The climate-triggered western shift of the confluence between the Dora Baltea and Po rivers (north-western Italy) during the late Holocene

Abstract

The study on the migrations of the confluence between the rivers Po and Dora Baltea was based on a detailed observation of aerial photographs, field surveys and sediment stratigraphy. The age of the sediments and morphological features was based on radiocarbon datings, on the presence of archaeological settlements and ancient artefacts and on historical data. The aerial photographs of the GAI 1954 flight of the Italian Air Force were used because in the year 1954, some low terraces were still clearly visible, while today are almost completely obliterated because of the works for the improvement of the rice fields. A succession of nine fluvioglacial and fluvial terraces and many abandoned riverbeds have been identified: the interpretation of the morphological features made it possible to identify the Po and Dora Baltea evolution during the late Holocene. The study established that during the last 3000 years, the confluence of the Dora Baltea into the Po has constantly migrated to the west and that this migration occurred during alluvial phases dating back to the Iron Age, 2nd century BC–1st century AD, 5th century AD, 6th–8th century AD and 15th–19th century AD. The alluvial phases occurred during periods of increased floods in northern Italy and advances of the Alpine glaciers in northwestern Italy and Switzerland. Neither tectonic deformations nor anthropic actions seem to have influenced fluvial evolution in the last 3000 years. The westward migration of the confluence between the rivers was therefore triggered by climatic changes, but caused by the different responses of the catchments of the Dora Baltea and the Po rivers to climatic changes, and by the greater slope of the Dora Baltea flood plain.

Keywords

alluvial phases fluvioglacial and fluvial terraces glacier fluctuations late-Holocene river migration NW Italy Po and Dora Baltea rivers evolution Po Plain

Introduction

Alluvial plains are often subject to changes in riverbeds and the areas occupied by rivers in the recent past may be at risk of flooding during exceptional rainfall events. Understanding and dating the phases of river evolution allow us to recognize the factors that influence the hydrological regime and to infer the causes that have triggered the migration of the riverbeds over the last millennia. The data on the river evolution can be used for the management of the alluvial plain, to avoid the improper use of some areas, and for the planning of the works needed for securing areas subject to flood hazard.

The correlation between climatic changes and variations in the hydraulic regime, in stream load and in riverbed morphology, also provides elements to evaluate the possible impact of climate changes assumed for the near future.

In this study, alluvial terraces and sediment successions are presented and discussed. The mapping of the terraces is also based on the use of aerial images taken in 1954. The images, though having a lower resolution than current aerial and satellite ones, represent, however, an environment much less subject to anthropic impact. In 1954, some low terraces were still clearly visible, while today, they are almost completely obliterated due to works for the improvement of the rice fields or to settlements and infrastructures.

The area under investigation is part of the western portion of the Po river floodplain (Piedmont, north-western Italy) and is included between the rivers Dora Baltea to the west, Sesia to the east, the Po river and the nearby Monferrato hills to the south (Figure 1).

Figure 1.

(a) The Po and Dora Baltea catchment basins, (b) topography and river network of the studied area and (c) position of the boreholes cited in the present paper.

The southern portion of this area consists of a succession of low terraces dating from the late Pleistocene to the present day, but the scarps that separate the surfaces of some terraces are often barely visible due to the low difference in altitude and to the levelling works for rice cultivation.

The aim of this work is to reconstruct the evolution of the river network during the late Holocene, both through the dating of sediments and by the analysis of terraces, in order to understand the causes and timing of the migration to the west of the confluence between the Dora Baltea and the Po highlighted in previous papers.

The stratigraphy of the sediments and the evolution of the terraces were used to verify whether the area was shaped only by alluvial and erosive phases linked to climatic variations or also by other factors, and if the late-Holocene river migration may have created the conditions for exposing older terrace surfaces to flood hazard.

Site setting

The area covered by this study corresponds to a part of the Po Plain located just north of the Monferrato hills. It is an area crossed by three Alpine rivers (the Po, the Dora Baltea and the Sesia) and by local streams (the Stura and the Cornasso) fed by springs and connected to a dense network of artificial canals.

The Po is the largest river, being the collector of the waters of most of the valleys of the north-western Alps (Figure 1a), at the head of which there are few glaciers of limited size. The Dora Baltea, which flows into the Po in the western part of the studied area (Figure 1a), is fed by the Valle d’Aosta, the valley that drains the Italian side of some of the highest mountains in the Alps (M. Bianco, 4810 m; M. Rosa, 4634 m; M. Cervino, 4478 m; M. Gran Paradiso, 4061 m) characterized by the presence of some of the largest glaciers on the southern side of the Alps.

Currently, each river flows in a single stream channel, mainly because of the works of hydraulic regulation; however, the historical cartography (Crosio and Ferrarotti, 1996; Pistan, 2003) shows that the rivers were formed by at least two channels that flowed around islands.

The cartography also shows the presence of still very evident ancient river beds indicated by narrow, elongated depressions, often in the form of meanders, and the presence of oxbow lakes. The depressions corresponding to the ancient river beds surround former fluvial islands.

As for the local streams, the Stura is the longest and is a tributary of the river Sesia (Figure 1b); it is fed by springs located NW of Fontanetto Po, and flows in a mildly depressed and elongated area. At the eastern end of the area, this depression, together with various others occupied by the tributaries of the Stura stream, was identified by Tropeano and Olive (1989a, 1989b) as an ancient abandoned Alpine riverbed. The Stura’s meanders have small wavelengths and their amplitude is equal to or less than the width of the elongated depression that follows the old riverbed, formed by meanders of greater amplitude and wavelength. Even if the Stura bed is used in the complex irrigation system, its course is only partially conditioned by anthropic measures.

The Cornasso stream, which flows in abandoned riverbeds of the river Po already reported in literature (Giraudi, 2014b), begins in the area of Trino and, heading east, flows into the Stura.

According to previous geological studies, the area under examination consists of terraces sloping south: middle and late Pleistocene terraces in the North (Montrasio et al., 1969), upper Pleistocene in the central area and Pleistocene-Holocene in the southern area (Dela Pierre et al., 2003a, 2003b).

According to Carraro et al. (1970), towards the eastern end of the study area, there are low terraces, some of which not previously highlighted, that can be dated to the upper Pleistocene and Holocene.

According to a more recent paper (Giraudi, 2014b), in which the ages of terraces are more clearly defined, a terrace dated as late Holocene, slightly lower than the late upper Pleistocene one, forms a depression cutting the older terrace. In correspondence with the same late-Holocene terrace, the presence is reported of alluvial cover made up of thin sediments deposited by local streams.

On the whole, apart from the oldest northern terraces, the scarps that separate the most recent terraces can be very low and the evaluation of their heights is made difficult by the considerable artificial levelling carried out in order to increase the size of the rice fields.

In the central portion of the study area and in the area near the Po river west of Casale Monferrato, the alluvial sediments do not exceed 10–15 m in thickness and lie on tertiary marine sediments similar to those that form the adjacent Monferrato hills.

