Sedimentary facies characterization through CPTU profiles: An effective tool for subsurface investigation of modern alluvial and coastal plains

Cone penetration tests, a method that is typically used to determine the engineering properties of soils, can be used as an effective tool for refined subsurface stratigraphic investigations of alluvial and coastal plains, aside from the geographic location. High‐resolution calibration of piezocone penetration tests (CPTU) with 20 sediment cores enabled the detailed characterization of alluvial, deltaic and coastal depositional systems of the Po Plain. Twelve cored facies associations, typical of alluvial and coastal plain environments, were characterized based on four distinct CPTU profiles: basic cone resistance (Qc), sleeve friction (Fs), water pore pressure (U) and friction ratio (FR). Sandy facies associations (fluvial/distributary channel, bay‐head delta, transgressive barrier, delta‐front/beach‐ridge) typically have high (>4 MPa) Qc, low‐to‐negative U and low (<2%) FR. Muddy deposits (well‐drained/poorly‐drained floodplain, swamp, lagoon and prodelta) exhibit opposite trends. Heterolithic facies associations (crevasse‐levée, offshore/delta‐front transition) display characteristic seesaw profiles. Plotting of late Quaternary deposits onto the latest version of the cone penetration test typical soil behaviour chart (Robertson, 2010) enables the identification of distinctive facies associations reflecting distinctive grain size. CPTU interpretation leads to sedimentary facies recognition well beyond the simple lithological differentiation and, in particular, allows the refined characterization of clay‐rich and silt‐rich depositional units (swamp clays and peats, central‐inner and outer lagoon, proximal/distal prodelta deposits) that exhibit only subtle differences in lithology. CPTU data can also serve for the accurate detection of key stratigraphic surfaces with potential engineering applications, such as the Pleistocene–Holocene boundary. This latter, a common feature of several alluvial and coastal plain successions, is commonly marked by an abrupt upward decrease of basic cone resistance and sleeve friction from Late Pleistocene, pedogenized, stiff strata to overlying Holocene, organic‐rich, soft deposits. This study offers an updated CPTU‐facies characterization method that could be suitable for subsurface investigations of modern alluvial and coastal plains worldwide.

Identification of soil type has been one of the primary applications of CPT (Robertson, 2010). Soil texture classification has been generally accomplished by utilizing charts that link CPT parameters (i.e. the resistance to penetration at the tip of the penetrometer or cone resistance Qc, the friction of the sediment along the sleeve of the tool or sleeve friction Fs, and their ratio or friction ratio FR) to soil type (Begemann, 1965;Schmertmann, 1969). CPTU additionally provides water pressure (U), which is particularly useful for sand detection (Missiaen et al., 2015). Following calibration with sediment core analysis from the same sites, CPT-CPTU may offer continuous (in situ) views of sedimentological data with great accuracy, repeatability and economical convenience for facies identification (Schokker & Koster, 2004;Truong et al., 2016), stratigraphic correlation (Moran et al., 1989;Beets et al., 1996;Curzi et al., 2017), mapping (Powell & Quarterman, 1995) and sequence stratigraphic interpretation (Styllas, 2014;Zhang et al., 2014). However, despite this relatively large body of research, little attention has been dedicated to sedimentary facies characterization through CPT-CPTU profiles, and very few studies have clearly documented CPT-CPTU signatures of 'facies' (Lafuerza et al., 2005;Choi & Kim, 2006;Amorosi et al., 2017a;Zhang et al., 2018;Bruno et al., 2019).
In Canada, Monahan (1999) showed that deltaic sedimentary 'facies' have distinct signatures on CPT data. In The Netherlands, Schokker & Koster (2004) identified seven 'facies units' and used CPT to improve the characterization of Pleistocene aeolian and fluvial deposits and to correlate stratigraphic data. In the Llobregat delta plain (Spain), using Qc and FR parameters, Lafuerza et al. (2005) characterized six 'sediment facies' within a delta depositional system. In the macrotidal setting of the Qiantang River estuary (China), Zhang et al. (2018) distinguished five facies associations by CPT profiles.
In the southern Po Plain (Italy), after the pioneering work of Amorosi & Marchi (1999), who tested a CPTU-based method for facies characterization, several studies used CPTU parameters to improve subsurface stratigraphic analysis. Amorosi et al. (2017a) focused on alluvial facies and, in particular, on palaeosol recognition through CPTU profiles. Bruno et al. (2019) characterized estuarine and delta plain facies associations, with a focus on peat-layer detection and mapping.
However, none of these studies has provided a comprehensive documentation of the wide range of post-Last Glacial Maximum (LGM) facies associations, from proximal to distal depositional systems, in terms of CPT/CPTU basic parameters (i.e. non-normalized Qc, Fs and their ratio FR) plus U (for CPTU only), and a thorough analysis of their engineering properties is still lacking in the worldwide literature. Mud-dominated facies within alluvial and coastal plain regions are especially poorly known and their sedimentologicalgeotechnical characterization is still lacking. This work aims at filling this gap.
Based on the simultaneous interpretation of all of the above-mentioned parameters, a robust methodology for facies identification is proposed besides the soil charts, because no single parameter is diagnostic of a specific facies association, but their assemblage is (Amorosi & Marchi, 1999). The use of basic parameters makes this method independent from information that typically is not available during drilling operations (soil unit weight, groundwater conditions, etc.). This aspect may be of particular interest for geotechnical engineers that are generally interested in the in situ soil behaviour (Robertson, 2010) and, more in general, to geologists during exploration campaigns in alluvial and coastal plain settings all over the world.
Through calibration of CPTUs with detailed stratigraphy from 20 continuous cores from the subsurface of the modern Po alluvial and delta plains (Fig. 1), and the integration of palaeontological and sedimentological data from previous studies, the specific objective of this study is to test the CPT/CPTU-based method for facies identification and characterization as an independent, fast, simple and economical stratigraphic tool.

