Page 22 - Vacchi_Marriner_2016

M. Vacchi et al. / Earth-Science Reviews 155 (2016) 172–197 193
our database. For this reason, Fig. 13A,C only depicts the index points any possible role of compaction in the late Holocene was not possible
collected at sites: (i) with long-term vertical movements ≤0.06 mm a - with the present dataset.
1
(Fig. 2, MIS 5e); (ii) with ongoing vertical motions between -0.5 and Preliminary analysis of the normal distribution of late Holocene ris-
0.5 mm a -1 (Fig. 2, GPS); and (iii) not affected by significant ing rates (Fig. 4D) shows that values ranging between 0.55 and 0.20 mm
-1
compaction-related subsidence (e.g., the Ebro Delta) or local fault activ- a are the most represented along the Mediterranean coast. Rates in-
-1
ity (e.g. the Var fault). crease up to ~0.85 mm a , but caution should be used in using these
With the sole exception of southern Tunisia, the data indicate a con- values because a possible compaction influence cannot be excluded in
tinuous rise in RSL which, at the basin scale, rose by ~45 m in ~12 ka some dataset from the southeastern sector of the western Mediterra-
(Fig. 13A). nean Sea (see above).
Even if affected by scatter, the data show a general slowdown in The Fig. 13D clearly shows that late Holocene sea-level rising rates
western Mediterranean sea-level rise starting from ~7.5 ka BP, consis- are slower than those recorded by most of the Mediterranean tide
tent with the final phase of the North American deglaciation gauges (Table 3) but still account for at least the 25-30% of the rate.
(e.g., Carlson et al., 2008; Lambeck et al., 2014) followed by a further de- Therefore, for better understanding of current rates of Mediterranean
crease in rising rates related to the progressive reduction in meltwater sea-level rise from tide gauges, correction for GIA is required
input (e.g., Peltier and Tushingham, 1991; Milne et al., 2005). (e.g., Engelhart et al., 2009; Church and White, 2011). Site-specificas-
Comparatively, the late Holocene record (i.e. the last 4 ka BP) is af- sessment of the GIA signal in the western Mediterranean is beyond
fected by less scatter. In this period, ice equivalent meltwater input is the scope of this paper. However, future application of spatio-
negligible (Peltier, 2004; Milne et al., 2005; Church et al., 2008). There- temporal statistical models (e.g., Parnell et al., 2011; Engelhart et al.,
fore, once tectonics and sediment compaction are factored out, any 2015; Parnell and Gehrels, 2015) to the Mediterranean RSL data may
change observed in RSL is entirely related to vertical land movements provide a long-term baseline against which to gauge changes in sea-
th
due to GIA (e.g., Engelhart et al., 2009, 2015). According to Engelhart level during the 20 century (e.g., Engelhart et al., 2009; Kemp et al.,
et al., 2009, we calculated the late Holocene RSL rising rates excluding 2011; Gehrels and Woodworth, 2013) and provide a framework for
the 20th century sea-level contribution and thus expressing all our more accurate 21 st century sea-level predictions (e.g. Church and
data with respect to MSL at A.D. 1900 (Fig. 13B). We further excluded White, 2011; Horton et al., 2014).
for this analysis the southern Tunisia index points describing a RSL his-
tory not comparable with the other datasets in our study area (see 7. Conclusions
Section 6.2). Thus, the late Holocene record in Fig. 13B most likely en-
compasses the total variability in isostatic contribution in the western In this paper, we reviewed 917 RSL proxies using, for the first time in
Mediterranean. Our estimates indicate that GIA-related land move- the Mediterranean basin, a standardized protocol to produce RSL index
ments in the western Mediterranean vary between ~3.4 m in the south- and limiting points. This allowed us to compare and contrast data from
eastern part of the basin up to ~1.6 m in the northwest. The latter value different literature sources in order to obtain basin-scale insights into
is mainly derived from region #14 and #22, where the Holocene record the processes driving Holocene RSL changes. The database in Appendix
is only based on intercalated index points. For this reason, assessment of A is dynamic and future users can add further data or subdivide the data

Fig. 13. A) Total plot of the western Mediterranean index points from areas that are tectonically stable and minimally affected by compaction-related subsidence. B) Variability of late
Holocene (last 4.0 ka BP) RSL index points in the western Mediterranean. Note that the x and y axis are forced to the MSL at A.D. 1900. C) Approximate location of the index points
plotted (red dots) and not plotted (gray dots) in panel a, see Section 6.3 for details. D) Normal distribution of late Holocene rising rates plotted against the 20th century RSL rise (with
error) derived from long-term Mediterranean tidal gauges (see Table 3). Number above squares denotes the number of years used to compute the trend. MG, Malaga; MS, Marseille;
GE, Genova; VE, Venezia P. Salute; TR, Trieste; RO, Rovinj; BK, Bakar; SP, Split Gradska; DB, Dubrovnik.