1 Foreword

The inner part of the Italian peninsula is occupied by the Apennines chain, an about 1200 km-long and up to 250 km-wide Mio-Pliocene ridge formed by overlapping tectonic units constituted by carbonatic and basin/terrigenous flysch deposits, which are bounded along the coasts by more recent sediments of marine, alluvial and volcanic origin. The flysch deposits consist of alternating and highly deformed or completely disarranged rock and argillaceous layers. Their intrinsic lithological heterogeneity and complex fabric, along with the deteriorability of the hard fine-grained component, make extremely difficult material characterization for engineering purposes.

To address the problem, in the Seventies and in the Eighties some important research projects funded by national agencies allowed to gain some knowledge about the general features of these formations and the hydraulic and mechanical properties of their fine-grained component (Esu 1977; A.G.I., 1979; Bilotta et al. 1985). The scientific initiatives undertaken by the Italian, Greek and French geotechnical societies, which promoted some targeted conferences on the subject (“The Geotechnics of Structurally Complex Formations” held in Capri in 1977, the “Geotechnical Engineering of Hard Soils – Soft Rocks”, held in Athens in 1993, and then in Naples in 1998) contributed to some improvement of the knowledge. In the same period several studies on fine-grained materials having an intermediate behaviour between pure soils and pure rocks were conducted also abroad (Underwood 1967; Morgenstern and Eigenbrod 1974; Banks et al 1975; Botts 1986; Medley 1994). In the following years some studies were still carried out in Italy (Picarelli et al. 2002; Vitone and Cotecchia 2011; Napoli et al. 2022) even though with rapidly declining financial supports. Today, despite the growing land use for construction of infrastructures in areas occupied by these complex and often highly unstable soils, the research is going on even more slowly.

Considering the problems posed by these deposits to the stability of ancient towns and villages of historic relevance and to the construction of roads, highways, tunnels or pipelines crossing the internal mountainous areas of the Italian peninsula, this paper intends to highlight their peculiar mechanical and hydraulic behaviour. To this aim, the results of an extensive investigation recently carried out in a wide area occupied by the so-called Flysch Rosso (“Red Flysch”, here referred as RF), a typical formation present in a wide part of Southern Apennines, will be presented. Even being affected by uncertainties depending on the well-known complexity of such materials, in the Authors' opinion the information presented in this paper represents a useful contribution to the knowledge of the behaviour of flysch deposits. In particular, reported data come from an investigation campaign that included a significant number of in-situ tests, thus involving soil volumes larger than those that are usually examined in the laboratory, meeting in this way the need to investigate the material response at a scale that is more suitable for the analysis of engineering problems. As it will be shown, the hydraulic and mechanical behavior of the material is based on the typical criteria of Soil Mechanics, since the response of the whole seems to be essentially governed by the fine-grained component, even though the contribution of rocky or coarse-grained cemented layers or blocks is not negligible as revealed by the results of field tests.

For the sake of clarity, in the following with the term microfabric we will indicate the arrangement of soil particles and pores (micro-scale) and with microstructure the combination of microfabric and interparticle bonding (Mitchell 1976); moreover, the terms mesofabric and macrofabric will be referred to the arrangement of aggregates of soil particles and/or rock elements and of discontinuities (fissures and fractures) at the scale of laboratory samples (centimeters to decimeters) and of field problems (meters to tens of meters) respectively (Picarelli 1991).

2 The Apennines Chain

The Apennines chain is a Neogenic belt formed by the overlapping of several tectonic units characterized by Mesozoic–Tertiary shallow water to slope-facies carbonates (Apennine and Apulian carbonate platforms) and pelagic basin (Lagonegro) successions, with the interposition of several Miocene–Pliocene terrigenous deposits (thrust-top and foreland basins). The belt is the result of the convergence, during Mio-Pliocene, of the African and the Euro-Asiatic plates. Their collision led to formation of a duplex-type orogenic system, about 10 km thick (Fig. 1), with a tectonic convergence towards the Apulia Foreland (Patacca and Scandone 2001, 2007; Cinque et al. 1993). During the Plio-Quaternary age, extensional tectonic phases predominated, leading to the uplift of the Southern Apennines chain and formation of the present landscape. The general geomorphological structure of the area is characterized by steep ridges or deep river gorges. However, where clays and marly terrains predominate a hilly landscape with often unstable gentle slopes prevails. Figure 1 shows the main thrust belt along the Southern Apennines and a cross section throughout the main tectonic units obtained from data provided by deep boreholes and geophysical surveys.

Fig. 1
figure 1

Tectonic setting of Southern Apennines (Picarelli et al. 2002)

Full size image

Both basin and terrigenous flysch successions are characterized by interlayered rock layers or slightly cemented sediments (sandstones, limestones, conglomerates, marls and clays) that were deposited in a marine environment. The primary macrofabric of these deposits (Fig. 2a) is thus characterized by regular layering; only locally it may display a chaotic arrangement caused by submarine landslides. During the late Miocene and Pliocene, a tectonic compressive regime strongly deformed such deposits (Fig. 2b); as a result, the secondary fabric of the fine-grained component consists of an aggregate of small hard fragments that are usually called “scales”. Finally, during the Pliocene and Quaternary tectonic uplift, faulting and folding processes led to formation of thick shear zones (Fig. 2c). In some cases, the fine-grained component experienced some metamorphism and mineralogic changes. The special fabric of these formations, which makes challenging any mechanical soil characterization, is the heritage of such a complex geological history. Italian geotechnical engineers generally recognize these materials as “structurally complex formations” (Esu 1977).