GEOLOGICAL SETTING
The Po Plain (Fig. 1A) is one of the largest alluvial plains in Europe and the most densely populated area in Italy. It is drained by the Po River (652 km) and by its Alpine and Apenninic tributaries (Fig. 1A) of ca 3000 km 2 , largely corresponds to the modern delta plain (Fig. 1B). The Po Plain is part of the rapidly-subsiding Po Basin that formed in response to the convergence between Africa and Eurasia since the Cretaceous (Carminati & Doglioni, 2012), leading to the formation of two orogens with opposite vergence: the southern Alps and the northern Apennines (Fig. 1A). These two mountain chains delimit the Po Basin to the north and to the south, respectively (Fig. 1A). South-Alpine buried structures consist of a single arc-shaped thrust system running from west to east, close to the Garda Lake (Vannoli et al., 2015). Northern Apennines buried structures include the Emilia arc to the west and the Ferrara arc to the east (Fig. 1A). These thrust systems became active in the Late Miocene (Boccaletti et al., 2011) and are considered to be still active.
The sedimentary infill of the Po Basin is characterized by a general shallowing-upward trend of Pliocene deep-marine to Quaternary shallowmarine and continental deposits (Ori, 1993). The Pliocene-Quaternary succession is up to 8 km thick in the main depocentres (Pieri & Groppi, 1981). Beneath the study area ( Fig. 1), Middle Pleistocene-Holocene strata, ca 70 m thick, are composed of alluvial deposits, which accumulated under sea-level fall and lowstand conditions during glacial periods, alternating with coastal to shallow-marine sediments deposited during the ensuing post-glacial transgressions and interglacial highstands (Amorosi et al., 2004). The uppermost coastal to shallow-marine interval is dated to the Holocene (Amorosi et al., 2017b). The Holocene succession of the Po Plain exhibits a maximum thickness of about 30 m and a large variety of sedimentary facies (Bruno et al., 2017;Campo et al., 2017). Sand bodies consist of bay-head delta, distributary-channel, transgressivebarrier, beach-ridge and delta-front facies; the muddy units are characterized by soft, paralic and shallow-marine deposits ( Fig. 2A and B). The Holocene deposits unconformably overlie Late Pleistocene alluvial units, composed of coarsegrained, (fluvial-channel, crevasse/lev ee) and finegrained (floodplain; Fig. 2A and B) facies associations. The boundary between Late Pleistocene and Holocene strata is represented by a weaklydeveloped palaeosol ( Fig. 2A and B) across a wide sector of the southern Po Plain (Bruno et al., 2022). The study units are invariably finer-grained than gravel size, which favours the application of the CPTU method. The groundwater level fluctuates from 1 to 4 m depth throughout the study area.
The Holocene succession has been subdivided into eight millennial-scale parasequences (Fig. 2A and B;Amorosi et al., 2017b). Early Holocene estuarine deposits form parasequences 1 to 3 that are stacked into a retrogradational pattern ( Fig.  2A and B) reflecting the stepwise post-glacial sealevel rise (Bruno et al., 2017). Aggradationally to progradationally-stacked parasequences 4 to 8 (Middle-Late Holocene; Fig. 2A and B) record the multi-phase Po delta upbuilding, with activation/ deactivation of distributary-channels and delta lobe switching .
Average sediment recovery was >90%. Cores were analysed for sedimentology (lithology, grain-size trends, sedimentary structures, upper and lower boundaries, thickness, colour, and accessory materials) and geotechnical properties, such as pocket penetrometer measurements on fine-grained sediments. Pocket penetrometers are accessible tools generally used for valuating consistency of soils and approximating their unconfined compressive strength during coring operations . Palaeontological data and 114 14 C dates, available in Amorosi et al. (2017aAmorosi et al. ( ,b, 2019 and Bruno et al. (2017), were considered.
The CPTU profiles used in this work are part of the subsurface database of the Emilia-Romagna Geological, Seismic and Soil Survey, freely accessible online at https://applicazioni.regione. emilia-romagna.it/cartografia_sgss/user/viewer. jsp?service=geologia.
A commercial standard piezocone device was adopted in conformity to the international standard (ISSMFE, 1989) and reference test procedures (ISOPT, 1988). The test equipment is made of a 60°cone, with a 10 cm 2 base area and a 150 cm 2 friction sleeve positioned above the cone. Cells used to measure Qc and Fs values are included within the system. The filter for pore pressure is located behind the cone tip. An average speed of about 2 cm/s was selected to realize CPTs. A hydraulic jacking and reaction system mounted on a heavy truck with screw anchors form the pushing equipment, with thrust capacity of 20 tonnes. The results of CPTU tests are depicted through four curves (Fig. 3). Three profiles represent the vertical variations of the following parameters: Qc (non-normalized cone tip resistance), Fs (sleeve friction) and U (pore water pressure). The fourth curve represents the friction ratio (FR = fs/ Qc 9 100), which is used for the evaluation of sediment type in the most common soil-charts (Robertson, 2010). As also documented by Devincenzi et al. (2004), the use of Qc is due to several reasons: (i) Qc ability to distinguish between sedimentary facies with contrasting mechanical response; (ii) the fact that the distinction between non-normalized and normalized tip resistance does not bring more accuracy, especially for soft facies distinction; and (iii) the possibility of also including CPT tests, as Qc is a common parameter for both. These latter are generally even more common than CPTUs in geological databases worldwide (Koster et al., 2018). Furthermore, normalized charts require the detailed knowledge of parameters that are difficult to estimate and do not provide practical benefits when used at depth <30 m (Robertson, 1990) which corresponds to the maximum depth of the investigated succession.
Geotechnical characterization of individual facies associations was based on sedimentological criteria, following facies analysis on cores. All CPTU parameters were simultaneously examined and compared with the corresponding core segment.
The Robertson et al. (1986) chart is considered one of the most effective tools for sediment   classification (Berry et al., 1998) and has become popular also among stratigraphers (Lafuerza et al., 2005;Styllas, 2014;Zhang et al., 2018;Helfensdorfer et al., 2020). In this study, Qc was plotted against FR in the updated, nonnormalized soil behaviour type (SBT) chart of Robertson (2010). In this latter, soil behaviour zones were reduced from 12 to 9 to match the normalized chart of Robertson (1990), and the dimensionless cone resistance (Qc/Pa) was plotted against FR, both on log scales to expand the portion where FR < 1%.

