Page 9 - Vacchi_Marriner_2016

180 M. Vacchi et al. / Earth-Science Reviews 155 (2016) 172–197
Lower Water MLW in order to encompass the whole tidal zone (Fig. 3B, 3.3. Archeological index and limiting points
Table 1). We added a further 0.5 m of additional error to account for the
environmental uncertainty. The long history of human occupation in the western Mediterranean
has left rich archaeological evidence along its coastlines, including, for in-
stance, harbours, fish tanks, slipways and coastal quarries. Auriemma and
3.1.3. Beachrock index points Solinas (2009) furnished a synthesis of archaeological sea-level proxies.
Beachrocks are a lithified coastal deposit where lithification is a
−2 Many different archaeological structures that were originally emerged,
function of CO 3 ion concentration in seawater, microbial activity and or in contact with seawater, today lie below MSL and therefore attest to
degassing of CO 2 from seaward flowing groundwater (Mauz et al., a relative change in the position of the sea surface and the structure.
2015b). Field experiments and coastal observations (e.g., Hanor, 1978; The “functional height” of an archaeological sea-level indicator corre-
Hopley, 1986; Neumeier, 1998) suggest that cementation occurs within sponds to the elevation of a specific architectural part with respect to
a few decades, in areas where suitable coastal morphology provides the MSL position at the time of its construction (e.g., Antonioli et al.,
sufficient accommodation space for soft sediment to settle (Mauz 2007). Ancient port interface structures (e.g., quays and jetties), fishtanks
et al., 2015b). The cement by which the loose sand and gravel are locked
and fishponds represent the most reliable sea-level indicators
into position is indicative of the nearshore zone between the shoreface (e.g., Marriner and Morhange, 2007; Auriemma and Solinas, 2009;
and the beach, at the interface between seawater and meteoric water
Morhange and Marriner, 2015) and have been used to produce index
(e.g., Neumeier, 1998; Vousdoukas et al., 2007). Issues on the use of points in our database. In Section 2.2, we described the differences in
beachrocks as accurate sea-level indicators are present in the literature
(e.g., Kelletat, 2006). However, recent studies have demonstrated that the interpretations of fishtanks and fishponds as sea-level index points.
There is considerable morphological variability in these structures
the cement is crucial for identifying the spatial relationship between (Higginbotham, 1997) and the definition of a standardized indicative
the coastline and the zone of beachrock formation (e.g., Vousdoukas meaning for the western Mediterranean scale is particularly challenging.
et al., 2007; Mauz et al., 2015b). For this reason, in our database the indicative meaning assigned to a
Thus, a definition of the indicative meaning depends largely on the
preservation of the original cement and its link with other sedimentary fishtank index point encompasses the different archaeological interpreta-
tions (maximum two, see Section 2.2) provided by the original papers.
information (e.g., Mauz et al., 2015b). In the Mediterranean region, A variety of coastal archaeological structures, such as coastal
beachrocks have been sampled and dated down to −45 m (e.g., De quarries, tombs, breakwaters and coastal roads can provide insights to
Muro and Orrù, 1998; Orrù et al., 2004). Many studies have determined
reconstruct the RSL in a given area (Auriemma and Solinas, 2009). How-
the vertical accuracy of beachrock samples through SEM, petrographic ever, due to the difficulties in establishing a relationship with a former
and cathodoluminescence analyses of cements with a precision up to
MSL, we used these archaeological markers as terrestrial limiting points.
±0.25 m in microtidal settings (e.g., Desruelles et al., 2009; Vacchi Submerged structures, including harbour foundations and wrecks as
et al., 2012b). Such vertical accuracy cannot be obtained without accu- well as the fine-grained sediments deposited inside the ancient ports
rate description of the chemistry, crystal form and fabric of the cement. (e.g., Marriner and Morhange, 2007), were used to produce marine lim-
In the intertidal zone, the metastable aragonite and High Magnesium iting points.
Calcite (HMC) form as irregularly distributed needles, isopachous fibres
or rims and micritic cement (Neumeier, 1998; Desruelles et al., 2009). 3.4. Altitude of former sea-level
Samples having these characteristics have an indicative range spanning
HAT to MLW (Fig. 3C, Table 1). However, 11 beachrock samples in our For each dated index point, RSL is estimated using the following
database did not meet these requirements because the original source equation:
did not contain enough information.
The beachrock formation zone (i.e. the mixing zone, Vousdoukas
RSL i ¼ A i RWL i ð1Þ
et al., 2007) can exceed the intertidal zone ranging from slightly subtidal
to supratidal (spray zone). The amplitude of this zone depends on wave (Shennan and Horton, 2002), where A i is the altitude and RWL i is the
exposure and the local geomorphological setting and is not symmetrical reference water level of sample i, both expressed relative to the same
with respect to the MSL because it is greater in the supratidal zone datum; MSL in our analysis.
(Mauz et al., 2015b). For these samples, we thus adopted a conservative The total vertical error is obtained by adding in quadratic individual
indicative range of +2 m MSL to −1 m MSL (Fig. 3C, Table 1). In our errors according to:
opinion, such approximation largely encompasses the mixing zone
amplitude for the microtidal coasts of the Mediterranean. e i ¼ e 1 þ e 2 þ e 3 þ e n … 1=2 ð2Þ
2
2
2
2

3.2. Sea-level limiting points (Shennan and Horton, 2002), where e 1 , … e n represent the sources
of error for each index point i including the indicative range (Fig. 3).
Terrestrial limiting points usually form at an elevation above HAT, The additional errors comprise an error associated with calculating the
but can form in the intertidal zone due to rising groundwater tables sample altitude. This can be as small as ±0.05 m with high precision
(e.g., Engelhart and Horton, 2012). Therefore, a conservative lower surveying (e.g., Shennan, 1986), but can increase to more than
limit of MSL has been employed in this analysis. In this database, terres- ±0.5 m when the altitude was estimated using the environment in
trial limiting points are typically samples deposited in freshwater which the sample was collected (e.g., saltmarsh; Engelhart and
marshes and swamps, alluvial plains, archaeological soils and on the Horton, 2012) or when the depths of the samples are measured with
subaerial part of beaches (Table 1). diving gauges (e.g., Rovere et al., 2010; Vacchi et al., 2012a). In southern
Marine limiting points are typically samples deposited in open France and Corsica, many data are expressed in relation to the local
marine or prodelta environments as well as lagoonal environments National Geodetic Datum (NGD) that is presently 0.1 m below MSL
that do not meet the requirements to be classified as index points. Fur- (e.g. Vella and Provansal, 2000; Morhange et al., 2013). All the data
ther, in situ marine benthos (e.g. Lithophaga sp, Mesophyllum sp.) living levelled to the NGD were therefore corrected.
in the infralittoral zone and with no direct relationship to a former A tidal error was included if the tidal information was based on mul-
midlittoral zone was converted into marine limiting point (Table 1). tiple tidal stations or on tidal modeling. In our database this error does
Reconstructed RSL points must fall below terrestrial limiting points not exceed ±0.3 m, due to the microtidal setting of the Mediterranean.
and above marine limiting points (example in Fig. 3B). We included a core stretching/shortening sampling error ranging from