Journal of Marine Science and Engineering

Article Sea Level Rise Scenario for 2100 A.D. in the Heritage Site of (, )

Marco Anzidei 1,*, Fawzi Doumaz 1 , Antonio Vecchio 2,3 , Enrico Serpelloni 1, Luca Pizzimenti 1, Riccardo Civico 1 , Michele Greco 4 , Giovanni Martino 4 and Flavio Enei 5

1 Istituto Nazionale di Geofisica e Vulcanologia, 56126 Pisa, Italy; [email protected] (F.D.); [email protected] (E.S.); [email protected] (L.P.); [email protected] (R.C.) 2 Radboud Radio Lab, Department of Astrophysics/IMAPP, Radboud University— Nijmegen, 6500GL Nijmegen, The Netherlands; [email protected] 3 Lesia Observatoire de Paris, Université PSL, CNRS, Sorbonne Université, Université de Paris, 92195 Meudon, France 4 School of Engineering, Università della Basilicata, 85100 Potenza, Italy; [email protected] (M.G.); [email protected] (G.M.) 5 Museo del Mare e della Navigazione Antica, 00058 Santa Severa, Italy; [email protected] * Correspondence: [email protected]

 Received: 29 November 2019; Accepted: 15 January 2020; Published: 21 January 2020 

Abstract: Sea level rise is one of the main risk factors for the preservation of cultural heritage sites located along the coasts of the Mediterranean basin. Coastal retreat, erosion, and storm surges are posing serious threats to archaeological and historical structures built along the coastal zones of this region. In order to assess the coastal changes by the end of 2100 under the expected sea level rise of about 1 m, we need a detailed determination of the current coastline position based on high resolution Digital Surface Models (DSM). This paper focuses on the use of very high-resolution Unmanned Aerial Vehicles (UAV) imagery for the generation of ultra-high-resolution mapping of the coastal archaeological area of Pyrgi, Italy, which is located near . The processing of the UAV imagery resulted in the generation of a DSM and an orthophoto with an accuracy of 1.94 cm/pixel. The integration of topographic data with two sea level rise projections in the Intergovernmental Panel on Climate Change (IPCC) AR5 2.6 and 8.5 climatic scenarios for this area of the Mediterranean are used to map sea level rise scenarios for 2050 and 2100. The effects of the Vertical Land Motion (VLM) as estimated from two nearby continuous Global Navigation Satellite System (GNSS) stations located as close as possible to the coastline are included in the analysis. Relative sea level rise projections provide values at 0.30 0.15 cm by 2050 and 0.56 0.22 cm by 2100 for the IPCC AR5 8.5 scenarios ± ± and at 0.13 0.05 cm by 2050 and 0.17 0.22 cm by 2100, for the IPCC Fifth Assessment Report ± ± (AR5) 2.6 scenario. These values of rise correspond to a potential beach loss between 12.6% and 23.5% in 2100 for Representative Concentration Pathway (RCP) 2.6 and 8.5 scenarios, respectively, while, during the highest tides, the beach will be provisionally reduced by up to 46.4%. In higher sea level positions and storm surge conditions, the expected maximum wave run up for return time of 1 and 100 years is at 3.37 m and 5.76 m, respectively, which is capable to exceed the local dune system. With these sea level rise scenarios, Pyrgi with its nearby Etruscan temples and the medieval castle of Santa Severa will be exposed to high risk of marine flooding, especially during storm surges. Our scenarios show that suitable adaptation and protection strategies are required.

Keywords: sea level rise; coastlines; 2100; storm surges; heritage sites; Pyrgi; Mediterranean; UAV; DSM

J. Mar. Sci. Eng. 2020, 8, 64; doi:10.3390/jmse8020064 www.mdpi.com/journal/jmse J. Mar. Sci. Eng. 2020, 8, 64 2 of 18

J. Mar. Sci. Eng. 2020, 8, 64 2 of 18 1. Introduction 1. Introduction ObservationalObservational and and instrumental instrumental data data collected collected worldwide worldwide since since the the last last two-three two-three centuries centuries show thatshow the globalthat the sea global level sea is continuouslylevel is continuously rising with rising an with accelerated an accelerated trend in trend recent in years, recent which years, coincideswhich withcoincides the rise with in globalthe rise temperaturesin global temperatures [1]. Global [1]. meanGlobal seamean level sea islevel expected is expected to rise to rise by aboutby about 75 to 20075 cm to by200 2100cm by in 2100 the worstin the scenariosworst scenarios [2–5], i.e.,[2–5], the i.e. most, the mos serioust serious effects effects of climate of climate change change that that might occurmight in future occur decades.in future Thesedecades values. These will values be even will larger be even in subsiding larger in coasts subsiding of the coasts Mediterranean, of the entailingMediterranean, widespread entailing environmental widespread changes, environmental coastal retreat,changes, marine coastal flooding, retreat, marine and loss flooding of land,, and which willloss be of disadvantages land, which will for be human disadvantages activities. for The human sea level activities. rise will The amplify sea level the rise impacts will amplify exerted the by a multitudeimpacts of exerted hazards by (i.e., a multitude storm surges, of hazards flooding, (i.e. coastal, storm erosion, surges, and flooding, tsunamis) coastal on the erosion infrastructure, and andtsunamis) building on integrity, the infrastructure people safety, and economicbuilding integrity, assets, and people cultural safety, heritage. economic assets, and cultural heritage.Therefore, it is important to mitigate these risks by providing multi-temporal scenarios of expected inlandTherefore, extension itof is marine important flooding to mitigate as a consequence these risks ofby the providing sea level multi rise-temporal for a cognizant scenarios coastal of managementexpected inland [6–8 ].extension of marine flooding as a consequence of the sea level rise for a cognizant coastalThis management is particularly [6– true8]. for the Mediterranean region where ancient civilizations were born and developedThis alongis partic itsularly coasts true [9,10 for]. the A large Mediterranean number of region heritage where sites ancient are located civilizations at the waterfront were born orand very closedeveloped to the sea along level its and coasts are [9,10]. exposed A large to marine number flooding of heritage under sites the eareffects located of ongoing at the waterfront climate change. or very close to the sea level and are exposed to marine flooding under the effects of ongoing climate A large part of these sites, which are dated back to the Greek, Roman, and Medieval ages, are exposed change. A large part of these sites, which are dated back to the Greek, Roman, and Medieval ages, at increasing risks to coastal hazards that are related to a sea-level rise [10]. are exposed at increasing risks to coastal hazards that are related to a sea-level rise [10]. Aerial photogrammetric surveys performed by small Unmanned Aerial Vehicles (UAVs) can Aerial photogrammetric surveys performed by small Unmanned Aerial Vehicles (UAVs) can provide accurate topography at low costs and, in short time, for small areas with respect to conventional provide accurate topography at low costs and, in short time, for small areas with respect to aerialconventional surveys [ 11 aerial]. Results, surveys when [11]. analyzed Results, in when combination analyzed with in sea combination level rise projections with sea level and vertical rise landprojections movements and (VLM), vertical can land support movements the realization (VLM), can of the support expected the sea-level realization rise of andthe stormexpected surge scenariossea-level for rise future and storm decades surge [12 scenarios]. for future decades [12]. InIn this this study, study, we we show show an an e effectiveffective applicationapplication for for the the coastal coastal archaeological archaeological area area of ofPyrgi, Pyrgi, Italy, Italy, whichwhich is is near near Rome Rome (Figure (Figure1 ).1). High High resolution resolution mapsmaps of expected flooded flooded areas areas and and coastal coastal positions positions forfor 2100, 2100, even even in in storm storm surge surge conditions, conditions, areare reportedreported in this this paper paper..

FigureFigure 1. 1.The The investigated investigated area area of of Pyrgi Pyrgi with with the the location location of of the the GNSS GNSS stations stations of of MAR8 MAR8 and and TOLF TOLF and theand nearby the nearby tide gauge tide gauge station station of . of Civitavecchia.

OurOur approach approach is to is apply to applya multidisciplinary a multidisciplinary methodology methodology previously previously tuned in the tuned savemedcoasts in the projectsavemedcoasts (www.savemedcoasts.eu project (www.savemedcoasts.eu), which includes), topography, which includes geodesy, topography, sea level geodesy, data, and sea climatic level projectionsdata, and toclimatic estimate projections realistic sea to levelestimate rise realistic scenarios sea for level targeted rise scenarios coastal areas. for targeted Our approach coastal providesareas. a usefulOur approach analytical provides tool to a identify useful analytical the best adaptation tool to identify and the defense best adaptation strategies againstand defense the sea strategie levels rise impact,against and the tosea protect level rise heritage impact, sites. and to The protect results heritage help sites. decision-makers The results help in thedecision selection-makers of in the the best practicalselection actions of the aimed best atpractical preserving actions the aimed archaeological at preserving and historical the archaeological sites located and in coastalhistorical areas sites that

J. Mar. Sci. Eng. 2020, 8, 64 3 of 18 J. Mar. Sci. Eng. 2020, 8, 64 3 of 18 locatedare subjected in coastal to sea areas level-related that are subjected risks. The toproposed sea level- methodologyrelated risks. The can beproposed exported methodology in other areas can of bethe exported Mediterranean in other region areas of and th beyonde Mediterranea its borders.n region and beyond its borders.

