1The southern Venetian Plain has been constructed by the two major Italian rivers, the Po and the Adige, and is adjacent to the Venice Lagoon and the present Po delta system. It is characterised by the presence of a complex network of alluvial ridges that are the sedimentary and geomorphological products of fluvial activity of the Adige and Po. Studies that attempt stratigraphic correlations between this area and the Venetian area are still lacking, despite several investigations of the late Quaternary succession in the Venice Lagoon (Tosi et al., 2007 a and b; Amorosi et al., 2008 b; Zecchin et al., 2008, 2009). Recent studies focus on the southern and central Po delta (Amorosi and Milli, 2001; Stefani and Vincenzi, 2005; Amorosi et al., 2008 a and b; Simeoni and Corbau, 2009) which, as evidenced by A. Amorosi et al. (2008b), is a “natural laboratory” where it is possible to trace stratigraphic surfaces from the basin margin to the coastal plain. In contrast, the relationships between the Adige and the Po alluvial systems have been little considered in previous literature. M. Bondesan et al. (1995) presented the evolution of the Po delta coastal plain, highlighting the architecture of buried and outcropping beach ridges formed by the Adige, the Po and Apennine rivers, but do not discuss the chronostratigraphic relationship between the alluvial ridges and the coastal sedimentary bodies. S.D. Mackey and J.S. Bridge (1995) presented three-dimensional models of alluvial stratigraphy and, among other examples, considered the avulsion sequence in the Po delta, without taking into account the Po northernmost branch and the interplay between Adige and Po systems. S. Piovan et al. (2010) discussed this interaction and its possible consequences for ancient human settlements but only focused on the alluvial system and did not consider coastal and deltaic evolution. The aim of this paper is to analyse the interactions between the Adige and Po alluvial systems during the late Holocene and to define how changes in the hydrographical network influenced deltaic sedimentation. The correlation between the late Holocene delta mouth sedimentary system recognised in the southern Venice Lagoon (Zecchin et al., 2009) and the alluvial succession in the continental area is also discussed on the basis of new geomorphological and geochronological data.
2The southern Venetian Plain is part of the foreland region located between the Southern Alps and the Apennines. During the early Pleistocene, the basin experienced rapid drowning to bathyal depths and sediment starvation. This was followed by the accumulation of a 750 m-thick succession up to the late Pleistocene, showing a generalised regressive trend from basinal turbidites to deltaic deposits and then an alternation between shallow-marine to continental sediments (Kent et al., 2002; Massari et al., 2004). This depositional history is documented by integrated studies based mostly on the 915 m-deep Venezia 1 core (e.g., Favero and Passega, 1980; Müllenders et al., 1996; Kent et al., 2002; Massari et al., 2004; Barbieri et al., 2007). The late Pleistocene-Holocene stratigraphy in the Venetian-Friulian Plain and in the Venice Lagoon area is documented by several multidisciplinary studies (e.g., Miola et al. 2006; Tosi et al., 2007 a and b; Amorosi et al., 2008b; Fontana et al., 2008; Zecchin et al., 2008; Fontana et al., 2010). During the Last Glacial Maximum (LGM; 24-15 ka BP), most of the North Adriatic shelf was exposed (Correggiari et al., 1996) due to the MIS 2 glacio-eustatic sea-level lowstand. The alluvial Po Plain extended for about 300 km south of the present Venetian coastal area. The large amount of debris produced in the alpine areas by LGM glacial and periglacial environmental conditions contributed to the general aggradation of alluvial megafans in the Venetian-Friulian alluvial plain (Mozzi, 2005; Fontana et al., 2008 and 2010). Between 17 and 9 ka cal. BP, fluvial activity in the plain was confined within incised valleys, through which the sediments passed and were deposited on the Adriatic shelf (Fontana et al., 2008). This produced a sedimentary hiatus recognisable across the entire Venetian-Friulian Plain (Tosi, 1994; Mozzi et al., 2003; Fontana et al., 2008). As shown by A. Amorosi et al. (2005), subsidence and sediment influx were important local factors for the timing of the maximum marine transgression which, in the southern sector of the Venetian Plain, was reached around 7-5.7 ka cal. BP (Favero and Serandrei Barbero, 1980; Amorosi et al., 2008b) and was located several kilometres landward of the present coastline (fig. 1). Post-glacial sea level rise induced a new phase of aggradation in the alluvial plain, with the upbuilding of major alluvial ridges of the Venetian Plain. The area is affected by a long-term subsidence of (mean value) 0.6 mm/a (Antonioli et al., 2009). The main components of natural subsidence in the Po Plain are tectonic and sediment loading, sediment compaction and post-glacial rebound (Carminati and Martinelli, 2002). The area of the Po delta is also strongly affected by a short-term (anthropogenic) subsidence of 10-20 mm/a, due to an increase in groundwater pumping for industrial, agricultural and domestic applications. This has been especially true in the second half of the 20th century, as rapid economic growth has increased the need for water. The elevated subsidence rate has determined an overall relative sea-level rise, which is presently significantly higher than that caused by eustatic factors.
