Strategies to stabilise a complex instability phenomenon in Central Italy

The instability phenomena that regards very large and complex landslides are a challenge to minimize social and economic losses. In this paper we present the studies to programme and design the stabilizing works of the complex slope instability phenomena in the town of Lettomanoppello (Central Italy). It is located in the central-eastern margin of the Apennines, along the north-western foothills of the relief of the Maiella Mountain (2793 m a.s.l.). The slope, on which the village of Lettomanoppello (350 m a.s.l.)is located, is historically known for its susceptibility to landslides. These phenomena involve geo-morphological aspects and geotechnical conditions and complex geo-materials characteristics. The area incudes Quaternary continental deposits (detritus deposits, limestone debris hetero-metrics, landslide body) and Mio-Pleistocene marine deposits (clay, clay, marl and chalk alteration; limestone). The hydrogeological conditions of the area are characterized by a complex drainage system that affects the stability of the slope. Among the stabilization works that are provided in design, the drainage works take a prominent role. The paper presents the geotechnical characterization for the design (location and dimensions) of the tunnel for deep water drainage. The location of drainage tunnel was design according to the hydraulic condition. The tunnel crosses a horseshoe shaped underneath the historical centre of Lettomanoppelllo and the elevation of its mouths are at 290 m a.s.l. The total length is about 1000 m, and the coverings vary from 60 m to 100 m. The operation of drainage tunnel improves the overall stability of the slope and the drainage waters will be conveyed to pipe towards a hydraulic turbine located at 190 m a.s.l. in the lower part of the slope to produce electric energy.


Introduction and history
The landslide of Lettomanoppello, located in Central Italy (south-west of the city of Pescara) is a large, very complex active landslide [1].Slope instability phenomena, often triggered by intense rainfall and seismic events, are recurrent in this region of the Apennines and frequently result in serious damage to the historical-environmental heritage and in dreadful economic losses.The small town of Lettomanoppello, located in the outermost north-west offshoots of the mount Maiella, on the right-hand side of the valley of the Lavino River, is involved by considerable traffic of vehicles heading for the tourist resorts in the nearby "Maiella National Park".The slope just below the town has been involved in the past, since ancient times, by several impressive landslides.Evidence and documentation exist of various recent landslides occurred in the period l920-1987.The most catastrophic 1249 (2023) 012010 IOP Publishing doi:10.1088/1755-1315/1249/1/012010 2 event, occurred in 1953, caused the complete destruction or irreparable damage of several buildings (some ruins can be seen still at present).Various stabilization works were executed along the slope at different times, since 1920, in order to mitigate the effects of the landslides and to protect the town from possible regression of the movements towards the top of the hill.However, these works did not result in any substantial stabilization of the slope.In 1987 evidences of active movements (long semi-circular cracks, settlements) were observed on the road pavement of Corso Vittorio Emanuele, the main street crossing the town.The downhill edge of the road was nearly coinciding with the crown of the latest landslides.Several nearby buildings were also damaged.Stabilization works were immediately executed limited to the uppermost portion of the slope, directly concerned by the incipient sliding.At the same time, a general plan of the works required to stabilize definitively the entire slope was outlined.This paper illustrates the basic geological, geomorphological, hydrogeological and geotechnical conditions of the slope.The baseline schematization was based on surveys pushed to the depth of 50-100 m.Also are illustrated the major features and causes of the landslide and so was assumed the mitigation hypothesis of movement.The geotechnical characterization is used for the design of tunnel and make same consideration for deep water drainage.Figure 1 show the study site.

Geomodel study
A geomodel study consists of five phases, namely: geology, geomorphology, hydrogeology, subsurface investigation, and geotechnical model.The first three phases explain geological units and groundwater conditions, however the investigation and the geotechnical model quantifies the parameters and provides data for geotechnical design.

Geological aspects
The Maiella Mountain is a large anticlinal fold formed by a succession of almost exclusively calcareous marine environment formations of Jurassic-Miocene age [2,3].The northern and northwestern sectors of this anticlinal structure are formed by extensive monoclinal dipping clods, with modest inclination, partly interrupted by faults.At the offshoots of one of these clods the town of Lettomanoppello is built.Behind the village outcrops the more recent rocks of the marine lithological succession (MCD) that are highly fractured and have generally high permeability.This limestone substrate is overlain by deposits of evaporitic environment from the Upper Miocene (MA) consisting of dark clay, layered and fractured marly shales with intercalated banks, layers, and lenses of gypsum (G).The thickness of this formation is quite variable because of the complex history of its deposition in the Lettomanoppello area.The evaporitic sediments rest on the limestone basement according to a morphologically complicated surface, and their thickness increases rapidly toward the west.The clayey-marly soils are strongly overconsolidated, often cemented, with high shear strength, low compressibility, and low permeability.The upper part of the deposit, only a few meters thick, is strongly altered and characterized by poor mechanical properties i.e. sand and silt between the clasts (DT).The thickness of the debris blanket varies between 20 and 50 m at the built-up area and between 10 and 15 m in the part immediately downstream.At the foot of the slope, the thickness decreases until it reaches the substrate.The area of the downstream slope at the Lavino stream is covered with reworked materials, ancient landslides (quiescent or active) that have involved the debris blanket or the clayey-marly formation (F).