2.2. Geo Geo-Archaeological-Archaeological S Settingetting

TThehe heritage heritage site site of of Pyrgi Pyrgi is is located located along along the the coasts coasts of of N Northernorthern , Latium, between between the the villages villages of of SantaSanta Severa Severa and and , Cerveteri, which which is is about about 50 50 km km north north of of Rome Rome (Italy) (Italy) ( (FigureFigure 1). The The area area includes includes thethe Castle of of Santa Santa Severa Severa that, that with, with Pyrgi, Pyrgi, is one is of one the of most the important most important heritage heritagesites of the sites Tyrrhenian of the Tyrrheniancoast. The coast. area has The been area settled has been since settled the V-IV since millennium the V-IV millennium B.C. [13] and B.C. continuously [13] and continuously developed developedduring the during Neolithic the Age, Neolithic during Age, the Bronze during Age the (II Bronze millennium Age (II B.C.), millennium and during B.C.) the, Ironand Ageduring (IX-VIII the Ironcentury Age B.C.),(IX-VIII thanks century to itsB.C.), good thanks environmental to its good conditions. environmental In the conditions. Etruscan In phase the Etruscan (VII-IV century phase (VIIB.C.),-IV Pyrgi century was B.C.), the port Pyrgi of the was ancient the port Etruscan of the city ancient of Etruscan andplayed city of an Caere important and played role in anthe importantmaritime commercerole in the being maritime frequented commerce by being andfrequented Phoenicians by Greeks ships. Theand areaPhoenicians includes aships. sanctuary The areaand includes the temples a sanctuary of Eileithyia-Leukothea and the temples and of , Eileithyia Cavatha,-Leukothea Suri, andand theApollo, Etruscan Cavatha, analogue Suri, and to thethe Etruscan Phoenician Uni Astarte analogue [13 ].to the Phoenician Astarte [13]. AfterAfter the the Romani Romanizationzation of of this this area area (III (III century century B.C.), B.C.), Pyrgi Pyrgi became became a a maritime maritime colony colony and and a a tall tall fortressfortress surrounded surrounded by by a a polygonal polygonal wall wall was was built built on on part part of of the the Etruscan Etruscan settlement. settlement. DuringDuring the Roman the Roman imperial imperial age, age,the thecity city of Pyrgi of Pyrgi continued continued to to be be frequented frequented until until the the 5t 5th-6thh-6th centurycentury A.D. A.D. when when a a Byzantine Byzantine castrum castrum was was built built on on its its remains remains with with the the early early Christian Christian church church of of SantaSanta Severa Severa inside. inside. Later,Later, the the medieval medieval and and renaissance renaissance village village became became a a large large farm farm located located in in a a strategic strategic position position betweenbetween the the main main harbors harbors of of R Romeome an andd Civitavecchia Civitavecchia [13 [13––16].16]. In In terms terms of of its its geological geological setting, setting, the the coastcoast of Pyrgi Pyrgi belongs belongs to to the the roman roman co-magmatic co-magmatic province province [17] that [17] underwent that underwent major volcanicmajor volcanic activity activityduring theduring Plio-Pleistocene the Plio-Pleistocene era (Figure era2 ().Figure The surface 2). The geology surface is geology characterized is characterized by a bedrock by a belonging bedrock belongingto the allochthonous to the allochthonous Outer Ligurids Outer [18 Ligurids], represented [18], represented by the by Formation, the Tolfa spanning Formation, from spanning the Late fromCretaceous the Late to Palaeogene Cretaceous [19], defined to Palaeogene as the Pietraforte [19], defined formation as and the “comprehensive Pietraforte formation succession” and [20]. “comprehensiveThe latter consists succe of assion” series [20]. of marl The limestonelatter consists and of grey a series marl of beds marl that limestone outcrop and between grey marl Rome beds and thatCivitavecchia outcrop between underlying Rome the and Pietraforte Civitavecchia unit. underlying Biogenic sandstones the Pietraforte of the unit. Early-Middle Biogenic sandstones Pleistocene ofPanchina the Early Formation-Middle Pleistocene partly overlies Panchina the Pietraforte Formation [21 partly]. overlies the Pietraforte [21].

.

FigureFigure 2. Simplified 2. Simplified geological map geological of the study area map (from of http: //dati.lazio.it the study/catalog / areait/dataset (from/carta- http://dati.lazio.it/catalog/it/dataset/cartageologica-informatizzata-regione-lazio-25000-geologica). Legend:-informatizzata (1) anthropic-regione debris,-lazio (2)-25000 coastal). Legend: and marsh (1) anthropicsands and debris recent, dunes,(2) coastal (3) calcareousand marsh marls sands and and clays, recent (4) flysch,dunes, (5)(3) clays calcareous with chalks, marls (6)and landslide clays, (4 de) flyschposits,, (5) (7) clays alluvial with deposits, chalks, (6) (8) landslide lavas, (9) de travertines, posits, (7) (10)alluvial sandy deposits deposits,, (8) andlavas (11), (9) sandy travertines deposits, (10) of sandymarine deposits facies. , and (11) sandy deposits of marine facies.

J. Mar. Sci. Eng. 2020, 8, 64 4 of 18

The Holocene deposits are represented by travertines, slope debris, alluvial, weathering deposits, gravels, and sandy beaches [22]. The neotectonic of this sector of the Tyrrhenian coast of Italy is marked by the elevation of the MIS 5.5 marine terraces that show stability and a weak uplift of the inland sector for the last 124 years [23,24], which is related to magmatic injections under the Vulsini and Sabatini volcanic complexes [24]. Reference [25] underlines that the long-term uplift may not be an appropriate description for all the past two millennia since some weak subsidence may have occurred at Pyrgi and along the nearby coasts.

3. Methods We applied a multidisciplinary approach using coastal topography, geodesy, and climatic-driven estimates of the sea-level rise to provide maps of flooding scenarios for the year 2100 A.D. for the coast of Pyrgi. Our study consists of three main steps: (1) the realization of UAV surveys to obtain an ultra-high-resolution orthophoto and a DSM model of the coastal area to map the current and the projected coastline positions including sea level data in the analysis, (2) the estimation of the current vertical land movements from the analysis of geodetic data from the nearest GPS stations, and, (3) by combining these data with the regional IPCC-AR5 projections (RCP-2.6 and RCP-8.5 scenarios), the calculation of the upper bounds of the expected sea levels for the targeted epochs of 2050 and 2100 A.D. and the corresponding expected inland extent of the marine flooding and shoreline positions were calculated. Lastly, storm surge scenarios were implemented for return times of 1 and 100 years, for sea level rise conditions.

4. Digital Terrain Model Reconstruction To realize the ultra-high-resolution Digital Surface Model (DSM), an aerial photogrammetric survey was performed using a radio-controlled multi-rotor Da Jiang Innovation (DJI) Phantom 4pro UAV system, equipped with a high resolution lightweight digital camera, to capture a set of aerial images of the investigated area (Table1).

Table 1. Survey features.

Survey and Camera Features Number of images 306 Camera stations 286 Flying altitude 79.8 m Tie points 1,158,646 Ground resolution 1.94 cm/pix Projections 3,891,753 Coverage area 0.226 km2 Reprojection error 0.528 pix Camera model FC6310 (8.8 mm) Focal length 8.8 mm Camera resolution 5472 3648 Pixel size 2.41 2.41 µm × ×

The UAV was controlled by an autopilot system using waypoints previously planned by PIX4D® capture IOs App as a Ground Control Station system. To optimize the photogrammetric spatial resolution and coverage of the surveyed area, a constant altitude of 70 m was maintained during the flight and 306 partly overlapping (70% of longitudinal and 70% lateral overlapping between subsequent photos) aerial digital photos were acquired during three successful flights of 13, 11 and 15 min of duration each. To scale the aerial images, we used a dual-frequency geodetic Global Positioning System Real Time Kinematic (GPS/RTK) receiver STONEX S900A® to measure the coordinates of a set of reference Ground Control Points (GCPs) falling in the investigated areas. GCPs positions were estimated in real time by the RTK technique with about 1 to 2 cm of accuracy, with respect to the reference GPS station TOLF. The latter is part of the GNSS network of the Istituto Nazionale di Geofisica e Vulcanologia (INGV) [26] and also pertains to the Leica SmartNet ItalPoS network (https://hxgnsmartnet.com/en-gb/) for real-time positioning services. We used the Agisoft PhotoScan® J. Mar. Sci. Eng. 2020, 8, 64 5 of 18

