Page 26 - 23

ARTICLE IN PRESS
1592 K. Lambeck et al. / Quaternary Science Reviews 23 (2004) 1567–1598
5° 10° 15° 20°
45° 45°
(i)
40° -0.28 -0.28
-0.45 -0.37
-1.09
(ii)
0.23 -0.11 0.08
0.15 0.35
(iv)
40°
0.87 (iii)
(iv) 0.63 1.15
1.97
1.72
35°
5° 10° 15° 20°
Fig. 11. Estimates of the average vertical tectonic uplift and
subsidence rates (mm/year) for the Holocene interval. Sites without
numbers correspond to localities where the tectonic signals are zero or
less than the estimated precision of the vertical rates. The shaded zone
in the North Adriatic corresponds to the core sites for which Holocene
subsidence is effectively zero. The tectonic signature falls into four
broad categories: (i) subsidence along the North Adriatic coast, (ii)
predominantly tectonic stability along the central and northern coast
of Tyrrhenian Italy but with local areas of some uplift or subsidence,
(iii) uplift in southern Italy, including eastern Sicily, and (iv) tectonic
stability in western Sicily and Sardinia.
pattern of the sea levels for all epochs is determined by
the hydro-isostatic contribution, with the sea basin floor
subsiding under the additional water load. Thus
observed sea levels from the island sites such as
Pantelleria, between Sicily and Tunisia, Marettimo and
Sardinia, should exhibit lower levels than sites on the
Italian Peninsula for the same epoch. Along the central
Tyrrhenian coast, between about Argentarola and
Palinuro, the isobases are nearly parallel to the shore
and data from different locations can be combined into
a composite regional sea-level curve if desired. But this is
not the case for the Adriatic coast where the glacio- and
hydro-isostatic contributions combine to produce a
well-defined north–south gradient such that levels in
the northern Adriatic lie persistently above those further
south. (See Fig. 7 in Lambeck and Johnston, 1995, for
the separate approximate evaluation of the two compo-
nents. Note that the contour labelled À1.5 in Fig. 7b for
T=2 ka should read À0.5 m.) The isobases also follow
the general outline of the North African coast but,
because of indentation in the coastal geometry, the site
at Djerba (Tunisia) becomes effectively an inland site
and the predicted levels in the Gulf of Ga"bes lie above
those for northern Tunisia. In fact, if Late Holocene sea-
level indicators have been preserved along the section of
coast between Tunis and Tripoli they should provide a
good test of the isostatic model. Nowhere within the
region are sea levels from stable coasts predicted to
occur above present sea level at any time during the
Holocene, and the mid-Holocene highstands character-
istic of equatorial latitudes of Africa (Faure et al., 1980)
are absent here because the glacio-isostatic effect is still
sufficiently large to counteract the hydro-isostatic effect.
The palaeo shorelines are obtained from the relation-
ship between water depth h(j, t) at time t and location j
relative to that h(j, t0) at time t0
hðj; tÞ ¼ hðj; t0Þ À Dzpredðj; tÞ;
where Dzpred(j, t) is the eustatic–isostatic prediction of
the change in sea level at time t since the time t0 (Fig. 7).
The palaeo shoreline location is then given by the
contour h(j, t)=0. The water depths used here are from
the DTM compilation on a 2.5’ grid and the results are
devoid of detail in the smaller and shallow embayments.
During the LGM, at 20 ka BP (Fig. 12) the predicted
35° isobases range from about À120 m in the Ga"bes Gulf
(Tunisia) to À150 m in the middle of the basin and
extensive exposure of the shelves occurs in the North
Adriatic as well as along parts of the North African
coast and marine LGM deposits will not occur on these
shallow parts of the shelf. Of potential importance is the
shallow basin, possibly isolated from the Mediterranean
at some time during the LGM, that forms at latitude
B43N. Cores from this feature may provide important
constraints on LGM sea levels and ice volumes in much
the same way as did the cores from the Bonaparte Gulf
of northwestern Australia (Yokoyama et al., 2000). The
sub-aerially exposed shelves in the northern Adriatic
persist for much of the Lateglacial period. The location
of the southernmost sample, dated at 13–11 ka BP, for
example, would have been about 13 m above coeval sea
level at its time of formation and would have remained
sub-aerially exposed until B10.9 ka. At 10 ka the
isobases range from about À55 to À35 m below present
level but the shallowest values occur where the sea has
not yet encroached. Only by about 8 ka does the
Adriatic take its present form (Fig. 12). At 6 ka BP the
sea levels begin to approach present-day values but
the spatial variability remains significant, coastal levels
ranging from BÀ2 m at Ga"bes to BÀ9 m in Sicily and
Calabria. At the Roman period, B2 ka BP, levels
fluctuated from about À0.5 in the north of the Adriatic
to as low as À1.8 m in parts of Sardinia.
7. Conclusion
Even in the absence of the tectonic processes, relative
sea level is spatially variable along the length of the