Fig. 2
figure 2

Schematic representation of the geological history of terrigenous deposits in the Apennines chain

Full size image

The “Red Flysch” formation that is dealt with in this paper is an excellent example of the tectonized flysch formations that outcrop along the Apennines chain.

3 The Structurally Complex Formations

The basic features of structurally complex formations are as follows: (i) an intrinsic lithological heterogeneity, (ii) a small porosity and (iii) some bonding affecting the microstructure of the fine-grained component, (iv) a cemented coarse-grained or fractured and sometimes destructured rocky component; (v) a widespread presence of both small and persistent discontinuities (cracks, joints, shear fissures and surfaces) that characterize the meso and the macrofabric of the whole.

The classification of structurally complex formations proposed by Esu (1977) identifies three distinct material types, which are schematically presented in Fig. 3a.

Fig. 3
figure 3

a The structurally complex formations (Esu 1977, redrawn by Picarelli et al. (2022); b typical size of scales; c definition of ratio RT

Full size image

The first type (A) consists of homogeneous weakly bonded fine-grained soils (stiff clays, claystones, clay shales) often generically indicated as clay shales. The A1 sub-type includes stiff fissured jointed clays. The A2 sub-type consists of small, millimeter to centimeter (Fig. 3b) fragments ("scales") bounded by polished, sometimes shiny, surfaces.

The type B includes alternating layers of rock/slightly cemented coarse-grained soil and of hard fissured fine-grained materials with a typical scaly fabric. The ratio RT between rock plus slightly cemented coarse-grained and total volume (Fig. 3c) is highly variable from case to case. This material is further subdivided into three sub-types, the differences depending again on material macrofabric, which may be quite regular or highly deformed or even chaotic. The sub-type B1 consists of regularly alternating layers. The rock/cemented layers present closely spaced joints while the clay shale interbeds are highly fissured with fissures almost parallel to the layers. The two sub-types B2 and B3 consist of highly deformed and completely disarranged materials, respectively. This configuration (secondary fabric) differs from the original one generated by the sedimentary processes (primary fabric) due to tectonic stresses that led to formation of a folded and discontinuous macrofabric characterized by deformed layers with joints filled with clay (B2 type), or even by isolate rock blocks scattered into a highly fragmented fine-grained matrix (B3 type).

The third type (C) is a mixture of rock fragments and of a clay matrix generated by erosion and re-deposition and/or by slope movements (colluvial soils). According to a different point of view, the sub-types B3 and C might be recognized as Bim-soils (Medley 1994; Napoli et al. 2022) or as soft Bim-rocks.

As illustrated above, a highly fragmented mesofabric is a typical feature of the clayey component of types A2 and B. In B1 formations, it consists of centimeter hard platy fragments (Fig. 3b) bounded by polished fissures parallel to the layer. In the A2, B2 and B3 formations, it consists of millimeter to centimeter “scales” bounded by polished surfaces (minor shears) and of principal shear surfaces that may cross such aggregates of fragments.

The Red Flysch, which is present in a wide area of the Southern Apennines, may be classified as a B structurally complex formation.

4 The Mesofabric of Tectonized Clay Shales (Types A2 and B)

The typical mesofabric of tectonized clay shales in type A2 and B formations is the result of the complex stress field experienced during the compressive (Fig. 2b) and uplift stages (Fig. 2c).

A schematic picture of the deformation mechanisms of such deposits due to compression, bending and folding is shown in Fig. 4.

Fig. 4
figure 4

Schematic deformation mechanism during the tectonic compressive phase in A2 and B formations: a a fold in the Red Flysch; b schematic picture of the fold; c effects of bending in a B deposit; d effects of bending in a A2 deposit

Full size image

A first obvious effect is fracturing of the competent rock layers in type B2 and B3 materials, as shown in Fig. 4a where highly fissured clayey interlayers are visible. This is schematically represented in Fig. 4b. In the fold hinges, compressive strains led to a bedding-parallel fissility, while in the limbs (Fig. 4c) strong shear strains produced the typical sheared ("scaly") mesofabric of these materials (Vannucchi et al. 2003). Similar phenomena take place also in terrigenous deposits lacking of a rocky component (type A2, Fig. 4d). Macroscopically, scaly surfaces are smooth and shiny or polished due to alignment of clay particles. The described fabric can affect large masses. Deformation mechanisms like that of the fold limbs characterize fault zones. The typical fissility of the fine-grained component of B1 materials, which present a mesofabric with plates more or less parallel to the layers, could be justified by compression mechanisms as in the hinges of folds.