softwareJ. Mar. Sci. packageEng. 2020, 8 (, http:64 //www.agisoft.com) based on the Structure-from-Motion photogrammetry5 of 18 technique [27] to process the acquired georeferenced images. The analysis included: (1) camera alignmentcloud, and with 3) generation image position of an and orthophoto orientation, covering (2) generation a land surface of a dense of 0.226 points km2 cloud,with a and ground (3) generationresolution of 1.94 an orthophoto cm/pixel as covering well as the a land DSM surface creation of in 0.226 the WGS84 km2 with-UTM32 a ground coordinate resolution system. of 1.94 cm/pixelThe as obtained well as the ortho DSM-rectified creation images in the WGS84-UTM32 (orthophotos) and coordinate digital system. elevation models were also managedThe obtained by Global ortho-rectified Mapper® imagesand ESRI (orthophotos) ArcGis® software and digital to elevation create cross models sections, were also slope managed maps, bysurfaces, Global and Mapper coastline® and positions ESRI ArcGis as well® software as calculate to create the dimension cross sections, of the slope poten maps,tial flooded surfaces, areas. and coastlineThe extracted positions Digital as wellSurface as calculate Model has the a dimension resolution of of the 15.5 potential cm/pixel flooded and a areas.point density The extracted of 41.5 Digitalpoints/m Surface2 Model has a resolution of 15.5 cm/pixel and a point density of 41.5 points/m2 AsAs targetstargets forfor GCPsGCPs measuredmeasured byby GPSGPS/RTK/RTK,, wewe usedused aa setset ofof i)i) natural natural markersmarkers belongingbelonging toto fixedfixed structuresstructures (e.g.,(e.g., thethe centercenter ofofmanholes, manholes, wallwall andand sidewalksidewalk corners,corners, smallsmall structures,structures, oror largelarge stones),stones), and and ii) ii) mobile mobile targets targets for for the the time time of theof the flight, flight, such such as thin as thin metal metal crosses crosses of 60 of 6060 cm× 60 of cm size, of × deployedsize, deployed in the in investigated the investigated area.All area. these All GCPs these wereGCPs chosen were chosen as areas as recognizable areas recognizabl on thee images on the duringimages data during analysis. data analysis. InIn total,total, 2222 Ground Ground Control Control Points Points (GCPs) (GCPs) andand CheckCheck PointsPoints (CPs)(CPs) werewereused used to to geo-reference geo-reference thethe orthophoto orthophoto with with the the ground ground control controlpoint point toolset toolset of of AGISOFT AGISOFT photo photo scan scan (Figure (Figure3 ). 3). The The 3D 3D coordinatescoordinates of of these these points points have have been been estimated estimated with a with mean a RMS mean of RMS 0.6 and of 0.9 0.6 cm and for 0.9 the cm horizontal for the andhorizontal vertical and components, vertical components, respectively, respectively, and were used and were to evaluate used to the evaluate vertical the accuracy vertical ofaccuracy our final of DSM.our final Figure DSM.3 shows Figure the 3 orthophotoshows the orthophoto while Figure while4 shows Figure the 4 DSM.shows the DSM.

FigureFigure 3.3. TheThe orthophoto orthophoto with with the the Ground Ground Control Control Points Points (red (red dots) dots) position position used used during during the UAV the UAVsurveys. surveys.

J. Mar. Sci. Eng. 2020, 8, 64 6 of 18 J. Mar. Sci. Eng. 2020, 8, 64 6 of 18

FigureFigure 4. 4.The The DigitalDigital SurfaceSurface Model (DSM) (DSM) of of the the coastal coastal sector sector of ofPyrgi, Pyrgi, from from the the analysis analysis of the of the aerialaerial photos. photos.

5.5. Tidal Tidal Correction Correction and and Coastline Coastline PositionPosition TheThe coast coast of of Pyrgi, Pyrgi, similar similar toto mostmost of the coasts of of the the Mediterranean Mediterranean Sea, Sea, is ischaracterized characterized by bya a micro-tidalmicro-tidal environment environment andand tidestides areare generally in in the the range range of of ±3030 cm. cm. Only Only in inthe the Gulf Gulf of ofGabes Gabes ± (Tunisia)(Tunisia) and and the the North North AdriaticAdriatic SeaSea (Italy), t tidesides may may reach reach ampl amplitudesitudes up up to toabout about 2 m 2 [27 m– [2730].–30 ]. WeWe used used the the tidal tidal data data collected collected by by the the Italian Italian tidal network network managed managed by by the the Italian Italian Institute Institute for for EnvironmentalEnvironmental Protection Protection and and Research Research (ISPRA)(ISPRA) (data(data areare freelyfreely availableavailable at www.mareografico.itwww.mareografico.it)) at theat sea the level sea level station station of Civitavecchia of Civitavecchia (located (located at LAT at 42LAT◦05 42°038.25”,05’38.25 LON’’, LON 11◦47 11°022.7347’22.730), which’), which is placed is nearplaced Pyrgi nea (Figurer Pyrgi1 ),(Figure to estimate 1), to theestimate tide level the tide (TL) level and (TL) the local and meanthe local sea mean level sea (LMSL) levelat (LMSL) the time at of theth UAVe time surveys of the UAV (26 April, surveys 2019 (26 at April 07:30, 2019 or 05:30 at 07:30 UTC). or We05:30 considered UTC). We theconsidered complete the time complete series to accounttime series for a to long-term account for linear a long trend,-term representative linear trend, representative of the mean seaof the level mean during sea level UAV during surveys, UAV with respectsurveys, to whichwith respect the TL to is which defined. the ThisTL is tide defined. gauge This station tide showsgauge station a valid shows recording a valid period recording of about eightperiod years of (2011–2019)about eight years with a(2011 sea– level2019) trend with ofa sea 0.25 level0.1 trend mm /ofa, which0.25 ± 0.1 is calculatedmm/a, which from is acalculated linear fit on from a linear fit on the monthly data (Figure 5a). From± the data analysis, a mean tide amplitude of the monthly data (Figure5a). From the data analysis, a mean tide amplitude of ~35 cm and a maximum ~35 cm and a maximum tidal range up to ~60 cm has been estimated (Figure 5b). To define a tidal range up to ~60 cm has been estimated (Figure5b). To define a reference level for elevation data, reference level for elevation data, the mean sea level was computed propagating the linear trend the mean sea level was computed propagating the linear trend from the time of surveys, assuming from the time of surveys, assuming the mean sea level as a reference value for the year 2018. A the mean sea level as a reference value for the year 2018. A mean sea level of 4.8 11.8 cm above the mean sea level of 4.8 ± 11.8 cm above the topographic benchmark was estimated± from the hourly topographictidal data. Since benchmark the UAV was surveys estimated have frombeen theperformed hourly during tidal data. a high Since tide the of 4 UAV cm, as surveys shown have by the been performedtidal recordings during a (www.mareografico.it high tide of 4 cm, as), shown then bythe the reference tidal recordings sea level ( www.mareografico.it at the time of surveys), then thecorresponds reference sea to a level position at the of timeonly of0.8 surveys cm above corresponds the mean sea to level a position for 2018. of only 0.8 cm above the mean sea levelThis for value 2018. is negligible compared to the DSM accuracy, so we did not consider it for the analysisThis valueof sea is level negligible rise scenarios. compared Last toly the, we DSM used accuracy, the local so mean we did sea not level consider calculated it for at the the analysis tide ofstation sea level of Civitavecchia rise scenarios. as Lastly,elevation we data, used given the localthe small mean distance sea level between calculated this location at the tide and stationPyrgi. of CivitavecchiaThe obtained as value elevation was used data, to given define the the small position distance of the between coastline this during location the and surveys Pyrgi. and The the obtained one valueexpected was usedfrom tothe define sea level the rise position scenarios of the to 2100 coastline A.D. during the surveys and the one expected from the sea level rise scenarios to 2100 A.D.

J. Mar. Sci. Eng. 2020, 8, 64 7 of 18 J. Mar. Sci. Eng. 2020, 8, 64 7 of 18

(a)

(b)

FigureFigure 5. 5. SeaSea level level data data analysis for the tidetide gaugegauge ofof Civitavecchia,Civitavecchia, which which is is located located a a few few km km north north of ofPyrgi Pyrgi (see (see Figure Figure1 for1 for location). location). ( a(a)) Monthly Monthly datadata ofof seasea levellevel recordings collected in the the time time span span 20112011–2019–2019 (about (about nine nine years). years). The The red red line line is is the the linear linear fit fit of of the the sea sea level level trend trend at at 0.25 0.25 ± 0.10.1 mm/yr mm/yr.. ± ((bb)) S Statisticaltatistical diagramdiagram of of sea sea level level heights heights (cm) (cm) versus versus time time (hours (hours/year)./year). The valuesThe values of sea of level sea height level heightfrequency frequency are reported are reported during during one year, one including year, including maximum maximum sea level sea heights level ofheights about of 45 about cm that 45 maycm thatexceed may the exceed tide amplitudes. the tide amplitudes. These can These be related can be to related storm surgeto storm events surge that events occur that only occur a few only hours a fe inw a hoursyear, whenin a year, water when is pushed water fromis pushed the sea from onto the the sea land onto due the to land a temporary due to a decrease temporary in atmosphericdecrease in atmosphericpressure and pressure wind. and wind.

