Development and decline of the ancient harbor of Neapolis

Received: 28 December 2016 Revised: 27 June 2017 Accepted: 29 June 2017 DOI: 10.1002/gea.21673 RESEARCH ARTICLE Development and decline of the ancient harbor of Neapolis Valentino Di Donato1 Daniela Giampaola2 Maria R. Ruello1 Mauro A. Di Vito3 Viviana Liuzza1 Vittoria Carsana2 Christophe Morhange4 Aldo Cinque1 Elda Russo Ermolli1 1 Dipartimento di Scienze della Terra, dell'Ambiente e delle Risorse, Università di Napoli Federico II, Naples, Italy Abstract Archaeological excavations, undertaken since 2004 for the construction of the new Naples sub- 2 Soprintendenza Archeologia Belle Arti e way, have unearthed the harbor basin of the Greco–Roman town of Parthenope–Neapolis, fur- Paesaggio per il Comune di Napoli, Naples, Italy nishing scientists with the opportunity to recover abundant archaeological remains and a thick 3 Istituto Nazionale di Geofisica e Vulcanologia, sezione Osservatorio Vesuviano, Naples, Italy 4 CNRS, IRD, Coll France, CEREGE, Université succession of diverse infill sediments. The latter underwent sedimentological, paleontological, and volcanological analyses. Compositional data analysis, applied to all three data sets, highlighted three main paleoenvironmental changes in the harbor basin from the Augustan Age up to the 6th Aix Marseille, Aix-en-Provence, France century A.D. The beginning of harbor activity is recorded during the 3rd century B.C. when sed- Correspondence Russo Ermolli, Dipartimento di Scienze della Terra, dell'Ambiente e delle Risorse, Università di Napoli Federico II, Naples, Italy. Email: [email protected] imentation was interrupted by intensive dredging of the sea-bottom. The impact of the A.D. 79 Paola Romano, in memorian to coastal progradation. The final closure of this part of the bay occurred at the end of the 5th to Scientific editing by Jamie Woodward the beginning of the 6th century A.D., due to increased alluvial input linked to both natural and Vesuvius eruption, recorded for the first time in the Neapolitan territory, led to a reduction in Posidonia meadows and to an ensuing phase of more restricted water circulation and pollution. At the beginning of the 5th century A.D., an open lagoon environment was established, attesting anthropogenic causes. KEYWORDS A.D. 79 eruption, compositional data analysis, Greco–Roman period, molluscs, geoarchaeology, Naples 1 INTRODUCTION history of harbor activity from the 3rd century B.C. to the 5th century A.D. Morphostratigraphy and pollen analysis revealed the natural and During the last 15 years, geoarchaeological investigations associated cultural landscapes of the harbor catchment from the Greco–Roman with the construction of the new Naples subway in Italy have shed up to the Late Roman periods, highlighting the presence of mixed oak new light on the presence and configuration of the Greco–Roman har- woods and tree crops (walnut, chestnut, and grapevines) on the slopes bor and its surroundings in an excavation area of more than 40,000 surrounding the town and of vegetable gardens with cabbages around m2 . This study offered the possibility to collect large amounts of the harbor area (Russo Ermolli, Romano, Ruello, & Barone Lamuga, stratigraphic and biosedimentological data, which have significantly 2014). improved our knowledge about the history of the Greco–Roman town Since 2012, the opening of new excavation areas (Line 6 and Area 4) in Municipio Square provided the opportunity to further improve and the evolution of its coastal landscape. The main phases of ancient harbor activity and the reconstruc- geoarchaeological knowledge regarding the history of the ancient tion of shoreline fluctuations have been debated in numerous publi- harbor and its paleoenvironmental evolution between the Hellenistic cations (Amato et al., 2009; Carsana et al., 2009; Cinque et al., 2011; period and 5th century A.D. This goal has been achieved through a Giampaola & Carsana, 2010; Giampaola et al., 2006; Ruello, 2008). In multimethodological approach, including sedimentology, archaeology, particular, the results of geoarchaeological field surveys in the excava- macro- and micropaleontology, volcanology, and geostatistics. tion area of Line 1, carried out between 2002 and 2007, confirmed the The geoarchaeological study of the ancient harbor of Naples is a hypothesis concerning the presence of a Greco–Roman harbor in the recent example of a new area of inquiry that considers ancient har- modern Municipio Square (Capasso, 1895). Investigation of the ancient bors as exceptional base-level archives in which the sedimentolog- harbor stratigraphic sequences allowed us to reconstruct a continuous ical record and the environment are the result of both geological Geoarchaeology. 2018;1–16. wileyonlinelibrary.com/journal/gea c 2018 Wiley Periodicals, Inc. ⃝ 1 2 DI DONATO ET AL . factors. The landscape of Naples is therefore primarily characterized by landforms (such as positive and negative volcanic forms and their remnants) linked to the late Quaternary activity of the PF (last eruption Monte Nuovo, A.D. 1538). In particular, the landscape is the result of the mantling of preexisting volcanic edifices by the 15 ka Neapolitan Yellow Tuff (Deino, Orsi, De Vita, & Piochi, 2004), of the collapse of the related caldera (PF depression) and, between 15 and 3.8 ka, of the accumulation of products from at least 70 explosive eruptions of variable magnitude which generated variably dispersed pyroclastic deposits and formed tuff cones, tuff rings, and lava domes (Di Vito et al., 1999). These morphologies are cut by SW-NE and NE-SW fault scarps, widely spread in the study area (Amato et al., 2009; Cinque et al., 2011; Romano et al., 2013) due to the intense volcano-tectonics (Fig. 1) detected within the PF caldera, the Bay of Naples, and the Geological and structural setting of the Campanian Plain and location of the archaeological sites cited in the text. PF: Phlegrean fields; 1) Quaternary sedimentary deposits; 2) Quaternary volcanic rocks and sediments; 3) pre-Quaternary units; 4) main faults FIGURE 1 Campania Plain (Bruno, Rapolla, & Di Fiore, 2003; Cinque et al., 2011; Di Vito et al., 1999). A sequence of depositional and erosional processes, accompanied by fault re-activation, further outline the morphology of this area, together with widespread anthropogenic deposits, reworked by a long-term human activity. The Naples area was also affected by deposition of two Plinian processes and human activity (e.g., Marriner & Morhange, 2007; Mar- eruptions of Vesuvius: the Avellino eruption during the Early Bronze riner, Morhange, Flaux, & Carayon, 2017; Morhange et al., 2003, 2015, Age (Di Vito et al., 2009 and references therein) and the Pomici di 2016; Sadori et al., 2015). In the present study, we adopt composi- Pompei eruption in A.D. 79 (Sigurdsson, Cashdollar, & Sparkes, 1982). tional data analysis methods (CoDA; Aitchison, 1986) to probe the These two eruptions were characterized by similar dynamics, with sedimentological and paleontological data. Despite the advanced sta- sequences of magmatic phases dominating the main part of the erup- tistical framework which is now available to analyze compositional tions and minor phreatomagmatic explosions occurring in the last data (Pawlowsky-Glahn & Buccianti, 2011), until now relatively few eruption phases (Cioni, Bertagnini, Santacroce, & Andronico, 2008). studies have adopted this approach on proxies for paleoecological The magmatic phases are characterized by deposition of widely dis- reconstructions (Di Donato, Martin-Fernandez, Daunis-i-Estadella, & persed pumice fallout sediments distributed in the eastern sectors of Esposito, 2015; Kaniewski et al., 2013; Rossi et al., 2015; Sgarrella, Di the Somma-Vesuvius, whereas the phreatomagmatic phases generated Donato, & Sprovieri, 2012). Sedimentological and paleontological data pyroclastic density currents forming thinly laminated-, plane parallel to acquired in the new archeological excavation area of Municipio Square dune bedded-, or massive-ash beds distributed in the plains surround- (Liuzza, 2014) provide the opportunity to carry out a paleoenviron- ing the volcano (Cioni et al., 2008; Di Vito, Castaldo, de Vita, Bishop, & mental reconstruction of harbor sedimentation based on an integrated Vecchio, 2013). compositional