In the same area, the marine bedrock was eroded starting from the final phases of the early Pleistocene (Giraudi, 2017), and below the late Pleistocene and Holocene sediments, pockets of much more ancient alluvial deposits could be preserved.

Methods

Improvement of our knowledge of the geological evolution of the area was based on a detailed observation of aerial photographs, field surveys, sediment stratigraphy and the presence of archaeological settlements and ancient artefacts.

Regarding aerial photographs, those relating to the GAI 1954 flight of the Italian Air Force were used. In that period, in fact, the rice fields were still being levelled by low-impact mechanical methods.

The rice fields, which had to be necessarily sub-horizontal in order to be submerged by just a shallow layer of water, were smaller in size than today and, as a consequence, they were far more conditioned by the morphological features, and the shape of the fields highlighted even scarplets a few decimetres high.

The interpretation of more recent aerial photographs no longer allows such detail to be achieved. Today, the area is much more anthropized and modern agricultural levelling by means of larger mechanical means has had a major impact on the alluvial plain, incorporating in a single rice field various smaller fields previously separated by scarplets, or by changing the slope of some terraces.

In the course of time, to enable the rice fields to be irrigated, the natural hydrographic network has been partially modified and connected to a dense network of canals and irrigation ditches: it is therefore impossible to represent the entire network of channels at the scale of the maps of this study. Hence, it was chosen to include in the figures only those stretches of the watercourses that still show natural trends (large and small curves, meanders on a small and large scale) that cannot be confused with artificial canals characterized mainly by straight or slightly curvilinear stretches.

To define the stratigraphy, we took into account the results of the boreholes drilled in the area in various periods (ENEL, 1977, 1984; ARPA Piemonte, 2018, which reports also the data from ENEL), the observations of artificial exposures highlighted by excavations, starting from the 1980s of the 20th century, and sediment outcrops exposed by river erosion.

The chronological framework of the sediments is based on radiocarbon datings performed on tree remains (Charrier and Peretti, 1977; Giraudi, 1998; Tropeano and Olive, 1989a, 1989b) and wooden poles (Giraudi, 1998) and on charcoal found during archaeological excavations (Negro Ponzi Mancini, 1989) and on archaeological artefacts (Giraudi, 1998, 2014b; Negro Ponzi Mancini, 1989) (Table 1).

Table 1.

Radiocarbon ages, reported in Negro Ponzi Mancini (1989), Tropeano and Olive (1989a) and Giraudi (1998). The calibration program used was not stated by the authors.

Material	Radiocarbon age	Calibrated age	References
Wood	1595 ± 50 BP	AD 346–575	Charrier and Peretti (1977)
Wood	1590 ± 60 BP	AD 380–600	Tropeano and Olive (1989a)
Wood	1515 ± 45 BP	AD 440–640	Tropeano and Olive (1989a)
Charcoal	1330 ± 60 BP	AD 660–777	Negro Ponzi Mancini (1989)
Charcoal	1175 ± 70 BP	AD 776–897	Negro Ponzi Mancini (1989)
Charcoal	1065 ± 55 BP	AD 974–1031	Negro Ponzi Mancini (1989)
Charcoal	1050 ± 60 BP	AD 997–1015	Negro Ponzi Mancini (1989)
Wood	710 ± 60 BP	AD 1225–1400	Giraudi (1998)
Wood	510 ± 50 BP	AD 1325–1460	Giraudi (1998)
Wood	460 ± 50 BP	AD 1410–1620	Giraudi (1998)

The set of data made it possible to highlight the phases of major alluvial sedimentation. The phases identified through geological methods were then compared with periods characterized by heavy floods, known through historical sources (Camuffo and Enzi, 1994), starting from 2nd century BC.

The comparison between the geological maps reported in this paper and the representation of the area in ancient maps then enabled the positions of some fluvial stretches and their variations starting from the 13th century to be dated.

The alluvial terraces

The area studied consists of a succession of fluvial and fluvioglacial terraces, shaped by the rivers Dora Baltea and Po (Dela Pierre et al., 2003a, 2003b; Giraudi, 2014b; Montrasio et al., 1969a).

In particular, the highest terraces (terraces A, B and C in Figure 2) are formed by fluvioglacial sediments connected to the Valle d’Aosta moraine amphitheatre, lying about 20 km upstream from the confluence of the Dora Baltea into the Po. The other terraces (D, E, F, G, H and I) are of fluvial origin.

Figure 2.

The fluvioglacial and fluvial terraces, and the abandoned riverbed in the southern Vercelli Plain.

The scarps between terraces C, D, E, F, G and H are always fairly low, generally between 1.5 and 0.5 m. The most evident scarps are present from the area just west of Fontanetto Po to the area between Palazzolo and Trino. Just east of Trino, the scarps become almost imperceptible, while towards the eastern edge of the area studied, the scarps are a little higher.

Only the scarps that separate terraces H and I are very clear in correspondence with the convex portion of the bends of the current riverbeds or abandoned meanders, and can reach heights of 3–4 m or more.

Four schematic N-S sections showing the series of terraces are drawn in Figure 3.

Figure 3.

North–south schematic sections in four different places of the studied area showing terraces, sediments, radiocarbon ages and archaeological findings.

Terraces A and B

These are the highest terraces forming the northern limit of the area studied (Figures 2 and 3). Terrace A consists mostly of fluvioglacial deposits connected to the moraines of the Valle d’Aosta and can be dated to the LGM, according to Gianotti et al. (2015), which in the Valle d’Aosta glacier occurred about 21,000 years ago (Gianotti et al., 2008). Terrace B is the result of the re-shaping of the previous terrace because of local streams (Giraudi, 2014b).

Terrace C

This is a terrace of fluvioglacial origin that can be followed upstream (Giraudi, 2014b) and connects with the moraines of the first phases of glacier retreat, reported by Gianotti et al. (2008), following the local LGM. The shaping of the terrace ceased, therefore, after the local LGM. The stratigraphy of the sediments that form terrace C has been highlighted by various boreholes (ARPA Piemonte, 2018; ENEL, 1977, 1984) (Figure 1c). The terrace is mainly composed of generally coarse sandy gravels, alternating with medium gravel, covered by a few decimetres of sandy and silty alluvial sediments.