CPTU CHARACTERIZATION OF FACIES ASSOCIATIONS
In the alluvial and coastal sectors of the Po Basin, Late Pleistocene-Holocene facies associations have been broadly described based on their sedimentological and palaeontological characteristics and, where available, using pollen data (Bruno et al., 2017;Campo et al., 2017;Cacciari et al., 2018;Amorosi et al., 2021). In these works, geotechnical data have been considered as useful, but secondary, information to support sedimentological Fig. 3. Example of calibration between cone penetration test with piezocone (CPTU) and continuous core (C3, Fig. 1 for location). Pocket penetrometer measurements on fine-grained sediments provide an almost continuous record of in situ unconfined compressive strength that can be used for facies characterization . Qc, Cone tip resistance; Fs, Sleeve friction; U, pore water pressure; u 0 , static equilibrium pore pressure; FR, Friction Ratio; FU, fining-upward trend. Red arrows highlight the simultaneous increase in Qc and Fs, and the abrupt decrease in U of palaeosols. interpretation (Amorosi et al., , 2017aBruno et al., 2019). This study, instead, focuses on the CPTU characteristics of 12 facies associations that were grouped into four depositional systems. These are described below from proximal (alluvial) to distal (shallow-marine) settings.
An example of sediment core-CPTU test calibration is shown in Fig. 3. Depth correspondence between cores and CPTU curves is about 95%, with <1 m offset (Fig. 3). As documented by previous work (Schokker & Koster, 2004;Missiaen et al., 2015), this is most likely caused by compression of soft deposits during core recovery (Koster, 2016). Distinctive sedimentological and CPTU characteristics are summarized in Table 1.

Alluvial plain deposits
This depositional system includes Late Pleistocene and Holocene facies associations (Figs 2 and 4). Fluvial-channel facies association Facies characteristics. This facies association is made of medium to coarse grey sands with erosional bases, fining-upward (FU) trends and sharp or transitional tops (Fig. 4A). Individual sand bodies, 2 to 10 m thick, can be vertically stacked to form multi-storey deposits ( Fig. 2). High-angle, unidirectional cross-stratification and horizontal bedding are locally preserved. Macrofaunal remains are rare and generally represented by shell fragments. The meiofauna includes only sporadic fragments of freshwater ostracods.
Based on the combination of lithology, erosional lower boundary, thickness and fossil content, this facies association is interpreted as a fluvial-channel deposit. The FU tendency and presence of unidirectional flow structures strongly support this interpretation (Miall, 2013).
CPTU characteristics. This facies association is typified by high Qc values, in a range of 5 to 20 MPa, locally >20 MPa ( Fig. 5A and Table 1). Fs is generally low (0.05 to 0.1 MPa), with very few exceptions (up to 0.2 MPa; Fig. 5). Water pore pressure (U) is commonly negative (<0) or lower than static equilibrium pore pressure u 0 . Mean FR is <0.5%. The lower boundary is generally characterized by a sharp increase of Qc, followed upward by progressively decreasing values (Figs 3 and 5).
High Qc values and low FR percentages are consistent with coarse to medium fluvial sand deposits (Styllas, 2014). Low to negative U indicates high permeability and tendency to dilate (Moran et al., 1989;Dafalla et al., 2020). Peaks in Qc at lower boundaries and upward decreasing values are consistent with erosional bases and FU grain-size trends, both characteristic of fluvialchannel deposits.
Crevasse/lev ee facies association Facies characteristics. This facies association includes three different lithofacies: lithofacies 1 depth (m)  is characterized by coarsening-upward (CU) trends from grey clayey silts to silty sands, with gradational lower boundaries and sharp tops (Figs 4A and 6). Lithofacies 2 shows many similarities with fluvial-channel sand bodies (FU trend, erosional base and gradational top) but a typically lower thickness (<1.5 m) and finer grain size (medium to silty sands; Fig. 4B). Lithofacies 3 includes the repetitive alternation of grey to varicoloured silty sand and silty clay (Fig. 4C). Plant debris and rootlets are locally abundant. Thickness is generally <2 m. This facies association reflects deposition in three distinct sub-environments close to river channels (Bridge, 1984). Given its CU trend, gradational lower boundary and sharp top, lithofacies 1 is interpreted as a crevasse splay deposit. Lithofacies 2 represents a crevasse channel deposit, based on grain-size, thickness and FU trend. The silty sand-silty clay alternations that typify lithofacies 3 most likely reflect traction plus fallout deposition. These lithological characteristics, along with indicators of subaerial exposure (plant debris and rootlets) are typical of natural lev ees.
Highly variable Qc (up to 15 MPa), FR (0.5-2.0%) and U (negative-to-positive) values are consistent with texturally heterogeneous facies ranging from medium sand to clayey silts (Styllas, 2014). Qc and Fs trends in CPTU-facies 1 are consistent with characteristic CU trends of crevasse splay deposits. On the other hand, Qc and Fs trends of CPTU-facies 2 are consistent with FU tendencies observed in crevasse-channel deposits in cores. Seesaw profiles in CPTU-facies 3 likely reflect sand-mud centimetre-thick alternations typical of natural lev ee deposits.
Well-drained floodplain facies association Facies characteristics. This facies association, up to >10 m thick, is composed of massive and bioturbated silts and clays, with carbonate nodules, scant plant fragments and root traces ( Fig. 4A and B). Locally, faint horizontal lamination is observed. Yellowish to brownish clay mottles of Fe and Mn oxides can be present (Fig. 4B). Dark organic-rich silty clay layers overlying light grey horizons with calcite nodules are common features in the Late Pleistocene succession ( Fig. 4A and B). Sparse fragments of freshwater gastropods can be found, whereas the meiofauna is generally absent.
This facies association has been interpreted as deposited in a well-drained floodplain. Bioturbation, lithology, macrofossils and accessory materials reflect a low-energy freshwater depositional environment dominated by suspension fallout, with prolonged episodes of subaerial exposures, as demonstrated by oxidized and pedogenically modified horizons (Buol et al., 2011). These latter have been interpreted as weakly-developed palaeosols (Inceptisols of Soil Survey Staff, 1999), each composed of an organic-rich, commonly dark, eluvial horizon (A), and an illuvial Bw or Bk horizon (Fig. 4B), typified by a higher content of secondary calcite. Whereas horizon Bw is generally characterized only by incipient weathering with colour ranging between greyish brown to yellowish brown, secondary calcite content is higher in the more whitish grey Bk. This horizon generally includes characteristic nodules   Table 1). Fs values are generally high, up to 0.18 MPa, and U >> 0. Friction Ratio ranges between 3.5% and 10% (Figs 3, 5A and 5B). Individual palaeosols are marked by the subtle, but consistent downward increase in Qc (up to 1.2 MPa) coupled with a sharp spike in Fs (up to 0.12 MPa) and commonly an abrupt decrease in U (Figs 3 and 5A).