approach. The coastal landscape is characterized by hilly terrains with alternating cliff promontories and small bays or narrow coastal plains. These conditions provided natural resources and protected landing 2 GEOLOGICAL AND ARCHAEOLOGICAL SETTING places favoring human settlements since prehistoric times (Amato et al., 2009; Carsana et al., 2009; Cinque et al., 2011; Romano et al., 2013). At the end ofthe 8th to beginning of the 7th century B.C., Greek The town of Naples lies in the Campania Plain, a tectonic depression colonists from Cumae founded at the foot of South Martino Hill and on located along the western Tyrrhenian margin of the Southern Apen- Megaride Island the settlement of Parthenope (Fig. 2). The necropo- nine chain (Fig. 1), which has been subsiding since the Early Pleis- lis of via Nicotera and ceramics recovered from Chiatamone and the tocene (Brancaccio et al., 1991; Cinque et al., 1997; Caiazzo et al., South Maria degli Angeli landfill, dated between the end of the 8th 2006). The urban territory currently spreads over a wide area of active and the 5th centuries B.C., are the only archaeological evidence of this volcanism and tectonics, between the eastern edge of the nested ancient center (Fig. 2). caldera of the Phlegraean Fields (PF) and the Vesuvius stratovolcano Farther toward the eastern edge of Parthenope, a new polis, Neapo- apron (Fig. 1). During the Late Pleistocene and the Holocene, the lis, was founded at the end of the 6th to beginning of the 5th cen- Naples area was affected by the deposition of thick tuff deposits and tury B.C. (D'Agostino & Giampaola, 2005) on the gently sloping area of sequences of pyroclastic fallout and density current deposits, mainly Pendino (Figs 1 and 2). Parthenope survived as a minor center named erupted by PF and subordinately by the largest Plinian eruptions of Palaepolis(old-polis) until the 4th to 3rd centuries B.C. The Neapolis Vesuvius. The geomorphological setting of the study area is the result of a complex interplay among endogenous, exogenous, and anthropogenic urban setting is well known and includes the road network (strigas), fortifications, necropolis, aqueduct, main public monuments, and villas (AA.VV., 1985; Napoli, 1959, 1967; Stazio, 1988; Zevi, 1995). 3 DI DONATO ET AL . F I G U R E 3 The sea-floor of the Imperial Age harbor with ceramics dated to the first half of the 1st century A.D. [Color figure can be viewed at wileyonlinelibrary.com] F I G U R E 2 Archaeological setting of Parthenope and Neapolis. The box indicates the harbor area as detailed in Figures 4 and 5 [Color figure can be viewed at wileyonlinelibrary.com] The excavations for Line 1 of the new Naples subway allowed reconstruction of the ancient coastal landscape, confirming the existence of a wide and sheltered bay between Neapolis and Parthenope (Fig. 2), and demonstrated that the Municipio embayment was the port since at least the 3rd century B.C. when intensive dredging of the sea bottom removed almost all the older sediments (Carsana et al., 2009; Giampaola & Carsana, 2005; Giampaola et al., 2006). The harbor was functional during the Roman Imperial age and up to the beginning of F I G U R E 4 Piazza Municipio area and plan of the excavations. The main archaeological findings of 3rd to 2nd century B.C. are indicated together with details of the dredging operations [Color figure can be viewed at wileyonlinelibrary.com] the 5th century A.D. when a lagoon environment became established in the bay. The final closure of this sector of the bay occurred during in the excavations of Line 1 (Carsana & Del Vecchio 2010; Carsana & Late Roman times due to increased sedimentation (Amato et al., 2009; Guiducci, 2013; Carsana, D'Amico, & Del Vecchio, 2007; Giampaola Carsana et al., 2009; Cinque et al., 2011). et al., 2006). The pottery, often well preserved, represents goods and ship equipment lost during loading and unloading operations in the harbor, or waste from the surrounding suburban areas (Fig. 3). Apart 3 NEW GEOARCHAEOLOGICAL DISCOVERIES from the chronological value, this large and diversified recovery provides exceptional evidence of the economy and commercial trades of the town. New geoarchaeological excavations carried out for Line 6 and Area A few pottery remains coeval with the foundation of Parthenope 4 expanded our previous knowledge about the harbor and ancient were found in the dredged marine sediments. This discovery, together coastal landscape. In particular, the ancient harbor extension was more with the recovery of sea-bottom deposits saved from the dredging precisely delineated in a suburban position, not far from the Neapolis phases, dated to the end of the 6th to beginning of the 5th century fortifications. The large inlet hosting the harbor extended from Piazza B.C., is perhaps evidence for the use of this bay as a harbor during the Bovio to Piazza Municipio, limited to the east by the relief of South Archaic period. The existence of the harbor basin is certainly attested Maria di Porto Salvo and to the west by the hill of Castel Nuovo. by archeological evidence from the 3rd to 2nd centuries B.C. when Archeological surveys and classification of a large number of ceram- intensive dredging phases of the sandy-mud infilling were undertaken ics (Fig. 3) ensured a very detailed chronology of the investigated sed- with the aim of deepening the sea bottom (Fig. 4). The same evidence iments, spanning from the 3rd century B.C. to the 6th century A.D. On was already identified in the excavations of Line 1 (Carsana et al., the whole, this chronology is in agreement with that already defined 2009). Important structures discovered in Line 6, such as a probable 4 DI DONATO ET AL . F I G U R E 5 The same area as shown in Figure 4 with the main Imperial Age findings. The red circles indicate the position of the investigated sections V2a, V2c and top. The dashed line represents the cross section of Figure 6A [Color figure can be viewed at wileyonlinelibrary.com] haulage ramp and containment walls for terracing the slope at the rear level (b.s.l.). The RSL was deduced from the upper limit of the marine of the western edge of the harbor (Fig. 4), are dated to the same period. biomarkers (e.g., serpulids, vermetids, balanids, ostrea, etc…) found in The survey also revealed the sea bottom stratigraphy from the Repub- the maximum concavity of the tidal notch, recovered along the quay lican to the late Imperial Age and two poorly preserved shipwrecks front (Fig. 6B). dated to the 2nd century B.C. Another two shipwrecks (F and G in The sediments that accumulated from the base of the Augustan Fig. 5), dated to the end of the 2nd century A.D., were unearthed in quay up to the beginning of the 6th century A.D. represent the main the central part of the basin. These discoveries expand upon what was object of the present work and were analyzed using sedimentologi- already unearthed in Line 1 where two shipwrecks dated to the end of cal, paleontological, and volcanic proxies. These sediments recorded the 1st century A.D. and one of the 2nd century A.D. (B) were found the history of the harbor bay, revealing both the natural and anthro- near a wooden pier dated to the end of 1st century A.D. (Carsana et al., pogenic events that impacted its environment. Ash deposits in the har- 2009; Giampaola & Carsana, 2005; Giampaola et al., 2006). bor, related to the A.D. 79 eruption, were also investigated. This is the In addition to the stratigraphy, shipwrecks, and pottery remains, the first evidence for the presence of these deposits in the urban Neapoli- new excavation unearthed part of the Augustan Age harbor infrastruc- tan territory, mainly due to the lack of stratigraphic information in a ture and important buildings located along the coastal strip. Along the densely urbanized context. The stratigraphic continuity of the studied southeastern edge of the inlet (Area 4), a pier composed of two arms sequences is only interrupted, in a very limited area immediately close built in cement mortar within a wooden formwork, extends across an to the quay, by a new dredging phase after the A.D. 79 eruption. This area of about 360 m2 (Fig. 5). It represented a complex system con- action, recognized through a series of parallel trenches (Fig. 6C), aimed ceived to create an artificial closing of the port to protect