ARTICLE

https://doi.org/10.1038/s41467-021-24250-1 OPEN

Climate pacing of millennial sea-level change

variability in the central and western
Mediterranean
Matteo Vacchi 1,2 ✉, Kristen M. Joyse3, Robert E. Kopp 3, Nick Marriner 4, David Kaniewski5 &
Alessio Rovere 6
1234567890():,;

Future warming in the Mediterranean is expected to significantly exceed global values with
unpredictable implications on the sea-level rise rates in the coming decades. Here, we apply
an empirical-Bayesian spatio-temporal statistical model to a dataset of 401 sea-level index
points from the central and western Mediterranean and reconstruct rates of sea-level change
for the past 10,000 years. We demonstrate that the mean rates of Mediterranean industrial-
era sea-level rise have been significantly faster than any other period since ~4000 years ago.
We further highlight a previously unrecognized variability in Mediterranean sea-level change
rates. In the Common Era, this variability correlates with the occurrence of major regional-
scale cooling/warming episodes. Our data show a sea-level stabilization during the Late
Antique Little Ice Age cold event, which interrupted a general rising trend of ~0.45 mm a−1
that characterized the warming episodes of the Common Era. By contrast, the Little Ice Age
cold event had only minor regional effects on Mediterranean sea-level change rates.

1 Dipartimento di Scienze Della Terra, Università di Pisa, Pisa, Italy. 2 CIRSEC—Centro Interdipartimentale di Ricerca per lo Studio degli Effetti del

Cambiamento climatico dell’Università di Pisa, Pisa, Italy. 3 Department of Earth & Planetary Sciences and Rutgers Institute of Earth, Ocean, and Atmospheric
Sciences, Rutgers University, New Brunswick, NJ, USA. 4 CNRS, ThéMA, Université de Franche-Comté, UMR 6049, MSHE Ledoux, Besançon, Cedex, France.
5 TRACES, UMR 5608 CNRS, Université Toulouse Jean Jaurès, Maison de la Recherche, Toulouse, Cedex 9, France. 6 MARUM, Center for Marine

Environmental Sciences, University of Bremen, Bremen, Germany. ✉email: [email protected]

NATURE COMMUNICATIONS | (2021)12:4013 | https://doi.org/10.1038/s41467-021-24250-1 | www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24250-1

C
limate and sea-level reconstructions for the pre-industrial points (SLIPs), defining the discrete position of past RSL in space
period (i.e., before 1850 CE) provide context for current and time18. We focus our analysis on the central and western
and future changes1–3. Determining the rates, mechan- Mediterranean (Fig. 1), which are less affected by neotectonic
isms, and geographic variability of relative sea-level (RSL) change processes19 than the eastern part. The results of our analysis
following the Last Glacial Maximum (LGM) is relevant to gau- constitute the first basin-scale assessment of sea-level variability
ging how climatic forcing may influence the rates of future sea- in the Mediterranean for the last 10,000 years and represent the
level change3,4. Compilations of sea-level proxies have facilitated natural and geological backgrounds against which modern
the quantification of the response of the solid Earth and geoid to Mediterranean sea-level rise should be quantified.
ice-mass redistribution5–7 and provided constraints for statistical
and geophysical models used to project future sea-level rise8. The
Mediterranean Sea is a semi-enclosed basin lying in a transition Results and discussion
zone between mid-latitude and subtropical atmospheric circula- Millennial variability of sea-level change rates. The SLIPs
tion regimes and is characterized by strong land-sea contrasts9. database (Supplementary Table 1) is composed of (i) 391
For this reason, it is considered a hotspot of current climate radiocarbon-dated geological samples from transitional brackish
change9–11, and future warming in the Mediterranean area is environments, fossil intertidal bioconstructions, beachrocks, and
expected to exceed global rates by ∼25%12. This may result in (ii) 12 archeologically dated marine structures whose relationship
high sea-level rise rates compared to global averages, leading to with the contemporary mean sea-level is robustly defined20. The
significant losses in the environmental, cultural and economic spatial distribution of the SLIPs covers a large portion of the
values of Mediterranean coasts13. Furthermore, semi-closed central and northern sectors of the central and western Medi-
basins are poorly resolved by the ~1° resolution typical of glo- terranean basin, while in the southern sector the available data are
bal climate models included in CMIP5/6, which are generally restricted to the coasts of Malta and the Gulf of Gabes (Fig. 1).
used to drive projections of local dynamic sea-level change14. The SLIPs database allowed us to reconstruct the rates of RSL
Finally, offset between data and glacio-isostatic adjustment (GIA) change in 48 sub-regions clustered according to their geographic
models were recently highlighted by extended sea-level records15. proximity (Supplementary Fig. 1a). The average vertical accuracy
This adds complexity to defining the magnitude and spatial of the different SLIPs is ±0.7 m (max ±1.3 m, min ±0.2 m) while
variability of the isostatic component, which affects both current the average age error is ±0.15 ka (max 0.62 ka, min 0.025 ka). All
and future sea-level changes. errors are reported at ±1σ. The age of the SLIPs spans the whole
The increasing availability of continuous and high-resolution Holocene period (Supplementary Fig. 1b). 10.3% of the SLIPs
Holocene and Common Era Mediterranean relative sea-level date to the early Holocene (−10,000 to −6000 CE), 32.4% to the
(RSL) reconstructions15,16 provides the opportunity to explore mid-Holocene (−6000 to −2000 CE), and 57.3% to the late
the role of climatic factors in mediating sea-level variability in the Holocene (−2000 to 1950 CE). Virtually uncompressible samples
Holocene (i.e., the last 11.7 ka). These data are of major impor- (see “Methods”) represent 36.5% of the entire record, while 39.4%
tance because regional climatic forcing can lead to significant of the SLIPs dates between −5000 and 1950 CE (Supplementary
departures from global mean sea-level projections10. Fig. 1b).
Here, we applied an empirical-Bayesian spatio-temporal sta- The spatio-temporal model allowed us to reconstruct sea-level
tistical model17 (see “Methods”) to a dataset of 401 sea-level index change rates since −8000 CE. There is a paucity of SLIPs for the

Fig. 1 Spatial distribution of the central and western Mediterranean sea-level index points (SLIPs) used for this analysis. Br is the Balearic Sea; Li is the
Gulf of Lion; Lg is the Ligurian Sea; nTy is the northern Tyrrhenian Sea; sTy is the southern Thyrrenian Sea; Gb is the Gulf of Gabes; Io is the Ionian Sea; nAd
is the northern Adriatic; sAd is the southern Adriatic Sea; Sr is Sardinia Island; Cr is Corsica Island; Si is Sicily Island.