Qc and FR values are interpreted to represent floodplain silt-clay sediments (Sarti et al., 2012). The simultaneous increase in Qc and Fs and the abrupt decrease in U reflect response to penetration of over-consolidated, pedogenized muds.

Estuary and delta plain deposits
This depositional system includes facies associations that formed in freshwater to brackish environments. This group dominates the Holocene succession in the proximal and central sectors of the study area, whereas it occurs at distal locations in thin, transgressive parasequences (Fig. 2).
Distributary-channel facies association Facies characteristics. This facies association shares several features with fluvial-channel deposits, including lithology, FU trend and erosional lower boundary (Fig. 6A). Differently, it displays lower thickness (<5 m) and width (<1400 m; Bruno et al., 2017), and is finer-grained (silty fine sand; Fig. 6B). Fragments of freshwater to low-brackish gastropods have been observed (Fig. 6A). Wood fragments are rare. This facies association is in lateral transition to organicmatter-rich muds.
Lithology and FU trend reflect deposition in active channels. However, based on the finer grain-size and lower thickness than fluvialchannel deposits, palaeontological data and lateral transition to organic matter-rich muds, this facies association is inferred to reflect distributarychannel deposits (Bhattacharya, 2006). CPTU characteristics. This facies association is characterized by relatively high Qc values (1-10 MPa), locally exceeding 20 MPa ( Fig. 7A; Table 1). Fs is generally <0.05 MPa, U < u 0 and FR < 1% (Fig. 7A). At the base, Qc may sharply increase to 6 MPa, whereas at the transition to overlying muds it decreases gradually (Fig. 7A).
Relatively high Qc, low FR and negative U are typical of sand bodies. Comparatively lower values of Qc and Fs, and sedimentary features of encasing deposits permit discrimination of distributarychannel deposits from fluvial-channel sand bodies. Distributary-channel-related facies (i.e. crevasse and lev ee deposits) are not described, because their sedimentological and geotechnical characters are similar to their fluvial counterparts.
Bay-head delta facies association Facies characteristics. This facies association is composed of medium to fine grey sand (Fig. 6B). Its base is either erosional or sharp, whereas its top commonly is a sharp surface. The thickness of this facies association is generally <5 m, and sand bodies display a CU trend (Fig. 6B). Wood fragments and vegetal remains are locally encountered at the top. Body fossils, though generally poorly preserved, exhibit a diagnostic mixture of brackish (Cyprideis torosa and Cerastoderma glaucum) and freshwater (Ilyocypris) species (Amorosi et al., 2017b). This facies association typically occurs at the transition from inner estuary (freshwater) to outer estuary (brackish) environments (Fig. 2).
Based on lithology, palaeontological data, grain-size trends and stratigraphic position, this facies association is interpreted as bay-head delta mouth deposits. The co-existence of brackish fossils and freshwater ostracods, along with the typical CU trend, are consistent with progradation of sand bodies into an estuarine (brackish) environment (Dalrymple et al., 1992;Schwarz et al., 2011), occasionally subject to fluvial input (Hijma et al., 2009). CPTU characteristics. This facies association shows relatively high Qc (up to >20 MPa; Table  1) and Fs (0.05-0.1 MPa) values. U is <u 0 (Fig.  7B). Its lower boundary is marked by the sharp increase of Qc (up to 7 MPA, Fig. 7B), which generally increases upward.
High Qc, low FR and U < u 0 are consistent with a coarse to fine sandy texture (Amorosi & Marchi, 1999). Upward increasing Qc values suggest a general CU grain-size trend, whereas the sharp increase in Qc at the base is consistent with the sharp lower boundary of bay-head delta deposits observed in cores.
Freshwater swamp facies association Facies characteristics. This facies association, up to 15 m thick (Fig. 2), is composed of very soft grey to dark-blackish clays ( Fig. 6A and B). Rare silt to silty sand intercalations, at the centimetrescale, are locally encountered. Wood fragments and vegetal remains are abundant at discrete horizons or scattered along decimetre-thick core segments ( Fig. 6A and B). Iron and manganese oxides are absent. Peat layers, 10 to 40 cm thick, and mostly made of poorly decomposed wood fragments and plant remains, occur at different stratigraphic levels (Figs 4A, 4B and 6B).
Molluscs (Bithynia tentaculate and Valvata macrostoma) and freshwater ostracods (Candona and Pseudocandona) typify this facies association (Amorosi et al., 2017b). Pollen data (Cacciari et al., 2018) show that clayey intervals include open to dense alder carr vegetation, with a mixed oak-holm oak ecological component in subordinate position. On the other hand, peat layers are characterized by pollen grains of herbaceous A B C Fig. 7. Identification of estuary and delta plain facies associations in CPTU profiles (Qc, Fs, U and FR) and adjacent cores C9 (A), C5 (B) and C8 (C) (Fig. 1 for location). Black circles highlight peaks on CPTU that correspond to a discrete peat horizon. FU, fining-upward trend; CU, coarsening-upward trend. Lagoonal (inner: Lg-I; outer: Lg-O) facies associations identification in CPTU profiles (Qc, Fs, U and FR) and adjacent core (C8) (Fig. 1 for  location). wetland community, with aquatic plants (hydrophytes), or sparse alder carrs, the latter being tolerant to prolonged intervals of radical drowning (helophytes).
Fine-grained and dark-coloured sediments with high organic-matter content and freshwater fossils suggest deposition in stagnant, semipermanently flooded wetlands (for example, swamps; Salel et al., 2016), as part of inner estuarine or upper delta plain environments. Peat layers likely accumulated under permanently waterlogged (high water table) and reducing conditions, as also demonstrated by the absence of traces of oxidation and presence of aquatic plants (Cacciari et al., 2018). Conversely, grey clays were deposited under a relatively lower water table, close to an active channel, as suggested by rare coarser beds reflecting weak overbank deposition.