Mechanical aspects of shearing processes were well described by some scholars in the past taking as basis Riedel studies (Riedel 1929). Riedel showed that soil failure is preceded by formation of short shear surfaces, called minor shears (or Riedel shears), inclined to the direction of shear of an angle that is a function of the local direction of principal stresses. The rupture process is then accomplished by formation of a major and persistent shear surface (the principal shear) in the imposed failure direction, which tends to interconnect minor shears (Fig. 5a). After failure the material then presents a shear zone including minor shears and the principal shear. Skempton (1967), Tchalenko (1970) and others noticed that the same features observed in laboratory tests may be found in the field at very different scales (macrofabric). Figure 5b, for example, shows a fault zone, which displays a set of inclined shears concentrated into a shear zone. As outlined by Picarelli et al. (1997) based on data from literature, shear zones can be generated by different natural phenomena. In particular, the slip surface of landslides should coincide with the principal shear that develops in the latest stage of the failure process. An example of shear zone including minor shears and a principal shear found at the base of a landslide in highly OC clays was reported by Skempton and Petley (1967) (Fig. 5c). Figure 5d, in turn, shows a complex system of conjugate shears (R and R′) observed by Davis et al. (2000) in a shear zone in the Sheets Gulch area (Utah, US): there, hundreds of R and R' shears with lengths from centimetres to tens of meters and thicknesses from millimetres to tens of centimetres, were recognized.

Fig. 5
figure 5

Shear zones in clay at different scales: a direct shear test (Skempton 1967); b a fault at Jari, Pakistan (Skempton and Petley 1967); c a shear zone found at the base of the Guildford slide (Skempton and Petley 1967); d a shear zone containing conjugate Riedel shears observed in Utah, US (modified after Davis et al. 2000)

Full size image

A theoretical interpretation of the mechanism of shear failure of an infinite slope subjected to groundwater table rising is provided by Urciuoli (2002), who, in particular, shows that the thickness of the shear zone depends on the scale of the problem at hand; other factors, such as the initial state of stress and both friction and dilation angle of the soil, play an important role as well (Urciuoli and Picarelli 2004). It is also evident that the accumulation of high strains can destroy or deeply modify the described soil fabric. Tsui et al. (1988), for example, describe the shear zone found at the base of a cut in an ice-thrusting rock in the Highvale mine, Alberta, which consists of remoulded clay including only isolated hard lumps (lithorelicts) and single shears. The same can be often noticed in tectonized clay shales that are quite deteriorable materials.

5 The Red Flysch Formation

The Red Flysch is a typical structurally complex flysch formation that outcrops in a vast area of Southern Italy. Since the Eighties it has been the object of studies (Iaccarino et al. 1995; Pellegrino et al. 2004; Picarelli et al. 2005; Urciuoli et al. 2016) regarding some active earthflows in an area nearby the town of Potenza (site A, Fig. 6a). Another important campaign of investigations (Lunardi et al. 2012) was carried out in the first decade of this century for the construction of tunnels along the Val Fortore (Benevento-Italy) road (site B, Fig. 6a). Further investigations have been recently carried out along a section of the Campanian Apennines (site C, Fig. 6a, b), where a high-speed railway that connects the Tyrrhenian and Adriatic coasts is being built. In the following some of the experimental data taken in this last campaign referring to some facies of the formation are synthesized and interpreted.

Fig. 6
figure 6

a Geological sketch of the Southern Apennines chain (Vitale et al. 2018) (notice that the Red Flysch formation is part of Lagonegro-Molise Basin); b typical profiles of the Red Flysch formation (Vitale and Ciarcia 2018); c Red Flysch facies investigated in the present study at site C (see Table 2)

Full size image

5.1 Geology of the Investigated Area

The Red Flysch belongs to the “Lagonegro-Molise” basin paleogeographic unit (Fig. 6a) (Vitale and Ciarcia 2018), which includes a meso-cenozoic deep basin sequence with basal carbonate deposits followed by siliceous and clay successions (Fig. 6b). The Red Flysch represents the upper part of the Unit and consists of Cretaceous and Miocene calcarenites and calcirudites. In the Red Flysch different facies may be recognized (Fig. 6c), which are described in Sect. 5.2. As a whole, the Red Flysch formation represents the typical escarpment succession of a deep basin receiving lateral input by big debris channels mostly fed by a limestone source. Sin-sedimentary channels consequent to slumping phenomena, which are visible at the outcrop scale, seem to fully confirm this hypothesis. The whole succession is covered by Miocene sandstones of the Numidic Flysch or Anzano formations.

During the late Miocene and the Pliocene, the Red Flysch was deformed by tectonic compressive phases that created folding systems both at the regional and local scales, which are evident in the clayey and marly portions (Figs. 2b and 4d). During the Plio-Quaternary, extensional tectonic phases prevailed, leading to the uplift of the Southern Apennines chain and to a reshaping of the landscape (Fig. 2c). The area is characterized by small horsts and deep river gorges including limestones and other hard rocks. Where clays and marly terrains crop out, the dominating hilly landscape displays a widespread slope instability.

5.2 The Lithological Components of the Red Flysch Formation

Table 1 summarizes the investigation campaign carried out along a portion of the railway route about 5 km long (Fig. 6a). The results of part of the tests listed in the table have been considered in this paper.