it from the at maintaining a functional water depth suitable for docking, after the southern winds. increased sediment supply due to the eruption. The southwestern part of the protected inlet is equipped with a quay unearthed in two areas, which represented the coastline profile of the ancient harbor until the early 5th century A.D. (Fig. 5). To the rear of the northern quay, a wide road develops, very probably the via per cryptam (Fig. 5), commissioned by the Emperor Augustus to connect 4 SEDIMENTOLOGICAL AND PALEOECOLOGICAL ANALYSES OF THE HARBOR SEDIMENTS Neapolis and the PF. At the beginning of the 1st century and continuing into the 2nd century A.D., thermal baths rose along its trace (Fig. 5). Sediments of Neapolis harbor dating from the Augustan Age up to The southern quay, 6.5 m wide and 3.5 m high, extends in a NW- the beginning of the 6th century A.D. were investigated by means of SE direction for 24.5 m. It is built with tuff stones in a cement mortar an integrated approach including granulometric, volcanological, micro- resting on two rows of squared blocks overlaying the top of tuffaceous and macropaleontological analyses. CoDA methods were applied to bedrock. The latter was reshaped by an artificial cut conceived for low- both granulometric and macrofaunal data, involving relative variation ering the base of the structure in order to create a higher water col- biplots (RVB; Aitchison & Greenacre, 2002), simplicial principal com- umn close to the quay (Fig. 6A). The depth was ≈3 m, based on the rela- ponents analysis (SPCA; Aitchison, 1983), and cluster analysis based tive sea-level (RSL) position at that time, which was ≈1.60 m below sea on Aitchison distances. Ostracod data were too scanty and incomplete 5 DI DONATO ET AL . F I G U R E 6 A) On the left side, the Augustan quay and artificial cut into tufaceous bedrock; in the background, the investigated infill sediments with position of the studied sections; B) frontal view of the Augustan quay with sea level position and water column as inferred from marine biomarkers; and C) general view of the Augustan quay and of the thermal baths with evidence of the drainage pipe. Dredging holes placed in front of the quay after deposition of the A.D. 79 eruptive event [Color figure can be viewed at wileyonlinelibrary.com] for this kind of approach. The paleoecological interpretation of fos- tion (see below). As in the sediments of Line 1, Posidonia oceanica sil assemblages was based on the modern ecological requirements of remains (Cennamo, Caputo, Stefano, Russo Ermolli, & Barone Lumaga, the identified mollusk taxa, as summarized in Supplemental Table 1. All 2014) are abundant throughout the sequence. methods are detailed in the Supplemental Text. 4.2 4.1 The analyzed sections Granulometric analysis Granulometric data are summarized into a ternary diagram (Folk, The analyses were carried out on two stratigraphic sections, V2a and 1954) in which gravels, sands, and muds (silt and clay) are considered as V2c, located 11 m and 35 m from the Augustan quay, respectively parts of the composition (Fig. 8A). In both sections, coarser samples are (Fig. 5). Section V2a is 3.20 m thick, extending 1.80–5.00 m b.s.l. those dated to the second half of the 1st to the 3rd centuries A.D. and (Fig. 6A). The beginning of sedimentation in this bay sector, exposed from the first half of the 5th to the beginning of the 6th century A.D. All in section V2a, directly lies above the artificial cut in the tuffaceous samples show a low degree of sorting (Fig. 7), indicating that the marine bedrock, related to the quay construction. Twenty-five samples were processes did not have the competence to rework the sediment that collected along this section, each sample corresponding to a differ- is often deposited rapidly and continuously at base level. Since sand is ent archeostratigraphic unit (AU). Section V2c is 3.05 m thick, extend- the major component of most samples, the variability of samples can ing 2.75–5.80 m b.s.l. (Fig. 6A). In this bay sector, the sediments ana- be better highlighted in a centered ternary diagram in which both sam- lyzed in section V2c lie above the last surface (2nd century B.C.) of ples and grids are centered with respect to the compositional center the dredged sea-bottom sands (Fig. 4). Moreover, the V2a–V2c tran- of the data (Eynatten, Pawlowsky-Glahn, & Egozcue, 2002; Fig. 8B). On sect is not affected by the 1st century A.D. dredging, which was lim- the basis of cluster analysis (see Supplementary Text for details), four ited to areas near the quay (Fig. 6C). Eighteen samples were collected groups of samples were distinguished (Fig. 9A). along this section and, as for section V2a, each sample corresponds to In order to define the characteristics of the clusters, an SPCA of a different AU. The sections cover a time interval spanning from the 1st granulometric data was undertaken (Fig. 9). The isometric-log-ratio century B.C. to the 6th century A.D. The uppermost interval of deposi- (ilr) (Egozcue et al., 2003) principal component 1, which accounts for tion, corresponding to the end of the 5th –beginning of 6th century A.D., most of the samples’ variability (about 83%), is almost parallel to ilr2 was removed before sampling in both sections for technical reasons axis, indicating that the main source of variability is related to rela- linked to the construction of engineering structures for the subway tive changes between gravel and mud (Fig. 10A). A secondary source station. Thus, samples covering this uppermost interval were obtained is related to relative changes of sand with respect to the other compo- from a different vertical profile (“Top” section in Fig. 7), in the south- nents. This result is also evident in the ternary diagram, showing the ern sector of the excavation site. In both sections, a 50 cm thick vol- back-transformed principal components axes (Fig. 10B). The cloud of caniclastic level was observed, with lenses of subrounded, centimet- samples is elongated in a direction forming a high angle with balance 1 ric to subcentimetric pumice lapilli, whose sedimentology and lithology axis. As shown by the confidence ellipses of Fig. 10A, samples belong- suggest that they are reworked deposits of the A.D. 79 Pompeii erup- ing to the sections V2a and V2c appear distinguished in the ilr space. 6 DI DONATO ET AL . F I G U R E 7 Granulometric data and Folk statistical parameters for log V2a (A), top section (A) and log V2c (B). Most samples show a low degree of sorting, ranging between 1 and 2𝜑 [Color figure can be viewed at wileyonlinelibrary.com] Indeed, a multivariate analysis of variance (MANOVA) performed on Towards the upper part of both sections, this assemblage shows a ilr coordinates indicates that these two groups of samples are signifi- decreasing trend and is replaced by the lagoon species Xestoleberis sp. cantly different at an alpha level of 0.05. (Fig. 11). The uppermost levels, only present in section V2a, are devoid Clusters G4 and G3 are characterized by high and low gravel to mud of ostracods. The general paucity of ostracods (i.e., 10 specimens in log-ratios, respectively. Clusters G1 and G2 are located closer to the 100 g) and their fragmentation did not allow the application of the origin of axes and are distinguished by the relative abundance of sand, CoDA approach to these data. which is higher in cluster G2 (Fig. 10B). Mollusk assemblages found in logs V2a and V2c are shown in Fig. 11. Bittium reticulatum is a fundamental component of the assemblages, 4.3 Paleontological analysis notwithstanding a decreasing trend, more marked in log V2a. In general, apart from the uppermost intervals, a large part of the assem- Ostracods show a general dominance of coastal assemblages (mainly blages is constituted by taxa related to bottom vegetation, mainly rep- Loxoconcha romboidea, Aurilia convexa, Lepthocythere rara and Pontho- resented by Posidonia oceanica meadow, as testified by its remains com- cythere elongata) all along the lower interval of both sections (Fig. 11). monly found in the sediments. Paphia aurea and Loripes lucinalis are rare 7 DI DONATO ET AL . F I G U R E 8 A) Granulometrical data reported as a gravel, sand, and mud ternary diagram (Folk, 1954). Orange dots (black in printed version): log V2c; green dots (dark grey in printed version): log