Low Qc and Fs values and U > 0 are characteristic of fine-grained swamp deposits (Sarti et al., 2012). Such values are significantly lower than those typically recorded in well-drained floodplain deposits. Distinctive peaks in Qc and Fs invariably associated with peat layers likely reflect post-burial peat consolidation .
Poorly-drained floodplain facies association Facies characteristics. This facies association is composed of grey clays and silty clays. Carbonate nodules are sparse and less frequent than in welldrained floodplain deposits. Iron-manganese oxides are very rare and evidence of pedogenesis (for example, horizonation) is lacking. Plant fragments and organic matter are locally abundant but invariably in lower proportion than in swamp deposits ( Fig. 4A and B). Shell fragments of freshwater molluscs (Planorbia) and freshwater to low-brackish ostracods (Candona) are found occasionally.
Lithology, the homogeneous grey colour, a lack of pedogenic features and the fossil content allow interpreting this facies association as accumulating in a low-energy setting, such as a poorly-drained floodplain, under the persistent influence of freshwater due to high water table or river floods . CPTU characteristics. This facies association displays Qc values narrowly constrained between 0.8 to 1.8 MPa (Table 1). Fs is in a range of 0.03 to 0.1 MPa, U > 0 and FR 4 to 6% (up to 10%; Figs 3 and 5). Increasing values characterize this facies association from shallow (Qc ca 0.8 MPa, Fs ca 0.02) to relatively deeper stratigraphic intervals (Qc ca 1.8 MPa and Fs 0.07 MPa; Figs 3 and 5). In general, CPTU parameters from this facies association are transitional between well-drained floodplain muds and swamp clays.
Sedimentological characteristics and the diagnostic fossil content are indicative of a brackish, semi-enclosed basin, such as a lagoon. Given its stratigraphic position, and the landward transition to freshwater, river-dominated deposits, this facies association could also be a part of a wider (outer) estuarine environment (Boyd et al., 2006). The increasing sand/mud ratio from lithofacies 1 to lithofacies 2 reflects transition from low-energy (inner lagoon) to increasingly higherenergy (outer lagoon) sub-environments. This interpretation is supported by the highly diversified (brackish-marine) meiofauna content of lithofacies 2, compared to the oligotypic fossil association of lithofacies 1. CPTU characteristics. In line with core data, this facies association shows two CPTU facies with distinct geotechnical characteristics. CPTUfacies 1 shares many characteristics with swamp deposits, such as very low Qc (0.5-0.8 MPa), Fs (0.01-0.04 MPa) and U > 0 (0.1-0.4 MPa; Fig.  7C). The Qc profile is rectilinear and FR is generally high (5-9%; Fig. 7C). CPTU-facies 2 is characterized by higher Qc values, fluctuating between 0.9 and 1.5 MPa, up to 5 MPa (Table 1), with seesaw-shaped profile (Fig. 8). Fs displays a similar profile, with values ranging between 0.03 and 0.08 MPa (Fig. 8). U is generally >0, but locally <u 0 . Mean FR is <4% (Table 1).
Qc and FR in CPTU-facies 1 are consistent with a fine-grained, soft clay deposit. In CPTU-facies 2, seesaw profiles, with variable U, locally <u 0 values, reflect millimetre to centimetre-scale mud-sand alternations. Despite similar characteristics (seesaw profile and Qc values) between CPTU-facies 2 and the lev ee facies association, which makes their distinction quite difficult based on CPTU profiles alone, lagoon sediments are generally softer, as suggested by lower Fs measurements and FR percentages (see Table 1).

Nearshore sands
This depositional system consists of two sanddominated facies associations that accumulated in nearshore environments, during the Holocene transgression and subsequent coastal progradation, respectively (Fig. 2).
Transgressive barrier facies association Facies characteristics. This facies association is characterized by grey silty to fine sands with FU trend, erosional lower boundary and gradational top (Fig. 6D). Its thickness generally is <2 m, with mean values of about 1 m (Fig. 5D). The base is marked by a veneer of shell fragments of marine and brackish taxa (Fig. 6D). Above the erosional fossil-rich boundary, the meiofauna include only abraded specimens of A. beccari, Elphidium Crispum and Pontocythere turbida (Amorosi et al., 2017b).
Below the shell lag, a silty to very fine sand body with characteristic CU trend is locally encountered (Fig. 5D). This sand body includes poorlypreserved euryhaline and shallow-marine taxa, such as C. torosa, Ammonia and Elphidium spp.
Based on the combination of lithology, sedimentological features, fossil content and thickness, this facies association is interpreted as deposited in a high-energy, coastal environment.
In particular, medium to fine sands with FU trend overlying the basal shell-rich lag are interpreted to represent transgressive barrier island deposits. The basal shell-rich lag reflects wave erosion and reworking of backshore and upper shoreface strata during barrier retreat (wave ravinement surface of Swift, 1968). The limited thickness suggests erosion of upper shoreface strata, probably cannibalized during shoreface retreat . The thin sand body that underlies the ravinement surface, with its particular meiofauna assemblage and CU trend (Fig. 6D), may be interpreted as a washover fan deposit .
CPTU characteristics. This facies association is characterized by Qc values of 4 to 10 MPa, mean Fs of about 0.05 MPa, U generally <u 0 and FR ranging between 0.5 to 1.5% (Fig. 8A). The lower boundary is sharply marked by an increase in Qc and U (Fig. 8A). Qc and Fs rapidly decrease upward, whereas U and FR increase in the same direction (Fig. 8A). Locally, the opposite tendency is recorded (Fig. 8A, about 22  Beach ridge/Delta front facies association Facies characteristics. This facies association is composed of grey to brownish fine to coarse sand, with gradational base, general CU trend and thickness ranging between 2.5 m and 10 m (Figs 2, 6C and 6D). It forms a laterally extensive (up to 100 km) sediment body parallel to the modern shoreline (Campo et al., 2020b). Due to poor preservation of sedimentary structures in cores, two main lithofacies can be identified, mostly on the basis of the fossil content and of accessory materials. Lithofacies 1 is made of highly fossiliferous sand (Fig. 6C); whereas lithofacies 2 includes only rare bioclasts and scattered plant debris (Fig.  6D). In lithofacies 1, molluscs exhibit a varied fauna including Tritia neritea and Chamelea gallina. On the other hand, the meiofauna is generally poorly-preserved and only few specimens of foraminifera (Ammonia beccarii and Pontocythere turbida) have been identified (Amorosi et al., 2017b). Lithofacies 2 is characterized by a highly specialized mollusc fauna, with Lentidium mediterraneum as the main specimen (Scarponi & Angeletti, 2008). The meiofauna is almost absent and represented by opportunistic foraminifera (Ammonia tepida and Palmoconcha turbida).