2 NATURE COMMUNICATIONS | (2021)12:4013 | https://doi.org/10.1038/s41467-021-24250-1 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24250-1 ARTICLE

period −10,000 to −8000 CE. From the model, it was possible to Mediterranean climatic proxies12 and is consistent with the data
calculate the “central and western Mediterranean Sea-Level” extracted from the longest central and western Mediterranean
(Med-SL), which represents the common signal found at all sites tide-gauge data, which indicate sea-level rise rates of about
included in the model runs (Fig. 2a and Supplementary Table 2). 1.2–1.3 mm a−1 23,24 for the period 1880–2012 CE. Even higher
The Med-SL, which is uniform over the entire central and rates (1.7–1.8 mm a−1) are observed for the second part of the
western Mediterranean, absorbs a majority of the sea-level signal, last century, indicated by a larger dataset of tidal gauges and
whereas the regional signals (Fig. 2b, Supplementary Fig. 2) satellite altimetry25,26. This indicates that, at the basin scale, the
explain the variability we can observe at the basin scale. The mean estimate of industrial-era sea-level rising rate has no
model also estimates RSL for locations and times where there are equivalent analog during the last 4000 years and that the rate of
no direct observations because it recognizes that an observation central and western Mediterranean industrial-era sea-level rise is
associated with a single point in space and time is informative unlikely (~25% probability) to be a random occurrence (Supple-
about sea-level at proximal locations and times. mentary Fig. 3).
Med-SL estimates (Fig. 2a) indicate that the central and western Our analysis highlights significant variability in the regional
Mediterranean sea-level rise decreased from 8.7 ± 0.9 mm a−1 to rates of sea-level change (Fig. 2b, Supplementary Fig. 2), which
3.1 ± 0.8 mm a−1 during the period −8000 to −5000 CE. The resulted in contrasting sea-level rising trends in the different
slowdown continued over subsequent millennia, with average portions of the Mediterranean basin analyzed in this study.
rates of 1.5 ± 0.8 mm a−1 between −5000 and −2000 CE, 0.5 ± Between −8000 and −5000 CE, faster sea-level rise rates were
0.7 mm a−1 between −2000 and 0 CE, and 0.45 ± 0.7 mm a−1 in observed in the mid-western portion of the basin (<~10° E) with
the last 2000 years. This stabilization trend reflects the general rates of 8.0 ± 0.4 mm a−1 (Fig. 3). During the same period, sea-
decrease in rates of global sea-level change consistent with the level rose much slower in the eastern portion of the basin (>~10°
final phase of North American deglaciation and the consequent E) with rates that did not exceed 5.0 ± 0.5 mm a−1 (Fig. 3). We
sudden reduction of meltwater input and with the stabilization of then observed an inversion in this rising pattern after sea-level
global mean surface temperature7,21. Between −2000 and 1850 stabilized around −5000 CE. Since that period, rates of sea-level
CE, the ice-equivalent meltwater input is either zero or rise were always slower in the western portion of the basin (<~5°
minimal21,22. During this period, Med-SL rise rates ranged E) compared to the mid-eastern part (>~5°E, Fig. 3). These
between 0.30 ± 0.7 mm a−1 and 0.55 ± 0.6 mm a−1, while in the differences are particularly significant between −5000 and −2000
industrial-era (e.g., post 1850 CE) rates increased up to 1.05 ± CE when low average rise rates (0.5 ± 0.2 mm a−1) are recorded in
0.6 mm a−1 (Fig. 2a, Supplementary Table 2). This acceleration the Balearic Sea while high average rates (2.0 ± 0.3 mm a−1)
closely mirrors post-industrial warming observed in several characterize the Ionian Sea. In the last 4000 years, we observed a
progressive decrease in the spatial variability of the rising with
maximum average rates (0.8 ± 0.2 mm a−1) still recorded in the
Ionian Sea, while minimal average rates (0.2 ± 0.2 mm a−1) are
recorded both in the Balearic Sea and in the Gulf of Gabes
(Fig. 3). We remark that the Gulf of Gabes, in Tunisia, represents
a unique setting within the Mediterranean, as it is the only
Mediterranean region where a purely isostatic mid-Holocene
highstand is recorded, mediated by continental levering effects27.
Sediment compaction and tectonics may have a role in
controlling the observed spatial variability of sea-level change
rates among regions. However, these components were mini-
mized in our SLIPs database by prioritizing samples that are
virtually incompressible or less prone to compaction, and by
excluding data from regions that are significantly affected by co-
seismic or volcano-tectonic vertical ground motions (see
“Methods”). For this reason, much of the observed spatial
variability of sea-level change rates is related to glacio- and hydro-
isostatic adjustment (GIA), which has been the dominant process
influencing the Mediterranean RSL evolution since the global
mean sea-level stabilization of the mid-Holocene7,23. Our data
indicate a general eastward increase of the GIA component in the
central and western Mediterranean basin with minimal isostatic-
driven subsidence recorded along the Spanish coast and in the
Balearic Islands and maximum rates recorded in the Ionian Sea.
This pattern differs from the one proposed by the available GIA
models7,28 which predict the maximal GIA-related sea-level
change, with comparable magnitude, in the Ionian Sea and in the
area comprising the Balearic, Sardinia and Corsica islands
(Supplementary Fig. 3). The offset between the data and models
Fig. 2 Rates of relative sea-level change for the central and western is probably related to lateral variations in mantle viscosity and/or
Mediterranean region in the last 10,000 years. a Common sea-level in the thickness of the lithosphere, which are currently not taken
signal (Med-SL). The inlet graph shows the Med-SL variation in the last into account by Mediterranean GIA models15,29. Lateral hetero-
4000 years. The solid line and shaded envelope denote the model mean geneity of the Earth’s mantle may significantly affect the Earth’s
and the 1σ uncertainty (see Supplementary Table 2). b variability of relative response to deglaciation30,31. Our results can thus be employed
sea-level (RSL) change rates in the 48 central and western Mediterranean for an improved tuning of geophysical predictions of RSL
regions included in the analysis. The solid line and shaded envelope are the evolution in the basin, which is characterized by significant
model mean and 1 s uncertainty. variability in lithospheric thickness and complex mantle

NATURE COMMUNICATIONS | (2021)12:4013 | https://doi.org/10.1038/s41467-021-24250-1 | www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24250-1

Fig. 3 Spatial and temporal variability of relative sea-level (RSL) changes and their uncertainties across the central and western Mediterranean basin
in the time periods 2000 to −2000 CE, −2000 to −5000 CE, and −5000 to −8000 CE. Note the changes of scale for the different time intervals.