Terrace D

Terrace D, observed exclusively in the central portion of the area (Figure 2), is represented by some discontinuous surfaces, isolated between terraces of lower altitude located both north and south. Terrace D does not show any continuity with terraces that extend towards the moraines of the Valle d’Aosta, and it is not possible to establish whether it is connected with the sedimentation of the Dora Baltea or the Po. The lithology of the sediments cannot help to solve the problem because, before reaching the study area, the Po cut into fluvioglacial sediments of the Dora Baltea basin: the sediments of the two basins are already mixed. Terrace D is slightly lower than terrace C (Figure 3). NW of Balzola, in the only area where the two terraces are close together, the difference in altitude can be estimated as 1–1.5 m. The sediments forming terrace D, observed in exposure and boreholes, consist of sandy gravels with interbedded gravel in a sandy matrix. On the surface of terrace D, as on all the most recent terrace surfaces, there are alluvial sandy, silty and clayey sediments deposited after the shaping of the terrace. The fine-grained sediments suggest that the top of the terrace was submerged during overbanks events reached during later alluvial phases. The alluvial sandy sediments that cover the top of terrace D of the Pobietto area are cut by some late early and middle Bronze Age graves (Venturino Gambari, 2006; Venturino Gambari and Perotto, 1996), and said graves are covered by other silt and silty-sandy alluvial sediments. There are no direct dating elements, but terrace D is more recent than LGM and older than the late early-middle Bronze Age, and its age is assumed to be late upper Pleistocene–early Holocene.

Terrace E

Terrace E occupies mostly the long, narrow depression that separates terraces C and D in the central area, but becomes much wider towards the east (Figures 2 and 3). The scarp between terraces D and E, in the western portion, reaches 1–1.5 m in height, but to the east of Trino, it disappears almost completely. In the area located south of terrace D, a remnant of terrace E is preserved near Palazzolo. The same terrace is continuous in the area east of Morano Po and is connected with the larger terrace described above. The stratigraphy of some boreholes shows that the sediments that form terrace E consist mainly of sandy gravels underlying silty sands. A few decimetres below the surface of terrace E, artefacts of the middle and late Bronze Age (14th–13th century BC) (Gambari, 1989) have been identified, covered by sandy gravels and silty sand. On the sediments overlying the Bronze Age artefacts, remains of settlements dating back to the 1st–2nd century AD have been found. The top surface of the terrace therefore seems to have been shaped in a period during the 1st millennium BC.

Terrace F

Terrace F forms part of the area between Crescentino and Palazzolo and does not extend north of terrace D (Figures 2 and 3). The maximum height of the scarplet that separates terraces E and F can be estimated as around 1 m. Where terrace F comes in contact with terrace D, however, the height of the scarp may be slightly higher. The sediments that form terrace F, observed in exposures and boreholes, are sandy gravels with interbedded lenses of gravelly sand, underlying a few decimetres-thick layer of silty sand.

In the area of Crescentino, the summit sediments of terrace F cover a Roman grave dating back to the 1st century AD, excavated in the same locality where there are also some graves of the 1st century BC (La Rocca, 2000). Other graves of the 1st century BC (La Rocca, 2000) were found near the surface of the same terrace, just east of Crescentino. The graves must, therefore, have been excavated on surfaces, probably the remains of an older terrace, surrounded by depressions corresponding to old abandoned riverbeds, filled in during the shaping of terrace F. The terrace top must have been shaped during a period after the 1st century AD. It is evident, therefore, that beneath the surface of terrace F, and probably of the younger ones, lie the sediments of older terraces. The older sediments can be hidden and never appear in the southern portion of the area because they were almost wholly obliterated by the most recent fluvial evolution.

Terrace G

Terrace G runs from Crescentino to the eastern edge of the area along a strip approximately parallel to the Po riverbed (Figures 2 and 3). The top portion of the terrace is about 1 m lower than the previous one. Some cores drilled in the sediments that form the terrace indicate that it consists of sandy gravels with interbedded layers of gravelly sand of alluvial origin.

Terrace H

Terrace H is not very extensive and is mostly formed by a series of abandoned river channels cut deeply into terrace G to the west of Trino (Figures 2 and 3). The remnants of the channels lie at lower altitudes and cut the older abandoned riverbeds. The alluvial sediments that form the terrace, observed on the freshly dug banks of a channel about 2 m deep SW of Trino (Figure 4), consist of gravel with little sand at the base, sand with stratified silt lenses and some lenses of sandy gravel and silty sand. The excavation of another channel, in the area SE of Pobietto, has instead shown a more complex stratigraphic succession (Figure 4). The sediments are formed, at the bottom, by weathered alluvial sandy gravels, underlying alternating layers of alluvial gravelly sand and sandy silt, and sometimes dark clay horizons of palustrine origin. The sedimentary bodies indicate a series of erosion and sedimentation phases. The oldest gravels could also be the remains of terraces that have now disappeared because of river evolution.

Figure 4.

(a) Stratigraphy of the sediments sampled in two boreholes near the archaeological site of San Michele, east of Trino, and geological section exposed on the banks of canals dug through terraces (b) E, F, (c) H and I since 1980.

Terrace I

This is the lowest terrace that runs along the Po and the Dora Baltea (Figures 2 and 3). The scarp that separates it from terrace H can reach a height of 3–4 m. Some boreholes have shown that the sediments forming the terrace are made up mostly of sandy gravel with interbedded sand lenses.

However, most of our knowledge on sediment stratigraphy derives from the survey of the banks of the river Po, in the stretch that runs along the Monferrato hills, and the excavation of a canal several hundred metres long observed in the area SE of Pobietto. In the exposure at Pobietto (Figure 4), the sedimentary succession is represented by layers and lenses of sandy gravels, sands, silty sands and clayey silts. Along the banks of the river Po, it can be observed that alluvial deposits lie on tertiary marine sediments similar to those that form the Monferrato hills (Giraudi, 2017). Moreover, as reported in Fozzati and Giraudi (1983) and Giraudi (1998), it was observed that about 100 wooden piles were planted in the tertiary sediments outcropping at the margins of the riverbed. Two wooden piles, dated by the radiocarbon method, provided ages of AD 1310 ± 90 and 1390 ± 70 cal. yr and were locally covered by gravelly sandy and then sandy alluvial deposits. The last sediments contained, at the base, the remnant of a tree dated AD 1510 ± 105 cal. yr. According to Giraudi (1998), the piles were fixed during a phase of prevailing erosion that ceased in the 15th century or earlier. Later, a phase of strong alluvial sedimentation began that led to the shaping of terrace I. North of Camino, during the flood of the year 2000, the remains emerged of medieval walls previously covered by alluvial sand and sandy silts forming terrace I. The banks of the Po, cutting terrace I, also highlight an extremely interesting fact that suggests the presence of more ancient alluvial deposits covered by younger alluvial sediments. In some places, south of Palazzolo, north of Camino, west of Morano Po, but also in a quarry, now abandoned, located just south of the Po riverbed near Coniolo, bronze artefacts dating back to the Bronze Age (early, north of Camino; middle, near Palazzolo and Coniolo and late, west of Morano Po) have been found (Facchin, 1997; Janigro D’Aquino, 1979; Viale, 1970). The artefacts were lying in correspondence with the contact between the alluvial sandy gravels and the tertiary marine sediments. The alluvial sediments that cover the artefacts appear to have been deposited in three different periods: the first following the early Bronze Age (i.e. in the first half of the second millennium BC or the end of the third), the second after the middle Bronze Age (i.e. in the second half of the second millennium BC) and the third following the late Bronze Age (i.e. in the first millennium BC). It is likely that the Po riverbed cuts alluvial deposits (and probably ancient riverbeds) sedimented after the early, middle and late Bronze Age. In the areas where the artefacts were found, the river water flowed directly over the bedrock. The alluvial sediments covering the artefacts of the early and middle Bronze Age correspond chronologically to those underlying the settlement of the late Bronze Age lying below the top of terrace E, while those subsequent to the late Bronze Age correspond chronologically to those that form the top of terrace E.