Fig. 1 – Geomorphological map of the southern Adige-Po alluvial plain from the interpretation of DTM
Fig. 1 – Carte géomorphologique de la plaine alluviale commune au Pô et à l’Adige établie sur la base de l’interprétation de MNT
The chronostratigraphic attributions are from this paper and from S. Piovan et al. (2010). The rectangle indicates the study area. 1: present-day Adige alluvial ridge; 2: present-day Adigetto alluvial ridge; 3: Montagana-Este alluvial ridge, active until Early Middle Age; 4: Sant'Urbano alluvial ridge, active during pre-Roman Age; 5: Pozzonovo alluvial ridge, unknown age; 6: Bangoli alluvial ridge, active during the Bronze Age; 7: Conselve alluvial ridge, active during Roman Age; 8: Terrassa alluvial ridge, unknown age; 9: Fratta alluvial ridge, active until Iron Age; 10: Adria alluvial ridge, active until Iron Age; 11: Saline-Cona alluvial ridge, active during the Bronze Age; 12: Brenta alluvial ridge, active during the Roman Age; 13: hydrography. Coastlines – 14: around 7.0-5.7 ka cal. BP (Rizzetto et al., 2003; Amorosi et al., 2008b; Simoni and Corbeau 2009); 15: 9th-4th century BC (Bondesan et al., 1995).
Les attributions chronostratigraphiques sont faites à partir du présent article et de S. Piovan et al. (2010). Le rectangle indique la zone d’étude. 1 : bourrelet alluvial actuel de lʼAdige ; 2 : bourrelet alluvial actuel de lʼAdigetto ; 3 : bourrelet alluvial de Montagnana-Este, actif jusquʼau début du Moyen-Âge ; 4 : bourrelet alluvial de SantʼUrbano, actif durant la période pré-romaine ; 5 : bourrelet alluvial de Pozzonovo, dʼâge indéterminé ; 6 : bourrelet alluvial de Bagnoli, actif pendant lʼÂge du Bronze ; 7 : bourrelet alluvial de Conselve, actif pendant la période romaine ; 8 : bourrelet alluvial de Terrassa, dʼâge indéterminé ; 9 : bourrelet alluvial de Fratta, actif jusquʼà lʼÂge du Fer ; 10 : bourrelet alluvial dʼAdria, actif jusquʼà lʼÂge du Fer ; 11 : bourrelet alluvial de Saline-Cona, actif pendant lʼÂge du Bronze ; 12 : bourrelet alluvial de la Brenta, actif pendant la période romaine ; 13 : hydrographie. Trait de côte – 14 : vers 7-5,7 ka BP (Rizetto et al., 2003 ; Amorosi et al., 2008b ; Simoni et Corbeau, 2009) ; 15 : entre le IXe et le IVe siècle av. J.-C. (Bondesan et al., 1995).
3The recent work of M. Zecchin et al. (2009) describes the Holocene succession in the southern Venice Lagoon (fig. 2). The Holocene sequence is bounded at the base by a subaerial exposure surface (S1) that represents an identifiable stratigraphic marker throughout the North Adriatic coastal plain and the sequence boundary at the transition between Pleistocene and Holocene. It corresponds to the indurated “caranto” palaeosol (Mozzi et al., 2003; Fontana et al., 2008; Donnici et al., 2011), which is considered as the bounding unconformity at the base of the post-glacial deposits, called the “Po Synthem” in the new 1:50.000 Geological Map of Italy (Tosi et al., 2007 a and b; Cucato et al., 2012). The sequence continues with incised valley fill, lagoonal and shallow-marine deposits corresponding to the lowstand and transgressive systems tracts (LST/TST) interval (Units H1a and H1b; Zecchin et al., 2009; fig. 2). In this location, deltaic to shallow-marine and beach-ridge deposits, as well as coeval alluvial and lagoonal deposits, represent the highstand systems tract (HST) (Units H2a and H2b; Zecchin et al., 2009; fig. 2). The upper part of the succession consists of tidal channel deposits and tidal and subtidal flat sediments (Unit H3; fig. 2). The conclusions of M. Zecchin et al. (2009) are that (i) this succession records the post-LGM transgression in the low-gradient topography of the north-Adriatic shelf, and (ii) the following highstand deposition, first characterised by the growth of a delta system, is then influenced by human impacts, which caused a localised transgression mostly related to river diversions and consequent delta abandonment during historical times. The southern Venetian Plain was built up during the late Holocene by the rivers Po and Adige. The rate of sea-level rise was insufficient to balance the supply of fluvial sediment, leading to progradation of the Po and Adige delta systems (Stefani and Vincenzi, 2005).