The lithology, general CU trend and geometry of this facies association are consistent with sheet sands formed in a high-energy coastal palaeoenvironment during Po Delta progradation. Fossiliferous sands (lithofacies 1), including a mixture of nearshore taxa, are inferred to have been deposited along beach ridges by wave action and longshore currents (Bhattacharya & Giosan, 2003). In lithofacies 2, the abundance of Lentidium mediterraneum and of an opportunistic fauna tolerant to high amounts of riverine organic matter is more consistent with highly-stressed environmental conditions, typical of a delta-front environment, where sedimentation is mostly influenced by river activity (Bhattacharya & MacEachern, 2009;Bohacs et al., 2014). Thick and amalgamated sand bodies are inferred to have been deposited close to the shoreline, where the energy of waves and river currents is stronger; basinward, where the transport energy weakens, sand bodies thin out and can be sandwiched within shallow-marine muds (Fig. 2).
CPTU characteristics. This facies association is typified by Qc values ranging between 4 and 11 MPa (locally up to 20 MPa; Fig. 8A, 8B and Table 1). Fs is generally <0.1 MPa, U < u 0 and FR < 1% (Fig. 8A). Qc values may show an overall upward increase ( Fig. 8A and B). The lower boundary is frequently marked by a consistent increase of Qc (+2/+10 MPa, Fig. 8A), whereas the top is generally sharp (Fig. 8A and B).
Qc and FR values are typical of both delta-front and beach-ridge facies (Lafuerza et al., 2005) and show no obvious distinction between riverinfluenced and wave-dominated systems. Increasing upward values of Qc reflect the overall CU grain-size trend, which is characteristic of coastal (deltaic and strandplain) progradation under highstand conditions.

Shallow-marine deposits
This group includes shallow-marine facies associations observed in the distal sector of the Holocene Po coastal Plain succession (Fig. 2).
Offshore/Delta front transition facies association Facies characteristics. This facies association is composed of grey silty clay with millimetre to centimetre-thick sand intercalations (Fig. 6D). Its thickness ranges between 1 m and 3 m. Bioturbation is common. Based on the fossil content and stratigraphic position, this facies association can be subdivided into two lithofacies. In lithofacies 1, which generally overlies transgressive sands (Figs 2 and 8A), the meiofauna is characterized by infralittoral and epiphytic species associated with molluscs, such as Chamelea gallina and Antalis inaequicostata. In the overlying lithofacies 2, which shows upward transition to deltafront deposits (Fig. 2), Lembulus pella and varicorbula gibba are the main components of the mollusc assemblage (Scarponi & Angeletti, 2008), whereas the meiofauna consists of rare Ammonia (tepida and parkinsoniana) and Polmoconcha turbida. Plant debris are locally encountered.
The characteristic alternation of sand-mud layers, the diagnostic fossil content, and stratigraphic position between transgressive barrier sands (below) and prograding nearshore sand bodies (above), suggest the transition between coastal and shallow-marine depositional environments. Palaeontological data allow the further distinction between shoreface-offshore transition deposits (lithofacies 1) and delta-front transition deposits (lithofacies 2). Opportunistic species, in particular, characterize delta-front transition setting because of highly stressed conditions due to riverine influence (Dasgupta et al., 2020).
CPTU characteristics. This facies association is characterized by scarcely rectilinear to seesawlike CPTU profiles (Fig. 8A). Qc ranges between 0.9 to 1.5 MPa, with peaks up to 8 MPa (Table 1). Fs and FR range between 0.02 to 0.1 MPa, and 1 to 7%, respectively. U has typically positive values, though locally negative (or <u 0 ) values can be encountered (Table 1).
Qc and FR values are indicative of a heterolithic facies association with dominant muds and subordinate fine sand (Robertson, 2010), in accordance with the interpretation of this facies association as an offshore-delta front transition. The distinction between offshore-transition and deltafront transition lithofacies is not possible based on CPTU data alone.
Lithology, sedimentological features and palaeontological data denote a shallow-marine environment (Dasgupta et al., 2020). In lithofacies 1, the dominance of clay, combined with a diversified meiofauna and the relative abundance of open-marine species are consistent with an offshore setting. The dominance of silt (lithofacies 2) or clay (lithofacies 3), in association with a mostly opportunistic meiofauna, is indicative of proximal versus distal prodelta environments, respectively. This further distinction is also supported by the relative abundance of Nonionella turgida, which is an indicator of distal prodelta settings (Barbieri et al., 2021).
CPTU characteristics. This facies association is characterized by low values of Qc and Fs, ranging between 0.3 to 1.0 MPa and 0.01 to 0.025 MPa, respectively (Fig. 8C). In particular, two distinct CPTU-facies can be recognized: CPTU-facies 1, with Qc of 0.6 to 1.0 MPa and Fs of about 0.025 MPa; and CPTU-facies 2, with slightly lower Qc and Fs values (Qc: 0.3-0.6 MPa and Fs: 0.01-0.025 MPa; Fig. 8C). Subordinate peaks in both of the curves are in the order of 1.2 to 5.0 MPa (Qc) and 0.03 to 0.05 MPa (Fs; Fig. 8C). Peaks in Qc are relatively more frequent in CPTUfacies 1, where profiles show a typical seesaw shape. A similar trend is recorded by the FR profile, with values ranging between 2.0 to 3.5% for CPTU-facies 1 and between 3% and 4% upward for CPTU-facies 2 (Fig. 8C). U commonly has positive values (>u 0 ), ranging between 0.2 to 0.3 MPa for CPTU-facies 2, and between 0.3 to 0.75 MPa for CPTU-facies 1 (Fig. 8C). The latter locally shows <u 0 (Fig. 8C and Table 1).