Morphology of the abandoned riverbeds affecting the alluvial terraces and stratigraphy of their sedimentary successions

The alluvial terraces are incised by various narrow depressions, from 1–2 km to 12–13 km long, that extend from west to east (Figure 2). In some cases, the depressions are still very evident, in others just perceptible. The depressions, some of which have already been reported and interpreted as abandoned river channels (Dela Pierre et al., 2003a, 2003b; Giraudi, 2014b), are clearly the remains of ancient riverbeds of the Dora Baltea and the Po. In the depressions flow the streams of the local hydrographic network, and therefore, both the coarse sediments of the Alpine rivers and those of the small streams can be preserved.

The presence of several abandoned riverbeds that affected the same terrace suggests that the morphology had to be similar to that still represented in the historical maps in the southern area, where remains of terraces were surrounded by river channels.

A useful criterion for understanding whether the abandoned riverbeds are to be attributed to the Po or the Dora Baltea is provided by the area in which they are located and by their size:

When they are present in the northern portion of the area and come from the area north of Fontanetto and Crescentino, and are just over 150–200 m wide, they can be attributed to the Dora Baltea;

When they are present in the southern part, starting from the area south of Crescentino and have a width that can reach about 500 m, they can be attributed to the river Po.

The shape, the origin and the grain size of the sediments filling the depressions that cut the terraces are reported below.

The depressions that highlight the former riverbeds that affect terrace C are at most 1.5–2 m deep, trending about W-E in the western part of the area and W-E and WNW-ESE in the eastern portion. According to all evidence, these are Dora Baltea riverbeds dating from LGM onwards. The stratigraphic sequences observed in these depressions show the presence of sandy and silty sediments with thin gravel horizons, up to 1.5 m thick, corresponding to the sedimentation of the local streams. Only in correspondence with the area north of Fontanetto Po, where springs are present (Figure 2), the bottom of the ancient riverbed consists of about 2 m of sandy gravels that cover silty-clayey and peaty sediments deposited in a marsh.

The depressions that reveal the ancient riverbeds eroded in terrace D run about W-E and NW-SE: it is not possible to establish whether they were produced by the Dora Baltea or the Po because they are short and discontinuous.

The depressions revealing the abandoned riverbeds that affect terrace E are several kilometres long, have considerable lateral continuity, and were formed by the Dora Baltea (Figure 2). Only in the area just east of Trino do all the ancient riverbeds flow into two of them. Just east of Balzola and in the area north of Morano Po, the ancient riverbeds of the Dora Baltea merge into other long depressions, corresponding to abandoned riverbeds of the Po coming from the area located south of terrace D. From Crescentino to the eastern portion of the studied area, the Stura stream and its tributaries flow in the abandoned riverbeds of the Dora Baltea. The stratigraphic succession at the bottom of the depressions consists of sandy gravels deposited by the Alpine river covered by clays, silts, sands and sand with gravel, deposited by local low-energy streams. A borehole (SG 2073b – ARPA Piemonte, 2018; ENEL 1984) drilled just east of Trino indicates that finer sediments reach the thickness of 4.2 m (Figure 4). In a core drilled to the north of Fontanetto Po, in an ancient bed of the Dora Baltea, an artefact dating back to the 2nd–1st century BC was found at a depth of 3.5 m (ENEL, 1984; Giraudi, 2014b). There is a difference in height of approximately 1 m between the surface of terrace E and the site where the borehole was drilled: it can be assumed that after the formation of the terrace, an erosion of at least 4.5 m of sediments took place. Other chronological information derives from the area just east of Trino, near the bed of the Stura, which flows in the past Dora Baltea riverbed (Figure 2). In this place, lying near or within the abandoned riverbed, Roman artefacts dating back to the 1st–2nd century AD were found, covered by a few decimetres of fine sediments. If the Dora Baltea riverbed had still been active in the 1st–2nd century AD, the Roman settlement would not have been possible because of the extremely high flood hazard. A particular case is represented by the site of San Michele, east of Trino, located within an abandoned channel of the Dora Baltea, where Negro Ponzi Mancini (1989, 1999) carried out archaeological excavations. In the Roman period, the place where the archaeological site lies was higher than the surrounding areas, but subsequent sedimentation covered the old morphological features making the area flat. At San Michele, Roman and early Middle Age remains have been excavated and some artefacts have been dated using radiocarbon and thermo-luminescence methods. Below and among the remains of the buildings of the archaeological site, there are at least three sandy and silty alluvial deposits due to the Stura stream (Figure 4). The earliest deposit contains the remains of a 1st century BC–1st century AD settlement, the second one is dated to the first half of the 5th century AD; and the third one is dated between the 6th and 8th centuries AD. According to Negro Ponzi Mancini (1999), between the 6th and the 10th century AD, the depression surrounding the San Michele archaeological settlement was nearly filled by thin alluvial sediments. Other information on the stratigraphy of the sediments that fill the ancient riverbeds derives from the observations made during the excavation of a canal, 2.5 m deep and about 200 m long, in the area between Trino and Palazzolo (Figure 4). The stratigraphic succession is formed by sandy gravel alluvial deposits of the Alpine rivers, exposed for a few decimetres, covered by about 2–2.25 m of fine gravel in a sandy-silty matrix and silty sand and silt, deposited by the Stura stream. The ancient riverbeds that cut terrace E indicate the variations of the Dora Baltea that occurred in the first millennium BC and ended between the 2nd–1st century BC and the 1st–2nd century AD.

The depressions that highlight the abandoned riverbeds that affect terrace F are almost continuous from Crescentino to Palazzolo, have a prevalent direction from west to east and then from NW to SE and were carved by the Dora Baltea. In the area of Palazzolo, one of the ancient riverbeds of the Dora Baltea joins an abandoned riverbed of the Po, trending around west-east, which continues towards Trino. East of Trino, the riverbed has been almost completely eroded and only a small stretch is preserved in the Morano Po area. Just west of Trino, at the boundary with terrace G (Figure 4), the excavation of a canal highlighted the stratigraphy of the sediments filling one of the abandoned riverbeds of the Po river. The southern bank of the riverbed is made up of coarse sandy gravels, about 3 m thick, cut to the north by sandy gravels with sand lenses dipping gently towards the north, about 1 m thick. The sandy gravel underlies a layer of medium-coarse sand of alluvial origin, about 1.5 m thick, covered by sandy silts, a little more than 1 m thick, the latter deposited by local streams. The coarse sediments that form the bank, being eroded by the sediments filling the riverbed, could be considerably older and represent the remnants of an older terrace eroded during the fluvial evolution.