4The Po is the largest Italian river by length (652 km) and catchment area (71,000 km²). It flows from the Western Alps through the Po Plain, receiving tributaries from both the Alps and the Apennines. It reaches the Adriatic Sea southeast of the study area, where it has formed a lobate delta mouth since 3500 cal. BP (Cremonini, 2007; Simeoni and Corbau, 2009). Ground elevations in the study area range between 5 and -3.5 m a.s.l.; the portions below sea level have mostly been reclaimed in the last two centuries (Consorzio di Bonifica Delta Po Adige, 1991). The Adige is the second longest Italian river (410 km) and the third in terms of catchment area (12,200 km2). The Adige “high plain”, corresponding to the piedmont portion of the Adige sedimentary system, is characterised by gravels and a braided river course. Downstream, in the “low plain”, the sediments are sandy and silty-clayey, and the river adopts a single meandering channel. The Adige reaches the Adriatic Sea a few kilometres south of the Venice Lagoon, through a cuspate delta mouth.
Fig. 2 – The Holocene succession in the extreme southern part of the Venice Lagoon
Fig. 2 – La stratigraphie holocène dans la partie la plus au sud de la lagune de Venise
See fig. 3 for location (modified from Tosi et al., 2007b and Zecchin et al., 2009). 1: Unit H3, Tidal channe land and modern lagoonal deposits; 2: Unit H2b, Prograding delta front/prodelta, shoreface and beach ridge deposits; 3: Unit H2a, Delta plain and adjacent alluvial and lagoonal deposits; 4: Unit H1b, Transgressive shallow marine deposits; 5: Unit H1a, Transgressive back-barrier deposits; 6: Pleistocene continental succession; 7: S2, maximum flooding surface; 8: S1, sequence boundary.
Voir fig. 3 pour la localisation (modifié à partir de Tosi et al., 2007b et Zecchin et al., 2009). 1 : Unité H3, chenaux de marée et dépôts lagunaires modernes ; 2 : Unité H2b, front deltaïque en progradation/prodelta, dépôts de plage et dʼavant plage ; 3 : Unité H2a, plaine deltaïque et dépôts fluviatiles et lagunaires adjacents ; 4 : Unité H1b, dépôts marins associés à la transgression ; 5 : Unité H1a, dépôts dʼarrière cordon associés à la transgression ; 6 : séquence continentale Pléistocène ; 7 : S2, niveau de submersion maximal ; 8 : S1, limite de séquence.
5The Brenta River megafan and the Po sedimentary system bound the late Holocene Adige sedimentary system to the north and to the south, respectively (Ministry of University Research and Technology, 1997; Fontana et al., 2008). However, the boundary between the Adige and Po alluvial systems is not well defined, as branches of the Po have occasionally diverted onto the Adige alluvial plain and vice versa. The geomorphological evidence of such interaction is a complex network of alluvial ridges pertaining to the Po and Adige (Veggiani, 1972; Peretto, 1986; Marcolongo, 1987; Castiglioni, 1999), which intersect and overlie one another (fig. 1). These alluvial ridges incorporate sandy and silty channel deposits, natural levees and minor, proximal crevasse splays. The interdistributary basins generally are silty-clayey, although they frequently show accumulations of peat in extensive swamps or are occupied by dense networks of sandy distal crevasse channels and splays. Some of these ridges have already been recognised and studied. The feature termed here the “Fratta ridge” (fig. 1) is a major alluvial ridge that was built by the Po when it was flowing from Castelnuovo Bariano through Fratta Polesine (Piovan et al., 2010). Downstream of Fratta Polesine, this ridge divides into two branches: the southern one, here called the “Adria ridge” (fig. 1), continues in an approximately W-E direction towards Adria, while the northern one, here called the “Saline-Cona ridge” (fig. 1), runs SW-NE through Rovigo, Saline and Cona to the Venice Lagoon (Castiglioni, 1978; Piovan et al., 2010). Radiocarbon dates integrated with archaeological stratigraphy prove that the “Saline-Cona ridge” was active between the second half of the 3rd millennium and the end of the 2nd millennium BC (Piovan et al., 2010). Petrographical analysis of the Saline-Cona channel body testifies that this alluvial ridge (which crosses the present-day Adige) belongs to the Po system (Piovan et al., 2010). Po sands are characterised by a predominant quartzolithic lithofacies, the abundance of schistose serpentinites and the presence of glaucophane while Adige sands have a larger content of dolostone fragments and acidic volcanic rocks. Another major alluvial ridge, here called the “Montagnana-Este ridge” (fig. 1), was formed by the Adige River at the foot of the Euganean Hills between the Bronze Age and Roman times (Marcolongo, 1987; Marcolongo and Zaffanella, 1987; Balista, 2004). It runs through Montagnana, Este and Monselice to Pernumia, where it splits into two