structure19. Nonetheless, it should be noted that our analysis has observed, and the oscillatory patterns observed have a period that
some geographic limitations due to the absence of SLIPs along is too short to be influenced by glacio-isostatic processes. The fact
much of the African coast and near the Gibraltar Strait. that these fluctuations were observed across all regions would also
exclude potential local tectonic influences and compaction-
Regional climatic influence on sea-level rise rates. Notwith- related subsidence. Instead, we suggest that regional climatic
standing regional differences, our spatio-temporal analysis shows forcings are the most likely mechanism driving the variability in
that the central and western Mediterranean regions were char- the sea-level change data.
acterized by several sea-level oscillations in the last 6 millennia In effect, while it is known that large ice melting was minimal
(Supplementary Fig. 2). Looking for the source of these sea-level after −2000 CE21,34, much less is known about the responses to
oscillations, isostatic processes can be excluded. Isostasy is, by shorter-term Mediterranean climatic changes35,36 such as the
definition, a progressive and slow viscoelastic response of the Roman Warm Period (RWP, ~−500 CE to ~500 CE), the Late
Earth to the redistribution of ice and ocean loads32,33. GIA Antique Little Ice Age (LALIA, ~536 to ~660 CE), the Medieval
modeling is unable to resolve the scale of sea-level fluctuations Climate Anomaly (MCA, ~860 to ~1250 CE) and the Little Ice

4 NATURE COMMUNICATIONS | (2021)12:4013 | https://doi.org/10.1038/s41467-021-24250-1 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24250-1 ARTICLE

Age (LIA, ~1250 to ~1850 CE). In the Common Era, Med-SL rise The rising trend only returned to values similar to the pre-
rates varied within a range of ~0.9 mm a−1 (Supplementary LALIA (0.45 ± 0.7 mm a−1) during the MCA which was char-
Table 2, Fig. 4). Rise rates up to 0.5 ± 0.7 mm a−1 characterized acterized by warmer climatic conditions (Fig. 4) and remained at
the warmer episode occurring at the RWP while we observed a similar values for much of the LIA (1250–1600 CE) suggesting a
deceleration of the sea-level change rates (0.3 ± 0.7 mm a−1) negligible influence of this cooling episode on central and western
during the LALIA (Fig. 4). The LALIA is recognized as one of the Mediterranean sea-level rise rates. In the remaining part of the
most important cooling episodes of the Common Era36. This pre-industrial period (1600 and 1800 CE) rates rose to 0.6 ± 0.6
cooling event is found in different proxies (Fig. 4) such as mm a−1 while we observed a progressive acceleration of sea-level
temperature anomalies in Europe37, and specifically in the central rise in the industrial-era, with rates up to 1.05 ± 0.6 mm a−1
and western Mediterranean38,39 and the European Alps36. During (Fig. 4), which are significantly faster than any warm climatic
this period, we also observe a decrease in sea-surface tempera- episode of the Common Era.
tures (SSTs) in the western Mediterranean40, as well as the Our spatio-temporal analysis shows a strong relationship
exceptional seventh-century solar minimum41. between Mediterranean temperatures and the rate of sea-level rise

Fig. 4 Reconstructed variability of common sea-level signal (Med-SL) for the central and western Mediterranean region in the Common Era. Solid line
and shaded envelope are the model mean and 1 s uncertainty. Med-SL is compared with a European Temperature anomalies37, b central Mediterranean
cooler climate pollen data38, c summer (JJA) temperature anomalies in the European Alps36, d total solar irradiance41, e sea-surface temperatures (SSTs)
in the western Mediterranean40. Temporal extension of the Roman Warm Period (RWP), the Late Antique Little Ice Age (LALIA), the Medieval Climate
Anomaly (MCA), the Little Ice Age (LIA), and the Industrial period are from refs. 35,36. The graded bar is the Med-SL model mean.

NATURE COMMUNICATIONS | (2021)12:4013 | https://doi.org/10.1038/s41467-021-24250-1 | www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24250-1

confirming, at the basin scale, the results locally obtained in the which represents the sub-regionally varying, slow, and temporally non-linear SL
northwestern Adriatic Sea42. These oscillations are thus con- field; and rf(x, t), which represents the sub-regionally varying, fast and temporally
non-linear SL field. As in Kopp et al.3, the data model represents observations (yi)
trolled by the differential response of Mediterranean sea-level to as:
cooling/warming episodes, as demonstrated by the variability of