V2a; blue dots (light grey in printed version): top section. Labels refer to the archaeostratigraphical units. B) The same data of diagram A, centered by perturbating the composition with the inverse of the compositional center of the dataset (see “Supplementary text”). Note that both sample points and grids are centered [Color figure can be viewed at wileyonlinelibrary.com] F I G U R E 9 A) Cluster analysis of granulometric data (Ward's method based on Aitchison distances, i.e., the Euclidean distances on ilr coordinates). The number of clusters is defined by means of Mojena index. B) Cluster analysis of mollusk assemblages [Color figure can be viewed at wileyonlinelibrary.com] or missing in the lower interval of both sections. Cerastoderma glaucum, the relationship among taxa and between taxa and samples, the three which is also rare in the lower intervals, increases in the uppermost obtained clusters, each of them further divided in two subclusters, part of both sections. were considered to distinguish row points into the RVB (Fig. 12). As with the granulometric data, mollusk assemblages were ana- The first three axes of the RVB account for about 70% of the total lyzed by means of cluster analysis (Fig. 9B). In order to better highlight variability. The relationships observed along the first axis of the biplots 8 DI DONATO ET AL . F I G U R E 1 0 A) Scatter plot of balances 1 and 2. Balance 1 expresses the log contrast between sand and other components, balance 2 the log contrast between gravel and mud (see “Supplementary text”). The continuous and dotted dark blue lines (black in printed version) are the principal components of ilr coordinates. The thin dotted lines (light blue; black in printed version) show the fields of the Folk diagram as they appear in ilr space. Colors and symbols of row points indicate the section and the clusters to which they belong; green (dark grey in printed version) samples: log V2a; orange (black in printed version) samples: log V2c.; blue (light grey in printed version) samples: top section. Section samples and clusters are contoured by 2𝜎 confidence ellipses. Dash-dot lines: sections; dashed lines: clusters. B) Simplicial principal components analysis of granulometrical data. Data are centered as in Fig. 8B. Directions of maximum variability in the ternary diagram (dark blue; black in printed version) are obtained by back transforming the principal components of ilr data shown in Fig. 10A. The first direction is related to changes of sand with respect to the other two components, the second to relative changes of gravel versus mud. Dashed and dash-dotted lines show compositional confidence ellipses (coherent with the Aitchison geometry) corresponding to those shown in Fig. 10A [Color figure can be viewed at wileyonlinelibrary.com] highlight a high variability in the log-ratios between Bittium reticulatum of a MANOVA based on the first 3 axis scores (Quinn & Keough, 2002) or Nassarius corniculus, located at the negative side, and Cerastoderma of the form biplot indicate that sample mean differences are not signif- glaucum, Cyclope neritea and Tapes decussatus, located at the positive icant at an alpha level of 0.05. side. The column point of Gibbula spp., which appears located near the origin of the biplot in a two-dimensional representation, shows an orientation toward the positive side of axis 3 when a third axis is included in the RVB. As a consequence, the links connecting its column point to other taxa form high angles with the links connecting taxa without 5 EVOLUTION OF THE ENVIRONMENTS IN THE HARBOR BAY a significant projection on axis 3 and in particular with those located near the origin of the biplot, such as Nassarius mutabilis and the group UCSA. Since the angle between the links connecting column points approximates the correlation between the corresponding log-ratios, this behavior indicates that the relative variation of Gibbula spp. with respect to the other taxa tends to be independent from their relative variations. Additional information was added by including balances 1 and 2, adopted in the sedimentological analysis, as supplementary vectors in the RVB of the mollusk assemblages. The vector of balance 2 is opposed to the Gibbula spp. column vector. This indicates that relative increases of this taxon tend to correspond to increases in the mud to gravel ratios. In effect, samples of cluster G3 (having low gravel to mud log-ratios) are mostly located near the Gibbula spp. column point, in The results of sedimentological and paleontological analyses, constrained by chronology, allowed three main paleoenvironmental units to be distinguished along the studied sections from the Augustan Age up to the Late Roman period: marine, transitional, and terrestrial (Fig. 13). This succession of environments reflects the natural evolution of a limited accommodation space, which is gradually filled with sediments (Marriner & Morhange, 2006). In general, the paleoenvironmental changes reconstructed in the two stratigraphic profiles are well correlated and show that the transition from the marine to the lagoon environment occurred at the beginning of the 5th century A.D. The transition from the lagoon to the continental environment occurred at the end of the 5th to beginning of the 6th century A.D., through a short phase of alluvial input into a shallow lagoon environment. opposition to the vector of balance 2. The vector of balance 1, which is related to increases in sand versus other components, seems related 5.1 The marine environment to high relative abundances of Cerastoderma glaucum or Cyclope neritea and low relative abundances of Bittium reticulatum. Unlike the granu- 5.1.1 lometric data, the mollusk assemblages belonging to the sections V2a In both sections, samples dated to the Augustan Age up to the first and V2c are not well differentiated in the biplot. In effect, the results half of the 1st century A.D. belong to clusters M3b (Figs 11 and 13). From the Augustan Age up to A.D. 79 9 DI DONATO ET AL . F I G U R E 1 1 Percentage distribution of selected mollusk taxa and ecological groups of ostracods in log V2a (A) and log V2c (B) [Color figure can be viewed at wileyonlinelibrary.com] As indicated by the location on the negative side of axis 1 of the RVB investigated area of the harbor was characterized by a vegetated (Posi- (Fig. 12), they are characterized by high relative abundance of Bit- donia meadows) low-energy shoreface environment, with a muddy- tium reticulatum and Nassarius corniculus. Coherently with the abun- sandy bottom. dant remains of Posidonia oceanica (Fig. 13), the mollusk and ostra- From the first half of the 1st century A.D., the growth of mollusk cod assemblages (mainly Aurila convexa) indicate that a well-developed species associated with a shallower Posidonia oceanica meadow is indi- meadow was present at the sea-bottom from the Augustan Age up to cated by the relative increase in Gibbula spp. in the samples of section the Pompeii eruption. Samples of cluster M3b mostly belong to granu- V2c assigned to cluster M3b (Fig. 11B).In particular, a general tendency lometric cluster G1 (Fig. 12), characterized by low log-ratios between toward a closing of the environment seems to be evident by a change in mud and gravel. The results indicate that, during this time interval, the the malacofauna (Fig. 11). As shown by the position of the row points 10 DI DONATO ET AL . F I G U R E 1 2 Relative variation biplot of mollusk assemblages, showing the first 2 axis (A) and the first three axis (B) obtained from singular values decomposition of clr data (see “Supplementary text”). Sample points are shown with symbols and colors (scaled grey in printed version) to show both the granulometrical and mollusk subcluster to which they belong (see Fig. 9). 