Page 15 - Sea-level change_2004

ARTICLE IN PRESS
K. Lambeck, A. Purcell / Quaternary Science Reviews 24 (2005) 1969–1988 1983

change wouldbecome more gradual if the ocean volume considered. An increase in lower-mantle viscosity to
22
continuedto increase slowly. The absence of such an values above 2 10 Pa s has the effect of lowering sea
abrupt change in any observational recordwould, level, but the dependence becomes small for this location
therefore, be indicative of ongoing ocean volume change as is illustratedin Fig. 10b. Likewise, the lithospheric
at this time. Significant is that for the range of models dependence is not large here (Fig. 10c) andhigher values
the Late Holocene predictions are consistently above for the effective thickness have only a small effect on the
present andhigher than the observed values (Fig. 9a) amplitude highstand.
unless the mantle viscosities lie outside the range
5.2. Peloponnisos, Greece

Evidence for Holocene sea-level change is available
from several localities at the heads of the Gulfs of
Argolid, Messini and Elos and at Navarine Bay (Kraft
andRapp, 1975; Kraft et al., 1977, 1980) where tectonic
stability is suggestedby the occurrence of the Last
Interglacial sea levels at a few meters above their present
level (Kelletat et al., 1976; van Andel, 1987). Some small
spatial variability between these sites is predicted
( 0.5 m at 6 ka) (Lambeck, 1995b) but the information
here has been combinedinto a single curve for
illustrative purposes only. The evidence includes very
shallow marine sediments, providing lower limits, but
most of the evidence for the past 7 ka is from shallow-
lagoonal and back-swamp sediments that provide upper
limits only. The comparisons in Fig. 9b indicate that the
predicted levels lie systematically above the observed
limits. At this site, the predicted levels can be lowered by
increasing the lithospheric thickness (Fig. 10c), unlike
for the Carmel coast where the dependence of the
predicted values on H 1 is much less. An increase in
lower-mantle viscosity also results in a lowering of the
prediction but the dependence on this parameter
decreases with increasing viscosity beyond the range
considered. Within the range of upper-mantle viscosity
used, the amplitudes of the predicted sea level decrease
with decreasing viscosity but beyond this range they
again increase, andmodifications of the earth-model
parameters do not bring the observations and predic-
tions into near congruence.

5.3. Versilia plain, Italy

A 70 m core from the Versilia plain of northern
Tuscany yielda quite complete Holocene sea-level
record(Antonioli et al., 1999b) andsea-level indicators
from 6 to 7 ka BP are 4–9 m below present level. The
predicted values here also lie above the observed values,
including the continental indicators that specify upper
limits only. Lithospheric dependence here is small but of
the opposite trendto that for the previous two sites and
to force agreement between observations andpredic-
tions through modification of the earth rheology
Fig. 10. Earth-model parameter dependence of relative sea-level requires an increase in either (or both) the upper- or
predictions for 6 ka BP at four Mediterranean sites. The grey-tone
areas define the parameter range of Table 1. (a) Dependence on upper- lower-mantle viscosity (Fig. 9b), although the former is
mantle viscosity, (b) dependence on lower-mantle viscosity, and (c) inconsistent with the inference from the Carmel coast or
dependence on lithospheric thickness. from Peloponnisos.