1Poseidonia (nowadays Paestum) was founded by Greek settlers at the turn of the 6^th century B.C. in the southern portion of the Sele Plain (fig.1), a short distance from the sea. The Greek city flourished from its foundation until about the second half of the 5^th century B.C. when it came under the rule of the Lucanians. From that moment on Poseidonia took the italic name Paistom. In 273 B.C., the Romans transformed Paistom into a Roman colony, named Paestum. Under Roman rule, the city enjoyed a new and prolonged phase of development and prosperity.

2The historical vicissitudes of Poseidonia are linked to complex and partially unknown palaeo-geographic and palaeo-environmental scenarios. Unresolved problems include, for example, the morphological setting of the territory of Paestum at the moment of the foundation of Poseidonia and the subsequent environmental modifications until Late Roman times; the position of harbour structures and/or dockings during the Archaic and Classical periods; the Roman canalization system; the dynamics and origin of travertine incrustation in Late Roman to Medieval times and its relation to the definitive abandonment of the city of Paestum.

3To contribute to the resolution of such problematic aspects, an interdisciplinary working group was formed in order to integrate historical and archaeological data with geologic and geomorphologic expertises. In the first study phase, particular attention was paid to the reconstruction of the physical landscape at the time when Poseidonia was founded, and to the environmental modifications which affected the territory of Paestum in historical times.

1 - Setting of the study area

4The study area comprises the Paestum plain which occupies the southern portion of the Sele river coastal plain (fig.1). The Sele plain is located within a triangular tectonic depression, which was generated by faulting during the Lower Pleistocene. The filling of this depression was first accompanied by a strong subsidence and then, from the Late Middle Pleistocene onwards, by a weak uplift (Cinque, 2006). This uplift raised, by up to 25 m above sea level, the beach and lagoon sediments deposited during the interglacial marine transgressions which formed the coastal units and palaeo-beach-dune ridges of Ponte Barizzo, Masseria Stregara and Gromola (fig.1) of Middle to Upper Pleistocene age (Cinque, 2006).

5The advance and growth of the Paestum plain was significantly favoured by the deposition of travertine which was caused by the carbonate-rich groundwater feeding the springs of Capodifiume (Cestari, 1969) located at the base of the western slope of the Mt. Soprano carbonate structure (fig.1).

6From the Middle to Upper Pleistocene (Cinque, 2006), three main travertine units were generated, forming terraced surfaces with respect to the surrounding plain: the Cafasso unit, located in a more internal position, and the Gaudo and Paestum units with a maximum height of about 20 m and located at short distances from the present shoreline (fig.1). During the last Upper Pleistocene Interglacial transgression, when the coast was partially submerged due to the glacio-eustatic sea-level rise, the Gromola beach ridge was deposited, the most seaward located travertine units were partially submerged and transformed into near-coast banks and low capes (Cinque, 2006). After this transgression the deposition of travertine recommenced in the area of Paestum leading to the surface on which Poseidonia was to be built. Stone tools found within the urban area of Paestum have been attributed to the Mousterian industry (Rosskopf, 2006) and suggest, for the Paestum unit at least, an Upper Pleistocene age.

7In the area of Paestum, travertine deposition continued until at least the late Holocene (Cinque, 2006). During the maximum Holocene transgression, the travertine terrace of Paestum became the most prominent cape. Successively, the advance of the plain was marked by the deposition of partially anastomosed dune-ridges, dating from the Middle Holocene onwards. These dune-ridges gave rise to two major dune alignments, respectively named the Laura and Sterpina ridges (fig.1) and dated by ¹⁴C from 5300 to 2500 years BP (Brancaccio et alii, 1988). The Holocene dune ridges closed the plain sea-wards and led to a lagoonal to fluvial-palustrine deposition within the topographically depressed back-ridge areas located in the area of Paestum (fig.2) and north to the Sele river (Barra et alii, 1998, 1999) which remained active until very recent times (Rosskopf, 2006).

2 - Methodological aspects

8To investigate the landscape evolution of the area of Paestum during the Late Holocene, a detailed geomorphologic study was carried out. This study was based on field surveys and the interpretation of aerial photos from 1943, the analysis of pre-existing well logs and the interpretation of a detailed topographic base with contour lines every meter, reconstructed using 1:5000 scaled topographic maps (Carta Tecnica dell’Italia Meridionale, 1975, and Carta Tecnica Regionale Campania, 2002).

9The study allowed us to distinguish a number of different coastal and fluvial landforms and to reconstruct their relative chronology on the basis of the relationships between singular landforms and that between landforms, archaeological remains and the local bedrock. In figure 2, which represents a synthesis of examined geomorphologic, stratigraphic and archaeological data, we report the main geomorphologic elements and the location of archaeological sites used in the reconstruction.