2𝜎 confidence ellipses are drawn to contour sections (dashdotted lines) and clusters (dotted lines). Bold red (black in printed version) vectors represent the balances obtained from CoDa of granulometrical data: [Color figure can be viewed at wileyonlinelibrary.com] belonging to cluster M2b (Fig. 12), the second half of the 1st century A.D. (prior to A.D. 79) is marked in both sections by a compositional switch (from M3b to M2b, Fig. 13). In particular, a decrease in the relative abundance of Bittium reticulatum and an increase in the relative abundance of opportunistic species, such as Rissoa ventricosa, Paphia aurea and Loripes lucinalis (Fig. 11) is recorded. At the present stage of investigation, it cannot be ascertained if this slight shallowing tendency was determined by natural phenomena, such as the accretion of infratidal bottom sediment, or if it was in some way related to the construction of artificial barriers within the harbor (piers of 1st century A.D.), as pointed out in the Roman harbor of Fréjus in France (Bony, Morhange, Bruneton, & Gébara, 2011). The apparent rate of sedimentation (ARS) recorded before the A.D. 79 (Fig. 13) shows rather high values, up to 2.5 cm/yr, which can be easily considered as a consequence of the dredging and of the construction of the Augustan quay. In fact, the building of this structure implied the regularization of both the coastline and the sea-bottom through the artificial cut of the tufaceous bedrock (Fig. 6A), in order to create a water column suitable for shipping. This cut created a new and wide accommodation space that was rapidly filled by sediments. Once the artificial space was filled, the mean ARS lowered to 0.45 cm/yr (mean value) until the A.D. 79 eruption (Fig. 13). This ARS value seems to be the usual rate in this bay, when no disturbing event (anthropic and/or natural) takes place. The intense commercial activity in the harbor during this period, testified by the construction of the quay and the large amount of pottery lost underwater, seems not to have been disturbed by the rapid accumulation of sediments (from ≈ 4.60 to ≈3.20 m b.s.l. close to the quay). Considering that the attested sea level was at 1.60 m b.s.l., the water column of ≈3.5 to ≈2.0 m was still suitable for loading/unloading operations (Boetto, 2010; Poveda, 2012). 5.1.2 The A.D. 79 eruption and its impact on the marine environment The A.D.79 Pompeii eruption generated widespread pyroclastic deposits with widespread impacts around the Somma-Vesuvius (Sigurdsson, Carey, Cornell, & Pescatore, 1985). In particular, the main phase of the eruption was Plinian and dominated by magmatic explosions which generated thick fallout deposits distributed southeastward of the volcano, including the Pompeii area, the eastern area of the Gulf of Napoli and the Sorrento Peninsula. These deposits are mainly composed of well sorted, white to gray pumice lapilli (Carey & Sigurdsson, 1987). The second main phase of the eruption was dominated by phreatomagmatic explosions which produced pyroclastic density currents radially distributed around the volcano up to a distance of 15 km. The currents emplaced cohesive, cross-laminated to massive, fine-grained ash beds. Archeological excavations (unpublished data by the authors) in the urban area of Naples have revealed that the pyroclastic density current deposits were distributed in many lowlands between the western apron of Vesuvius and the Parthenope area (Fig. 1). This distribution is in agreement with the reconstruction described by Gurioli et al. (2010), abruptly truncated in the urban territory of Naples for absence of data. These deposits have also been recognized in the Municipio area where they are represented by two cohesive and massive ash layers containing accretionary lapilli, thickening in the preexisting lowlands and depressions. The SEM analyses revealed that the majority of the juvenile fragments of the fine ash deposits (Fig. 14) show characteristics such as poorly- to nonvesicular blocky morphologies, alteration skins, pitting and minor quench cracks, and adhering particles, that suggest a predominantly phreatomagmatic origin. This ash deposit is barren. These characteristics corroborate the correlation of these deposits with those of 11 DI DONATO ET AL . F I G U R E 1 3 Chronostratigraphic sequence of paleoenvironments along the two studied sections compared to the apparent rate of sedimentation (ARS) and to the composition of the coarse fraction (> 2 mm). ARS was calculated by taking into account, for each century, the years represented in the sediments and the corresponding thickness [Color figure can be viewed at wileyonlinelibrary.com] the phreatomagmatic final phases of the eruption which emplaced (Fig. 13). The changes in the gravel fraction are mainly driven by pyroclastic density currents, likely flowing inland. Furthermore, the potsherd inputs from inland and by pumice arrivals (Fig. 13) from both presence of soft accretionary lapilli suggests the absence of reworking inland and wind-sea currents accumulated near the quay, especially of the ash after its deposition. The ash layer is overlain in the two in concurrence with the A.D. 79 eruption deposits and their long- studied sequences by a deposit composed of an alternation of lenses of term reworking. The primary deposition coupled with the arrival of subrounded to rounded, well sorted pumice lapilli and ash, for a total reworked volcanic material, caused the increase in the ARS, which thickness of about 40 cm (Fig. 15A). The high degree of roundness reaches values around 1.35 cm/yr (mean value) until the end of the 1st of lapilli and the structure of this deposit suggest that the pumice century A.D. (Fig. 13). lapilli, erupted during the Plinian phase of the eruption (before the Just after A.D. 79 and up to the end of the 3rd century A.D., the Posi- phreatomagmatic phase) and distributed southeast of Vesuvius, were donia oceanica meadow appears poorly developed (Fig. 13), indicating transported by marine currents and deposited on the beach after a the occurrence of unfavorable ecologic conditions at the basin bottom. long phase of reworking. These reworked deposits were interpreted The ostracods, totally absent in V2a between A.D. 79 and the first half by Delile, Goiran, Blichert-Toft, Arnaud-Godet, and Romano (2016a) of the 2nd century A.D., are still dominated by coastal assemblages and as being of a tsunami origin. record a significant amount of Xestoleberis spp., a lagoon taxon (e.g., sec- On land, the A.D. 79 eruption is only represented by the fine- tion V2c, Fig. 11B). In the same interval, a further decrease in Bittium grained ash deposit (Fig. 15B). The absence of the reworked pumice reticulatum and peaks of Rissoa ventricosa and Paphia aurea are recorded rich deposits and/or of marine remains, even at short distances and (Fig. 11). In general, the mollusk assemblages belonging to clusters low elevation from the Roman coastline in the whole excavation area, M2a and M2b are indicative of reduced oxygenation, which may also allows us to exclude the speculation of the tsunami driven deposi- be related to an increase in trophic level and/or pollution within the tion suggested by Delile et al. (2016a). The hypothesis is also excluded harbor. Bottom vegetation may be damaged by thick deposits of float- by the absence of damage to buildings and structures in the coastal ing tephra and/or by the subsequent erosion of volcanic material from area. the catchment, which could enhance water turbidity (Ayris & Delmelle, After A.D. 79, the textural features of the studied sections show 2012). If the impact of the A.D. 79 eruption, even if greatly diluted by increasing variability. As evidenced by the compositional analysis of distance, has caused a temporary change in the sea-water chemistry granulometric data, the main source of variation in V2a section is and turbidity, or even directly induced damage to the bottom vege- related to changes in the gravel versus mud ratio, from G1 to G4 tation in the harbor of Neapolis, this effect should have been flanked 12 DI DONATO ET AL . F I G U R E 1 4 SEM images of selected pyroclasts of A.D. 79 ash deposits (AU 7476, Fig. 15B). A) whole sample dominated by nonvesicular clasts; B) weakly vesicular to nonvesicular clasts with pitted surface; C) and D) blocky clasts with adhering particles, alteration skin and quench cracks. All features are typical of fragments generated by phreatomagmatic fragmentation (Heiken, G.H., 1974; Heiken, G. & Wohletz, K., 1985) and prolonged by other causes. Probably, the concomitant protective stones from harbor structures) are an important component of the effect of piers, increasing sedimentation, and polluting effects of ther- coarse sediment fraction during the 2nd century A.D. (Fig. 13) when mal bath drainpipes could have enhanced environment