Similar Qc and FR have been reported from fine-grained prodelta sediments (Lafuerza et al., 2005;Zhang et al., 2018). Characteristic Qc, Fs and FR values between CPTU-facies 1 and 2 are consistent with proximal and distal prodelta lithofacies, respectively, identified in cores. On the other hand, the distinction between offshore and distal-prodelta deposits cannot be accomplished on the basis of CPTU measurements alone.

Facies associations and the soil behaviour type chart
The 12 facies associations characterized in terms of all basic CPT parameters (plus U) were plotted onto the SBT chart of Robertson (2010;Fig. 9) to test if this chart, based on two variables only, can be useful for facies identification well beyond the simple differentiation between sand and mud.
The SBT chart includes nine different soil behaviour zones, each defined by particular Qc and FR thresholds (Fig. 9). Each zone groups soil types using traditional lithological descriptions (i.e. clay and sand). This updated chart can be helpful for real-time data processing and interpretation during CPT and/or CPTU procedures, especially in predominantly silica-based, young (i.e. Upper Pleistocene-Holocene) and uncemented soil (i.e. ideal soil).
Plotting sedimentary facies into SBT charts results in the clear lithological distinction between coarse-grained and fine-grained deposits. Coarse-grained facies show typically high (>4 MPa) Qc and low (<2%) FR, and generally The SBT chart may provide accurate prediction of sediment type, but this is not straightforward, and more parameters need to be considered to accurately characterize facies associations from different depositional systems (Styllas, 2014). Besides lithology, the concept of sedimentary facies involves additional features, which are strictly related to their geological history. Since facies associations are genetically linked to their depositional system, each deposit depends on depositional conditions, but also on postdepositional processes (Truong et al., 2011;Maliva, 2016). This 'genetic link' makes the use of sequence stratigraphic concepts as an ideal complementary tool for CPTU interpretation. In this regard, the comparison of SBT charts locally shows significant overlap (see dashed lines in Fig. 9) of CPTU characteristics among facies associations from different depositional systems. This is due to the fact that different facies have similar engineering properties and are made of the same lithology. However, sedimentary facies can be characterized by distinctive grain-size trends (FU, CU), U ranges (<0, >0), thickness, gradational/sharp/erosional contacts and diagnostic peaks in CPTU curves. For this reason, the SBT charts cannot be utilized for accurate facies identification if sedimentary facies have the same lithology (and similar mechanical behaviour). For example, fluvial-channel and bay-head delta sands both plot on class 6 ( Fig. 9A and B). In this case, FU or CU trends may be helpful for their distinction. These trends are very well-expressed by CPT/CPTU curves (Figs 3, 5, 7 and 8), but unavailable in the SBT charts (Fig. 9). Similarly, for swamp deposits, concomitant peaks in Qc, Fs, U and FR curves are diagnostic of peat deposits (Fig. 7B). Thus, interpretation of CPT/CPTU curves using sedimentological concepts is a much more effective tool for high-resolution facies identification than SBT charts alone. Where distinct sedimentary facies exhibit remarkable overlap onto the SBT chart, such as for swamp, lagoon and prodelta deposits ( Fig. 9B and C), CPTU test interpretation is largely insufficient to refine facies characterization. Even though U values may be slightly different (but generally positive) for these three facies associations, lagoon and prodelta deposits locally showing <u 0 (Table 1), variations of pore pressure alone do not allow for their recognition. In such instances, calibration of CPTU data with detailed facies analysis from sediment cores and sequence stratigraphic concepts can be used to predict proximal to distal facies relationships. As facies accumulation is mostly a function of the interaction between accommodation space and sediment supply (Posamentier et al., 1988;Van Wagoner et al., 1990;Muto & Steel, 1997) and facies distribution is strongly related to systems tract architecture (for example, estuarine deposits are typically developed in TST; prodelta and delta front in HST/LST; amalgamated fluvial bodies in LST; Shanley & McCabe, 1991;Allen & Posamentier, 1994;Cattaneo & Steel, 2003;see Fig. 2) the sequence stratigraphic approach can enhance interpretation of CPTU data, allowing reliable interpretation of deposits that may show nearly identical lithological characteristics.
The CPTU data may also reveal specific characteristics undetected in cores because of recovery issues. Grain-size trends and centimetre to decimetre mud intercalations within sand deposits can easily get lost during core recovery with the rotary wash drilling method. Furthermore, sand recovery is not easy to accomplish during routine drilling operations, unless special and expensive equipment is utilized. For example, silty clay intercalations can be present within beach-ridge deposits (or within other sand bodies, such as the bay-head delta in Figs 2 and 7B). This stratigraphic information could be lost during core recovery. Amorosi et al. (2017b) utilized CPTU profiles to identify flooding surfaces within Holocene strata (Fig. 2), proving that similar information can be precious for high-resolution stratigraphic analysis. In this regard, Monahan (1999) noticed that sharp versus gradational lithological boundaries, CU/FU sequences and dips in foreset beds could be recognized beyond the limits of borehole control. Similarly, T€ ornqvist et al. (2000) showed the potential of CPT data to identify marked facies transitions not recovered in the core.
As a whole, CPTU interpretation is as an effective tool for the characterization of fine-grained alluvial, paralic and shallow-marine facies associations for which sedimentological characterization can be difficult. Since CPT and CPTU are fast and economical tests, they can be used for preliminary investigations of late Pleistocene-Holocene strata in order to evaluate and plan additional (more expensive) geological tests.
This CPT/CPTU approach can reasonably be exported to similar settings worldwide as several delta and coastal plain successions share comparable post-LGM stratigraphic architectures and stratal thickness (Hori et al., 2002;Tanabe et al., 2006;Hijma et al., 2009;Styllas, 2014; in response to eustatic sea-level rise, with only minor differences between microtidal and macrotidal areas (Zhang et al., 2018).