Table 1 Experimental program carried out for the construction of high-speed railway on the analysed Red Flysch facies
Full size table

This rich campaign of investigations allowed to identify the stratigraphic successions along the investigated section showing that the Red Flysch, which in the professional practice is often treated as a homogeneous formation, is instead a very tricky set of distinct structurally complex deposits (facies). The basal zone (Fig. 6b) consists of calcirudite slope deposits with frequent slumpings and channels. This calcareous member, which is about 150 m-thick, gradually turns into marly–calcarenite layers and then to varicoloured clayey members. Frequent decimetre calcarenites layers testify cyclic sedimentological variations in the ancient deposition basin. In some boreholes, the clayey member includes some metres of silt and sandstones. The marly and clay successions are up to 200 m thick.

Along the investigated section, seven different facies have been recognized in the Red Flysch based on the following factors: (i) the ratio RT, (ii) the arrangement and continuity of the rocky or cemented component, (iii) the nature and the uniaxial strength of such a component and (iv) the nature of the clayey component. The main features of the single facies are summarised in Table 2 and their acronyms are reported in Fig. 6c. Table 2 synthetically describes the macro-fabric of each facies based on the classification of structurally complex formations (Fig. 3); the range of values of both thickness and spacing of rock or cemented coarse-grained layers is also reported. These last, which significantly prevail in four facies (70% more or less of the investigated depth), consist of highly fractured marls (M in the table and in the following figures), sandstones (S), limestones (L) or conglomerates (C).

Table 2 Main features of the Red Flysch facies
Full size table

Drilling has been often complicated due to both complexity of soil structure and tendency of some clayey facies to swell because of wetting. For these reasons, it has been carried out following a continuous-cycle system. Sampling has then been carried out by a wireline modality; water has been generally used as a drilling fluid but also polymers have been seldom adopted.

Different sampling techniques have been used depending on the local soil features. Where possible (shallow depths, softer layers), undisturbed soil samples have been recovered through Shelby samplers; at higher depths, where the soil stiffness increases, Mazier e Triplex samplers have been often preferred. Laboratory samples have been prepared using a cylindrical iron steel cutter gently pushed into the sample core and then extruded with the extrusion dolly.

Some data about the uniaxial strength of the rock/cemented component, as measured in laboratory compression tests or indirectly, by point load tests, are presented in Fig. 7.

Fig. 7
figure 7

Uniaxial strength of the lapideous component of single facies

Full size image

The differences are significant; in particular, while the uniaxial strength of conglomerates and limestones is fairly high, the other lithotypes are much weaker, in some cases consisting in practice of a slightly cemented soil, as in the case of “sandstone”, whose uniaxial strength (not reported in Fig. 7) ranges between 0.03 and 1.60 MPa.

The fine-grained component varies from facies to facies; in particular, this study distinguishes the varicoloured “scaly” clay shales (indicated with the symbol “cs”) from the hardest grey “scaly” marly clays (“mc”). In addition to types “cs” and “mc”, further clayey components, such as red clays and marls (“rc”), have been recognized.

No data are available about the mineralogic composition of the fine-grained component of the investigated facies. However, previous studies indicate that the mineralogy of clay shales of the Red Flysch formation should be dominated by kaolinite and illite. Hematite should be present only in some facies (especially in reddish clays), whilst it is absent in the others (Fiore et al. 2000).

Figure 8 reports the envelopes of grain size and plasticity of the fine-grained component of the single facies. As shown in the figure, minor differences may be recognized, at least in the area at hand. In any case, the scattering of data, which depends on the intrinsic variability of the different fine-grained components, is so high that referring to the average values may be misleading. The void ratio of the tested materials, as obtained from conventional oedometer tests, essentially ranges between 0.55 and 0.7, quite low values consistent with the complex stress history of the formation. The overconsolidation ratio obtained from oedometer tests seems to not exceed a value of 5, but probably these data are affected by a too low value of the maximum applied pressure (5 MPa) and/or by some disturbance in the sampling and specimen preparation phases.

Fig. 8
figure 8

a Grain size  and, b plasticity of the fine-grained components of the Red Flysch facies

Full size image

As discussed above, in the following the seven facies have been recalled with an acronym: capital letters refer to the lapideous/coarse-grained cemented component, small letters to the fine-grained one. The deepest facies (RF-CLrc), essentially includes up to 5–8 m thick limestone and conglomerate layers with metric intercalations (up to 10 m) of red marls and clays. This facies is followed by a sequence of thin layers (a few decimetres thick) of limestone, marl and varicoloured clay (RF-LMcs) and by a deposit of poligenic conglomerates with lapideous limestone elements and a clayey matrix (RF-CLcs).

These facies are overlapped by a sandstone facies, with marl and varicoloured clays (RF-MScs) and 5–10 m thick rocky layers. The shallow facies is essentially pelitic and consists of varicoloured clay shales (RF-Lcs) including calcareous layers, which turn into scaly marly clays with sandstone (RF-MSrc) and to a facies with a marly component that includes calcareous layers (RF-Lmc) as well.