Most of the depressions that highlight the abandoned riverbeds of the Dora Baltea that affect terrace G start north of the current confluence with the Po and flow from WNW to ESE and west-east. Only in the area south and SE of Crescentino, some abandoned riverbeds of the Po river are preserved. Starting from the area south of Palazzolo, the ancient riverbeds trend west-east, except for a short stretch SE of Trino where they have a NNW-SSE direction, and have been shaped by the Po. To the south of Trino, it has been possible to recognize several abandoned riverbeds of the Po that continue up to the eastern end of the study area: the most recent of these can be followed, also to the west up to the area of Crescentino. Remains of tree trunks, up to 10 m long and with a diameter of up to 1.2 m, have been found in some quarries located in an old river bed near Terranova (Figures 2 and 3). The trunks were lying in a sub-horizontal position between alluvial sediments suggesting that the trees had been overwhelmed during some floods. The radiocarbon dating of some wood samples made by Charrier and Peretti (1977) and Tropeano and Olive (1989a, 1989b) indicated an age between the 4th and 7th centuries AD. It can be assumed that terrace G, in which the quarries are cut, was shaped during the early Middle Ages. A meandering stretch of an abandoned riverbed, located near Trino, was indicated in documents of the 12th–13th century AD as ‘lacus (lake)’: it must be assumed that there was an oxbow lake, or that the memory of the lake remained in that area, at least until the late Middle Ages.

The depressions that highlight the ancient riverbeds that affect terrace H are younger than terrace G and therefore more recent than the 7th century AD. According to historical data, one of the riverbeds forming terrace H was active up to a part of the 13th century AD in the SW area of Trino, while at Palazzolo, it was active until 1292 (Pistan, 2003). At that time, the Po abandoned its old bed, elongated in an west-east direction, and headed south. Pistan (2003) reports that on terrace H, north of Pontestura, in the 13th century, there was a settlement, now disappeared. The abandoned riverbeds that form terrace H were therefore active between the 9th and 13th centuries AD.

The depressions that reveal the abandoned riverbeds that affect terrace I correspond to river channels drawn in historical maps dated between the 16th and 19th centuries. The maps show the changes of the Po riverbed (Crosio and Ferrarotti, 1996) and indicate that the riverbed in the same period could be formed by several branches (generally two). Only with the development of large hydraulic control works during the 19th, 20th and the beginning of the 21st century was there strong anthropic conditioning that brought the riverbed to its current state. According to Ogliaro (1976), the Dora Baltea migrated westwards during a major flood in 1473.

The depth at which the sediment infilling of abandoned riverbeds lies suggests that during the erosive phases subsequent to the modelling of the tops of terraces E and F, the river beds deepened at least 4 or 5 m. From the first centuries of the Current Era up to the beginning of the sedimentation of terrace G, the area must have been formed by a series of discontinuous terraces bounded by scarps 4 m high (or more), surrounded by depressions in which the main rivers and local streams flowed. The plain, therefore, was quite different from the very flat current one.

During the major floods, the waters returned to occupy the old riverbeds and could flow without submerging the surfaces of terraces D and E and the settlements of the first centuries of the Current Era. As a consequence, the buildings and the roads present on these terraces were not reached by the floodwaters.

Since then, the depressions have been almost completely filled by sediments deposited by local streams and by the fine sediments laid down by the alluvial waters of the main streams entering them during the major flood events. The sediments reduced the sections of the depressions that allowed the floodwaters to flow between the remnants of terraces. For this reason, during floods, the water level increased in the course of time, and today also some portions of the top of terraces D and E are exposed to flood hazard.

Alluvial phases: Comparison between geological data and historical sources

The morphology of the area and the stratigraphy of alluvial sediments show a succession of depositional and erosive events. The alluvial phases are evidenced by the shaping of terraces, by the partial filling of riverbeds by gravelly deposits, and by sedimentation of fine deposits on the surfaces of terraces or in the abandoned riverbeds.

The size of terraces and the thickness of the sediments that form them indicate a volume of sediments much greater than those that fill the abandoned riverbeds: this could suggest that morphological and sedimentary features were the product of events of different orders of magnitude.

However, we must take into account that in a large area of the southern part of the plain, there are very recent sediments covering remains of older deposits. Therefore, the sediments and morphological features currently preserved provide only a partial knowledge of the area’s evolution, so that we are not able to estimate the real amount of sediments deposited during the alluvial phases.

The synthesis of the alluvial phases recognized in the study area indicates a series of events that are well correlated with the results of the study on historical data (Figure 5). A number of papers analyzing historical data on floods report the events that affected the area under study and the periods of greater frequency of floods in the Po basin. According to Camuffo and Enzi (1994), reliable record of the floods relating to the Po Plain begins in the 2nd century BC. The succession of alluvial events clearly indicates alternating periods, which lasted some centuries, characterized by a greater or lesser frequency, or even the absence, of alluvial events. In particular (Figure 5), the Po floods were frequent from the 2nd century BC to the 1st century AD (Roman alluvial phase), from the late 6th to the 9th century AD (early Middle Ages alluvial phase), and from the end of the 12th to the 19th century AD (‘Little Ice Age’ – LIA alluvial phase). During the early Middle Ages alluvial phase, a large flood event occurred that caused the flooding of large areas of the Po Plain, the so-called ‘Paul the Deacon Deluge’, from the name of the Longobard chronicler who was alive at the time of the event and who described it. Shorter periods during which an increase in floods occurred are dated between the 4th and 5th centuries (late-Antique alluvial phase) and the 11th–12th centuries AD (late Middle Ages alluvial phase).

Figure 5.

Comparison between the age of the alluvial phases recorded by the stratigraphy and morphology of the alluvial deposits in the Vercelli Plain, the alluvial phases known through historical data (from Camuffo and Enzi, 1994), and the phases of glacial advances and retreats in the Western Italian Alps (from Deline and Orombelli, 2005; Giraudi, 2009) and in the Swiss Alps (from Holzhauser et al., 2005; Joerin et al., 2006).

However, since the Po Plain also extends downstream of the study area, historical flood data could also refer to different areas and, in theory, could indicate alluvial phases that did not produce large-scale effects in the area under investigation. Nevertheless, the age of alluvial sediments and the river avulsions in the area under examination reveal a clear correlation with the alluvial phases reported in Camuffo and Enzi (1994).

In fact, based on the data reported above (Figure 5), we can observe the following:

The most recent sedimentation phase of the Dora Baltea north of terrace D can be dated between the 2nd–1st century BC and the 1st century AD, and corresponds to the early Roman alluvial phase.

Most of the alluvial sediments of terrace F date back to the time of the diversion of the Dora Baltea (1st century BC–1st century AD), that is to say, the late Roman alluvial phase. The sediments of the local stream that incorporate the Roman remains of the 1st century BC–1st century AD at the archaeological site east of Trino are dated as the same phase.