3 - Landscape modifications during the historical period

10Collected data enabled us to demonstrate that the travertine deposit of Paestum (fig.1) comprises more than one unit. Among the various units that can be distinguished, thanks to geomorphologic and archaeological-stratigraphic data, the travertine on which the city of Poseidonia was built, represents the most ancient one. This travertine, which preserves the name Paestum unit (see figure 2), also shows a certain continuity south of the urban area. In particular it includes, in the south-west, the northern parts of Torre di Paestum and Licinella (locality Lupata) and southwards a zone extending from Santa Venere to Tempa del Prete (fig.2). Within the zone located immediately south of the urban area of Paestum this unit is nearly outcropping and frequently only covered by some tens of centimetres of soil as evidenced by various archaeological sites located at Santa Venere and Licinella (sites 21, 22, 23 and 24, fig.2, tab.1), dating from the fifth to the second centuries B.C.

11At these sites, the majority of the tombs are represented by rectangular graves cut directly into the travertine bank. The artificial cuts and holes within the travertine deposit at site 21 (Greco et alii, 1987) also indicate a prehistoric age for this portion of the Paestum unit. Proceeding along the south-eastern portion of the Paestum unit, various stratigraphic logs around Spinazzo (located east of Santa Venere) yield evidence of beach and dune deposits reaching a maximum height of nearly 11 m above sea level which can be attributed, in our opinion, to the Gromola beach-dune ridge (fig.2).

12In the area immediately north of the city of Paestum, between the Paestum unit and Gaudo unit, there is clear morphologic evidence for a distinct phase of travertine deposition which generated a distinct unit, herein named the Arcione unit. This unit is of pre-Archaic age as it hosts, besides the famous little temple (“Tempietto”) around Gaudo (site 4, fig.2, tab.1), dating to the 4^th century B.C., many tombs dating to the 6^th to third centuries B.C. and to the 1^st to 2^nd centuries A.D. (sites 5-7, fig.2, tab.1), frequently directly cut in travertine. The deposition of the Arcione unit was most likely accompanied by the transformation of the easterly located Laghetti zone into a marshy area, as testified by the presence of peaty-clay sediments within the first few meters of the subsoil. The absence of any archaeological findings within these sediments suggest their deposition in pre-Archaic times. The Arcione unit appears slightly dissected by a palaeo-channel of the Capodifiume river, oriented parallel to the northern wall of the city (fig.2). This palaeo-channel, most likely, was active in the period elapsing between the deposition of the Laura dune ridge and the Sterpina ridge (Sterpina 1, Guy, 1990), and connected to a lagoonal area located east of the internal border of the Laura ridge (area of Terra del Tesoro and Porta Marina, fig.2). Before the 6^th century B.C., the Capodfiume stream succeeded in eroding and locally levelling the Laura ridge. This is evidenced by the tombs of the Archaic necropolis of Ponte di Ferro (site 3, fig.2) which are located at shallow and homogeneous depths within the dune sands of the Laura ridge. In that period, in the area in front of Porta Marina, a slightly saline waterbody existed, fed by the spring waters of Porta Marina (Guy, 1990). This stretch of water remained active until at least the end of the first century A.D. as evidenced by the presence, within the filling, of an undisturbed ash and pumice layer attributed to the Vesuvian eruption
of 79 A.D. (Guy, 1990).

13In the area around the city of Paestum, the modifications during the Roman period are limited to the infilling of the ancient moat present, outside the city walls (two thirds of the moat were already infilled by 79 A.D., Schläger, 1962) and to the local re-activation of travertine deposition as, for example, around Torre di Paestum (site25, figs.2 and 3) giving rise to a thick superficial crust, separated from the underlying Paestum unit by about 1 m of sandy and debris rich dark clays.

14From (Late?) Roman to Medieval times (at least until the 7^th century A.D.), the city of Paestum was affected by travertine deposition which seems partially related to waters coming from the relicts of the moat, but whose origin is still uncertain. The resulting travertine crust, normally only some decimetres thick, locally reached several meters in thickness (Violante, 2000).