The photos of some calcareous/marly facies and a schematic picture of their macrofabric are reported in Fig. 9. This shows the following:(i) the basal RF-CLrc facies, which presents ordered limestone layers separated by thin layers of marl and red clay; (ii) the RF-LMcs facies that consists of twisted and fractured calcareous layers separated by thicker layers of marls and of varicoloured clay; (iii) the RF-CLcs facies consisting of conglomerate with limestone clasts in a varicoloured clay matrix.

Fig. 9
figure 9

Some outcrops of the Red Flysch formation: a schematic pictures and b photos

Full size image

A synthetic information about the degree of fracturing of the single facies and the role played by tectonics is provided by the RQD factor, which is shown in Fig. 10 (see also Table 1). The severe impact of tectonics on the RF-LMcs, RF-Lcs and RF-Lmc facies, which are characterized by a disordered or chaotic macrofabric (B2-B3 types), seems to be well captured by low RQD values. In contrast, the higher values of the same parameter that have been measured in the RF-CLrc, RF-MScs, RF-MSrc facies is associated with a more regular fabric (B1). Finally, a lower degree of fracturing characterizes the chaotic RF-CLcs facies, whose fabric is essentially a result of the sedimentation conditions (primary fabric).

Fig. 10
figure 10

RQD values measured in Red Flysch facies

Full size image

Some examples of the mesofabric of the argillaceous component of clayey/marly facies of the Red Flysch are shown in Fig. 11, which highlights scales, shear fissures and footprints of principal shears widespread in the matrix.

Fig. 11
figure 11

Clayey samples from Red Flysch in the investigated area: a, b, c varicoloured clays (“cs”), d, e, f grey marly clay (“mc”)

Full size image

The complex geology of the investigated area is illustrated in Fig. 12, which clearly shows the alternating facies present in a long section of the investigated area and their contacts with more recent formations.

Fig. 12
figure 12

Cross section of the investigated area at site C (see Fig. 6)

Full size image

6 Challenges in the Red Flysch Characterization

6.1 The Major Role of Fabric

A fundamental step in the analysis of geotechnical problems in tectonized flysch formations is the material characterization through measurement of their hydraulic and mechanical properties. According to the principles of Soil Mechanics that consider the soil as a porous medium, such properties should depend on intrinsic and state properties, such as grain size distribution, plasticity, porosity, water content, current stresses and history. In the domain of Rock Mechanics, when the medium has to be treated as an assemblage of rigid blocks separated by well-defined sets of discontinuities (discontinuum model), material characterization should focus on the mechanical and physical properties of both the rock matrix and the discontinuities. However, experience shows that flysch characterization is generally much more complicated than using one or the other model due to the major role of fabric, whose impact is much greater than that of any other factor.

Figure 13 is a typical example of the materials that engineers have to deal with in areas occupied by tectonized flysch. The picture shows an about 10 m2 vertical section of a B3 outcrop in Red Flysch, in which blocks from calcarenite layers are plunged into layers of scaly marl or stiff clay. The bedding surfaces are interrupted at several points locally reaching lengths not greater than 2 m, while shorter tectonic discontinuities (essentially principal shears) cross fine-grained layers in various directions. Due to the short length and variable spatial orientation, such macro-discontinuities play a complex and not negligible role on the overall mass behaviour, which often cannot be easily modelled following the classical criteria of the Rock Mechanics.

Fig. 13
figure 13

a Section of an outcrop in Red Flysch (RF-Lmc facies); b magnification

Full size image

As mentioned above, it is clear that the hydraulic and mechanical response of these formations depends on the combined influence of the following: (1) the basic properties of the single components, (2) the mesofabric of the fine-grained component and (3) the macrofabric of the whole. In principle, the behaviour of the formations that are dealt with in this paper is essentially governed by the fine-grained component, the rock or cemented coarse-grained component having lost any continuity. As a matter of fact, the investigations have been just focused on the fine-grained component. In general, however, laboratory tests can investigate only volumes up to, say, 10–3 m3, thus highlighting only the role of basic properties and of the mesofabric of the fine-grained component. This suggests that in many cases laboratory investigations might be not representative because larger material volumes should be considered. Field tests, in turn, can provide useful information about the response of the whole, thus accounting for the role of macrofabric, up to a volume of about 1 m3, thus losing the role of discontinuities and of rock or of coarse-grained cemented blocks or layers at a higher scale.

The results of the present investigation, which includes both laboratory and in situ tests (as mentioned above, not all results of the tests indicated in Table 1 have been used in this paper), provide some data about the hydraulic and mechanical properties of clays and marls featuring some facies of the Red Flysch highlighting at the same time inherent challenges in material characterization. Due to the obvious sampling difficulties, the data presented below concern only a part of the material recovered from boreholes and thus they could not be fully representative of the investigated facies; they, however, provide useful information.