The fine sediments, dated to the first half of the 5th century AD, which cover the archaeological site east of Trino were deposited during the late-Antique alluvial phase. Also the sediments that form the top of terrace F, which cover the 1st century AD graves and are older than the sediments deposited during the alluvial phase following the 6th century AD, were probably deposited during the late Antique phase.

The sediments containing wood dated between the 5th and 7th centuries AD of terrace G and the fine alluvial deposits of the archaeological site east of Trino, dated between the 6th and 8th centuries AD, were deposited during the early Middle Ages alluvial phase. The radiocarbon age (Charrier and Peretti, 1977; Tropeano and Olive, 1989a, 1989b) of some of the large trees felled by one or more floods during the early Middle Ages is compatible with that of the ‘Paul the Deacon Deluge’.

The changes in the river bed known from historical documents date back to the late Middle Ages and ‘Little Ice Age’ floods.

Both the late Antique and early Middle Ages alluvial phases, which produced the sedimentation of fine materials on terrace F, contributed to the filling of the depressions affecting terrace E.

Alluvial phases recorded in the Po Plain in the Emilia Region and the easternmost Po Plain, contemporary to that recorded in the study area, are reported by Veggiani (1979, 1983, 1985), Cremaschi and Gasperi (1989), Marchetti (2002), Cremonini et al. (2013) and Giraudi (2014a).

The alluvial sediments of the Po river also bear witness to an alluvial phase which is more ancient than those known through historical sources. This is the alluvial phase that produced the sedimentation of the sandy gravel overlying the 8th century BC artefacts, dating back to the Iron Age. An alluvial phase occurred also during the first half of the 1st millennium BC in some basins of the Northern Apennines flowing towards the Po Plain (Giraudi, 2014a).

The comparison between the geological and historical data shows that the alluvial sedimentation periods recorded in the study area were contemporaneous with phases during which frequent flooding events occurred in the Po basin.

Correlations between alluvial phases and oscillations in Alpine glaciers

The alluvial phases that have affected the area covered by the present study, even though some possible interference due to anthropic actions starting from the Roman Age is not excluded a priori, are mainly linked with particular climatic periods. Looking at Figure 5 in fact, it can be seen that in four cases out of five, the alluvial phases are contemporaneous with glacial advances recorded in the Italian Western Alps during the last 2500 years. Deline and Orombelli (2005) reported the glacier advances in the Mount Bianco, in the Dora Baltea basin, while Giraudi (2009) discusses the advances and retreats of the glaciers in the Valle Orco, in the western Po basin.

Also the oscillations of the glaciers on the Swiss Alps (Holzhauser et al., 2005; Joerin et al., 2006) correlate with the periods of frequent floods in the Po basin.

The alluvial phases contemporaneous with glacial advances are dated to the Iron Age, to a period dating back to the 5th century AD, to the periods between the 6th and the 9th century AD and between the 14th and 19th century AD.

The glacial variations influenced to a greater extent the hydraulic regime of the Dora Baltea because of the larger extension of the glaciers in its basin.

Only the Roman alluvial phase corresponds to a period of strong glacial retreat recorded in Valle Orco (Giraudi, 2009) and in the Swiss Alps (Holzhauser et al., 2005; Joerin et al., 2006).

Causes of the migration of the confluence between the Dora Baltea and Po rivers

The identification of the abandoned riverbeds of the main rivers, shown in Figure 2, the position of terraces (Figures 2 and 6) and the age and characteristics of the alluvial sediments, reported in chapter 5, provided the information needed in order to assume the presence of ancient confluences of the Dora Baltea and Po rivers, at least starting from the shaping of terrace E.

Figure 6.

Terraces forming the (a) southern Vercelli Plain, (b) the abandoned riverbeds and (c) the western migration of the confluence of the rivers Dora Baltea and Po.

Terrace E (formed by the Dora Baltea river during the 1st millennium BC) in the area between terraces D to the south, and C, to the north (Figure 6), is connected with a portion of the same terrace lying between Morano Po and the area north of Casale Monferrato, formed by the Po river.

The confluence of the two rivers, therefore, was located north of Casale Monferrato.

North of terrace D, the ancient riverbeds that shaped the terrace E were carved by the Dora Baltea up to the 2nd–1st century BC, while between Morano and Casale Monferrato, there are coeval riverbeds of the Po (Figure 6b). The Dora Baltea, at least until the 2nd–1st century BC, flowed into the Po in the area west of Villanova. Very likely, at a later stage, the confluence of the rivers occurred near Balzola. In the same period, in the eastern portion of the study area, the Po flowed as far as Motta dei Conti, in an area never reached later. The 1st millennium BC terraces and abandoned riverbeds suggest that some river diversions occurred which produced the migration of the confluence westwards.

Also during the Roman alluvial phase, the Dora Baltea suffered a diversion that caused it to flow into the Po to the west of Palazzolo, many kilometres to the west of the older confluence.

The ancient river channels carved in terrace F indicate that during the late-Antique alluvial phase, the confluence between the Dora Baltea and the Po was near Palazzolo, and then moved south of Fontanetto Po.

The ancient riverbeds carved in terrace G indicate that the Dora Baltea, during the early Middle Ages alluvial phase, flowed into the Po south of Fontanetto Po and that the confluence then migrated 2–3 km towards the west.

Finally, the Dora Baltea riverbed (known as Dora Morta, Figure 2), abandoned in 1473 (Ogliaro, 1976), indicates the position of the confluence in the Po before the start of a migration of about 10 km westwards which was completed during the alluvial phase of the LIA.

It is therefore clear that the repeated westward migrations of the confluence between the Dora Baltea and Po rivers in the last 3000 years are closely linked to the phases of increased floods, that occurred mainly during glacial advances in the western Italian Alps and in the Swiss Alps, and consequently were mainly due to climatic factors.

Part of the study area, according to some authors (Giraudi, 2014b; Michetti et al., 2012a), was affected by differential tectonic uplift during the Holocene. Michetti et al. (2012) hypothesize late Pleistocene and Holocene activity of the Monferrato thrust front. Giraudi (2014b), observing the change in the altitude of the top of late Pleistocene and Holocene river terraces, assumed the uplift of the area between Fontanetto Po and Trino.

As a consequence, the possibility of a partial tectonic influence on the river mobility must be considered.

Assuming that the tectonic uplift was sufficiently strong to influence the flow of the Alpine rivers in the last 3000 years, then it should have been able to influence also the course of the small streams. But the small streams, that run in a very flat alluvial plain inside the abandoned Alpine riverbeds, do not show diversions or other morphological features that can be produced by tectonic deformations. It follows that, if the tectonics were not able to influence the flow of the small streams, they cannot have been able to divert the Alpine rivers. Therefore, we can assume the lack of influence of tectonics on the westward migration of the confluence of the rivers in the last 3000 years.

