Dzobsðj; tÞ À DzIðj; tÞ À DzTðj; tÞ
¼ Dzobsðj; tÞ À fDzpredðj; tÞ À DzeðtÞg À DzTðj; tÞ interval B10.3 to B9.4 ka BP that lie above the lowest
ð5Þ above the revised predicted levels. This may be a
with
sDzÃe ¼ ðs2t:c:obs þ s2dzIÞ; ð6Þ
where DzeÃðtÞ is the observationally derived estimate of
the eustatic sea level and Dzpred(j, t) is the sea-level
prediction based on the nominal eustatic sea-level
function. The variance of the tectonically corrected
observed sea level is given by (3) and s2szI is the variance
of the isostatic correction (2). The function DzeÃðtÞ is
compared with the nominal function (dashed line) in
Fig. 6a. A notable Holocene discrepancy between the
two estimates occurs in the interval from B7 to B9.5 ka
where a number of marine-based estimates lie system-
atically above the nominal function.
The property of the eustatic sea-level function is that
it is independent of location and the results for the
Italian coastline can be directly compared with the
results from other localities. The most complete records
of Holocene sea-level change remain those from
Barbados (Fairbanks, 1989; Bard et al., 1993), Huon
(Chappell and Polach, 1991) and Tahiti (Bard et al.,
1996) to which have been added data from
Christchurch, New Zealand (Gibb, 1986; Lambeck
et al., 2002). This ‘global’ result, corrected for iso-
static and tectonic effects, is illustrated in Fig. 6b where
the eustatic sea-level function is defined as the upper
limit to the isostatically corrected coral indicators from
these three localities. These estimates also lie a few
meters above the nominal function in the mid-late
Holocene interval. Thus we adopt as best estimate for
the interval 10.5–7 ka BP the weighted mean of the
Italian and global estimates, where the former is
represented by the smoothed function illustrated in
Fig. 6a.
Modification of the eustatic sea-level function has the
consequence of also modifying the isostatic corrections
but, as the corrections to the former are relatively small
and the isostatic corrections for the Italian sites
represent about 10–15% of the total sea-level signal,
these further corrections are second-order effects. As to
which ice sheet needs to be modified to accommodate
the change in the global ice volume cannot be ans-
wered from sea-level observations far from the former
ice margins. Thus, we have made the adjustment
by partitioning the change between the distant ice
sheets of North America and Antarctica. With the
so-modified ice models the predicted sea levels at the
Italian sites have been revised and compared with the
tectonically corrected observations as before (Figs. 5b
and 6c).
Despite these modifications to the isostatic model the
previously identified isolated anomalous points persist,
such as the marine and transitional data points in the
continental points within the same time interval and well
consequence of misidentification of the depositional
environments—that the marsh deposits were terrestrial
deposits rather than salt marsh, for example—but
unfortunately many of the older cores are not available
for re-examination.
6.1.2. Tectonic vertical movement estimates
A second group of anomalous points occurs in the
interval 0–5 ka where a number of marine data points lie
higher than coeval terrestrial data points. These
correspond to the tectonically uplifted sites of Taormina
(15), Scilla (14) and Capo Rizzuto (22). In all three cases
the predicted rates of uplift are based on the elevation of
the MIS 5.5 shorelines and on the assumption of
uniform rates of uplift but the results here suggest that
recent uplift rates may have been higher than the long-
term averages. The difference between the tectonically
corrected data—based on the long-term averages—and
the predicted values are illustrated in Figs. 7a and b for
Taormina and Scilla. For both sites the differences for
the individual data points exhibit systematic trends that
indicate that uplift rates for the past 6000 years
have been about 100% greater than the long-term
averages. This will be investigated separately and here
we use the revised rates estimated from Fig. 7 of,
respectively 1.7270.22 and 1.9770.33 mm yrÀ1. The
Capo Rizzuto evidence suggests a recent uplift rate of
1.1570.25 mm yrÀ1 that may have been about 50%
greater than predicted from the MIS 5.5 data although
this is based on a single data point.
Estimates for the vertical tectonic rate at Volturno
(10) are ambiguous as discussed above, with estimates of
the elevation of the MIS 5.5 shoreline fluctuating
between 750 m (Romano et al., 1994), depending on
whether the location is within the coastal plain or on the
adjacent ridges. The comparison of observed and
predicted values, with the former based on the assump-
tion of an absence of vertical tectonics, suggests in fact
that the area has been subject to some uplift over the
past 8000 years (Fig. 7c). The observations here are from
peat and marsh samples and are assumed to mark the
upper limit of sea level. From this evidence the estimated
rate of uplift is 0.3570.29 mm yrÀ1, consistent with the