Application of CPTU-facies identification to subsurface investigation of alluvial and coastal plains The concept of sedimentary facies has become increasingly common among geotechnical engineers. Because of the spatial variability of subsurface geomechanical behaviour and its potential influence on differential ground settlements, a sedimentary facies approach can be highly useful for mapping, modelling and predict consolidation of late Quaternary deposits, especially beneath highly-populated coastal areas, from Europe to Asia (Ricceri et al., 2002;Choi & Kim, 2006;Dipova, 2011). The soil behaviour can be highly influenced by depositional and postdepositional processes (Truong et al., 2011;Sarti et al., 2012). As different facies have distinct geotechnical and geomechanical characteristics (Truong et al., 2013;Lorenzo et al., 2014), facies characterization through CPTU could be utilized to make predictions about the 3D spatial distribution of geomechanically 'weak' versus 'strong' deposits, aside from any specific geographic location. Through the analysis of CPTU profiles, for example, soft, unconsolidated Holocene deposits with low Qc and Fs can be easily distinguished from stiff (high Qc and Fs) and more consolidated Pleistocene sediments, as shown in Fig. 10. Since Holocene, soft (and organic-rich) deposits typically overlie hardened pedogenized horizons also in several delta plains of Asia, America and Europe (Blum & Aslan, 2006;Choi & Kim, 2006;Bos et al., 2012;Hijma et al., 2017;Fan et al., 2019), this approach can have important applications from an engineering perspective also outside the Po Plain at the global scale.
In geotechnical surveys, the convenience of a facies approach is also suggested by studies on soil liquefaction (Watts et al., 1992;Amorosi et al., 2016) and sediment compaction assessment . The latter mostly concerns soft (for example, paludal, lagoonal and prodelta) clays, and may be one of the major factors controlling subsidence, as documented by recent research on organic-matter-rich Holocene strata of The Netherlands and Italy (van Asselen et al., 2009;Teatini et al., 2011;Bruno et al., 2020). Considerable volumes of these paludal, estuarine and A B Fig. 10. Recognition of the Pleistocene-Holocene boundary based on CPTU data: (A) core C3; and (B) core C1 ( Fig. 1 for location). Stiff, well-drained floodplain Pleistocene facies are overlain by soft poorly-drained to swamp Holocene facies associations. CPTU data are supported by pocket penetration information suggesting a similar geotechnical behaviour. Qc, Cone tip resistance; Fs, Sleeve friction; U, pore water pressure; u 0 , static equilibrium pore pressure; FR, Friction Ratio; MPa, Megapascal unit. prodelta-offshore facies associations are present within Holocene Po Plain strata (Campo et al., 2020a). Similar soft deposits are very common also beneath other modern delta/coastal plains worldwide, from Europe, Asia and America (T€ ornqvist et al., 2008;Tanabe et al., 2013Tanabe et al., , 2015Zhang et al., 2014Zhang et al., , 2021Sarti et al., 2015;Koster et al., 2018;Zoccarato et al., 2018Zoccarato et al., , 2020Liu et al., 2020). Thus, their recognition and mapping based on simple CPTU investigations might be economically practical and convenient.
Even though basic CPT parameters plus U turned out to be an excellent tool for in situ facies characterization, additional data from cone penetration testing could be considered for this purpose. For example, dissipation tests could be used to estimate soil permeability and to assess the correct equilibrium piezometric pressure. As seismic CPTU (SCPTU) tests are also becoming increasingly common, evaluating the potential of additional parameters, such as in situ shear wave velocity or compression wave velocity, might be very useful for future CPTU-based facies characterization.

CONCLUSIONS
Engineering geological investigation of unconsolidated late Quaternary deposits through piezocone penetration tests (CPTU) is commonly limited to graphical representation of soil lithology into soil behaviour type (SBT) charts. In the southern Po Plain, where an abundance of sedimentological, palaeontological and chronological data is available, the calibration of CPTU profiles with 20 adjacent continuously cored boreholes demonstrates the strong potential of CPTU analysis for recognition and geotechnical characterization of alluvial, deltaic and coastal facies associations that are common to comparable settings all around the world. In order to visualize the engineering properties of the study units and to test the high potential of the proposed CPTUbased methodology for facies analysis, facies associations were plotted onto the SBT charts using non-normalized (i.e. basic) cone resistance (Qc) and friction ratio (FR).
Major results can be summarized as follows: • Sand-rich (fluvial/distributary-channel, bayhead delta, transgressive barrier, delta-front/ beach-ridge) and mud-dominated (well-drained/ poorly-drained floodplain, swamp, lagoonal and prodelta) facies associations have contrasting cone resistance (Qc) and water pore pressure (U) signatures that clearly reflect strong differences in grain size. On the other hand, plotting similar lithologies onto the SBT chart of Robertson (2010) resulted in local overlap between facies associations.
• Despite only subtle differences in grain size, clay-rich and silt-rich facies associations have markedly distinct engineering properties. Within well-drained floodplain deposits, palaeosols exhibit sharp peaks in Qc and Fs, and a general abrupt decrease in U that can be easily detected on CPTU profiles. The swamp facies association can be subdivided into two main CPTUlithofacies: swamp clays, with characteristic rectilinear CPTU profiles; and swamp peats, typified by concomitant peaks in Qc, Fs, U and FR. Although swamp and lagoon facies associations have similar engineering characteristics, swamp clays and central lagoon facies plot towards the 'sensitive fine-grained' zone 1, whereas swamp peat locally falls into the 'organic-soil' zone 2 and outer lagoon deposits into the 'sand mixtures' zone 5.
• Although the SBT chart offers good prediction of sediment type, parameters other than basic Qc and FR values can improve facies characterization, including Qc and Fs trends, negative/positive pore pressures, thickness values, gradational versus sharp bases/tops and characteristic peaks in CPTU profiles.
• In some instances, distinct facies associations may overlap onto the soil behaviour type chart and cannot be identified through CPTU analysis. In such instances, accurate facies analysis on sediment cores and sequence stratigraphic concepts are required to predict proximal to distal facies relationships and guide stratigraphic correlation.
• CPTU data may also reveal specific characteristics undetected in cores because of recovering issues, such as grain-size tendencies and centimetre to decimetre thick mud intercalations within sand deposits that may represent important stratigraphic surfaces.
• This study offers a methodology for CPTU profiles interpretation that can be confidently used for identification of facies association and for a variety of applications (stratigraphic correlation, hydrogeological, geotechnical and engineering investigations) in alluvial, coastal and deltaic settings worldwide because of similar late Quaternary stratigraphic architecture. After calibration with a limited number of cores, CPTU can be useful to increase data coverage at reasonable costs. This methodology can be complementary to the soil texture classification of the SBT charts and is particularly recommended for in situ facies association recognition, no matter where alluvial and/or coastal unconsolidated successions are located in the world.