WeWe prefe preferredrred to to adopt this this local local vertical vertical datum datum instead instead of of the the value value of of the the geoid geoid elevation elevation for for thethe Italian Italian region region provided provided by the by International the International Service for Service the Geoid-ISG-ITG2009 for the Geoid-ISG (http:-ITG2009//www. (isgeoid.polimi.ithttp://www.isgeoid.polimi.it/Geoid/Geoid/Europe/Italy//Europe/Italy/ITG2009_g.htmlITG2009_g.html)[31] since it is) an [31] independent since it is and an moreindependent accurate andelevation more datum.accurate The elevation International datum. Service The International for the Geoid Service (ISG) for estimates the Geoid a geoid (ISG) height estimates in the a Italiangeoid heightGeoid (ITG)in the 2009 Italian for theGeoid coastline (ITG) of2009 Pyrgi for at the 48.319 coastline m. This of elevation Pyrgi at corresponds 48.319 m. This to the elevation contour correspondsline equal to to zero the in contour the Italian line heightequal to reference zero in frame.the Italian The height DSM height reference reference frame frame. The DSM is 0.42 height m, as referenceestimated frame by GPS is 0.42/RTK m, data as estimated at a GCP locatedby GPS/RTK at the data sea levelat a GCP along located the coastline. at the sea Its level elevation along wasthe coastlinecorrected. Its for elevation the tidal rangewas corrected at the time for of the surveys.tidal range The at reference the time mean of the sea surveys. level estimated The reference by the meantidal analysissea level provided estimated a by local the mean tidal seaanalysis level provided with an uncertainty a local mean of sea11.8 level cm. with an uncertainty of ± ±11.8 cm.

J. Mar. Sci. Eng. 2020, 8, 64 8 of 18 J. Mar. Sci. Eng. 2020, 8, 64 8 of 18

6.6. Vertical LandLand MotionMotion (VLM)(VLM) atat PyrgiPyrgi TheThe current current rate rate of VLMof VLM at Pyrgi at Pyrgi was estimated was estimated by the analysis by the analy of thesis available of the GPS available data collected GPS data at thecollected nearest at GNSS the nearest stations GNSS of TOLF, stations belonging of TOLF, the INGVbelonging Rete Integratathe INGV Nazionale Rete Integrata GPS (RING) Nazionale network GPS (DOI:10.13127(RING) network/RING) (DOI:10.13127/RING) and MAR8, belonging and to MAR8, the Topcon-NetGeo belonging to network the Topcon (http:-//NetGeowww.netgeo.it network). These(http://www.netgeo.it). stations, which are These located stations, at about which 6.5 km are and located 0.3 kmof at distance about 6.5 from km the and study 0.3 site,km respectivelyof distance (Figurefrom the1), study have a site, robust respectively time series (Figure that span 1), forhave 2004–2019 a robust for time TOLF series (15.23 that years) span and for 2012–20192004–2019 forfor MAR8TOLF (15.23 (7.38 years)years) (Figureand 20126).–2019 for MAR8 (7.38 years) (Figure 6).

(a)

(b)

FigureFigure 6.6. The vertical components (UP) of the time series of the GNSS station for (aa)) TOLFTOLF (time(time spanspan 2004–2019,2004–2019, aboutabout 1515 years)years) andand (b(b)) MAR8 MAR8 (time (time span span 2012–2019), 2012–2019), both both located located near near Pyrgi Pyrgi (see (see Figure Figure1 for1 for location). location).

GPSGPS data data analysis analysis has hasbeen beencarried carried out following out following the procedures the procedures already described already in Reference described [ 32 in] andReference updated [32] in Reference and updated [33] by in adopting Reference a three-step [33] by procedureadopting using a three the-step GAMIT procedure/GLOBK usingV10.7 [34 the] andGAMIT/GLOBK QOCA software. V10.7 This [34] is and part QOCA of a continental-scale software. This GPSis part solution of a continental including->scale3000 GPS stations solution [34]. Theincluding daily positions>3000 stations of TOLF [34]. and The MAR8 daily havepositions been estimatedof TOLF and in the MAR8 GPS have realization been estimated of the ITRF2008 in the frameGPS realization [35], i.e., the of IGb08the ITRF2008 reference. frame The [35] position, i.e., timethe IGb08 series reference. have been The analyzed position in ordertime toseries estimate have andbeen correct analyzed offsets in orderdue to to station estimate equipment and correct changes offsets while, due simultaneously, to station equipment estimating changes annual while and, semi-annualsimultaneously periodic, estimating signals annual and a linear and semi velocity-annual term, periodic whereas signals velocity and uncertainties a linear velocity have term,been estimatedwhereas velocity adopting uncertainties a power law have+ white been noise estimated stochastic adopting model, a as power in Reference law + white [36]. The noise results stochastic show thatmodel, both as sites in areRefere relativelynce [36]. stable The in results the IGb08 show reference that both frame sites with are a vertical relatively velocity stable of in0.061 the IGb080.135 − ± mmreference/year for frame TOLF with and a vertical0.456 0.344 velocity mm of/year −0.061 for MAR8.± 0.135 We mm/y remarkear for that TOLF uncertainties and −0.456 associated ± 0.344 − ± onmm/y theea verticalr for MAR8. velocities We remark are about that uncertainties0.5 mm/year associated and are barely on the significant vertical velocitie in views ofare unresolved about ±0.5 ± questionsmm/year and about are the barely GPS referencesignificant frame in view stability of unresolved and additional questions factors about [30]. the GPS reference frame stabilityIn addition and additional to GPS data,factors the [30]. tectonic stability of this region is also inferred from the low level of seismicityIn addition deduced to GPS from data, historical the tectonic data stability [37] and instrumentalof this region recordings is also inferred of earthquakes from the (www.ingv.it low level o),f whichseismicity do not deduced report the from occurrence historical of significant data [37] events and for instrumental the last 3000 recordingsyears BP. of earthquakes (www.ingv.it),Lastly, assuming which do that not the report area willthe occurrence continue in of the significant near future events to have for the the last same 3000 tectonic years BP. trend shownLast inly the, assuming past, it is that reasonable the area to will neglect continue the contribution in the near offuture VLM to in have the sea the level same rise tectonic projections trend andshown flooding in the scenariospast, it is forreasonable 2100 A.D. to neglect the contribution of VLM in the sea level rise projections and flooding scenarios for 2100 A.D.

J. Mar. Sci. Eng. 2020, 8, 64 9 of 18 J. Mar. Sci. Eng. 2020, 8, 64 9 of 18

7. Relative Sea Sea-Level-Level RiseRise ProjectionsProjections and FloodingFlooding ScenariosScenarios for 2050 and 2100 A.D.A.D. To estimate the the sea sea-level-level rise rise for for 2050 2050 and and 2100 2100 A.D. A.D. at Pyrgi, at Pyrgi, we referred we referred to the to regional the regional IPCC IPCCAR5 sea AR5-level sea-level projections projections discussed discussed in the in the Fifth Fifth Assessment Assessment Report Report of of the the IPCC-AR5IPCC-AR5 [ [3,3], www.ipcc.chwww.ipcc.ch ](data (data available available from from the the Integrated Integrated Climate Climate data data Center-ICDC Center-ICDC of of the the University University of of Hamburg, http:http://icdc.cen.uni//icdc.cen.uni hamburg.dehamburg.de/1/daten/ocean/ar5/1/daten/ocean/ar5-slr.html).-slr.html). These These data data consist consist of the sea-levelsea-level ensemble ensemble mean mean values values and and upper upper/lower/lower 90% 90% confidence confidence bounds bounds of the of sea the level sea on level a global on a gridglobal (spatial grid (spatial resolution resolution 1 1 ), 1°×1°) obtained, obtained by adding by adding the contributions the contributions of several of geophysicalseveral geophysical sources ◦ × ◦ drivingsources long-term driving long sea-level-term sea changes:-level changes: (1) the thermosteric 1) the thermosteric/dynamic/dynamic contribution contribution (from 21 (from CMIP5 21 coupledCMIP5 coupled atmosphere-ocean atmosphere general-ocean circulation general circulati modelson AOGCMs), models AOGCMs), (2) the surface 2) massthe surface balance mass and dynamicbalance and ice sheetdynamic contributions ice sheet contributions from Greenland from and Greenland Antarctica, and (3) Antarctica, the glacier and3) the land glacier water and storage land contributions,water storage contributions, (4) the Glacial 4) Isostatic the Glacia Adjustmentl Isostatic A (GIA),djustment and (5)(GIA), the and inverse 5) the barometer inverse barometer effect [1]. Projections,effect [1]. Proje whichctions, are basedwhich on are two based diff erent on two Representative different Representative Concentration Concentration Pathways RCP Pathways 2.6 and RCP 8.5 2.6 while and providing RCP 8.5 thewhile least providing and most the amounts least of and future most sea amounts level rise, of respectively, future sea were level used. rise, Therespectively, IPCC regional were used. sea-level The rateIPCC at regional the grid sea point-level closest rate at to the the grid location point of closest the tide to t gaugehe location station of (Civitavecchia)the tide gauge station was considered. (Civitavecchia) By accounting was considered. for VLM By from accounting GPS data, for veryVLM high-resolutionfrom GPS data, DSMvery andhigh regional-resolution IPCC DSM sea and level regional projections IPCC at sea the level grid projections point corresponding at the grid to point the investigatedcorresponding area, to thethe firstinvestigated marine floodingarea, the scenarios first marine for Pyrgiflooding for 2050scenarios and 2100for Pyrgi A.D. havefor 2050 been and realized. 2100 A.D To. include have been the VLMrealized. effect To in include sea-level the projections, VLM effect we insubstituted sea-level the projections, modelled we GIA substituted contribution the to modelledthe IPCC rates GIA withcontribution the GPS vertical to the IPCC velocities, rates which with includes the GPS both vertical GIA andvelocities tectonic, which components. includes Uncertainties both GIA and for thetectonic sea-level compone estimationsnts. Uncertainties were calculated for the by sea combining-level estimations lower and were upper calculated sea level by bounds combining from lower IPCC projectionand upper andsea level errors bounds from GPS from measurements. IPCC projection In and any errors case, givenfrom GPS the tectonic measurements. stability In of any the case area,, thegiven VLM the havetectonic a null stability contribution of the area, in the the analysis. VLM have The a null relative contribution sea-level in rise the in analysis. RCP2.6 andThe RCP8.5relative scenariossea-level at rise 2050 in and RCP2. 21006 and A.D. RCP8.5 with respect scenarios to the at chosen 2050 and reference 2100 epochA.D. with 2017 respect are shown to thein Figure chosen7, andreference numerical epoch values 2017 are are shown reported in inFigure Table 72,. and numerical values are reported in Table 2.

