Page 16 - Sea-level change_2004

ARTICLE IN PRESS
1984 K. Lambeck, A. Purcell / Quaternary Science Reviews 24 (2005) 1969–1988

5.4. Co ˆte d’Azur, France the sea-level predictions during the Late Holocene
interval. Examples are illustratedin Fig. 11 for
Perhaps the best Late Holocene recordfrom this predictions along the Carmel coast and the Coˆ te d’Azur,
locality is that inferredfrom the fossil vermetid and two sites where the dependence on ice thickness is,
coralline algae record(Laborel et al., 1994) which respectively, least andgreatest. In these examples, the
14
indicates that at 4 C ka BP levels were about 1.5 m predictions for the preferred ice models for the two
lower than today, and that from at least this time northern regions (NE-2 andNA-2) (i) are compared
onwards, sea level rose nearly uniformly to its present with model predictions in which (ii) NE-2 is replaced by
level. Vertical tectonic movement appears to be negli- NE-1 only, (iii) NA2 is replacedby NA-1 only, and (iv)
gible on the time scale of concern here (Dubar et al., both NE-2 andNA-2 are replacedby the larger-volume
1992). The predicted Holocene values agree well with the ice loads of NE-1 and NA-1, respectively. At the Carmel
observed values and no modification of earth-model coast, changes in predictions during the Late Holocene
parameters appears warranted. The earth-model depen- remain small, o1 m at 6 ka BP, andinsufficient to
dence here is similar to that for Versilia and any change change the general observation that the predictions are
in earth-model parameters that may be suggested by the systematically higher than the observedvalues. At the
latter data would degrade the agreement with the Coˆ te d’Azur site, the sensitivity of the prediction on ice
observedrecordfor the past 3000 years. loadis larger, consistent with the site occurring closer to
the ice loads, and reaches 2 m at 6 ka BP for the
modification of both ice sheets but here we see a
5.5. Ice-load dependence potential for trade-off between lower-mantle viscosity
andice thickness (compare Figs. 9dand11b): increasing
Across the region, increasing the ice loads over ice thickness or increasing lower-mantle viscosity lower
Europe and/or North America has the effect of lowering the Late Holocene predictions at this site. However,
modifications of the ice sheets that lead to an substantial
increase in ice volume during the LGM and Late Glacial
intervals are inconsistent with the outcomes of inver-
sions of the rebounddata from the areas of former
glaciation.

5.6. Modification of the esl function

What the preliminary analysis indicates is that with
Dz esl ¼ 0 for to6.8 ka no choice of earth parameters
satisfies all the observational data sets. The modifica-
tions requiredfor the sites in the eastern Mediterranean
(Carmel coast andPeloponnisos) are at least partly
orthogonal to those requiredfor the western Mediterra-
nean sites (Versilia Plain andCoˆ te d’Azur). The
alternative solution is to attribute the discrepancies
between observations andpredictions noted above to
inadequacies of the esl function following the analysis
suggested by Eq. (5). The data discussed above is
inadequate to make precise estimates of the corrections
to this function as well as of the rheological parameters
but the trendis consistent with analyses carriedout for
other areas with the esl at 6.5 ka some 2–3 m less than
today and reaching the present-day levels at about 3000
years ago. This esl function (Fig. 12) implies that the
total ocean volume did not reach its present value until
about 3000 years ago andthat between 7000 and
3000 BP the ocean volume increasedsuch as to raise
the global level by 2.5–3 m (Lambeck, 2002). Whether
this is from melting of mountain glaciers or from
Fig. 11. Predicted sea levels, at (a) the Carmel coast and (b) the Coˆ te
d’Azur, for the nominal earth models and different combinations of Antarctic melting, or whether it includes a thermal
the two principal northern hemisphere ice sheets. The Antarctic and expansion component cannot be establishedfrom this
Alpine contributions are the same in all cases. kindof analysis alone, but improvedanalyses of the