15The sector of the Paestum plain located south of the city was affected by important modifications from pre-Archaic to Late Roman times related to further phases of travertine deposition and deviations of the Capodifiume stream. In particular, a phase of travertine deposition affected the area of Mancone, Tempa del Prete and the southern portion of Linora, most likely during the Middle Holocene. The resulting travertine unit, herein named the Mancone unit, only outcrops where the recent and extensive travertine excavations have led to the removal of the overlaying Spinazzo-Linoraunit (see below). The Mancone unit is characterised by a variable thickness of about 2 to 5.5 m and overlies grey, sandy clays and clayey sands reaching a maximum height of ~4 m above sea level. Within this clayey-sand sequence, some drillings yield evidence of an arenaceous, mollusc-rich interlayer which reaches a maximum height of -1.5 m above sea level. This particular sequence is very different from that elucidated by drillings which intercept the Paestum unit (typically formed by alternated travertine and sandy to marly-clayey layers) and, in our opinion, can be related to a palaeo-environment which was characterised, before the travertine deposition, by coastal and then lagoonal to fluvial-palustrine dynamics during the maximum Holocene marine ingression. This hypothesis is supported by the geomorphological evidence in the scarp, also locally remodelled by successive episodes of travertine deposition, that extends from Licinella to Tempa del Prete and to Mancone and which is interpreted as a relict of the Holocene cliff formed during the marine ingression (fig.2). On the basis of these considerations a Middle Holocene age is proposed for the Mancone unit. This phase of travertine deposition is most likely linked to a palaeo-course of the Capodifiume stream which followed, at least partially, the present course of the Cisterna channel (fig.2), located east of the Paestum unit. The Mancone unit hosts archaeological sites dating from the 6^th century B.C. to the 3^rd century A.D. (sites39, 42-44 and 48, fig.2) which refer to the Chora meridionale of Paestum.

16The sector east and south of the city was affected by an important phase of travertine formation during Roman to Late Roman times which generated an extensive deposit herein called the Spinazzo-Linora unit. This unit buried part of the ancient travertine surfaces ascribed to the Paestum and Mancone units and the archaeological sites of Archaic to Roman age located at Linora and Spinazzo (fig.2). South of Paestum, this unit (Linora sub-unit, fig.2) is well evidenced from a geomorphological point of view as it is characterised by a typical fan-like surface, topographically depressed with respect to the travertine surfaces located to the west and east. This unit reaches a maximum thickness of 4 m within the central sector (site 42) becoming thinner towards its southern and eastern border. This is evident, for example, within the zone of Licinella-Tempa del Prete where a large archaeological area is located (sites 27-29, fig.2 and tab.1) including the necropolis (site 27) which hosts the famous tomb of the diver (“tomba del tuffatore”,  Greco et alii, 1987). This tomb is located near the eastern border of the Spinazzo-Linoraunit which overlies the Paestum unit (fig.2). The findings are superficial and partially covered by the Spinazzo-Linora unit which reachesa variable thickness of between 0.6 and 2 m. The deposition of the Spinazzo-Linora unit can be related to a palaeo-course of the Capodifiume stream very similar to the present one. In the area of Linora, this phase of incrustation started after the 3^rd century A.D. and was preceded by the generation of a palustrine environment as testified in its central sector by the presence of a clayey layer which separates the Spinazzo-Linora unit from the underlying Mancone unit. For the deposition of this unit in the area of Spinazzo (Spinazzo sub-unit, fig.2) archaeological data suggest a beginning after the third century B.C. (sites 15 and 19, fig.2, tab.1), raising some doubt about a later beginning in the post-Imperial period (site 16, tab.1).

17A further travertine deposit (Licinella unit) is present within the zone of Licinella. Its deposition is related to the present course and mouth of the Capodifiume stream. Its western border is prominent with respect to the Sterpine ridge. From a morphological point of view it can be divided into two sub-units (Licinella I and Licinella II subunits, fig.2). For these sub-units, respectively, a medieval and a sub-present age can be proposed; this hypothesis is supported both by geomorphological considerations and the absence of archaeological remains both on their surfaces and within excavations of the first metres.

4 - The palaeo-geography at the moment of the foundation of Poseidonia

18With reference to the palaeo-geography of the area of Paestum at the arrival of the Greek settlers, the collected data indicate the following situation:

19The shoreline was less advanced with respect to the present and characterised by dunes attributed to the Sterpina ridge. The back-ridge area was topographically depressed and locally occupied by a slightly saline waterbody, present both in the Porta Marina area (fig.2) and north of the present mouth of the Solofrone stream. Furthermore, the Laghetti zone was still characterised by an ephemeral fluvial-palustrine stretch of water.

20Collected data do not allow us to ascertain a possible interference between the course of the Capodifiume stream and the Greek settlement.

21Nearly all human activities were localised on travertine which outcropped extensively and was represented, from north to south, by the Gaudo, Arcione, Paestum and Mancone units. These travertine units were topographically elevated with respect to the coastal plain and the internal and back-ridge depressions. The western border of the travertine unit of Paestum was aligned to the shoreline.

22In conclusion, the collected data and the performed reconstructions have allowed us to add important new data to the reconstruction of the environmental changes that affected the Paestum plain in historical times. In particular, these data have allowed a number of hypotheses on its geomorphologic evolution to be developed. These can be verified by future research, notably addressing the geological setting and a detailed re-examination of archaeological-stratigraphic data.