Figure 7. Relative sea level with respect to the 2017 level as obtained from the regional IPCC sea-levelsea-level projections, AR5AR5 RCP2.6 RCP2.6 (blue (blue line), line) and, and RCP8.5 RCP8.5 (red (red line) line) for for a null a null VLM. VLM. Color Color bands bands represent represent the 90% the confidence90% confidence interval. interval. The small-scale The small variations-scale variations observed observed in the data in are the related data are to the related ocean to component the ocean contributioncomponent contribution accounting foraccounting the effects for of the dynamic effects Seaof dynamic Surface HeightSea Surface (SSH), H theeight global (SSH) thermosteric, the global SSHthermosteric anomaly, and SSH inverse anomaly barometer, and eff inverseects (Church barometer et al., 2013a, effects b, http: (Church//icdc.cen.unihamburg.de et al., 2013a, b/)., Givenhttp://icdc.cen.unihamburg.de/). the vertical tectonic stability Given of the the area, vertical the VLM tectonic have stabilit a nully contributionof the area, the in theVLM projections. have a null contribution in the projections.

J. Mar. Sci. Eng. 2020, 8, 64 10 of 18

J. Mar. Sci. Eng. 2020, 8, 64 10 of 18 Table 2. Relative Sea Level Rise (RSLR, cm) above the current mean sea level at Pyrgi for 2050 and 2100 forTable the IPCC2. Relative 2.6 and Sea 8.5 Level climatic Rise scenarios,(RSLR, cm) in above the mean the current and maximum mean sea high level tide at conditions.Pyrgi for 2050 Given and the vertical2100 for tectonic the IPCC stability 2.6 and of the8.5 climatic area, the scenarios, VLM were innot the consideredmean and inmaximum the analysis. high tide conditions. Given the vertical tectonic stability of the area, the VLM were not considered in the analysis. Sea Level 2050 IPCC 8.5 (cm) Sea Level 2050 IPCC 2.6 (cm) Sea Level 2050 IPCC 8.5 (cm) Sea Level 2050 IPCC 2.6 (cm) RSLR + Mean high RSLR + Max high RSLR + Mean high RSLR + Max high RSLR RSLR + Mean high RSLR + Max high RSLR RSLR + Mean high RSLR + Max high RSLR tide (30 cm) tide (45 cm) RSLR tide (30 cm) tide (45 cm) tide (30 cm) tide (45 cm) tide (30 cm) tide (45 cm) 17 0.07 47 62 13 43 58 17 ± ±0.07 47 62 13 43 58 SeaSea level level 2100 IPCCIPCC 8.58.5 (cm) (cm) SeaSea level level 2100 2100 IPCC IPCC 2.6 2.6 (cm) (cm) RSLR ++ MeanMean highhigh RSLRRSLR+ Max+ Max high high RSLRRSLR + Mean+ Mean high high RSLRRSLR + MaxMax high RSLRRSLR RSLR tidetide (30(30 cm) cm) tidetide (45 (45 cm) cm) tidetide (30 (30 cm) cm) tidetide (45 (45 cm) 5656 ± 0.220.22 8686 101101 30 60 60 75 ±

TheThe projected projected coastline coastline positions positions for for 2100 2100 A.D., A.D., corresponding corresponding to the to RCP2.6 the RCP2.6 and RCP8.5 and RCP8.5 scenarios, scenarios, are obtained from the DSM for the sea levels listed in Table 2 and shown in Figure 8. The are obtained from the DSM for the sea levels listed in Table2 and shown in Figure8. The computed computed and the represented scenarios correspond to the local mean sea level (estimated with the and the represented scenarios correspond to the local mean sea level (estimated with the uncertainty of uncertainty of ±11.8 cm), and are obtained by neglecting the periodical contribution due to diurnal 11.8 cm), and are obtained by neglecting the periodical contribution due to diurnal and semidiurnal ±and semidiurnal tides. To account for it, in order to estimate the maximum inundation scenarios, tides. To account for it, in order to estimate the maximum inundation scenarios, the time series of the time series of daily hydrometric sea levels at Civitavecchia tidal station from January 2010 to daily hydrometric sea levels at Civitavecchia tidal station from January 2010 to December 2018 were December 2018 were included in the analysis. The average half amplitude of the daily tides was includedestimated in as the high analysis. as about The 30 average cm (Figure half amplitude5b). Consequently, of the daily in the tides RCP was 8.5 estimated scenario for as high2100, asif aboutwe 30tak cme into (Figure account5b). Consequently,both the sea-level in theriseRCP (56 cm) 8.5 scenarioand the mean for 2100, daily if tides we take (30 intocm), accountwe can infer both a the sea-levelmaximum rise water (56 cm) level and of 86 the cm. mean Since daily the tideshighest (30 sea cm), levels we canmay infer reach a maximum45 cm for a water few hours level a of year 86 cm. Sinceduring the temporary highest sea met levelseorological may reach conditions 45 cm for when a few a hours decrease a year in atmospheric during temporary pressure meteorological and wind conditionspush water when froma the decrease sea onto in the atmospheric land (Figure pressure 5b), the and maximum wind push expected water water from level the sea may onto reach the up land (Figureto 101 5cmb), a thebove maximum the current expected level. water level may reach up to 101 cm above the current level.

FigureFigure 8.8. Projected coastline coastline positions positions for for 2100 2100 A.D. A.D. for forthe the2.6 and 2.6 and8.5 IPCC 8.5 IPCC scenarios scenarios (see legend (see legend for fordetails). details). Coastline Coastline positions positions include include the the uncertainty uncertainty of ±11.8 of 11.8 cm cm in the in the reference reference mean mean sea sea level level ± positionposition as as estimatedestimated fromfrom the analysis of of tide tide gauge gauge data. data. The The cross cross sections sections along along the the shore shore are are also also shownshown in in the the figure.figure.

J. Mar. Sci. Eng. 2020, 8, 64 12 of 18

Table 4. Return values RT (years) of the significant wave height HS (m) and the related peak wave period TP (s).

RT (Years) HS (m) TP (s) 1 5.28 9.80 5 6.63 10.98 10 7.19 11.43 20 7.75 11.87 50 8.48 12.41 100 9.02 12.80 500 10.27 13.66 1000 10.80 14.01

J. Mar. Sci. Eng. 2020, 8, 64 11 of 18 The climate wave data for ordinary and extreme storm wave conditions in terms of the return period (RT) for RT = 1 year and RT = 100 years in a sea level rise condition for 2100 in the RCP8.5 climaticFor thescenario RCP 2.6reported scenario in forTables 2100, 2 consideringand 3 were aused sea-level to estimate rise of 30wave cm andsetup the and same wave mean run daily-up tidesduring (30 the cm), event a maximum for both sea-levelreturn times rise of(Figure 60 cm 1 can0). be inferred. During the rare maximum highest sea levelsIn (45 the cm) present (Figure analysis5b), the, maximumwind set up expected was neglected water level and would the precautionary reach 75 cm aboveassumption the 2019 that level. the directionThe seaof wave level risetravel scenarios does not for produce 2100 A.D. refraction depicts the was impact considered of the marine. The ex floodingpression for derived the coast by surroundingWeggel [41] and the castlethe relation of Santa of Severa, Komar the and temples Gaughan of Pyrgi, estimated andnearby breaker beaches. depth index The cross and sectionsbreaker shownheight index in Figure [42].9 tracedThe breaker along type a direction was correlated that is normal with the to thesurf coast similarity and across parameter the archaeological, which allows areathe run of Pyrgi,-up estimation which highlights by the theuse presence of the pred of aictive soft dune equations system [43,44]. running Based parallel on to our the topographic shore. This systemsurveys, is a well mean developed beach slope south of of3.7% the for castle the for investigated the whole baycoast, of was Pyrgi, assumed while, infor the the north wave side, run it-up is buriedassessment. by the modern buildings of the Santa Severa village.