Also human intervention on the rivers can be excluded before the LIA. In point of fact, only since the 15th century are any strong interventions on the Alpine rivers recorded in the area, and also today, the works for the prevention of floods and river bank erosion are, in some cases, not completely effective.

Floods create a scenario in which stream flow and velocity are unusually high because of the increased amount of water. During the alluvial phases, the river increases its capacity, that is, the total load of sediments that it can transport, and its competence, that is, the measure of the larger sized particles it can transport. As stream velocity and discharge increase, so do competence and capacity.

During a flood, the bedload can fill part of a riverbed reducing its section, forcing water to spill out over the riverbanks facilitating the change of the riverbed.

Within a morphologically homogeneous area, such as that represented by each terrace of the area in question, the changes in the river bed should be random, whereas the migration of the confluence between the Dora Baltea and the Po in the last two millennia has always been from the east to west.

The cause of the migration towards the west of the confluence of the Dora Baltea into the Po can be hypothesized by observing the topography of the study area and that of the Po and Dora Baltea plains just upstream. Figure 6c shows the topography and the slope of the alluvial plains of the two rivers upstream of Crescentino.

The Po Plain has an average slope of 3.1 m per km, while that of the Dora Baltea has a slope of 4.8 m per km. Instead, the average slope downstream of the confluence of the two rivers is much lower (1.7 m per km). The stream capacity is dependent upon the stream’s velocity, the amount of water flow and the gradation (because streams that occur on steeper slopes tend to have a greater velocity) (Strahler and Strahler, 2006).

In the river Dora Baltea, the variation in the speed of the water, its capacity and competence were definitely higher than those of the river Po, due to the greater slope of the riverbed, the greater impact of climatic variations on the higher mountain and the presence of a more extensive glacial cover in the catchment basin. It follows that the Dora Baltea stream load could also have been greater.

In correspondence with the change of slope (from 4.8 and 3.1 to 1.7 m per km), that takes place just after the confluence of the rivers, the speed of the river waters, the capacity and competence decrease and a larger amount of alluvial deposits is sedimented. This is demonstrated by the lower terraces of the Dora Baltea that, near the confluence zone, form a sort of flat alluvial fan (Figure 6c).

During the floods, when the in-stream debris filled the run-off channels, the Dora Baltea water levels were higher, the floodwaters spilled out of the riverbed, spreading laterally and cutting new channels directed towards the Po Plain, that lies at a lower altitude, namely, towards the SW and W.

Conclusion

The detailed study of the morphology of the southern Vercelli Plain, the stratigraphy of the alluvial sediments, the datings by means of radiocarbon methods of remains of trees and the presence of archaeological artefacts have made it possible to recognize several alluvial phases dating back to the Iron Age, 2nd century BC–1st century AD, 5th century AD, 6th–8th century AD and LIA.

In four out of five cases, the alluvial phases occurred during periods of advance of the Alpine glaciers. The climatic variations affected more the Dora Baltea catchment basin, both because the mountains are higher and the glacier cover was larger than in the Po basin.

However, the alluvial phase dating back to the 2nd century BC–1st century AD took place in a period of strong retreat of the Alpine glaciers, and must indicate a period characterized by heavy rainy events.

The study established that during the past 3000 years, the confluence of the Dora Baltea into the Po has constantly migrated to the west and that this migration occurred during the alluvial phases mentioned above.

It is probable that the migration to the west was linked to the greater increase in sedimentation in the Dora Baltea riverbed compared with that in the Po riverbed. Due to the greater slope, the stronger impact of the climatic variations on the higher mountains and the presence of a larger glacial cover in the catchment basin, the Dora Baltea stream load was greater.

When the in-stream debris filled the run-off channels, the flood levels were higher and the Dora Baltea floodwaters spilled out, spreading laterally and cutting new channels directed towards the lower altitude Po Plain, that is, towards the SW and west.

During the period just after the Dora Baltea diversion of about 2000 years ago, the morphology of the studied area was quite different from the current one. The stratigraphic data show that the riverbed of the Alpine rivers reached about 4 m below the top of the D and E terraces.

During the early centuries AD, the local streams, that is, the Stura and its tributaries, which followed the course of the ancient riverbeds, flowed at similar depths, inside broad depressions that cut the terraces.

During the major floods, the waters returned to occupy the old riverbeds and could flow without submerging the buildings and the roads of the first centuries of the Current Era present on the surfaces of terraces D and E.

Since then, the old riverbeds have been almost completely filled, the section of the depressions is strongly reduced, and today also some portions of the top of terraces D and E are exposed to flood hazard.

Footnotes

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

References

ARPA Piemonte (2018) Banca Dati Geotecnica – Sondaggi Geognostici e campioni di terreno. Available at: https://webgis.arpa.piemonte.it/Geoviewer2D/index.html?title=Arpa+Piemonte+-+Banca+Dati+Geotecnica+-+Sondaggi+geognostici+e+campioni+di+terreno&resource=agsrest%3Ahttp%3A%2F%2Fwebgis.arpa.piemonte.it%2Fags101free%2Frest%2Fservices%2Fgeologia_e_dissesto%2FBanca_dati_Geotecnica%2FMapServer

Camuffo

Enzi

(1994) Cambiamenti climatic negli ultimo 2000 anni. Il Quaternario-Italian Journal of Quaternary Science 7(1): 257–266.

Carraro

Dal Piaz

Govi

et al . (1970) Studio geologico di un settore di pianura a sud di Vercelli. In: Carraro

Dal Piaz

Sacchi

(eds) Studi geologici nel Vercellese ed in Valle Strona. Torino: Sirea Print, pp. 121–140.

Charrier

Peretti

(1977) Ricerche sull’evoluzione del clima e dell’ambiente durante il Quaternario nel settore delle Alpi Occidentali Italiane. Allionia, Bollettino dell’istituto ed Orto Botanico dell’università di Torino 22: 157–192.

Cremaschi

Gasperi

(1989) L’alluvione alto-medioevale di Mutina (Modena) in rapporto alle variazioni climatiche oloceniche. Memorie Società Geologica Italiana 42: 179–190.

Cremonini

Labate

Curina

(2013) The late-antiquity environmental crisis in Emilia Region (Po river Plain, Northern Italy): Geoarchaeological evidence and paleoclimate consideration. Quaternary International 316: 162–178.

Crosio

Ferrarotti

(1996) Trino: gli anni del diluvio. Studi Trinesi 13: 1–253.

Dela Pierre

Piana

Boano

et al . (2003a) Carta Geologica d’Italia alla scala 1:50.000, Foglio 157 Trino. Roma: APAT, Agenzia per la Protezione dell’Ambiente e per i Servizi Tecnici – Dipartimento Difesa del Suolo.

Dela Pierre

Piana

Fioraso

et al . (2003b) Note Illustrative della Carta Geologica d’Italia alla scala 1:50,000, Foglio 157 Trino. Roma: APAT, Agenzia per la Protezione dell’Ambiente e per i Servizi Tecnici – Dipartimento Difesa del Suolo.