6.2 The Hydraulic Response

As discussed above, the hydraulic response of flysch deposit depends on the combined influence of grain size, porosity and micro/mesofabric of the fine-grained component, and of macrofabric of the whole. Previous experience on this subject is quite random and scarcely organized not allowing to build a consistent and reliable framework. However, available data confirm previous considerations about the major role that is played by fabric. Urciuoli (1988), for example, shows that the hydraulic conductivity of stiff fissured clays characterizing a B3 structurally complex formation, as measured through falling head tests on undisturbed specimens placed in oedometer cells, is about one order of magnitude higher than the value measured on reconstituted specimens having the same overall void ratio (K = 10–3.5×10–11 m/s against 10–3×10–12 m/s); this highlights the influence of opened fissures. The Author also shows that the hydraulic conductivity measured by in situ tests (K = 8 × 10–11–5 × 10–9 m/s) is even two orders of magnitude higher, stressing in this way the role of major discontinuities on the overall permeability. Similar data are reported by other Authors (see for example Jappelli and Valore 1980). In contrast, constant head tests carried in a triaxial cell on some undisturbed specimens of the high plastic Bisaccia clay shale (A2 type, Ip = 60–150%) under confining effective stresses close to the field values and on reconstituted specimens having the same void ratio, provided similar rather low values (k = 7×10–13–2×10–12 m/s) thus suggesting that in the examined case the fissures were closed and that soil permeability depends on grain size and plasticity only (Picarelli et al. 2002).

The data reported in this paper concern four distinct facies of the Red Flysch, i.e. the RF-Lcs, the RF-Lmc, the RF-CLrc and the RF-MSrc types. These last two types are characterized by a higher RT value (70–85% and 60–70%, respectively) than the first two facies. Laboratory investigations consisted in a series of one-dimensional consolidation tests on small specimens of the fine-grained component (h = 2 cm) subjected to vertical effective stresses reproducing the values that are presumably operative in the field. The hydraulic conductivity was then calculated based on measured values of the coefficient of consolidation.

A number of Lefranc and Lugeon tests were carried out on site. The Lefranc tests were conducted on 50 cm-long sections located at the bottom of boreholes: the duration of each test was about one hour. The Lugeon tests were performed on 3 m long portions of boreholes closed at the edges by pneumatic packers, thus allowing to consider the overall response of a significant soil volume. Each test included five stages, in which the water pressure was increased and decreased between zero and a maximum value not higher than 0.3 MPa to avoid hydraulic soil fracturation, as established based on some preliminary field tests. The average value of the hydraulic conductivity was calculated at each stage of the test.

All results are summarized in Fig. 14. The response of the highly tectonized RF-Lcs facies (B3 type), which includes not more than 15% of limestone elements (Table 1) and is characterized by the lowest values of RQD (Fig. 10), is summarized in Fig. 14a. The laboratory tests were carried out on samples taken within a maximum depth of about 35 m. As shown in the figure, the coefficient of hydraulic conductivity, K, measured through laboratory tests, is comprised between 2×10–9 and 4×10–11 m/s, a range of values that is well below the one obtained from in situ tests down to a depth of 90 m, which covers the interval 2×10–5–5×10–9 m/s. Not hiding the problem of correctly interpreting the results of field tests, such strong differences seem to be mostly due to the influence of the network of major opened discontinuities (cracks and principal shears) that soil samples cannot include thus the laboratory tests cannot capture. The scattering of data is, however, relevant and the decreasing trend with depth is not so marked, probably because the degree of fracturing of this facies remains the same down to great depths. This seems to be confirmed by the results of Lugeon tests, which provide the highest values of K.

Fig. 14
figure 14

Results of permeability tests: a RF-Lcs facies, b RF-Lmc facies; c RF-CLrc and RF-MSrc facies

Full size image

The hydraulic response of the RF-Lmc facies (B2 type), which consists of marly clays with a slightly higher calcareous component (RT = 15–25%) than the RF-Lcs facies, is presented in Fig. 14b. Lefranc and Lugeon tests were mainly carried out up to a maximum depth of about 40 m except for one Lefranc test performed at a depth of about 90 m. The values of K provided by laboratory tests, which fall in the interval 8 × 10–9–5 × 10–11 m/s, are again much lower than those obtained from field tests, which range between about 1x10–5 and 7x10–9 m/s.

The differences however tend to decrease with depth despite the adoption of Lugeon tests, which examine the response of larger soil volumes, possibly due to a less widespread network of discontinuities in this less tectonized formation, and to a decreasing aperture of the macro-discontinuities.

The last figure (Fig. 14c) puts together the ensemble of RF-CLrc and RF-MSrc facies (B1 type) that present a richer rocky or cemented coarse-grained component (RT = 70–85% and 60–70%, respectively) and a less disturbed macrofabric than the other two facies. In this case the coefficient of hydraulic conductivity obtained through in situ tests only, performed down to depths of about 120 m, ranges between 7 × 10–8 and 2 × 10–9, which are the lowest values that have been measured in situ. Such a result might be due to a mutual sealing effect at the interfaces between rock and clay due to the higher presence of the first component that here is less tectonized.