Figure 9. Return t timeime (years) values of th thee significantsignificant wave heights (m).

Since the Etruscan time (2450 years BP), the coastline has retreated at about 80 m, at a mean rate of 3.3 cm/yr [38]. The temple of Pyrgi was originally placed at more than 120–125 m from the present-day shore and well protected from the sea by the dune system. Today, it is placed at about 40–45 m from the coastline and is undergoing water intrusion and frequent flooding, especially during storm surges (Figures8 and9). The proximity of the sea and the accelerated erosion of the dune system, which is highly exposed to waves especially during storm surges, are causing a continuous coastal erosion and beach retreat. Based on our sea level rise projections, a beach retreat of about 10 m is expected for 2100

A.D. (see cross sections B, C, and D in Figures8 and9). This rapid retreat will accelerate the erosion process and the likely dismantling of the soft dune system with the consequent direct exposure of the temples to the sea. At the same time, the beach located along the north-west side of the castle, which is lacking a dune system and characterized by a gentle slope, is expected to retreat up to about 25 m (see cross section A), exposing the foot of the castle to severe erosion and flooding. Lastly, Table3 shows the expected flooded areas in 2050 and 2100 A.D. for the RCP 8.5 and RCP 2.6 climatic scenarios, together with the corresponding percentage of beach loss, which may reach up to 46% during the highest tides. J. Mar. Sci. Eng. 2020, 8, 64 12 of 18

Table 3. IPCC scenarios, relative sea level rise projections for 2100, expected land loss (m2), and percentage of land loss with respect to the current surface for the given Relative Sea Level Rise (RSLR) projections. Given the vertical tectonic stability of the area, the VLM were considered in the analysis.

Percent of Land Loss IPCC Scenario RSLR 2100 (cm) Land Loss (m2) (16113.55 m2) IPCC 2.6 17 0.07 2033.33 12.6 ± IPCC 2.6 + max high tide 75 0.22 5387.87 33.4 ± IPCC 8.5 56 0.22 3790.47 23.5 ± IPCC 8.5 + max high tide 101 0.22 7480.02 46.4 ±

8. Storm Surge Scenarios To evaluate the storm surge scenario for Pyrgi, a local climate wave assessment was performed using the Weibull distribution and the Equivalent Triangular Storm model originally proposed by Boccotti [39]. The return values of the significant wave heights and related peak wave periods (Table4) have been evaluated on the parameters of the Weibull distribution [40] using data collected at the wave buoy of Ponza (wave data are freely available at http://dati.isprambiente.it/), which is located in the central Tyrrhenian sea in front of the coasts of Latium (at coordinates Latitude 40.867◦N and Longitude 12.950◦E).

Table 4. Return values RT (years) of the significant wave height HS (m) and the related peak wave period TP (s).

RT (Years) HS (m) TP (s) 1 5.28 9.80 5 6.63 10.98 10 7.19 11.43 20 7.75 11.87 50 8.48 12.41 100 9.02 12.80 500 10.27 13.66 1000 10.80 14.01

The climate wave data for ordinary and extreme storm wave conditions in terms of the return period (RT) for RT = 1 year and RT = 100 years in a sea level rise condition for 2100 in the RCP8.5 climatic scenario reported in Tables2 and3 were used to estimate wave setup and wave run-up during the event for both return times (Figure 10). In the present analysis, wind set up was neglected and the precautionary assumption that the direction of wave travel does not produce refraction was considered. The expression derived by Weggel [41] and the relation of Komar and Gaughan estimated breaker depth index and breaker height index [42]. The breaker type was correlated with the surf similarity parameter, which allows the run-up estimation by the use of the predictive equations [43,44]. Based on our topographic surveys, a mean beach slope of 3.7% for the investigated coast, was assumed for the wave run-up assessment. J. Mar. Sci. Eng. 2020, 8, 64 13 of 18 J. Mar. Sci. Eng. 2020, 8, 64 13 of 18

(a)

(b)

FigureFigure 10. 10.Projected Projected flooded flooded areas areas corresponding corresponding to the to maximum the maximum run-up run levels-up forlevels storm for surgesstorm withsurges returnwith times return for: times (a) 1 for: year (a (in) 1 blueyear at (in 3.37 blue m) at and 3.37 (b m)) 100 and years (b) (in100 blue years at (in 5.76 blue m). at The 5.76 coastline m). The position coastline onposition 26 April on 2019 26 isApril in black, 2019 whichis in black is estimated, which withis estimated an uncertainty with an of uncertainty11.8 cm from of ±11.8 the analysis cm from of the ± tideanalysis gauge data.of tide gauge data.

TheThe results results of of the the wave wave run-up run-up analysis analysis in in ordinary ordinary and and extreme extreme storm storm conditions conditions (Table (Table5 and 5 and FigureFigure 11 ),11), were were estimated estimated at 3.37at 3.37 m andm and 5.76 5.76 m form for return return time time of 1of and 1 and 100 100 years, years respectively., respectively.

J. Mar. Sci. Eng. 2020, 8, 64 14 of 18

Table 5. Values of maximum Rmax and medium Rmed wave run-up (m) estimated in the offshore wave conditions (HS and TP), beach slope (s m/m), breaking depth db (m), and self-similarity parameter (ξ0).

RT (Years) Hs (m) Tp (sec) s(m/m) db (m) ξ0 Rmax (m) Rmed (m) 1 5.28 9.80 0.037 6.27 0.187 3.37 1.46 100 9.02 12.80 0.037 10.71 0.187 5.76 2.50 J. Mar. Sci. Eng. 2020, 8, 64 14 of 18

FigureFigure 11. 11. Cross sections sections of of the the coastline. coastline. See SeeFigure Figure 7 for7 crossfor cross section section positions. positions. A-A’ (Beach A-A’ North (Beach– North–Castle)Castle) north northof the of castle the castleof Santa of Santa Severa; Severa; B-B’, B-B’,C-C’ C-C’and D and-D’ D-D’(Beach (Beach South South–Temples)–Temples) across across the thebeach beach and and the thethree three structures structures of th ofe temple the temple of Pyrgi. of Pyrgi. The expected The expected sea levels sea for levels 2100 forin the 2100 RCP8.5 in the RCP8.5scenario scenario and in andthe inmaximum the maximum RCP8.5 RCP8.5 high tide high condition, tide condition, are shown are shownby the yellow by the yellowand red and lines, red lines,respectively. respectively. Light Light blue is blue the iscurrent the current mean meansea level sea position. level position. For storm For surge storm scenarios surge scenarios for return for returntime 1 time and 1100 and years 100 in years sea level in sea rise level conditions, rise conditions, the area theis almost area is entirely almost flooded entirely with flooded the poss withible the possibleexception exception across across D-D’, D-D’,which which remains remains protected protected by the by the high high dune dune system system.. Projected Projected coastlines coastlines includeinclude the the uncertainty uncertainty of of 11.8±11.8 cm cm in in the the mean mean sea sea level level position position as as estimated estimated from from the the analysis analysis of tideof ± gaugetide gauge data. data.

9. DiscussionTable 5. Values and Conclusions of maximum Rmax and medium Rmed wave run-up (m) estimated in the offshore wave conditions (HS and TP), beach slope (s m/m), breaking depth db (m), and self-similarity parameter In this study, we assessed the marine flooding scenarios for 2050 and 2100 for the relevant heritage (ξ0). sites of Pyrgi on the basis of (a) a high-resolution DSM, (b) rates of VLM from GPS data, and (c) tidal data fromRT the(Years) nearby H tides (m) gauge Tp and (sec) the RCPs(m/m) 2.6 anddb 8.5 (m) climatic ξ0 scenariosRmax (m) released Rmed by (m) the IPCC. The results highlight1 how5.28 the flooding9.80 might0.037 impact both6.27 the beach0.187 and archaeological3.37 1.46 sites with a potential local100 sea-level 9.02 rise for the12.80 IPCC 8.5 scenario0.037 of10.71 about 560.187 cm for 21005.76 A.D. During2.50 the highest sea levels estimated by the tidal data at the nearby Civitavecchia station, the sea-level could potentially reach9. Discussion more than and 86 Conclusions cm and up to 101 cm. In addition, 2450 years ago, the shoreline was extended seawardIn ofthis about study, 80 mwe more assessed than the today marine while, flooding during thescenarios ancient for Roman 2050 period,and 2100 the for sea the level relevant was at aboutheritage 1.2 msites below of Pyrgi the current on the position,basis of (a) as a estimated high-resolution by the fishDSM, tank (b) datarates [34of ].VLM Hence, from a continuousGPS data, seaand level (c) risetidal is data occurring from the since nearby the last tide millennia gauge and with the an RCP acceleration 2.6 and 8.5 that climatic started scenarios about 100 released53 years by ± ago,the atIPCC. thebeginning The results of highlight the Industrial how the Era flooding [25]. Due might to the impac potentialt both significant the beach impactsand archaeological on both the sites with a potential local sea-level rise for the IPCC 8.5 scenario of about 56 cm for 2100 A.D. During the highest sea levels estimated by the tidal data at the nearby Civitavecchia station, the sea-level could potentially reach more than 86 cm and up to 101 cm. In addition, 2450 years ago, the shoreline was extended seaward of about 80 m more than today while, during the ancient Roman period, the sea level was at about 1.2 m below the current position, as estimated by the fish tank data [34]. Hence, a continuous sea level rise is occurring since the last millennia with an acceleration that started about 100±53 years ago, at the beginning of the Industrial Era [25]. Due to the potential significant impacts on both the coast and the heritage sites of the Temple of Pyrgi and the Santa