main ridges (Piovan et al., 2010), here termed the “Conselve” and “Bagnoli” alluvial ridges (fig. 1). The latter runs in a NW-SE direction from Pernumia to Agna, where it joins the Saline-Cona alluvial ridge. The Bagnoli alluvial ridge is attributed to an Adige palaeochannel on the basis of calcimetric analysis performed on a channel body sand sample. This palaeochannel was active around 3698-3321 cal. BP, on the basis of a radiocarbon dating performed in a buried tree trunk found in the same sand body sampled for the calcimetry (Zoletto, 1991). After an important avulsion near Bonavigo (La Cucca site), the Adige changed from its northern Montagnana-Este route to its present-day course (fig. 1). Archaeological and historical data suggest that this event possibly occurred in the late Roman-early Middle Ages (Zaffanella, 1979; Zerbinati, 2003; Balista, 2004). The alluvial ridge of the modern Adige River runs from Bonavigo through Legnago, Badia Polesine, San Martino di Venezze and Pettorazza Grimani to the cuspate delta mouth. Until the second half of the 18th century, the Adige was characterised by numerous meanders, subsequently cut off by the Venetian Government. The geomorphological and stratigraphic architecture of the late Holocene Adige ridges have yet to be studied in detail, despite their importance for the understanding of the interactions with the Po alluvial system and the coastal plain inland from the southern Venice Lagoon. This study focuses on the major Adige and Po alluvial ridges, i.e. the Conselve, Saline-Cona and the present-day Adige ridges, which are interconnected with and feed into specific delta lobes.
6A Digital Terrain Model (DTM) was processed (fig. 3) based on contours with a spacing of 0.5 m and a cell size of 5 m. Contours were derived from manual interpolation of spot heights from the “Carta Tecnica Regionale del Veneto” (1:5000 scale) in order to eliminate most anthropogenic constructions (e.g., artificial banks, railway embankments). Due to the elevation of alluvial ridges on the floodplain, it was possible to recognise them on the DTM and to map them in a GIS platform to obtain a geomorphological sketch (fig. 1). The interpretation of aerial orthophotographs (REVEN 90, scale 1:17000; GAI 1954-55, scale 1:33000/1:66000; Rossi-CRG 1997, scale 1:20000) and satellite images (from Google Earth) analysis allowed the recognition and mapping of the most important palaeochannels (fig. 1). In order to study the architecture of the alluvial ridges and the chronostratigraphic evolution of the alluvial plain, 28 manual boreholes up to 9 m in depth (about 6 m mean) were performed with an Edelman combination-type auger and a cylinder sampler. The cores were extracted, analysed and sampled in the field. The sedimentological field coring analysis included grain size, Munsell colour, mottles, redoximorphic traces, HCl effervescence for CaCO3 determination, nodules and concretions, plant or shell fragments, sedimentary structures and bed boundaries. The boreholes where positioned along three transects across the Conselve and modern Adige alluvial ridges, in order to provide geological cross sections approximately perpendicular to the supposed main flow direction. The cross sections were located where the morphology of the ridges is best defined, on the basis of the DTM and remote sensing analyses. In particular, they were located southeast of Conselve, east of Santa Margherita and across the Pettorazza Grimani palaeomeander (fig. 3).
7The logs were precisely positioned in altitude using the DTM. For this study, 6 carbonaceous layers (peat) and 3 wood fragments from the core samples were selected using the guidelines proposed by H.J.A. Berendsen and E. Stouthamer (2000) and submitted for radiocarbon dating (tab. 1). Peats were sampled with a cylinder sampler, which prevented contamination. The samples were collected from the inner part of the cores to prevent the possibility of external contaminations, stored in aluminium foil and dated by Angström Laboratory (Uppsala University, Sweden) or by Geochron Laboratories (Billerica-Cambridge, USA). Radiocarbon dates were calibrated using Calib 6.0 software (Stuiver and Reimer, 1993; Stuiver et al., 2005) and the calibration curve IntCal09 (Reimer et al., 2009). The full 2 sigma ranges are presented in tab. 1. Radiocarbon dates in the text refer only to 2 sigma ranges. As H.J.A. Berendsen and E. Stouthamer (2000) suggested for a similar alluvial context (the subsiding Rhine-Meuse delta plain), in order to determine the beginning of fluvial activity, a radiocarbon date from the top of organic beds underlying clayey overbank deposits was obtained, constraining the beginning of channel sedimentation. The radiocarbon date from an organic bed included in the natural levee was also used to date the channel activity.