deterioration. thermal baths were constructed behind the quay. Similarly, the construction of harbor facilities both in Marseille (France) Once the large sedimentary input linked to the A.D. 79 eruption and and Elaia (Turkey) led to destruction of the original Posidonia ecosys- its reworking was expended and the beach profile was regulated, the tem (Morhange et al., 2003; Seeliger et al., 2013). ARS gradually came back to its usual values (mean value 0.35 cm/yr in the 2nd and 3rd centuries A.D., Fig. 13), similar to those recorded before the eruption. At the beginning of the 4th century A.D., open 5.1.3 From the 2nd to the 4th century A.D conditions re-established with the recovery of the bottom vegetation The first tendency toward the establishment of a closed environment, as testified by the increase in Posidonia oceanica remains and the highlighted in the previous period, is enhanced from the first half of occurrence of mollusk assemblages belonging to cluster M3a. The the 2nd century A.D., as testified by mollusk assemblages of the outer 4th century A.D. is characterized by very low ARS, with mean values V2c section, included in cluster M1a, and by the presence of lagoon ranging around 0.17 cm/yr, which in the internal section may be partly ostracods (Fig. 11) up to the end of the 4th century A.D. The increase related to the relative decrease in the coarse fraction supplied by the in the high trophic level recorded in the bay until the end of the 3rd anthropic wastes (Fig. 13). At that time, the quay was no more in use century A.D., begins in concurrence with the construction of the due to the decreasing water column (maximum 1.20 m close to the Roman baths at the end of the 2nd century A.D. (Figs 5 and 6C), whose quay). A stagnation of port activities was recorded in the 3rd–4th drainage pipe flowed into the harbor probably inducing pollution, such centuries A.D. (Giampaola et al., 2006) both in Line 6 and Line 1, where as lead enrichment highlighted by Delile et al. (2016b). During this it occurred in concomitance with a phase of abandonment highlighted time interval, more or less similar assemblages (M2a and M2b) are by the decrease of horticultural activities around the harbor and the found in both sections (Fig. 13). Differently, the results of CoDA show development of wild vegetation (Russo Ermolli et al., 2014). Both the that samples of section V2a are included in the granulometric clusters lower ARS and the reduced use of the harbor could have favored the G4-G2, characterized by a relative abundance of sand and gravel environment recovery. (mainly composed of pumices and potsherd), while those of section By comparison, Delile et al. (2016a) evaluated a drastic decrease in V2c fall into clusters G2–G3, characterized by a reduced coarse the rate of sedimentation between the 3rd and the beginning of the component and finer sediments. In effect, both sections equally record 5th century A.D., as well as the absence of the 4th century A.D. levels the pumice and ash arrivals, whereas only in the internal section (V2a), in their sampling. In order to explain these anomalies, they hypothe- closer to the quay, the anthropic inputs (mostly potsherds and building size numerous dredging operations carried out in the 3rd and in the 13 DI DONATO ET AL . F I G U R E 1 6 The final closure sequence of the harbor: the basal lagoon deposits (first half 5th century A.D.) are overlain by alluvial inputs in a shallow lagoon, showing multiple erosional surfaces. At the top, the arrival of debris flows (end of 5th to beginning of 6th century A.D.) in a subaerial environment is marked by a basal undulated, erosional contact. Spotty marsh deposits seal the sequence [Color figure can be viewed at wileyonlinelibrary.com] an open lagoon, with a poorly vegetated bottom, as testified by the strong decrease in Posidonia remains. The relative abundance of fine sand in the inner section (cluster G2) and the relative increase in the finer fraction in the outer one (cluster G3), which typifies the samples of this interval, is another evidence of the transition toward a more sheltered environment, characterized by depositional conditions favorable to decantation. At that moment, the lagoon occupied the whole area of Piazza Municipio. A) Evidence of the A.D. 79 eruptive event in the marine environment: the ash layer directly deposited by pyroclastic density currents is overlain by reworked ash, pumice, lapilli, and Posidonia remains; B) the A.D. 79 ash primary deposit on land [Color figure can be viewed at wileyonlinelibrary.com] FIGURE 15 The uppermost levels of the lagoon environment are only represented in log V2a and in the Top section. They are characterized by polygenic, heterometric, and poorly sorted sediments (cluster G4, Fig. 7A) gently inclined toward the open sea, by fragmented and reworked mollusk shells, and by the absence of ostracods (Fig. 11). The multiple erosional surfaces (Fig. 16) marking the deposition of these 5th century A.D. However, during the archeological excavations car- sediments are indicative of runoff and alluvial input into the shallow ried out in both Lines 1 and 6 areas for more than 15 years, no chrono- lagoon environment. The drastic increase in ARS (mean value 2 cm/yr, logical gap was recorded in the 4th century A.D. and no evidence of Fig. 12) recorded in this period testifies to the beginning of rapid infill- dredging was identified between the 3rd and the 5th century A.D. ing processes, causing the shoreline progradation and the consequent end of harbor activity in this bay sector, according to what was already 5.2 The Lagoon environment and final harbor closure (beginning of the 5th to beginning of the 6th century A.D.) recorded in Line 1. In fact, the water column at the beginning of the The beginning of the 5th century A.D. marks an important paleoeco- continued farther eastward, as indicated by the archaeological remains logical swing, which was already highlighted in the excavations for unearthed in the Università excavation (Giampaola et al., 2006). lagoon phase (top of lagoon sediments at 2.30 m b.s.l.) was reduced to ≈50 cm close to the quay. The latter was completely buried by sediments at the end of the 5th century A.D. Port activities migrated and Line 1 (Amato et al., 2009; Carsana et al., 2009; Russo Ermolli et al., The topmost layers are represented by brownish, very poorly sorted 2014). The progressive closure of the harbor basin and the consequent sediments, devoid of mollusk remains, capped by a thin dark-grey silty- transition toward a muddy-sand lagoon context is recorded by an sand layer. The base of these units is characterized by an undulated increase in the percentage of lagoon ostracods in section V2c (Fig. 11) erosional surface (Fig. 16). These characters suggest alluvial sedimen- and by a change in the composition of mollusk assemblages, which in tation in a subaerial environment, i.e., debris flow, which marks the the upper interval of both sections belong to clusters M1a and M1b final closure of the lagoon (Fig. 13). A large part of the Neapolitan- (Figs 11 and 13). In the RVB (Fig. 12), samples belonging to cluster Vesuvian territory was affected in this period by debris flows, gen- M1a are located close to the column points of Cerastoderma glaucum erated by the emplacement of ash deposits related to the A.D. 472 and Cyclope neritea, indicating high relative abundances of these sub-Plinian Vesuvius eruption. This event resulted in a strong hydro- taxa, which corresponds a decreasing relative abundance of Bittium geological destabilization of an area larger than that directly affected reticulatum. These assemblages indicate low energy conditions within by the emplacement of the pyroclastic deposits (Di Vito et al., 2016). 14 DI DONATO ET AL . Subsequent ephemeral marshy episodes are testified by the spotty investigations on the ash samples. All figures, photographs and data occurrence of thin dark silty layers at the top of the alluvial deposits included in this article have officially been authorized by the Soprint- (Figs 13 and 16). endenza Archeologia Belle Arti e Paesaggio of Naples. Two anonymous reviewers and the editors are thanked for their comments that greatly improved the original manuscript. 