J. Mar. Sci. Eng. 2020, 8, 64 15 of 18 coast and the heritage sites of the Temple of Pyrgi and the Santa Severa Castle with a beach retreat up to about 25 m, the expected scenario reported in this study can support adaptation plans at different time scales, which are in agreement with the Protocol on Integrated Coastal Zone Management (ICZM, https://eur-lex.europa.eu/eli/prot/2009/89/oj) in the Mediterranean. Our results detail previous studies for the Italian [7–9,12,45–47] and the Mediterranean [10,48] regions and can contribute to raise awareness of policymakers and heritage managers toward the coastal hazard by highlighting the need to adapt actions to protect Pyrgi from marine flooding and erosion under the current conditions due to the expected sea level rise and storm surge scenarios. We remark that the unceasing coastal erosion since the Etruscan time [38] and the retreat of the soft cliffs characterizing the Pyrgi coastline mostly occurs during high energy marine events. During storm surges, the waves approaching the coast from the Northwest, West, and Southwest marine sectors are particularly dangerous for Pyrgi. The waves approaching the coast with fetches up to about 400 km are weakly slowed down by the seafloor morphology, which results in an increase of wave energy in sea level rise conditions [49] and leads to a maximum wave run-up of 5.76 m for a return time of 100 years. The impact of extreme events may be significantly amplified by the effects of climate change. In this scenario, the intensities of the strongest future storms will exceed the strength of any in the past in the Mediterranean Sea and the oceans [50–53]. The effects of the sea-level rise in the projected scenarios include severe coastal erosion and potential enhanced damages to the heritage site of Pyrgi, which, as many other coastal areas of the Mediterranean, deserves rapid actions for their future preservation.

Author Contributions: Conceptualization, M.A. and A.V. Methodology, M.A. Software; validation, M.A., A.V., and M.G. Formal analysis, A.V., F.D., E.S., M.G., and G.M. Investigation, L.P., R.C., M.A., and F.D. Resources, M.A. Data curation, M.A., A.V., F.D., and M.G. Writing—original draft preparation, M.A., A.V., and F.E. Writing—review and editing, M.A. and M.G. Visualization, M.A. and L.P. Supervision, M.A. Project administration, M.A. Funding acquisition, M.A. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Acknowledgments: The SAVEMEDCOASTS Project, funded by the EU (Agreement Number: ECHO/SUB/2016/742473/PREV16, coordinator Marco Anzidei) and Istituto Nazionale di Geofisica e Vulcanologia partially supported this study. We are thankful to Vincenzo Sepe of the INGV, for his contribution during preliminary GPS/RTK surveys. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Church, J.A.; Clark, P.U.; Cazenave, A.; Gregory, J.M.; Jevrejeva, S.; Levermann, A.; Merrifield, M.A.; Milne, G.A.; Nerem, R.S.; Nunn, P.D.; et al. Sea-level rise by 2100. Science 2013, 342, 1445. [CrossRef] [PubMed] 2. Kopp, R.E.; Kemp, A.C.; Bittermann, K.; Horton, B.P.; Donnelly, J.D.; Gehrels, W.R.; Hay, C.C.; Mitrovica, J.X.; Morrow, E.D.; Rahmstorf, S. Temperature-driven global sea-level variability in the Common Era. Proc. Natl. Acad. Sci. USA 2016.[CrossRef][PubMed] 3. Church, J.A.; Clark, P.U.; Cazenave, A.; Gregory, J.; Jevrejeva, S.; Levermann, A.; Merrifield, M.; Milne, G.; Nerem, R.S.; Nunn, P.; et al. Sea Level Change Supplementary Material. In Climate Change 2013: The Physical Science Basis; Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change; Stocker, T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M., Eds.; Cambridge University Press: Cambridge, UK; New York, NY, USA, 2013. 4. Bamber, J.L.; Oppenheimer, M.; Koppd, R.E.; Aspinallf, W.P.; Cooke, R.M. Ice Sheet Contributions to Future Sea-Level Rise from Structured Expert Judgment. 2019. Available online: www.pnas.org/cgi/doi/10.1073/ pnas.1817205116 (accessed on 8 April 2019). 5. Veermer, M.; Rahmstorf, S. Global sea level linked to global temperature. PNAS 2009, 106, 21527–21532. [CrossRef][PubMed] 6. Casella, E.; Rovere, A.; Pedroncini, A.; Mucerino, L.; Casella, M.; Cusati, L.A.; Vacchi, M.; Ferrari, M.; Firpo, M. Study of wave runup using numerical models and low altitude aerial photogrammetry: A tool for coastal management. Estuar. Coast. Shelf Sci. 2014, 149, 160–167. [CrossRef] J. Mar. Sci. Eng. 2020, 8, 64 16 of 18

7. Antonioli, F.; Anzidei, M.; Amorosi, A.; Lo Presti, V.; Mastronuzzi, G.; Deiana, G.; De Falco, G.; Fontana, A.; Fontolan, G.; Lisco, S.; et al. Sea-level rise and potential drowning of the Italian coastal plains: Flooding risk scenarios for 2100. Quat. Sci. Rev. 2017, 158, 29–43. [CrossRef] 8. Lambeck, K.; Antonioli, F.; Anzidei, M.; Ferranti, L.; Leoni, G.; Scicchitano, G. and Silenzi, S. Sea level change along the Italian coast during the Holocene and projections for the future. Quat. Int. 2011, 232, 250–257. [CrossRef] 9. Ravanelli, R.; Riguzzi, F.; Anzidei, M.; Vecchio, A.; Nigro, L.; Spagnoli, F.; Crespi, M. Sea level rise scenario for 2100 AD for the archaeological site of Motya. Rend. Lincei. Sci. Fis. e Nat. 2019.[CrossRef] 10. Anzidei, M.; Lambeck, K.; Antonioli, F.; Furlani, S.; Mastronuzzi, G.; Serpelloni, E.; Vannucci, G. Coastal structure, sea-level changes and vertical motion of the land in the Mediterranean. Geol. Soc. Lond. Spec. Publ. 2014, 388, 453–479. [CrossRef] 11. Colomina, I.; Molina, P. Unmanned aerial systems for photogrammetry and remote sensing: A review. ISPRS J. Photogramm. Remote Sens. 2014, 92, 79–97. [CrossRef] 12. Anzidei, M.; Bosman, A.; Carluccio, R.; Casalbore, D.; D’Ajello Caracciolo, F.; Esposito, A.; Nicolosi, I.; Pietrantonio, G.; Vecchio, A.; Carmisciano, C.; et al. Flooding scenarios due to land subsidence and sea-level rise: A case study for Lipari Island (Italy). Terra Nova 2017, 29, 44–51. [CrossRef] 13. Enei, F. Santa Severa tra Leggenda e Realtà Storica. Pyrgi e il Castello di Santa Severa Alla Luce Delle Recenti Scoperte; WordPress: Grotte di Castro, Municipality of , Italy, 2013; pp. 1–413. 14. , G. Il santuario di Pyrgi dalle origini mitistoriche agli altorilievi frontonali dei Sette e di Leucotea; Scienze dell’Antichità; Poligrafico dello Stato: Roma, Italy, 2000. 15. Enei, F. Pyrgi Sommersa. Ricognizioni archeologiche subacquee nel porto dell’antica Caere; Museo civico Santa Marinella: Santa Marinella, Italy, 2008; pp. 1–115. 16. Enei, F. Dal sito di Pyrgi, antico porto di Caere, nuovi dati per lo studio della linea di costa di epoca etrusca; Serra, F., Ed.; Archaeologia Maritima Mediterranea: Pisa, Italy, 2013; pp. 165–176. 17. Mattei, M.; Conticelli, S.; Giordano, G. The Tyrrhenian margin geological setting: From the Apennine orogeny to the K-rich volcanism. In The Colli Albani Volcano; Special Publications of IAVCEI; Funiciello, R., Giordano, G., Eds.; Geological Society: London, UK, 2010; Volume 3, pp. 7–27. 18. Sestini, G.; Bruni, P.; Sagri, M. The Flysch Basins of the Northern Apennines: A Review of Facies and of Cretaceous-Neogene Evolution. Memorie della Società Geologica Italiana 1986, 31, 87–106. 19. Abbate, E.; Sagri, M. The eugeosynclinal sequences. Sediment. Geol. 1970, 4, 251–340. [CrossRef] 20. Alberti, A.; Bertini, M.; Del Bono, G.L.; Nappi, G.; Salvati, L. Note illustrative della Carta Geologica d’Italia alla scala 1: 100000; -Civitavecchia Fogli 136–142; Poligrafica & Cartevalori: Ercolano, Italy, 1970. 21. Chiocchini, U.; Gisotti, G.; Macioce, A.; Manna, F.; Bolasco, A.; Lucarini, C.; Patrizi, G.M. Environmental geology problems in the Tyrrhenian coastal area of Santa Marinella, , central Italy. Environ. Geol. 1997, 3, 1–8. [CrossRef] 22. Ferranti, L.; Antonioli, F.; Mauz, B.; Amorosi, A.; Dai Pra, G.; Mastronuzzi, G.; Monaco, C.; Orrù, P.; Pappalardo, M.; Radtke, U.; et al. Markers of the last interglacial sea-level high stand along the coast of Italy: Tectonic implications. Quat. Int. 2006, 145, 30–54. [CrossRef] 23. Ferranti, L.; Antonioli, F.; Anzidei, M.; Monaco, C.; Stocchi, P. The timescale and spatial extent of vertical tectonic motions in Italy: Insights from coastal tectonic studies. Rendiconti Online Societa Geologica Italiana 2010, 11, 683–684. 24. Karner, D.; Marra, F.; Florindo, F.; Boschi, E. Pulsed uplift estimated from terrace elevations in the coast of Rome: Evidence for a new phase of volcanic activity? Earth Planet. Sci. Lett. 2001, 188, 135–148. [CrossRef] 25. Lambeck, K.; Anzidei, M.; Antonioli, F.; Benini, A.; Esposito, A. Sea level in Roman time in the Central Mediterranean and implications for recent change. Earth Planet. Sci. Lett. 2004, 224, 563–575. [CrossRef] 26. Avallone, A.; Selvaggi, G.; D’Anastasio, E.; D’Agostino, N.; Pietrantonio, G.; Riguzzi, F.; Serpelloni, E.; Anzidei, M.; Casula, G.; Cecere, G.; et al. The RING network: Improvement of a GPS velocity field in the central Mediterranean. Ann. Geophys. 2010, 53, 39–54. 27. Ullman, S. The interpretation of structure from motion. Proc. R. Soc. Lond. Ser. B 1979, 203, 405–426. 28. Sammari, C.; Koutitonsky, V.G.; Moussa, M. Sea level variability and tidal resonance in the Gulf of Gabes, Tunisia. Cont. Shelf Res. 2006, 26, 338–350. [CrossRef] J. Mar. Sci. Eng. 2020, 8, 64 17 of 18