8In order to reconstruct the palaeogeographical and stratigraphic evolution in the area inland from the southern Venice Lagoon, correlation between the stratigraphic successions presented in this paper and that of M. Zecchin et al. (2009) has been performed by comparing the sedimentary facies. The Cona cross-section, taken through the Saline-Cona Po alluvial ridge between Agna and Cona (Piovan et al., 2010; fig. 3), is reported and discussed in order to enhance chronostratigraphic correlations across the study area.
Fig. 3 – DTM of the study area with location of the studied cross-sections
Fig. 3 – MNT du secteur étudié avec localisation des coupes stratigraphiques
The DTM has been carried out in the Department of Geography – University of Padova, in collaboration with the Regional Agency for Environmental Protection and Prevention of the Veneto (ARPAV). 1: hydrography; 2: cross section; 3: borehole from Zecchin et al. (2009).
Le MNT a été effectué au Département de Géographie – Université de Padoue, en collaboration avec l’Agence Régionale pour la Prévention et la Protection de l’Environnement de la Vénétie (ARPAV). 1 : hydrographie ; 2 : transect ; 3 : sondage décrit in Zecchin et al. (2009).
Tab. 1 – Results of the radiocarbon dating
Tab. 1 – Datations par le radiocarbone
*Data recalibrated from S. Piovan et al. (2010).
*Données recalibrées par S. Piovan et al. (2010).
9DTM analysis aimed to obtain a detailed map of the alluvial ridges, with accurate determination of the morphology along the cross sections. Figure 3 shows the present-day Adige alluvial ridge from San Martino di Venezze to Cavarzere. The largest palaeomeander of the Adige at Pettorazza Grimani, rectified in 1783, is also readily recognisable in the DTM. Here, the modern Adige alluvial ridge is up to 4 m higher than the surrounding plain and it is more than 1 km wide, due to deposits from repeated crevasse splays, which led it to grow markedly.
10The minor ridges (Bagnoli, Conselve, and Terrassa ridges), which branch off in southward directions from the Adige ridge of Montagnana-Este (fig. 1), can be seen in the DTM (fig. 3). In its proximal part, the Bagnoli ridge is about 2 km wide, whereas before joining the Saline-Cona alluvial ridge at Agna it becomes narrower, up to 500 m wide (fig. 3). The Conselve alluvial ridge runs from Pernumia to Concadalbero in a NNW-SSE direction, where it joins the Saline-Cona alluvial ridge (fig. 3). The feature termed here the “Terrassa alluvial ridge” (fig. 1 and fig. 3) is the northernmost ridge of the Adige alluvial system, at the border with the Brenta River system. It runs in a W-E direction from Pernumia and, ca. 2 km east of Terrassa Padovana, it divides into two ridges: the northern ridge goes in a NW-SE direction and joins the Conselve ridge just upstream of Concadalbero. The southern branch of the Terrassa alluvial ridge joins the Conselve ridge about 2 km east of Arre. Both the Terrassa and Conselve alluvial ridges are up to 600 m wide and up to 3 m higher than the surrounding plain (fig. 3). The Saline-Cona ridge (shown in fig. 3 only from Saline to the margin of the lagoon) is also well outlined in the DTM, having an average width of 600 m and up to 3 m elevation above the surrounding plain.
11The Conselve cross section is located across the Conselve alluvial ridge between Conselve and Arre. It shows, at the base of boreholes CV04 and CV06, about 0.5 m of thick clayey silt, overlain by a whitish horizon with high percentage of CaCO3 nodules at a mean height of -4 m a.s.l. (fig. 4). This latter is interpreted as the calcic “caranto” palaeosol (Mozzi et al., 2003; Fontana et al., 2008). The section continues with an alternation of clay and organic-rich material up to about -2.3 m a.s.l. (CV06). This unit is truncated by an erosion surface that represents the bottom of a sand body that has been recognised in other cores (e.g., CV04). This sand body is interpreted as a channel body, which is about 200 m wide. In CV05, the sands are at least 5 m thick; they also appear at the bottom of borehole CV01, between -3 and -2 m a.s.l. The channel deposits are also recorded in CV03 and CV05 up to 3.5 m a.s.l. In CV04, CV03, CV02 and CV01, above ca. 1 m a.s.l., intercalations of fine sands and silty sand deposits organised in “wings shaped architecture” have been recognised adjacent to the main channel body. This succession is related to the natural levee complex of the latest phase of aggradation. The alternation of fine, organic and sandy sediments in CV02, CV01, and CV04 indicates that the fluvial activity probably continued with crevasse splay events that contributed to the growth of the alluvial ridge.