6 CONCLUSIONS ORCID The Neapolis harbor excavations represent a unique opportunity to enrich our knowledge about the history of this important Roman town. Elda Russo Ermolli http://orcid.org/0000-0003-1275-6158 The recovery of structures, such as piers, quays, thermal baths, and roads, has shed new light on the ancient town setting, coastal profile, and sea-level position from the 1st century B.C. up to the 6th cen- REFERENCES tury A.D. The archeological excavation together with the large amount AA.VV. (1985). Napoli Antica.Catalogo della mostra. Macchiaroli, Napoli. of remains found in the harbor sediments provided a precise chronol- Aitchison, J. (1983). Principal component analysis of compositional data. Biometrika, 70, 57–65. ogy of the infilling and allowed a detailed environmental history to be drawn on the basis of sedimentological and paleontological data. The evolution of the harbor basin was determined by the interplay between natural and human-induced phenomena. In particular, the classic regressive trend which characterizes the natural evolution of a protected bay was enhanced by: 1. the dredging phases conducted until the 2nd century B.C., which engendered hypersedimentation due to the creation of a wide accommodation space; 2. the sedimentary input related to the A.D. 79 Pompeii eruption and the reworking of its sediments; 3. the construction of piers within harbor basin in the 1st century A.D., which induced restricted circulation and consequent hypersedimentation; 4. the increased alluvial input related to strong hydrogeological destabilization, connected to the A.D. 472 sub-Plinian Vesuvius eruption, which caused the final closure of the harbor bay. The same causes, together with pollution induced since the 1st century A.D. by thermal bath effluent, contributed to environmental deterioration at the sea-bottom, engendering the drastic decrease in the Posidonia meadows. Our investigation also identified primary deposits of the A.D. 79 Vesuvius eruption for the first time in the Neapolitan territory. These were attested through the recognition, both at the sea-bottom and on land, of an ash layer directly deposited by currents of the final, phreatomagmatic phases of the eruption. Finally, the compositional data analysis approach proved valuable for interpreting the sedimentological and paleontological record of the ancient harbor of Neapolis. This valuable geoarchaeological tool can be applied to multiple paleoenvironmental proxy records in other cultural and geomorphic settings. ACKNO W LE D GE ME NTS This article is dedicated to the memory of our friend and colleague Paola Romano. C. Morhange thanks IUF, Labex OT-Med (ANR11-LABX-0061) and the A*MIDEX project (ANR-11-IDEX-0001-02). Daniela Mele of the University of Bari is acknowledged for the SEM Aitchison, J. (1986). The statistical analysis of compositional data. Monographs on statistics and applied probability (pp. 416). London: Chapman & Hall Ltd. (Reprinted in 2003 with additional material by The Blackburn Press) Aitchison, J., & Greenacre, M. (2002). Biplots of compositional data. Applied Statistics, 51(4), 375–392. Amato, L., Carsana, V., Cinque, A., Di Donato, V., Giampaola, D., Guastaferro, C., …Russo Ermolli, E. (2009). Ricostruzioni morfoevolutive nel territorio di Napoli: L'evoluzione tardo pleistocenica-olocenica e le linee di riva in epoca storica. Méditerranée, 112, 23–31. Ayris, P. M., & Delmelle, P. (2012). The immediate environmental effects of tephra emission. Bulletin of Volcanology, 74, 1905–1936. Boetto, G. (2010). Le port vu de la mer: L'apport de l'archéologie navale à l’étude des ports antiques. In M., Dalla Riva (Ed.), Meetings between cultures in the ancient mediterranean. Proceedings of the 17th International Congress of Classical Archaeology (pp. 112–128). Rome: Ministero per i Beni e le Attività Culturali. Bony, G., Morhange, C., Bruneton, H., & Gébara, C. (2011). 2000 ans de colmatage du port antique de Fréjus (Forum Julii), France. Une double métamorphose littorale. Comptes Rendues Geoscience, 343, 701–715. Brancaccio, L., Cinque, A., Romano, P., Rosskopf, C., Russo, F., Santangelo, N. & Santo, A. (1991). Geomorphology and neotectonic evolution of a sector of the Tyrrhenian flank of the southern Apennines (region of Naples, Italy). Zeitschrift für Geomorphologie, Suppl. Bd, 82, 47–58. Bruno, P. P. G., Rapolla, A., & Di Fiore, V. (2003). Structural setting of the Bay of Naples (Italy) seismic reflection data: Implications for Campanian volcanism. Tectonophysics, 72, 193–213. Caiazzo, C., Ascione, A. & Cinque, A. (2006). Late Tertiary–Quaternary tectonics of the Southern Apennines (Italy): New evidences from the Tyrrhenian slope. Tectonophysics, 421, 23–51. Capasso, B. (1895). Topografia della città di Napoli nell'XI secolo. Napoli: A. Forni. Carey, S., & Sigurdsson, H. (1987). Temporal variations in column high and magma discharge rate during the AD 79 eruption of Vesuvius. Geological Society American Bulletin, 99, 303–314. Carsana, V., D'Amico, V., & Del Vecchio, F. (2007). Nuovi dati ceramologici per la storia economica di Napoli tra tarda antichità ed altomedioevo. In M., Bonifay, & J.C., Trèglia (Eds.), Late Roman coarse wares, cooking wares and amphorae in the Mediterranean, archaeology and archaeometry (pp. 423–437). Oxford: BAR International Series, 1662. Carsana, V., Febbraro, S., Giampaola, D., Guastaferro, C., Irollo, G., & Ruello, M.R. (2009). Evoluzione del paesaggio costiero tra Parthenope e Neapolis: Una sintesi geoarcheologica per l'area dell'antico porto. Méditerranée, 112, 15–22. DI DONATO ET AL . Carsana, V., & Del Vecchio, F. (2010). Il porto di Neapolis in età tardo antica: Il contesto di IV secolo d.C. In S., Menchelli, S., Santoro, M., Pasquinucci, & G., Guiducci (Eds.), Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean, Archaeology and Archaeometry.Comparison between western and eastern Mediterranean (pp. 459– 470). Oxford: BAR International Series, 2185. Carsana, V., & Guiducci, G. (2013). I contesti ceramici di età medio imperiale dal porto di Neapolis. In L., Giron, M., Lazarich, & M., Conceicao Lopes (Eds.), Actas del I Congreso Internacional sobre Estudios Ceramicos. Cadice. (pp. 1007–1040), ISBN 978-84-9828-401-0. Cennamo, P., Caputo, P., Stefano, M., Russo Ermolli, E., & Barone Lumaga, M. R. (2014). Epiphytic diatom communities on sub-fossil leaves of Posidonia oceanica in the Graeco-Roman harbor of Neapolis: A tool to explore the past. American Journal of Plant Science, 5, 549–553. Cinque, A., Aucelli, P. P. C., Brancaccio, L., Mele, R., Milia, A., Robustelli, G., …Sgambati, D. (1997). Volcanism, tectonics and recent geomorphological change in the bay of Napoli. Supplementi Geografia Fisica e Dinamica Quaternaria, 3(2), 123–141. Cinque, A., Irollo, G., Romano, P., Ruello, M. R., Amato, L., & Giampaola, D. (2011). Ground movements and sea level changes in urban areas: 5000 years of geological and archaeological record from Naples (Southern Italy). Quaternary International, 232(1), 45–55. Cioni, R., Bertagnini, A.,Santacroce, R., & Andronico, D. (2008). Explosive activity and eruption scenarios at Somma-Vesuvius (Italy): A review. Journal of Volcanology and Geothermal Research, 178, 331–346. D'Agostino, B., & Giampaola, D. (2005). Osservazioni storiche e archeologiche sulla fondazione di Neapolis. InW. V., Harris & E., Lo Cascio (Eds.), Noctes Campanae, Studi di storia antica e archeologia dell'Italia preromana e romana in memoria di Martin W. Frederiksen (pp. 63–72). Napoli: Luciano Editore. Deino, A. L., Orsi, G., De Vita, S., & Piochi, M. (2004). The age of the Neapolitan Yellow Tuff caldera-forming eruption (Campi Flegrei caldera, Italy) assessed by 40 Ar/39 Ar dating method. Journal of Volcanological and Geothermal Research, 133, 157–170. Delile, H., Goiran, J.-P., Blichert-Toft, J., Arnaud-Godet, F., & Romano, P. (2016a). A geochemical and sedimentological perspective of the life cycle of Neapolis harbor (Naples, southern Italy). Quaternary Science Reviews, 150, 84–97. Delile, H., Keenan-Jones, D., Blichert-Toft, J., Goiran, J.