Summing up, all results show a much higher permeability than expected based on the nature and index properties of the finest component. This is well known from the literature concerning fissured clays. Working on the well-known London Clay, Garga (1988), for example, reports values of K, as measured from in situ tests, at least one order of magnitude higher than those obtained through laboratory tests, showing in addition that the differences decrease with depth. In these soils the rapid fluctuations of the water level that are usually measured through piezometers seem to confirm their rapid response to changing hydraulic boundary conditions (Skempton and Henkel 1957). The results obtained on the Red Flysch are certainly due to the presence of major tectonic discontinuities, which are widespread in the fine-grained component in the form of opened fissures, cracks and principal shears. In particular, the Lugeon tests, which are carried out on long water inlet sections (3 m), allow to investigate volumes of some cubic meters, in which major discontinuities can play a significant role. By the way, this consideration can justify the high values of K that have been measured even at great depth. Interestingly, the field tests carried out on facies with a high and well-ordered rock component (RF-CLrc + RF-MSrc, B1) provide smaller values of K than those measured in the clayey facies (RF-Lmc and RF-Lcs, B2 and B3). This suggests that the water flow might be favoured by the spreading and networking of discontinuities in the clayey matrix, which in B1 formations is balanced by the presence of rock layers, which play a sealing effect where fissures are interrupted by rock elements and vice versa. The fact that in B3 highly tectonized formations with a low rocky/cemented component the influence of discontinuities on permeability does not decrease significantly with depth seems to confirm such a consideration.

6.3 The Mechanical Response

The available technical literature mostly includes laboratory tests on the fine-grained component of this complex formation (A.G.I., 1979; Bilotta et al. 1985; Olivares and Picarelli 1999; Picarelli et al. 2002; Vitone and Cotecchia 2011). According to published data, as far as the axial strains remains within values less than 0.5%, fissuring plays a minor role and the soil behaves as a continuum; then, with the increase of the stress level, a sort of strain localization takes place along minor shears (Olivares and Picarelli 1999) (Fig. 15).

Fig. 15
figure 15

Deformation mechanisms of "scaly" clays in triaxial tests (from Olivares and Picarelli 1999)

Full size image

At failure, the material behaviour can be considered equivalent to that of a granular soil with scales taking the place of grains. Following this idea, the friction angle could be interpreted through the Patton theory (Patton 1966) as the combination of a basic friction angle, which in granular soils is associated with the critical friction angle, and of a "dilation angle". Accounting for the features of scales, which are bounded by minor shears, in this case the basic friction angle is probably close to the residual value; the dilation angle, in turn, depends on soil mesofabric, and, in particular, on size and arrangement of scales. The failure envelope is non-linear; due to the strong influence of the low friction angle that is operative along minor shears, it may be for most located below the critical failure envelope as obtained on reconstituted specimens (Picarelli et al. 2002). In undrained tests, of course, the soil response depends also on induced excess pore water pressures. Working on the A2 Bisaccia clay shale, Olivares (1997) measured essentially positive values of excess pore pressures in the phase preceding soil failure.

In this study the mechanical response of the Red Flysch was investigated in the laboratory, by undrained triaxial compression tests (CIU and UU tests, the CIU tests being conducted on 35.7 mm specimens in diameter, consolidated at an effective stress equal to the average lithostatic value) as well as in situ, by pressuremeter tests. These last allow to investigate a greater soil volume than in the laboratory. Of course, the length of the device compared to the spacing of discontinuities or of rock blocks size is an important factor to assess the representativeness of data (Mair & Wood 1987). Pressuremeter tests were carried out by two different devices. The first one (type 1) is a classic 60 cm long Menard equipment, which is constituted by three distinct cells. The instrument is set into a borehole at the selected depth and then the intermediate cell is pushed against the soil; the test includes a loading–unloading cycle. The second device (type 2) consists of a single 50 cm-long cylinder, which is equipped with three radial probes for high-precision volumetric deformation readings. In this case the test consists of three loading–unloading-reloading cycles.

Accounting for the adopted strain rate, pressuremeter tests in clayey soils are generally interpreted as undrained tests. In this investigation the undrained strength Su has been computed as the difference between the applied pressure required to double the cavity size and the initial pressure, divided by a numerical coefficient, N, which according to Amar et al. (1975) should range between 5.5 and 10 as a function of depth. Such a correlation is considered to be reliable at least for homogeneous soils (Isik et al. 2015). In the case at hand, N has been assumed to be equal to 5.5 for depths less than 30 m, and equal to 8 for higher depths. Moreover, the deformation modulus has been calculated along the first loading branch because the aim of the study was the comparison among strength values measured by pressuremeter tests and those obtained by laboratory CIU and UU tests taking into account for the effect of tested volumes.

All test results are shown in Fig. 16. The triaxial tests performed on samples taken at depths down to about 40 m provide quite consistent data, which follow a linear trend even though with some scattering. In particular, the average undrained strength of clays of the RF-Lcs facies (Fig. 16a) is about 80 kPa at a depth of 10 m and about 280 kPa at a depth of 40 m; the strength of the RF-Lmc facies (Fig. 16b) is slightly lower, ranging between about 80 and 200 kPa at the same depths as in the RF-Lcs facies. The excess pore pressures measured in the CIU tests are quite low being characterized by a Skempton coefficient of pore pressure Af at failure ranging between 0.05 and 0.15. Despite the low void ratio of such a clayey fine-grained material (Fig. 8c), the measured undrained strength is not so much higher than the value that would be obtained on normally consolidated clays with the same grain size and plasticity index, as computed by well-known empirical relationships. This remark seems to indirectly outline the major role played by mesofabric, and by minor shears, which tend to hide the role of the overconsolidation ratio.