12A peat layer, from -2.7 to -2 m a.s.l., has been sampled and radiocarbon dated at the bottom and the top in CV04. The results have given 3557-3699 cal. BP and 3201-3354 cal. BP, respectively (tab. 1). A peat intercalation in the natural levee complex has been sampled in CV01 at 0.9 m a.s.l. The radiocarbon dating yielded an age of 1950-2119 cal. BP. This shows that the channel was active after 3354-3201 cal. BP and during 1950-2119 cal. BP.
Fig. 4 – Conselve cross-section
Fig. 4 – Coupe stratigraphique de Conselve
See fig. 3 for location. 1: medium and coarse sand; 2; fine sand; 3: silty sand – alternation of silt and sand; 4: sandy silt, silt and clayey silt; 5: clay and silty clay; 6: humic silt and clay; 7: peat.
Pour la localisation, voir fig. 3. 1 : sables moyens à grossiers ; 2 : sables fins ; 3 : sables limoneux – alternance de lits sableux et limoneux ; 4 : limons sableux, limons et limons argileux ; 5 : argiles et argiles limoneuses ; 6 : limons et argiles humifères ; 7 : tourbe.
13The cross section shown in fig. 5 cuts the Conselve-Santa Margherita alluvial ridge about 3 km inland from the present-day coastline, southwest of Santa Margherita (fig. 3). Here, the bottom of the cross section is characterised by the presence of a clayey-organic sediment sequence. This fine-grained sediment sequence continues to about -2.5 m a.s.l. in SM02 and in SM05 with clay sediments. A sand body in SM04, SM03, and SM01, about 400 m wide, has been identified and interpreted as cutting the organic sequence. Silty sand and fine-sandy deposits characterise the wings shaped architecture of the natural levee deposits of the channel in SM05 and SM02 from about -3 to -0.5 m a.s.l. Three radiocarbon dates have been undertaken on SM02. The first two have been carried out at the bottom (-7.3 m a.s.l.) and the top (-5.2 m a.s.l.) of the thickest organic layer and yielded ages of 5754-5595 and 3383-3158 cal. BP, respectively. The third radiocarbon date, on a sample from the organic layer intercalated within the natural levee deposits at -2.5 m a.s.l., has given the age 2157-1995 cal. BP (tab. 1). The results demonstrate that the channel was active after 3383-3158 cal. BP and during 2157-1995 cal. BP.
Fig. 5 – Santa Margherita cross-section
Fig. 5 – Coupe stratigraphique du Santa Margherita
See fig. 3 for location. See fig. 4 for caption.
Pour la localisation, voir fig. 3. Pour la légende, voir fig. 4.
14The Pettorazza cross section (fig. 6) is located across the modern Adige River alluvial ridge, in correspondence with the palaeomeander cut off in 1783 by the Venetian Republic government. The bottom of the section revealed a sequence of silts, clays and organic material. In the southeastern part of the cross section there is evidence of a sand body, buried at -1.5 m a.s.l. It is characterised by a tabular shape, between 1 and 2 m thick in P18, P15, P05, P16, P04, and P03, while it is thicker in P01, where the base has not been reached. It has been interpreted as a crevasse splay deposit from the main channel. This latter is represented by a prominent body of sand, at least 6.5 m thick and more than 400 m wide at the higher elevation of the hill, recognised from boreholes P14 to P05.
15Three samples have been radiocarbon dated: in borehole P14 two samples were taken at depths of -3.55 m a.s.l. and -1.77 m a.s.l. The first was sampled at the bottom of a fine-grained organic sequence and the second from a peat layer buried by sandy natural levee deposits of the main channel recognisable between -1 and 1.75 m a.s.l. The samples yielded ages of 4299-3637 cal. BP and 2142-1560 cal. BP, respectively (tab. 1). The third radiocarbon date, performed on P16 from -1.5 to -1.4 m a.s.l., yielded an age of 702-699 cal. BP for peat overlying a sandy tabular crevasse-splay deposit. This demonstrates that the channel was active between 2142-1560 cal. BP and 702-699 cal. BP.
Fig. 6 – Pettorazza cross-section
Fig. 6 – Section transversale du Pettorazza
See fig. 3 for location. See fig. 4 for caption.