-P., Arnaud-Godet, F., Romano, P., & Albaréde, F. (2016b). A lead isotope prespective on urban development in ancient Naples. PNAS, 113, 6148–6153. Di Donato, V., Martin-Fernandez, J. A., Daunis-i-Estadella, J., & Esposito, P. (2015). Size fraction effects on planktonic foraminifera assemblages: A compositional contribution to the golden sieve rush. Mathematical Geosciences, 47(4), 455–470. Di Vito, M. A., Isaia, R., Orsi, G., Southon, J., de Vita, S., D'Antonio, M., …Piochi, M. (1999). Volcanism and deformation since 12,000 years at the Campi Flegrei caldera (Italy). Journal of Volcanological and Geothermal Research, 91, 221–246. Di Vito, M. A., Zanella, E., Gurioli, L., Lanza, R., Sulpizio, R., Bishop, J., …Laforgia, E. (2009). The Afragola settlement near Vesuvius, Italy: The destruction and abandonment of a Bronze Age village revealed by archaeology, volcanology and rock-magnetism. Earth and Planetary Science Letters, 277, 408–421. Di Vito, M. A., Castaldo, N., de Vita, S., Bishop, J., & Vecchio, G. (2013). Human colonization and volcanic activity in the eastern Campania Plain (Italy) between the Eneolithic and Late Roman periods. Quaternary International, 303, 132–141. Di Vito, M. A., de Vita, S., Rucco, I., Bini, M., Zanchetta, G., Boenzi, G., …Stanco, E. (2016). Volcaniclastic debris flows related to 472 A.D. eruption at Vesuvius: Social and environmental impact from stratigraphic 15 and geoarcheological data. Cities on Volcanoes, 9, 20–25 November, Puerto Varas (Chile). Egozcue, J. J., Pawlowsky-Glahn, V., Mateu-Figueras, G., & Barceló-Vidal, C. (2003). Isometric logratio transformations for compositional data analysis. Mathematical Geology, 35(3), 279–300. Eynatten, H. V., Pawlowsky-Glahn, V., & Egozcue, J. J. (2002). Understanding perturbation on the simplex: A simple method to better visualise and interpret compositional data in ternary diagrams. Mathematical Geology, 34(3), 249–257. Folk, R. L. (1954). The distinction between grain size and mineral composition in sedimentary rock nomenclature. Journal of Geology, 62(4), 344– 359. Giampaola, D., & Carsana, V. (2005). Neapolis Le nuove scoperte: La città, il porto e le macchine. In E., Lo Sardo (Ed.), Eureka II degli antichi (Catalogo Mostra 2005) (pp. 116–122). Napoli: Museo Archeologico Nazionale di Napoli. Giampaola, D., & Carsana, V. (2010). Fra Neapolis e Parthenope: Il paesaggio costiero ed il porto. In Blackman, D.J., & Lentini, M.C. (Eds.), Ricoveri per navi militari nei porti del Mediterraneo antico e medievale (pp. 119–129). Bari: Edipuglia. Giampaola, D., Carsana, V., Boetto, G., Crema, F., Florio, C., Panza, D., …Pizzo, B. (2006). La scoperta del porto di Napoli: Dalla ricostruzione topografica allo scavo e al recupero dei relitti. Marittima Mediterranea. International Journal of Underwater Archaeology, 2, 47–91. Gurioli, L., Sulpizio, R., Cioni, R., Sbrana, A., Santacroce, R., Luperini, W., & Andronico, D. (2010). Pyroclastic flow hazard assessment at Somma– Vesuvius based on the geological record. Bulletin of Volcanology, 72, 1021–1038. Heiken, G. H. (1974). An atlas of volcanic ash. Smithsonian Earth Science Contributions, 12, 1–101. Heiken, G. H., & Wohletz, K. H. (1985). Volcanic ash (pp. 246). Berkeley: University of California Press. Kaniewski, D., Van Campo, E., Morhange, C., Guiot, J., Zviely, D., Shaked, I., …Artzy, M. (2013). Early urban impact on Mediterranean coastal environments. Nature Science Reports, 3(3540), 1–5. Liuzza, V. (2014). Ricostruzione paleogeografica e paleoambientale della città di Napoli: Un'indagine geoarcheologica. Tesi di Dottorato in Scienze della Terra, XXVI ciclo. Università degli Studi di Napoli Federico II, 363 pp., Retrieved from https://www.fedoa.unina.it. Marriner, N., & Morhange, C. (2006). The ‘ancient harbour parasequence’: Anthropogenic forcing of the stratigraphic highstand record. Sedimentary Geology, 186, 13–17. Marriner, N., & Morhange, C. (2007). Geoscience of ancient Mediterranean harbours. Earth Science Reviews, 80, 137–194. Marriner, N., Morhange, C., Flaux, C., & Carayon, N. (2017). Harbors and ports. In A.S., Gilbert (Ed.), Encyclopedia of geoarchaeology (pp. 382–403). Berlin: Springer. Morhange, C., Blanc, F., Bourcier, M., Carbonel, P., Prone, A., Schmitt, S., …Hesnard, A. (2003). Bio-sedimentology of the late Holocene deposits of the ancient harbor of Marseilles (Southern France, Mediterranean Sea). The Holocene, 13, 593–604. Morhange, C., Marriner, N., & Carayon, N. (2015). The geoarchaeology of ancient Mediterranean harbours. In G.,Arnaud-Fassetta & N., Carcaud (Eds.), French geoarchaeology in the 21st century (pp. 281–289). Paris:CNRS editions, Alpha Morhange, C., Marriner, N., & Carayon, N. (2016). Eco-history of ancient Mediterranean harbours. In T., Bekker-Nielsen, & R., Gertwagen (Eds.), The inland seas, towards an ecohistory of the Mediterranean and the Black Sea (pp. 85–106). Stuttgard: Franz Steiner, Verlag. Napoli, M. (1959). Napoli greco-romana. Napoli: Fausto Fiorentino Editore. 16 DI DONATO ET AL . Napoli, M. (1967). Topografia e archeologia. In: Storia di Napoli 1, 373–507. Napoli: Fausto Fiorentino Editore. Roman harbours. Review of Palaeobotany and Palynology, 218, 217– 230. Pawlowsky-Glahn, V., & Buccianti, A. (2011) (Eds.). Compositional data analysis: Theory and applications (400 pp.). New York: Wiley. ISBN: 978-0-47071135-4. Seeliger, M., Bartz, M., Erkul, E., Feuser, S., Kelterbaum, D., Klein, C., …Brückner, H. (2013). Taken from the sea, reclaimed by the sea - The fate of the closed harbour of Elaia, the maritime satellite city of Pergamum (Turkey). Quaternary International, 312, 70–83. Peres, J.M. (1982). Major benthic assemblages, in: O. Kinne (ed.), Marine Ecology, Ocean management. Part 1, ch. 8 (Vol. V, pp. 373–522). New York: John Wiley & Sons. Poveda, P. (2012). Le navire antique comme instrument de commerce maritime: Restitutions 3D, tonnage, qualités nautiques et calculs hydrostatiques des épaves: Napoli A, Napoli C, Dramont E et Jules Verne 7 (PhD Thesis). Université Aix Marseille, Marseille. Quinn, G., & Keough, M. (2002). Experimental Design and Data Analysis for Biologists (pp. 557). Cambridge: Cambridge University Press. Romano, P., Di Vito, M. A., Giampaola, D., Cinque, A., Bartoli, C., Boenzi, G., …Schiano di Cola, C. (2013). Intersection of exogenous, endogenous and anthropogenic factors in the Holocene landscape: A study of the Naples coastline during the last 6000 years. Quaternary International, 303, 107–119. Rossi, V., Sammartino, I., Amorosi, A., Sarti, G., De Luca, S., Lena, A., & Morhange, C. (2015). New insights into the palaeoenvironmental evolution of Magdala ancient harbour (Sea of Galilee, Israel) from ostracod assemblages, geochemistry and sedimentology. Journal of Archaeological Science, 54, 356–373. Sgarrella, F., Di Donato, V., & Sprovieri, R. (2012). Benthic foraminiferal assemblage turnover during intensification of the Northern Hemisphere glaciation in the Piacenzian Punta Piccola section (Southern Italy). Palaeogeography, Palaeoclimatology, Palaeoecology, 333, 59–74. Sigurdsson, H., Cashdollar, S., & Sparkes, S. R. J. (1982). The eruption of vesuvius in A. D. 79: Reconstruction from historical and volcanological evidence. American Journal of Archaeology, 86(1), 39–51. Sigurdsson, H., Carey, S., Cornell, W., & Pescatore, T. (1985). The eruption of Vesuvius in A.D. 79. Natural Geographic Research, 1, 332–387. Stazio, A. (Ed.). (1988). Neapolis. Atti XXV Convegno di Studi sulla Magna Grecia, Taranto 3–8 ottobre 1985. Taranto: Istituto per la Storia e l'Archeologia della Magna Grecia. Wentworth, C. K. (1922). A scale of grade and class terms for clastic sediments. Journal of Geology, 30, 377–392. Zevi, F. (Ed.). (1995). Neapolis. Napoli: Edizioni Banco di Napoli. SUPPORTING INFORMATION Ruello, M. R. (2008). Geoarcheologia in aree costiere della Campania: I siti di Neapolis ed Elea -Velia. Tesi di Dottorato in Scienze della Terra, XX ciclo. Università degli Studi di Napoli Federico II, Retrieved from https://www.fedoa.unina.it. Additional Supporting Information may be found online in the support- Russo Ermolli, E., Romano, P., Ruello, M. R., & Barone Lamuga, M. R. (2014). The natural and cultural landscape of Naples (southern Italy) during the Graeco-Roman and the Late Antique periods. Journal of Archaeological Science, 42, 399–411. How to cite this article: Di Donato V, Ruello MR, Liuzza Sadori, L., Allevato, E., Bertacchi, A., Boetto, G., Di Pasquale, G., Giachi, G., …Mariotti Lippi, M. (2015). Archaeobotany in Italian ancient ing information tab for this article. V, et al. Development and decline of the ancient harbor of Neapolis. Geoarchaeology.2018;1–16. https://doi.org/10.1002/ gea.21673