y
sea-level rise rates observed in the Common Era. yi ¼ f xi ; t i þ w xi ; t i þ y0 xi þ εi ð2Þ
Therefore, our findings suggest that a deeper exploration of the where xi and ti are the location and time, respectively, of observation i, w(x, t) is a
regional climatic parameters, controlling the variability of rise white noise term that captures sub-decadal changes in RSL and vertical errors
beyond those nominally represented in the database, and yo(xi) is a site-specific
rates, is required to produce robust predictions of Mediterranean y
datum offset. εi represents errors in sea-level observations, and the term for time t i
sea-level evolution in a changing climate. Regionally, global is the sum of the mean observed age and an error term for time. As in Kopp et al.3,
predictions of sea-level rise may be worsened by the expected geochronological uncertainties are incorporated using the noisy-input GP
increased warming of the Mediterranean9. This may have major method138, which translates errors in the independent variable into comparable
implications for the near-future resilience of both natural and errors in the dependent variable.
Hyperparameters that define prior knowledge of the amplitude and spatio-
human-modified Mediterranean coasts, characterized by a temporal scales for sea-level in each term of the process model were optimized via
narrow littoral area with high concentrations of people and maximum likelihood. Optimized prior standard deviations for the common, slow,
assets and by rapid demographic, social, economic, and fast, and white noise terms were 10.4 m, 1.72 m, 2 cm, and 7 mm, respectively. The
environmental change43,44. optimized temporal scales for the common, slow, and fast components were 12200,
3590, and 14.1 years, respectively, while the optimized spatial scales for the slow
and fast components were 14° and 2° angular distances, respectively.
Methods
SLIPs database. We collated a database of SLIPs following the most recent
standards in sea-level studies18,45. The analysis of a wide range of geological and Data availability
archeological proxies46–133 resulted in a database of 401 SLIPs (Supplementary Data related to this article can be found in Supplementary Table 1 (SLIP database) and
Table 1) that identify the position of former RSL in the central and western Supplementary Table 2 (Med-SL rates). The references of the original papers used to
Mediterranean coasts from 12 ka to present (Supplementary Fig. 1b). Mediterra- produce the SLIP database are provided at the end of Supplementary Table 1. These data
nean SLIPs are commonly derived from cores on coastal and alluvial plains, coastal are available under a CC-BY 4.0 license at the following https://doi.org/10.5281/
marshes, and lagoons. For these samples, the definition of the indicative meaning is zenodo.4737120.
based on paleoecology and, in particular, on the macro-and micro-faunal assem-
blages (i.e., malacofauna, foraminifera, and ostracod assemblages). Furthermore,
fossil intertidal bioconstructions (e.g., vermetids and Lithophyllum byssoides rims)
Code availability
The code used for analysis in this article is archived on GitHub (https://github.com/
and beachrocks have also yielded SLIPs for the Mediterranean region15,16. Finally,
bobkopp/CESL-STEHM-GP).
we produced SLIPs using maritime archeological structures whose functioning
height is related to the former mean sea level such as fishtanks20 or when found
covered by fossil biological encrustations which clearly define the relationship with Received: 18 January 2021; Accepted: 26 May 2021;
the former tidal frame134. We did not include samples from sectors manifesting
major evidence for co-seismic land-level changes and/or crustal movements con-
trolled by volcanic activity135,136, which can generate significant departures from
climatic-driven sea-level changes. Sediment compaction can be an important issue
because it can lower the SLIP relative to the initial depositional elevation, resulting
in an overestimation of the sea-level rise18,137. In the database compilation, we References
prioritized “base of basal” samples, e.g., those recovered from sedimentary units 1. Church, J. A. & White, N. J. A 20th century acceleration in global sea‐level
overlying incompressible substrates such as Pleistocene sands/gravels or rocky rise. Geophys. Res. Lett. 33, https://doi.org/10.1029/2005GL024826 (2006).
basements. These samples are, therefore, less prone to compaction6. Where basal 2. Kemp, A. C. et al. Climate related sea-level variations over the past two
samples were not available, we used intercalated samples which are derived from millennia. Proc. Natl Acad. Sci. USA 108, 11017–11022 (2011).
facies of low-density, organic-rich sediment within a sequence of higher density, 3. Kopp, R. E. et al. Temperature-driven global sea-level variability in the
clastic units18,137, and are, therefore, susceptible to compaction. However, in the Common Era. Proc. Natl Acad. Sci. USA 113, E1434–E1441 (2016).
presence of two or more coeval intercalated SLIPs in the same region, we excluded 4. Kench, P. S. et al. Climate-forced sea-level lowstands in the Indian Ocean
those found at lower elevation as being affected by post-depositional during the last two millennia. Nat. Geosci. 13, 61–64 (2020).
compaction137. This procedure was of particular importance for the Rhone, Ebro, 5. Mitrovica, J. X. et al. On the robustness of predictions of sea level fingerprints.
and Tiber deltas, for Venice lagoon, and for the Romagna and Versilia coastal Geophys. J. Int. 187, 729–742 (2011).
plains. This practice allowed us to significantly reduce the effects of compaction on 6. García-Artola, A. et al. Holocene sea-level database from the Atlantic coast of
the SLIPs dataset. The fossil intertidal bioconstructions, beachrocks, and arche- Europe. Quat. Sci. Rev. 196, 177–192 (2018).
ological structures are not subject to compaction. To perform a high-resolution 7. Roy, K. & Peltier, W. R. Relative sea level in the Western Mediterranean basin:
assessment of variability in sea-level rise rates we clustered the data into 48 sub- a regional test of the ICE-7G_NA (VM7) model and a constraint on late
regions, based on SLIPs collected no more than ~50 km apart (Supplementary Holocene Antarctic deglaciation. Quat. Sci. Rev. 183, 76–87 (2018).
Fig. 1a). 8. Horton, B. P. et al. Mapping sea-level change in time, space, and probability.
Annu. Rev. Environ. Resour. 43, 481–521 (2018).
Spatio-temporal statistical model. We employed an empirical spatio-temporal 9. Cramer, W. et al. Climate change and interconnected risks to sustainable
hierarchical model16 to reconstruct the common sea-level change across the central development in the Mediterranean. Nat. Clim. Change 8, 972–980 (2018).
and western Mediterranean Basin, sub-regional variability in RSL changes, and 10. Adloff, F. et al. Mediterranean Sea response to climate change in an ensemble
spatial-temporal variability patterns in rates of RSL change across the basin over of twenty first century scenarios. Clim. Dyn. 45, 2775–2802 (2015).
the last 10 ka. The height and timing (with vertical and temporal errors) of paleo- 11. Guiot, J. & Cramer, W. Climate change: the 2015 Paris Agreement thresholds
RSL from the 401 SLIPs from 48 sub-regions in the central and western Medi- and Mediterranean basin ecosystems. Science 354, 465–468 (2016).
terranean database were fed into the model. Additional model inputs came from 12. Lionello, P. & Scarascia, L. The relation between climate change in the
tide-gauge records taken from the Permanent Service for Mean Sea Level (PSMSL) Mediterranean region and global warming. Regional Environ. Change 18,
(see Kopp et al.3 for more details). Spatio-temporal variabilities of RSL change and 1481–1493 (2018).
their uncertainties (Fig. 3) are calculated through a linear transformation of the 13. Anzidei, M. et al. Sea level rise scenario for 2100 AD in the heritage site of
RSL predictions. Pyrgi (Santa Severa, Italy). J. Mar. Sci. Eng. 8, 64 (2020).
The hierarchical model has three levels: (i) a process level, which models RSL 14. Eyring, V. et al. Overview of the Coupled Model Intercomparison Project
through space and time; (ii) a data level, which models how RSL from the process Phase 6 (CMIP6) experimental design and organization. Geoscientific Model
model is recorded by geological proxies; and (iii) a hyperparameter level, which Dev. 9, 1937–1958 (2016).
describes the prior expectations for spatial and temporal RSL variability. 15. Vacchi, M. et al. New relative sea-level insights into the isostatic history of the
The process model represents the RSL field as the sum of three components, Western Mediterranean. Quat. Sci. Rev. 201, 396–408 (2018).
each with a Gaussian Process (GP) prior: 16. Vacchi, M. et al. Multiproxy assessment of Holocene relative sea-level changes
f ðx; t Þ ¼ g b ðtÞ þ r s ðx; t Þ þ r f ðx; tÞ ð1Þ in the western Mediterranean: sea-level variability and improvements in the
definition of the isostatic signal. Earth-Sci. Rev. 155, 172–197 (2016).
where x represents a spatial location and t represents time. The three components 17. Ashe, E. L. et al. Statistical modeling of rates and trends in Holocene relative
that comprise the RSL field are gb, which is the common Med-SL signal; rs(x, t), sea level. Quat. Sci. Rev. 204, 58–77 (2019).

6 NATURE COMMUNICATIONS | (2021)12:4013 | https://doi.org/10.1038/s41467-021-24250-1 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24250-1 ARTICLE