10.

Deline

Orombelli

(2005) Glacier fluctuations in the western Alps during the Neoglacial, as indicated by the Miage morainic amphiteatre (Mont Blanc massif, Italy). Boreas 34: 456–467.

11.

ENEL (1977) Studio geologico della zona di Trino Vercellese. Unpublished. Roma: ENEL, Direzione delle Costruzioni.

12.

ENEL (1984) Rapporto per la localizzazione di una Centrale Elettronucleare nella Regione Piemonte. Area Po 1, Unpublished, XIX vols. Roma: ENEL, Direzione delle Costruzioni.

13.

Facchin

(1997) Ascia dell’Età del Bronzo. Notiziario di studi e ricerche 1: 56.

14.

Fozzati

Giraudi

(1983) Camino Monferrato, fiume Po. Quaderni della Soprintendenza Archeologica del Piemonte 2: 148.

15.

Gambari

(1989) Note per l’avvio di una ricerca sulla preistoria e la protostoria del territorio Trinese. In: Michele di Trino

(ed.) Studi Trinesi/8. Casale Monferrato: Eurograf, pp. 7–13.

16.

Gianotti

Forno

Ivy-Ochs

et al . (2008) New chronological and stratigraphical data on the Ivrea amphitheatre (Piedmont, NW Italy). Quaternary International 190: 123–135.

17.

Gianotti

Forno

Ivy-Ochs

et al . (2015) Stratigraphy of the Ivrea morainic amphitheatre (NW Italy): An updated Synthesis. Alpine and Mediterranean Quaternary 28(1): 29–58.

18.

Giraudi

(1998) L’evoluzione del Po nei pressi di Trino (Prov. di Vercelli - Piemonte) tra l’Eta’ del Bronzo ed il XVII secolo. Il Quaternario-Italian Journal of Quaternary Science 11(2): 227–232.

19.

Giraudi

(2009) Late-Holocene glacial and periglacial evolution in the upper Orco Valley, northwestern Italian Alps. Quaternary Research 71: 1–8.

20.

Giraudi

(2014a) Coarse sediments in Northern Apennine peat bogs and lakes: New data for the record of Holocene alluvial phases in peninsular Italy. The Holocene 24(8): 932–943.

21.

Giraudi

(2014b) Quaternary studies as a tool to validate seismic hazard potential of tectonic structures: The case of the Monferrato thrust front (Vercelli Plain, NW Italy). Alpine and Mediterranean Quaternary 27(1): 5–28.

22.

Giraudi

(2017) The morphology of the base of the alluvial sediments of the Vercelli Plain (Piedmont, NW Italy): An integration of the knowledge on the Quaternary geological evolution. Alpine and Mediterranean Quaternary 30(2): 93–103.

23.

Holzhauser

Magny

Zumbühl

(2005) Glacier and lake-level variations in west-central Europe over the last 3500 years. The Holocene 15(6): 789–801.

24.

Janigro D’Aquino

(1979) Reperti dell’Età del Bronzo a Morano sul Po. La Provincia di Alessandria 26(3): 42–43.

25.

Joerin

Stocker

TF and

Schl

chter

(2006) Multicentury glacier fluctuations in the Swiss Alps during the Holocene. The Holocene 16(5): 697–704.

26.

La Rocca

(2000) Crescentino, fraz. Li Galli, loc. Madonna del Palazzo. Necropoli romana. Quaderni della Soprintendenza Archeologica del Piemonte 17: 223–226.

27.

Marchetti

(2002) Environmental changes in the central Po Plain (northern Italy) due to fluvial modifications and anthropogenic activities. Geomorphology 44: 361–373.

28.

Michetti

Giardina

Livio

et al . (2012) Active compressional tectonics, Quaternary capable faults, and the seismic landscape of the Po Plain (northern Italy). Annals of Geophysics 55(5): 969–1001.

29.

Montrasio

Ragni

Bonsignore

et al . (1969) Foglio 57, Vercelli, della Carta Geologica d’Italia alla scala 1:100,000. 2nd Edition. Roma: Servizio Geologico d’Italia.

30.

Negro Ponzi Mancini

(1989) Campagne di scavo 1980–1988. In: Michele di Trino

(ed.) Studi Trinesi/8. Casale Monferrato: Eurograf, pp. 23–57.

31.

Negro Ponzi Mancini

(1999) San Michele di Trino (VC). Dal villaggio romano al castello medioevale. Ricerche di Archeologia Altomedioevale e Medioevale 25–26 (I-III). Firenze: All’Insegna del Giglio.

32.

Ogliaro

(1976) Le origini di Crescentino. Vercelli: SETE.

33.

Pistan

(2003) Per singula plebs. Le istituzioni pievane nella dinamica delle trasformazioni del territorio rurale nel Medioevo. Aree dell’antica diocesi di Vercelli a confronto. Trino: Tridinum, Associazione per l’Archeologia la Storia e le Belle Arti.

34.

Strahler

(2006) Introducing Physical Geography. Boston, MA: Wiley and Sons.

35.

Tropeano

Olive

(1989a) Eventi geomorfologici nelle Alpi Italiane e nella pianura occidentale del Po: inquadramento cronologico in base a radiodatazioni 14C. Il Quaternario-Italian Journal of Quaternary Science 6(2): 189–204.

36.

Tropeano

Olive

(1989b) Vitesse de la sedimentation holocéne dans la Plaine Occidentale du Po (Italie). Bulletin de l’association Francaise pout l’etude du Quaternaire 2: 65–71.

37.

Veggiani

(1979) Prove di un ciclo climatico di piovosità nell’Alto Medioevo nel Cesenate. Studi Romagnoli 30: 87–101.

38.

Veggiani

(1983) Degrado ambientale e dissesti idrogeologici indotti dal deterioramento climatico nell’Alto Medioevo in Italia. I casi riminesi. Studi Romagnoli 34: 123–146.

39.

Veggiani

(1985) Il delta del Po e l’evoluzione della rete idrografica padana in epoca storica. In: Atti Tavola Rotonda ‘Il delta del Po’, 24 Novembre 1982. Bologna: Academia delle Scienze dell-Istituto di Bologna, pp. 37–68.

40.

Venturino Gambari

(2006) La necropoli a tumuli di Morano sul Po. In: Venturino Gambari

(ed.) Navigando lungo l’Eridano- L necropoli Protogolasecchiana di Morano sul Po. Casale Monferrato: Stampa Azimuth Grafica s.r.l., pp. 9–40.

41.

Venturino Gambari

Perotto

(1996) Morano Po, loc. Pobietto. Necropoli dell’Età del Bronzo Finale. Quaderni della Soprintendenza Archeologica 14: 211–212.

42.

Viale

(1970) Vercelli ed il Vercellese nell’antichità. Vercelli: Cassa di Risparmio di Vercelli.