29. Lambeck, K.; Woodroffe, C.D.; Antonioli, F.; Anzidei, M.; Geherls, W.R.; Laborel, J.; Wright, A. Palaeoenvironmental records, geophysical modelling and reconstruction of sea-level trends and variability on centennial and longer time scales. In Understanding Sea Level Rise and Variability; Wiley-Blackwell: Oxford, UK, 2010; pp. 61–121. 30. Antonioli, F.; Lo Presti, V.; Anzidei, M.; Deiana, G.; de Sabata, E.; Ferranti, L.; Furlani, S.; Mastronuzzi, G.; Orru, P.E.; Pagliarulo, R.; et al. Tidal notches in Mediterranean Sea: A comprehensive analysis. Quat. Sci. Rev. 2015, 119, 66–84. [CrossRef] 31. Corchete, V. The high-resolution gravimetric geoid of Italy: ITG2009. J. Afr. Earth Sci. 2010, 58, 580–584. [CrossRef] 32. Serpelloni, E.; Faccenna, C.; Spada, G.; Dong, D.; Williams, S.D.P. Vertical GPS ground motion rates in the Euro-Mediterranean region: New evidence of velocity gradients at different spatial scales along the Nubia-Eurasia plate boundary. J. Geophys. Res. Solid Earth 2013, 118, 6003–6024. [CrossRef] 33. Serpelloni, E.; Pintori, F.; Gualandi, A.; Scoccimarro, E.; Cavaliere, A.; Anderlini, L.; Belardinelli, M.E.; Todesco, M. Hydrologically Induced Karst Deformation: Insights from GPS Measurements in the -Eurasia Plate Boundary Zone. J. Geophys. Res. Solid Earth 2017, 123, 4413–4430. [CrossRef] 34. Herring, T.; King, R.W.; McClusky, S. GAMIT Reference Manual, Release 10.4; Massachussetts Institute of Technology: Cambridge, MA, USA, 2010. 35. Altamimi, Z.; Collilieux, X.; Métivier, L. ITRF2008: An improved solution of the international terrestrial reference frame. J. Geod. 2011, 85, 457. [CrossRef] 36. Devoti, R.; D’Agostino, N.; Serpelloni, E.; Pietrantonio, G.; Riguzzi, F.; Avallone, A.; Cavaliere, A.; Cheloni, D.; Cecere, G.; D’Ambrosio, C.; et al. The mediterranean crustal motion map compiled at INGV. Ann Geophys. 2017.[CrossRef] 37. Guidoboni, E.; Comastri, A.; Traina, G. Catalogue of Ancient Earthquakes in the Mediterranean Area up to the 10th Century; Istituto Nazionale di Geofisica: Roma, Italy, 1994. 38. Rovere, A.; Antonioli, F.; Enei, F.; Giorgi, S. Relative sea level change at the archaeological site of Pyrgi (Santa Severa, Rome) during the last seven millennia. Quat. Int. 2011, 232, 82–91. [CrossRef] 39. Boccotti, P. Wave Mechanics for Ocean Engineering; Elsevier Science: Amsterdam, The Netherlands, 2000; Volume 64. 40. Arena, F.; Laface, V.; Malara, G.; Romolo, A.; Viviano, A.; Fiamma, V.; Sannino, G.; Carillo, A. Wave climate analysis for the design of wave energy harvesters in the Mediterranean Sea. Renew. Energy 2015, 77, 125–141. [CrossRef] 41. Weggel, J.R. Maximum breaker height. J. Waterw. Harb. Coast. Eng. Div. 1972, 98, 529–548. 42. Komar, P.D.; Gaughan, M.K. Airy wave theory and breaker height prediction. In Proceedings of the 13th Conference on Coastal Engineering, Vancouver, BC, Canada, 29 January 1972. [CrossRef] 43. U.S. Army Corps of Engineers. Coastal Engineering Manual, Part II; U.S. Army Corps of Engineers: Washington, DC, USA, 2006. 44. Mase, H. Random wave run-up height on gentle slope. J. Waterw. Port Coast. Ocean Eng. 1989, 115, 649–661. [CrossRef] 45. Lambeck, K.; Anzidei, M.; Antonioli, F.; Benini, A.; Verrubbi, V. Tyrrhenian sea level at 2000 BP: Evidence from Roman age fish tanks and their geological calibration. Rend. Fis. Acc. Lincei 2018, 29, 69–80. [CrossRef] 46. Anzidei, M.; Scicchitano, G.; Tarascio, S.; De Guidi, G.; Monaco, C.; Barreca, G.; Mazza, G.; Serpelloni, E.; Vecchio, A. Coastal retreat and marine flooding scenario for 2100: A case study along the coast of Maddalena Peninsula (southeastern ). Geogr. Fis. Dinam. Quat. 2018, 41, 5–16. [CrossRef] 47. Vecchio,A.; Anzidei, M.; Serpelloni, E.; Florindo, F.Natural Variability and VerticalLand Motion Contributions in the Mediterranean Sea-Level Records over the Last Two Centuries and Projections for 2100. Water 2019, 11, 148. [CrossRef] 48. Reimann, L.; Vafeidis, A.T.; Brown, S.; Hinkel, J.; Tol, R.S.J. Mediterranean UNESCO World Heritage at risk from coastal flooding and erosion due to sea-level rise. Nat. Comm. 2018, 9, 4161. [CrossRef] 49. Masselink, G.; Hughes, M.G. Introduction to Coastal Processes and Geomorphology; McCann, S.B., Ed.; Edward Arnold: London, UK, 1980; p. 354. 50. Sobel, A.H.; Camargo, S.J.; Hall, T.M.; Lee, C.Y.; Tippett, M.K.; Wing, A.A. Human influence on tropical cyclone intensity. Science 2016, 353, 242–246. [CrossRef] 51. Rahmstorf, S. Rising hazard of storm-surge flooding. PNAS 2017, 114, 45. [CrossRef] J. Mar. Sci. Eng. 2020, 8, 64 18 of 18

52. Menéndez, M.; Woodworth, P. Changes in extreme high water levels based on a quasi-global tide-gauge data set. J. Geophys. Res. 2010, 115, C10011. [CrossRef] 53. Lowe, J.A.; Gregory, J.M. The effects of climate change on storm surges around the United Kingdom. Philos. Trans. R. Soc. 2005, 363.[CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).