18. Shennan, I., Long, A. J. & Horton, B. P. Handbook of Sea-level Research Ch. 1 48. Dupré, M., Fumanal, M. P., Sanjaume, E., Santisteban, C. & Usera, J.
(John Wiley & Sons, 2015). Quaternary evolution of Pego coastal lagoon (Southern Valencia, Spain).
19. Faccenna, C. et al. Mantle dynamics in the Mediterranean. Rev. Geophysics 52, Palaeogeogr. Palaeoclimatol. Palaeoecol. 68, 291–299 (1988).
283–332 (2014). 49. Ruiz, J. M. & Carmona, P. La llanura deltaica de los ríos Júcar y Turia y la
20. Benjamin, J. et al. Late Quaternary sea-level changes and early human societies Albufera de Valencia. Geomorfologia i Quat. litoral Homenatge al Dr. V.
in the central and eastern Mediterranean Basin: an interdisciplinary review. Rosselló 1, 399–419 (2005).
Quat. Int. 449, 29–57 (2017). 50. Marco-Barba, J., Holmes, J. A., Mesquita-Joanes, F. & Miracle, M. R. The
21. Peltier, W. R. Global glacial isostasy and the surface of the ice-age Earth: the influence of climate and sea-level change on the Holocene evolution of a
ICE-5G (VM2) model and GRACE. Annu. Rev. Earth Planet. Sci. 32, 111–149 Mediterranean coastal lagoon: evidence from ostracod palaeoecology and
(2004). geochemistry. Geobios 46, 409–421 (2013).
22. Milne, G. A., Long, A. J. & Bassett, S. E. Modelling Holocene relative sea-level 51. Sanjaume, E., Segura, F., Garcia, M. L. & Pardo, J. Recent sedimentation in
observations from the Caribbean and South America. Quat. Sci. Rev. 24, Valencia Lagoon: preliminary results. J. Coast. Res. 8, 688–698 (1992).
1183–1202 (2005). 52. Carmona, P. & Ballester, J. P. Geomorphology, geoarchaeology and ancient
23. Marcos, M. & Tsimplis, M. N. Coastal sea level trends in Southern Europe. settlement in the Valencian Gulf (Spain). m.éditerranée. Rev. g.éographique
Geophys. J. Int. 175, 70–82 (2008). des. pays m.éditerranéens./J. Mediterranean Geogr. 117, 61–72 (2011).
24. Zerbini, S. et al. Sea-level change in the Northern Mediterranean Sea from 53. Rodríguez-Pérez, A., Blázquez, A. M., Guillem, J. & Usera, J. Maximum flood
long-period tide gauge time series. Earth-Sci. Rev. 167, 72–87 (2017). area during MIS 1 in the Almenara marshland (western Mediterranean):
25. Wöppelmann, G. & Marcos, M. Coastal sea level rise in southern Europe and Benthic foraminifera and sedimentary record. Holocene 28, 1452–1466 (2018).
the nonclimate contribution of vertical land motion. J. Geophys. Res. Oceans 54. Carmona, P. et al. Environmental evolution and mid–late Holocene climate
117, https://doi.org/10.1029/2011JC007469 (2012). events in the Valencia lagoon (Mediterranean coast of Spain). Holocene 26,
26. Anzidei, M. et al. Coastal structure, sea-level changes and vertical motion of 1750–1765 (2016).
the land in the Mediterranean. Geolog. Soc. Lond., Spec. Publ. 388, 453–479 55. Blázquez, A. M., Rodríguez-Pérez, A., Torres, T. & Ortiz, J. E. Evidence for
(2014). Holocene sea level and climate change from Almenara marsh (western
27. Stocchi, P., Colleoni, F. & Spada, G. Bounds on the time–history and Holocene Mediterranean). Quat. Res. 88, 206–222 (2017).
mass budget of Antarctica from sea-level records in SE Tunisia. Pure Appl. 56. Giaime, M. et al. In search of Pollentia’s southern harbour: geoarchaeological
Geophysics 166, 1319–1341 (2009). evidence from the Bay of Alcúdia (Mallorca, Spain). Palaeogeogr.,
28. Lambeck, K. & Purcell, A. Sea-level change in the Mediterranean Sea since the Palaeoclimatol., Palaeoecol. 466, 184–201 (2017).
LGM: model predictions for tectonically stable areas. Quat. Sci. Rev. 24, 57. Yll, R., Pantaleón-Cano, J., Pérez-Obiol, R. P. & Roure, J. M. Cambio climático
1969–1988 (2005). y transformación del medio durante el Holoceno en las Islas Baleares. Congrés
29. Stocchi, P. et al. MIS 5e relative sea-level changes in the Mediterranean Sea: ́ a la Peninsula
del Neolitic ́ Ibérica. Saguntum Extra 2, 45–51 (1999).
contribution of isostatic disequilibrium. Quat. Sci. Rev. 185, 122–134 (2018). 58. Burjachs, F., Pérez-Obiol, R., Roure, J. M. & Julià, R. Dinámica de la
30. Spada, G., Antonioli, A., Cianetti, S. & Giunchi, C. Glacial isostatic adjustment vegetación durante el Holoceno en la isla de Mallorca. Trabajos de Palinología
and relative sea-level changes: the role of lithospheric and upper mantle básica y aplicada 199–210 (1994).
heterogeneities in a 3-D spherical Earth. Geophys. J. Int. 165, 692–702 (2006). 59. Cearreta, A., Benito, X., Ibáñez, C., Trobajo, R. & Giosan, L. Holocene
31. Kuchar, J., Milne, G. & Latychev, K. The importance of lateral Earth structure palaeoenvironmental evolution of the Ebro Delta (Western Mediterranean
for North American glacial isostatic adjustment. Earth Planet. Sci. Lett. 512, Sea): evidence for an early construction based on the benthic foraminiferal
236–245 (2019). record. Holocene 26, 1438–1456 (2016).
32. Milne, G. A. & Mitrovica, J. X. Postglacial sea-level change on a rotating Earth. 60. Riera-Mora, S. & Esteban-Amat, A. Vegetation history and human activity
Geophys. J. Int. 133, 1–19 (1998). during the last 6000 years on the central Catalan coast (northeastern Iberian