Pour la localisation, voir fig. 3. Pour la légende, voir fig. 4.
16The results of this study enable reconstruction of the Holocene evolution of the study area. The “caranto” paleosoil, already recognised in the Cona cross section (Piovan et al., 2010), has been found in the Conselve cross section at a depth of about 7 m (fig. 7). It represents a sedimentary hiatus between 17 and 9 ka cal. BP in the whole Friulian-Venetian Plain (Mozzi et al., 2003; Fontana et al., 2008; Zecchin et al., 2009). The top of the caranto corresponds to the S1 surface in the southern Venice Lagoon (Zecchin et al., 2009; fig. 2). This confirms the presence of a major regional unconformity that separates late Pleistocene from Holocene deposits in the vicinity of the Venice Lagoon. Holocene alluvial aggradation started above this unconformity as a consequence of post-glacial sea-level rise. In Conselve and Cona, the lowest part of the Holocene deposits above the unconformity is characterised by an organic succession. This can be stratigraphically correlated with the lowest peat levels in the Santa Margherita cross section. Here, the beginning of organic deposition has been radiocarbon dated to 5754-5595 cal. BP (age of the lowest peat sample). The radiocarbon dating at the top of the peat bed in Santa Margherita yielded an age 3383-3158 cal. BP, which is coeval with the top of the peat layer dated at 3354-3201 cal. BP in the Conselve cross section. This chronostratigraphic correlation suggests that this sector of the alluvial plain inland from the coastline was subject to extensive organic deposition between ca. 5.7 and 3.2 ka cal. BP. Peat formation was probably related to high groundwater level during the maximum marine transgression, similar to what has been suggested by A. Amorosi and S. Milli (2001) in the Po and Tevere delta plains. Swampy conditions and sediment starvation in the study area seem to have persisted during the highstand phase until around 3 ka cal. BP, as indicated by the deposition of the peat layer.
17Above this fine-grained organic succession, a major aggradation of sandy channel belts has been recognised, which can be related to the formation of alluvial ridges by both Po and Adige palaeochannels. The main geomorphological products attributable to palaeochannels activity correspond with the Conselve ridge (recognised in the Conselve and Santa Margherita sections), the Saline-Cona ridge (Cona section) and the present-day Adige ridge (section at Pettorazza). This evidence allows a better understanding of the complex relationship between the Po and Adige alluvial systems.
18First, the comparison between the Cona section and dating of the Bagnoli ridge formation (dated to 3698-3321 cal. BP; Zoletto, 1991) demonstrates that both the Saline-Cona and Bagnoli ridges were active between ca. 4 and 3 ka cal. BP. This shows that the Adige and Po joined at Agna and shared an outflow into the lagoon area during this time period (fig. 8). This phase was followed by the deactivation of the northernmost Po palaeochannel (Saline-Cona branch), dated about 3 ka cal. BP (Piovan et al., 2010). From this time fluvial activity in the alluvial plain inland from the southern Venice Lagoon was related only to the Adige. The extension of the Conselve ridge to the coast coincides with the terminal tract of the Saline-Cona Po palaeochannel, as shown in the Santa Margherita cross section, where significant channel aggradation is recorded from Roman times.
19The activity of the composite Po-Adige system during 4-3 ka cal. BP and its continuation as the Adige alone between 3 and 1.5 ka cal. BP was coeval with sediment deposition in the delta and related lagoonal and littoral deposits in the southern Venice Lagoon, as evidenced by the radiocarbon dating of units H2a and H2b (lower part) by M. Zecchin et al. (2009; fig. 2). This chronology and the geomorphological continuity with the Po-Adige alluvial plain indicates that the delta was constructed by these two rivers (fig. 8).
20The chronostratigraphy of the Pettorazza cross section testifies that the Adige River started to flow along the southern course of Legnago-Badia Polesine and Pettorazza after 2142-1560 cal. BP. It was fully active at around 724-935 cal. BP (fig. 6), as proved by radiocarbon dating of a peat layer above crevasse splay deposits in the southeastern part of the cross section. At that time, the Montagnana-Este Adige palaeochannel and the related Bagnoli and Conselve palaeochannels were no longer active. As a consequence, the delta in the southern lagoon was also deactivated (fig. 8). This led to the drowning of the abandoned delta system due to continuous relative sea-level rise and to the consequent deposition of the lagoonal tidal flats and channels that characterise unit H3 of M. Zecchin et al. (2009).
Fig. 7 – Cona cross-section
Fig. 7 – Coupe stratigraphique du Cona
For the location, see fig. 3. For the legend, see fig. 4.
Pour la localisation, voir fig. 3. Pour la légende, voir fig. 4.