Fig. 16
figure 16

Undrained strength of the RF-Lcs (a) and RF-Lmc facies (b)

Full size image

The results of the in-situ tests display again a high scattering, which is probably due to the random presence of lapideous or cemented coarse-grained elements in the soil volume around the instrument and to the influence of discontinuities, which are not persistent and variously oriented in the space. However, in the RF-Lcs facies (Fig. 16a) the results of pressuremeter tests, which were mainly performed at higher depths than those investigated in the laboratory, cover the extrapolation of the line fitting the results of triaxial tests. This could be due to the low RT of this facies (RT = 5–15%), whose mesofabric might substantially govern the mass behaviour also at the macro-scale. On the other hand, considering the low values of excess pore pressures that have been measured in CIU triaxial tests, the difference in permeability measured in laboratory and in situ tests should play a minor role on the drainage conditions at failure. In the RF-Lmc facies (Fig. 16b) the value of Su obtained from pressuremeter tests, which ranges in between 200 kPa and more than 700 kPa, is with some exceptions higher than that obtained from laboratory tests. In the literature this result is variously interpreted being attributed either to high values of the strain rate (Wroth 1984), or to partial drainage occurring during the test (Wroth 1984; Pyrah et al. 1988), or to non-cylindrical expansion (Yeung and Carter 1990; Houlsby and Carter 1993), or to stress paths that are different from those that are imposed in triaxial tests (Wroth 1984). In this case however, the observed difference between laboratory and in situ test results is probably due the high value of RT. Moreover, in B2 formations as in the case at hand, in which the mutual distance of rocky/cemented layers ranges between 2 and 5 m (and even more in B1 formations), such effect is certainly more relevant than in chaotic B3 formations.

Soil stiffness in terms of undrained secant modulus (Eu-sec) of the facies RF-Lcs and RF-Lmc (RT = 5 to 25%), is presented in Figs. 17a and b, respectively. In particular, the values obtained in UU and CIU tests have been computed for a deviator stress equal to 30% of that at failure; the undrained modulus obtained from pressuremeter tests has been calculated from the quasi-linear section of the loading branch. As shown, the stiffness of the fine-grained component as from laboratory tests does not change significantly with depth keeping values of some tens of MPa in both facies. However, considering the sensitivity of the undrained modulus to sampling disturbance, especially at low to medium strains (Jardine et al. 1984), laboratory data probably underestimate its actual value. The scattering of data and the larger values obtained in field tests may be again attributed to the role of cemented elements.

Fig. 17
figure 17

Comparison between undrained modulus, Eu-sec, from CIU and UU tests and pressuremeter tests for a RF-Lcs and b RF-Lmc facies; c sample taken at test depth

Full size image

7 Conclusions

The knowledge about properties and behavior of flysch deposits that are widespread all along the Italian peninsula and in wide areas around the Mediterranean basin, is still very poor even though papers on this subject do not miss, especially in the geological literature. Often the technical problems are analyzed assuming that the overall mechanical behavior is essentially controlled by the fine-grained component. Therefore, the investigations mostly concern this component that is usually tested in the laboratory. In general, the test results are affected by obvious difficulties in obtaining good quality samples. However, these problems and the uncertainties related to the role of the clay macro-discontinuities that are widespread in clay and of the lapideous/cemented component make doubtful any material characterization.

The availability of a rich set of data obtained through both laboratory and in situ tests performed on materials present in a long section of the Southern Apennines chain where different facies of the Red Flysch formation outcrop, was a good opportunity to focus on a problem that is often disregarded, showing that the principles of Soil Mechanics or of Rock Mechanics alone generally do not fully apply to such formations.

In fact, in this case the overall material behavior not only depends on the features of the single components, but also on the macrofabric of the whole, which is in turn a result of the arrangement of the lapideous component and of the system of discontinuities that characterize both components. Such features can strongly govern the overall material mechanical response making it more pervious and more resistant and stiffer than indicated by laboratory tests.

The data that have been reported in this paper provide useful information, even though obviously they cannot cover the wide range of problems that could be encountered in the practice. However, they have shown that the network of macro-discontinuities, both in the fine-grained component and in the lapideous/coarse-grained cemented one, can generally favor water flow giving to the mass a higher permeability than measurable on small clayey specimens. On the other hand, both strength and stiffness of the material as a whole seem to be higher than those measured in laboratory tests on small fine-grained specimens. Probably this is related to the presence of the lapideous/coarse-grained cemented component, whose influence prevail respect to that played by the macro-discontinuities; as a matter of fact, in general these do not play a significant role unless their spatial orientation is consistent with the mechanism of failure that is theoretically expected in the problem at hand (kinematic compatibility).

Certainly, a way to run when dealing with these materials is patiently combining the results of careful laboratory and field tests with those obtained from back analysis of field experiments and from the monitoring of man-made works.