33. Melini, D. & Spada, G. Some remarks on Glacial Isostatic Adjustment Peninsula). Vegetation Hist. Archaeobotany 3, 7–23 (1994).
modelling uncertainties. Geophys. J. Int. 218, 401–413 (2019). 61. Daura, J. et al. Palaeoenvironmental record of the Cal Maurici wetland
34. Lambeck, K., Rouby, H., Purcell, A., Sun, Y. & Sambridge, M. Sea level and sediment archive in Barcelona (NE Iberian Peninsula) between c. 6000 and
global ice volumes from the Last Glacial Maximum to the Holocene. Proc. 4000 cal. yr BP. Holocene 26, 1020–1039 (2016).
Natl Acad. Sci. USA 111, 15296–15303 (2014). 62. Ejarque, A. et al. Coastal evolution in a Mediterranean microtidal zone: Mid to
35. Margaritelli, G. et al. Marine response to climate changes during the last five Late Holocene natural dynamics and human management of the Castelló
millennia in the central Mediterranean Sea. Glob. Planet. Change 142, 53–72 lagoon, NE Spain. PLoS ONE 11, e0155446 (2016).
(2016). 63. Laborel, J. et al. Biological evidence of sea-level rise during the last 4500 years
36. Büntgen, U. et al. Cooling and societal change during the Late Antique Little on the rocky coasts of continental southwestern France and Corsica. Mar.
Ice Age from 536 to around 660 AD. Nat. Geosci. 9, 231–236 (2016). Geol. 120, 203–223 (1994).
37. Luterbacher, J. et al. European summer temperatures since Roman times. 64. Sabatier, P. et al. Holocene evolution of a Languedocian lagoonal environment
Environ. Res. Lett. 11, 024001 (2016). controlled by inherited coastal morphology (South of France). Bull. de. la Soc.
38. Kaniewski, D. et al. Solar pacing of storm surges, coastal flooding and Geologique de. Fr. 181, 27–36 (2010).
agricultural losses in the Central Mediterranean. Sci. Rep. 6, 25197 (2016). 65. Vella, C. & Provansal, M. Relative sea-level rise and neotectonic events during
39. Kaniewski, D. et al. Recent anthropogenic climate change exceeds the rate and the last 6500 yr on the southern eastern Rhône delta, France. Mar. Geol. 170,
magnitude of natural Holocene variability on the Balearic Islands. 27–39 (2000).
Anthropocene 32, 100268 (2020). 66. Amorosi, A., Rossi, V. & Vella, C. Stepwise post-glacial transgression in the
40. Sicre, M. A. et al. Sea surface temperature variability in the North Western Rhône Delta area as revealed by high-resolution core data. Palaeogeogr.,
Mediterranean Sea (Gulf of Lion) during the Common Era. Earth Planet. Sci. Palaeoclimatol., Palaeoecol. 374, 314–326 (2013).
Lett. 456, 124–133 (2016). 67. Morhange, C., Laborel, J. & Hesnard, A. Changes of relative sea level during
41. Steinhilber, F., Beer, J. & Fröhlich, C. Total solar irradiance during the the past 5000 years in the ancient habour of Marseilles, Southern France.
Holocene. Geophysi. Res. Lett. 36, https://doi.org/10.1029/2009GL040142 Palaeogeogr., Palaeoclimatol., Palaeoecol. 166, 319–329 (2001).
(2009). 68. Vella, C. et al. Evolution of the Rhône delta plain in the Holocene. Mar. Geol.
42. Faivre, S., Bakran-Petricioli, T., Barešić, J., Horvatić, D. & Macario, K. Relative 222, 235–265 (2005).
sea-level change and climate change in the Northeastern Adriatic during the 69. Allinne et al. Archéologie et paléoenvironnement sur le site du pont romain
last 1.5 ka (Istria, Croatia). Quat. Sci. Rev. 222, 105909 (2019). des Esclapes (Fréjus, Var). ArcheoSciences. Rev. d’archéom.étrie 30, 181–196
43. Wolff, C. et al. A Mediterranean coastal database for assessing the impacts of (2006).
sea-level rise and associated hazards. Sci. Data 5, 180044 (2018). 70. Devillers, B., Excoffon, P., Morhange, C., Bonnet, S. & Bertoncello, F. Relative
44. Furlan, E. et al. Development of a Multi-Dimensional Coastal Vulnerability sea-level changes and coastal evolution at Forum Julii (Fréjus, Provence).
Index: assessing vulnerability to inundation scenarios in the Italian coast. Sci. Comptes Rendus Geosci. 339, 329–336 (2007).
Total Environ. 772, 144650 (2021). 71. Morhange, C. et al. Relative sea-level changes during Roman times in the
45. Khan, N. S. et al. Inception of a global atlas of sea levels since the Last Glacial northwest Mediterranean: the 1st century AD fish tank of Forum Julii, Fréjus,
Maximum. Quat. Sci. Rev. 220, 359–371 (2019). France. Geoarchaeology 28, 363–372 (2013).
46. Ferrer-García, C. & Blázquez-Morilla, A. M. The evolution of the Albufereta 72. Colombaroli, D., Marchetto, A. & Tinner, W. Long‐term interactions between
lagoon (western Mediterranean): climate cycles and sea-level changes. J. Coast. Mediterranean climate, vegetation and fire regime at Lago di Massaciuccoli
Res. 28, 1617–1626 (2012). (Tuscany, Italy). J. Ecol. 95, 755–770 (2007).
47. Brisset, E., Burjachs, F., Navarro, B. J. B. & de Pablo, J. F. L. Socio-ecological 73. Carboni, M. G. et al. Palaeoenvironmental reconstruction of late Quaternary
adaptation to Early-Holocene sea-level rise in the western Mediterranean. foraminifera and molluscs from the ENEA borehole (Versilian plain, Tuscany,
Glob. Planet. change 169, 156–167 (2018). Italy). Quat. Res. 74, 265–276 (2010).