Modified from Piovan et al., 2010.
Modifié à partir de Piovan et al., 2010.
Fig. 8 – Simplified map showing the late Holocene evolution of the Adige and Po delta systems in the southern Venetian Plain
Fig. 8 – Carte simplifiée montrant l’évolution des systèmes deltaïques de l’Adige et du Pô au cours de l’Holocène dans la plaine du sud de la Vénétie
1: A0, Adige, until Roman Age; 2: A1, Adige, 4-3 ka cal. BP; 3: A2, Adige, Roman Age; 4: A3, Adige, Mediaeval – Modern Age; 5: PO, Po, until Iron Age; 6: PI, Po, 4-3 ka cal. BP; 7: P2, Po, until Iron Age; 8: P3, Po, Mediaeval – Modern Age; 9: Saline-Cona Po and Conselve Adige deltas; 10: Po Mediaeval – Modern delta; 11: Adige Mediaeval – Modern delta; 12: Iron and Roman Age coastal dunes (Bondesan et al., 1995; Stefani and Vincenzi, 2005). Coastline – 13: around 7.0-5.7 ka cal. BP (Rizzetto et al., 2003; Amorosi et al., 2008b; Simoni and Corbeau, 2009).
1 : A0, Adige avant la période romaine ; 2 : A1, Adige vers 4-3 ka cal. BP ; 3 : A2, Adige durant la période Romaine ; 4 : A3, Adige durant les périodes médiévale et moderne ; 5 : P0, Pô avant lʼÂge du Fer ; 6 : P1, Pô vers 4-3 ka cal. BP ; 7 : P2, Pô avant lʼÂge du Fer ; 8 : P3, Pô durant les périodes médiévale et moderne ; 9 : delta commun au Pô de Saline-Cona et à lʼAdige de Conselve ; 10 : delta du Pô aux périodes médiévale et moderne ; 11 : delta de lʼAdige aux périodes médiévale et moderne ; 12 : dunes littorales de lʼÂge du Fer et de lʼépoque romaine (Bondesan et al., 1995 ; Stefani et Vincenzi, 2005). Trait de côte – 13 : vers 7-5,7 ka cal. BP (Rizetto et al., 2003 ; Amorosi et al., 2008b ; Simoni et Corbeau, 2009).
21This research contributes to regional geomorphological and palaeogeographical reconstruction of the southern Venetian alluvial plain. In particular, it enhances the work of S. Piovan et al. (2010) in the definition of the role of the Po and Adige alluvial systems in the construction of the alluvial plain and delta lobes. On the basis of new geomorphological and stratigraphic data, it is possible to assign the delta system recognised by M. Zecchin et al. (2009) in the southern Venice Lagoon to the Saline-Cona branch of the Po, which was active during 4-3 ka cal. BP. This implies that the Po delta extended as far as 30 km north of the present-day position of the river. In that time, the Adige did not directly reach the sea, as it was a tributary of the Po at Agna, flowing through the Bagnoli ridge. The northernmost lobe of the Po delta was thus fed by sedimentary input from both the Po and Adige. The deactivation of the Saline-Cona Po branch through avulsion just upstream of Rovigo, around 3 ka cal. BP, led to a southwards shift of the Po delta system (fig. 8). The Adige maintained its mouth in the same area and, from 3 ka cal. BP to Roman times, built its own delta in the southern Venice Lagoon, prograding onto the previous northern Po delta lobe. Avulsion by the Adige at Bonavigo during the early Middle Ages led to the abandonment of the Montagnana-Este-Conselve-Santa Margherita course along the southern foot of the Euganean Hills. As a consequence, the delta in the southern Venice Lagoon was abandoned and the Adige started to build another delta in its present-day position, about 15 km further to the south (fig. 8).
22The general seaward migration of the coastline since the mid-Holocene indicates that sediment supply from the Po and Adige overwhelmed relative sea-level rise during the highstand. The results show that major avulsive events in the upstream tracts of the Po and Adige Rivers forced the migration of delta lobes. Detailed reconstruction of the directions of the major branches of the Po and Adige in the alluvial plain allows a better understanding of deltaic evolution along the Adriatic coast. A general southward shift of the river systems and related delta lobes (fig. 8) is recorded during the last 4 ka, suggesting a migration of the depocentre towards the present Po delta area.
The research was carried out with the support of projects “Via Annia, an ancient Roman road”, funded by ARCUS s.p.a., and “Padova underground: a geoarchaeological investigation of the city”, funded by Fondazione Cariparo. The authors thank Sandra Primon for the precious discussion, the two anonymous referees and the guest editors Stéphane Cordier and David Bridgland for their careful review and comments.