NATURE COMMUNICATIONS | (2021)12:4013 | https://doi.org/10.1038/s41467-021-24250-1 | www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24250-1

74. Kaniewski, D. et al. Holocene evolution of Portus Pisanus, the lost harbour of 98. Barra, D. et al. New data on the evolution of the Sele River coastal plain
Pisa. Sci. Rep. 8, 11625 (2018). (southern Italy) during the Holocene. Il Quaternario 11, 287–299 (1998).
75. Ghilardi, M. et al. Reconstructing the landscape evolution and the human 99. Antonioli, F., Chemello, R., Improta, S. & Riggio, S. The Dendropoma
occupation of the Lower Sagone River (Western Corsica, France) from the (Mollusca Gastropoda, Vermetidae) intertidal reef formations and their
Bronze Age to the Medieval period. J. Archaeological Sci.: Rep. 12, 741–754 paleoclimatological use. Mar. Geol. 161, 155–170 (1999).
(2017). 100. Basso, D., Bernasconi, M. P., Robba, E. & Marozzo, S. Environmental
76. Ghilardi, M. et al. Enregistrements d’événements extrêmes et évolution des evolution of the Marsala sound, Sicily, during the last 6000 years. J. Coast. Res.
paysages dans les basses vallées fluviales du Taravo et du Sagone (Corse 24, 177–197 (2008).
occidentale, France) au cours de l’âge du Bronze moyen à final: une 101. Spampinato, C. R., Costa, B., Di Stefano, A., Monaco, C. & Scicchitano, G. The
perspective géoarchéologique. Geomorphol. Relief, Process. Environ. 23, 15–35 contribution of tectonics to relative sea-level change during the Holocene in
(2017). coastal south-eastern Sicily: new data from boreholes. Quat. Int. 232, 214–227
77. Vacchi, M., Ghilardi, M., Spada, G., Currás, A. & Robresco, S. New insights (2011).
into the sea-level evolution in Corsica (NW Mediterranean) since the late 102. Gerardi, F. et al. Geological record of tsunami inundations in Pantano
Neolithic. J. Archaeological Sci.: Rep. 12, 782–793 (2017). Morghella (south-eastern Sicily) both from near and far-field sources. Nat.
78. Poher, Y., Ponel, P., Médail, F., Andrieu-Ponel, V. & Guiter, F. Holocene Hazards Earth Syst. Sci. 12, 1185–1200 (2012).
environmental history of a small Mediterranean island in response to sea-level 103. Marriner, N., Gambin, T., Djamali, M., Morhange, C. & Spiteri, M.
changes, climate and human impact. Palaeogeogr., Palaeoclimatol., Palaeoecol. Geoarchaeology of the Burmarrad ria and early Holocene human impacts in
465, 247–263 (2017). western Malta. Palaeogeogr., Palaeoclimatol., Palaeoecol. 339, 52–65 (2012).
79. Lambeck, K., Anzidei, M., Antonioli, F., Benini, A. & Verrubbi, V. Tyrrhenian 104. Furlani, S. et al. Submerged speleothem in Malta indicates tectonic stability
sea level at 2000 BP: evidence from Roman age fish tanks and their geological throughout the Holocene. Holocene 28, 1588–1597 (2018).
calibration. Rendiconti Lincei. Sci. Fisiche e Nat. 29, 69–80 (2018). 105. Zaîbi, C. et al. Évolution du trait de côte à l’Holocène supérieur dans la Sebkha
80. Lambeck, K., Anzidei, M., Antonioli, F., Benini, A. & Esposito, A. Sea level in El-Guettiate de Skhira (Golfe de Gabès, Tunisie) à travers sa faune
Roman time in the Central Mediterranean and implications for recent change. d’ostracodes et de foraminifères. Geobios 44, 101–115 (2011).
Earth Planet. Sci. Lett. 224, 563–575 (2004). 106. Zaïbi, C. et al. Evolution of the sebkha Dreîaa (South-Eastern Tunisia, Gulf of
81. Evelpidou, N. et al. Late Holocene sea level reconstructions based on Gabes) during the Late Holocene: response of ostracod assemblages. Rev. de.
observations of Roman fish tanks, Tyrrhenian Coast of Italy. Geoarchaeology micropaléontologie 55, 83–97 (2012).
27, 259–277 (2012). 107. Morhange, C. & Pirazzoli, P. A. Mid-Holocene emergence of southern
82. Marra, F., Bozzano, F. & Cinti, F. R. Chronostratigraphic and lithologic Tunisian coasts. Mar. Geol. 220, 205–213 (2005).
features of the Tiber River sediments (Rome, Italy): implications on the post- 108. Strasser, A., Davaud, E. & Jedoui, Y. Carbonate cements in Holocene
glacial sea-level rise and Holocene climate. Glob. Planet. Change 107, 157–176 beachrock: example from Bahiret et Biban, southeastern Tunisia. Sediment.
(2013). Geol. 62, 89–100 (1989).
83. Salomon, F. Géoarchéologie du delta du Tibre: évolution géomorphologique 109. Lakhdar, R., Soussi, M., Ben Ismail, M. H. & M’Rabet, A. A Mediterranean
holocène et contraintes hydrosédimentaires dans le système Ostie-Portus Holocene restricted coastal lagoon under arid climate: case of the sedimentary
(Italie). Thèse de doctorat en géographie, option géoarchéologie. Physio-Géo. record of Sabkha Boujmel (SE Tunisia). Palaeogeogr., Palaeoclimatol.,
Géographie Phys. et. Environ. 8, 1–6 (2014). Palaeoecol. 241, 177–191 (2006).
84. Goiran, J. P., Tronchère, H., Collalelli, U., Salomon, F. & Djerbi, H. 110. Serandrei-Barbero, R., Albani, A., Donnici, S. & Rizzetto, F. Past and recent
Découverte d’un niveau marin biologique sur les quais de Portus: le port sedimentation rates in the Lagoon of Venice (Northern Italy). Estuar., Coast.
antique de Rome. Méditerranée 112, 59–67 (2009). Shelf Sci. 69, 255–269 (2006).
85. De Muro, S. & Orrù, P. Il contributo delle Beach-Rock nello studio della 111. McClennen, C. E. & Housley, R. A. Late-Holocene channel meander
risalita del mare olocenico. Le Beach-Rock post-glaciali della Sardegna nord- migration and mudflat accumulation rates, lagoon of Venice, Italy. J. Coast.
orientale. Il Quaternario 111, 19–39 (1998). Res. 224, 930–945 (2006).
86. Melis, R. T. et al. 8000 years of coastal changes on a western Mediterranean 112. Zoppi, U. et al. Preliminary estimate of the reservoir age in the Lagoon of
island: a multiproxy approach from the Posada plain of Sardinia. Mar. Geol. Venice. Radiocarbon 43, 489–494 (2001).
403, 93–108 (2018). 113. Madricardo, F. & Donnici, S. Mapping past and recent landscape
87. Melis, R. T., Depalmas, A., Di Rita, F., Montis, F. & Vacchi, M. Mid to late modifications in the Lagoon of Venice through geophysical surveys and
Holocene environmental changes along the coast of western Sardinia historical maps. Anthropocene 6, 86–96 (2014).
(Mediterranean Sea). Glob. Planet. Change 155, 29–41 (2017). 114. Tosi, L. Foglio geologico 128 Venezia e note illustrative. Carta geologica
88. Di Rita, F. & Melis, R. T. The cultural landscape near the ancient city of d’Italia alla scala 1, 000 (2007).
Tharros (central West Sardinia): vegetation changes and human impact. J. 115. Donnici, S., Serandrei-Barbero, R. & Canali, G. Evidence of climatic changes
Archaeological Sci. 40, 4271–4282 (2013). in the Venetian Coastal Plain (Northern Italy) during the last 40,000 years.
89. De Falco, G. et al. Early cementation and accommodation space dictate the Sediment. Geol. 281, 139–150 (2012).
evolution of an overstepping barrier system during the Holocene. Mar. Geol. 116. Galassi, P. & Marocco, R. Relative sea level rise and subsidence in the Caorle
369, 52–66 (2015). Lagoon (Northern Adriatic Sea), Italy, during Holocene. Il Quaternario 12,
90. Ruiz, J. M., Carmona, P., Gómez-Bellard, C. & van Dommelen, P. 249–256 (1999).
Geomorfología y cambio ambiental en el entorno de los yacimientos púnicos 117. Fontana, A. et al. Lagoonal settlements and relative sea level during Bronze age
de la llanura de Terralba (Golfo de Oristano, isla de Cerdeña, Italia). Boletín in Northern Adriatic: geoarchaeological evidence and paleogeographic
Geol.ógico y. Min. 129, 331–352 (2018). constraints. Quat. Int. 439, 17–36 (2017).
91. Orrù, P., Solinas, E., Puliga, G. & Deiana, G. Palaeo-shorelines of the historic 118. Fontana, A. & Zanferrari. A. Foglio geologico 86 San Vito al Tagliamento e note
period. Sant’Antioco Island, south-western Sardinia (Italy). Quat. Int. 232, illustrative. Carta geologica d’Italia alla scala 1:50,000 (2006).
71–81 (2011). 119. Bondesan, A., Meneghel, M., Rosselli, R. & Vitturi. A. Geomorphological Map
92. Orrù, P. E. et al. Sea level changes and geoarchaeology between the bay of of the Province of Venice, scale 1:50,000 (LAC, Firenze, 2004).
Capo Malfatano and Piscinnì Bay (SW Sardinia) in the last 4 kys. Quat. Int. 120. Furlani, S. et al. Tectonic effects on Late Holocene sea level changes in the Gulf
336, 180–189 (2014). of Trieste (NE Adriatic Sea, Italy). Quat. Int. 232, 144–157 (2011).
93. Orrù, P. E., Antonioli, F., Lambeck, K. & Verrubbi, V. Holocene sea-level 121. Covelli, S., Fontolan, G., Faganeli, J. & Ogrinc, N. Anthropogenic markers in
change in the Cagliari coastal plain (southern Sardinia, Italy). Quaternaria the Holocene stratigraphic sequence of the Gulf of Trieste (northern Adriatic
Nova 8, 193–212 (2004). Sea). Mar. Geol. 230, 29–51 (2006).
94. Barra, D. et al. The Versilian transgression in the Volturno river plain 122. Melis, R. et al. Sea level and paleoenvironment during roman times inferred
(Campania, Southern Italy): Palaeoenvironmental history and chronological from coastal archaeological sites in Trieste (northern Italy). Alp.
data. Il Quaternario 9, 445–458 (1996). Mediterranean Quat. 25, 41–55 (2012).
95. Sacchi, M. et al. Late-Holocene to recent evolution of Lake Patria, South Italy: 123. Ogorelec, B., Faganeli, J., Mišič, M. & Čermelj, B. Reconstruction of
an example of a coastal lagoon within a Mediterranean delta system. Glob. paleoenvironment in the Bay of Koper (Gulf of Trieste, northern Adriatic).
Planet. Change 117, 9–27 (2014). Annales 11, e200 (1997).
96. Barra, D. et al. Depositional history and palaeogeographic reconstruction of 124. Faivre, S. et al. Relative sea level change in western Istria (Croatia) during the
Sele coastal plain during Magna Grecia settlement of Hera Argiva (Southern last millennium. Quat. Int. 232, 132–143 (2011).
Italy). Geologica Romana 35, 151–166 (1999). 125. Cibin, U. & Stefani, M. Foglio geologico 187 Codigoro e note illustrative.
97. Amato, V. et al. Relative sea level changes and paleogeographical evolution of Carta geologica d’Italia alla scala 1:50,000 (2009).
the southern Sele plain (Italy) during the Holocene. Quat. Int. 288, 112–128 126. Correggiari, A., Roveri, M. & Trincardi, F. Late Pleistocene and Holocene
(2013). evolution of the north Adriatic Sea. Il Quaternario 9, 697–704 (1996).

8 NATURE COMMUNICATIONS | (2021)12:4013 | https://doi.org/10.1038/s41467-021-24250-1 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24250-1 ARTICLE

127. Sarti, G., Centineo, M. C. & Calabrese, L. Foglio geologico 205 Comacchio e Global Changes (PAGES), which in turn received support from the Swiss Academy of
note illustrative. Carta geologica d’Italia alla scala 1:50,000 (2009). Sciences and the Chinese Academy of Sciences.
128. Marriner, N. et al. Post-Roman sea-level changes on Pag Island (Adriatic Sea):
dating Croatia’s “enigmatic” coastal notch? Geomorphology 221, 83–94 (2014).
129. Faivre, S., Bakran-Petricioli, T., Horvatinčić, N. & Sironić, A. Distinct phases
Author contributions
M.V., K.J., R.E.K., and A.R. designed the research approach. M.V. led the research and
of relative sea level changes in the central Adriatic during the last 1500 years—
wrote the first draft of the paper. M.V., N.M., D.K., and A.R. contributed to compiling
influence of climatic variations? Palaeogeogr. Palaeoclimatol. Palaeoecol. 369,
the sea-level database. K.J. and R.E.K. performed the statistical analysis. M.V., K.J., R.E.
163–174 (2013).
K., N.M., D.K., and A.R. contributed to the writing of the manuscript.
130. Caldara, M., Caroli, I. & Simone, O. Holocene evolution and sea-level changes
in the Battaglia basin area (eastern Gargano coast, Apulia, Italy). Quat. Int.
183, 102–114 (2008). Competing interests
131. Caldara, M. & Simone, O. Coastal changes in the eastern Tavoliere Plain The authors declare no competing interests.
(Apulia, Italy) during the Late Holocene: natural or anthropic? Quat. Sci. Rev.
24, 2137–2145 (2005).
132. Di Rita, F., Simone, O., Caldara, M., Gehrels, W. R. & Magri, D. Holocene Additional information
environmental changes in the coastal Tavoliere Plain (Apulia, southern Italy): Supplementary information The online version contains supplementary material
a multiproxy approach. Palaeogeogr., Palaeoclimatol., Palaeoecol. 310, available at https://doi.org/10.1038/s41467-021-24250-1.
139–151 (2011).
133. Primavera, M., Simone, O., Fiorentino, G. & Caldara, M. The Correspondence and requests for materials should be addressed to M.V.
palaeoenvironmental study of the Alimini Piccolo lake enables a
Peer review information Nature Communications thanks Marco Anzidei and the other,
reconstruction of Holocene sea-level changes in southeast Italy. Holocene 21,
anonymous, reviewer(s) for their contribution to the peer review of this work.
553–563 (2011).
134. Morhange, C. & Marriner, N. Handbook of Sea-level Research Ch. 9 (John
Reprints and permission information is available at http://www.nature.com/reprints
Wiley and Sons, 2015).
135. Scicchitano, G. et al. Uplifted Holocene shorelines at Capo Milazzo (NE Sicily,
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
Italy): evidence of co-seismic and steady-state deformation. Quat. Int. 232,
201–213 (2011). published maps and institutional affiliations.
136. Aucelli, P. P., Cinque, A., Mattei, G. & Pappone, G. Late Holocene landscape
evolution of the gulf of Naples (Italy) inferred from geoarchaeological data. J.
Maps 13, 300–310 (2017). Open Access This article is licensed under a Creative Commons
137. Shennan, I. & Horton, B. Holocene land‐and sea‐level changes in Great Attribution 4.0 International License, which permits use, sharing,
Britain. J. Quat. Sci. 17, 511–526 (2002). adaptation, distribution and reproduction in any medium or format, as long as you give
138. McHutchon, A. & Rasmussen, C. Gaussian process training with input noise. appropriate credit to the original author(s) and the source, provide a link to the Creative
Adv. Neural Inf. Process. Syst. 24, 1341–1349 (2011). Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
Acknowledgements article’s Creative Commons license and your intended use is not permitted by statutory
M.V. is funded by the Rita Levi Montalcini program of the Italian Ministry of University regulation or exceeds the permitted use, you will need to obtain permission directly from
and Research (MUR). A.R. thanks the Deutsche Forschungsgemeinschaft (Excellence the copyright holder. To view a copy of this license, visit http://creativecommons.org/
Cluster “EXC 2077: The Ocean Floor—Earth’s Uncharted Interface”, project no. licenses/by/4.0/.
390741603) for support. R.E.K. and K.J. were supported by US National Science Foun-
dation grants DGE-1633557 and OCE-1804999. We further acknowledge PALSEA, a
working group of the International Union for Quaternary Sciences (INQUA) and Past © The Author(s) 2021

NATURE COMMUNICATIONS | (2021)12:4013 | https://doi.org/10.1038/s41467-021-24250-1 | www.nature.com/naturecommunications 9