Geomorphological aspects
The northwestern slopes of the Maiella slope down toward the valley with a slope of about 8-10°.At the height of the village the slope reduces to become, in some parts, sub-horizontal.In this area, the thick blanket of debris materials and over more limited areas, masses of limestone rock outcrop.Immediately downstream of the built-up area the mild slope is interrupted by an amphitheater, which extends for a length of 400 m right next to the built-up area (elevations 325-340 m a.s.l.), characterized by an escarpment averaging 30 m high and with a rather pronounced acclivity (30-45°).The escarpment represents the main edge of detachment of major landslides that also occurred recently and affected the underlying section of the slope up to the Lavino riverbed (i.e.DT).Downstream of this amphitheatre the slope is largely made up of the MA, in places outcropping, in places still covered by the thick debris flows.From the foot of the escarpment up to the Lavino River (elevation 190 meters above sea level), the slope is moderately steep (15° on average) but characterized by a whole series of small morphological features (bumps, undulations, depressions, counter-slope, and steep escarpments of varying heights) due to the existence of extensive and important landslide phenomena.

Hydrogeological aspects
The limestone rocks that form the deep bedrock in the area are home to an important aquifer whose supply basin extends considerably eastward to include part of the large limestone Maiella anticline.The following can be stated: (i) is certain the existence of an aquifer within the DT resting on the MA; (ii) a base water aquifer is contained in the MCD of the substrate (where perched water levels are also likely to exist.).The elevations of (ii) aquifer have a particular influence on the current landslide phenomena.

Subsurface investigations and geotechnical model
In the 1989 geotechnical investigation campaign, 5 continuous core borings were carried out along the tunnel route.Details are shown in the Table 1.Readings of water levels from piezometers, monitored up to January 1990, indicate that the free surface of the water table along the tunnel route is between 30.00 meters and 38.00 meters below ground level within the limestone debris.Since the tunnel is no less than 60 meters below ground level, the water head above the tunnel is 20 to 30 meters.
Table 2. Mechanical parameters for each material.However, an adequate number of samples, necessary to construct an adequate model, is not available, as in the case of treated studies.On the other hand, an extensive bibliography is also devoted to describing the range of variability of the most important parameters that characterize the materials that constitute soils.Therefore, synthetic data based on experimental values with the associated variability taken from the bibliography and given in Table 2 were considered [4,5].Figure 2 show a geological maps asset with the position of borehole and figures 3-4 shows the longitudinal section and the cross-section respectively.In particular, section B-B' was used to analyse the filtration process induced by the tunnel drainage.

Probable causes and hypotheses for stabilization intervention
The stability conditions of the slope are governed by the complex hydrogeological situation, in particular by the groundwater circulation sustained by the impermeable marly claystone deposit.The waters emerging at the ground surface and flowing down the hill cause softening of the clayey soils and contribute to the progress of the soil movements (the evolution are "multiple retrogressive slumps").The stabilization of the Lettomanoppello landslide is a challenging engineering task, due to the large dimensions of the landslide, the concurrence of various causes and the complex situation existing in the slope [6].Drainage works, effective in the medium-long term, aimed at reducing/controlling the major causes of the landslide (chaotic groundwater circulation, excessive porewater pressures in case of intense/long-lasting rainfall).
The deep drainage work in project is a tunnel consisting of catch the deep groundwater circulating in the (limestone and possibly also in the gypsum layers interbedded in the marly claystone deposit) in order to "cut" the water supply to the debris layer involved by the landslide directly in the rock substratum and also to drain the infiltration from shallow groundwater circulating from the debris layer.
We assume an excavation section with a diameter of about 4.50 m and a horseshoe-shaped path below the built-up area for a total length of about 1 Km, with entrances at an elevation of about 290 m.a.s.l.Figures 3 and 4 shows the section along the trace the and a perpendicular section placed about halfway along the tunnel.The tunnel will involve about 70 percent argillitic-chalky soils and, for the remainder, more or less fractured and limestone rocks, up to completely crushed.

Filtration problem and results
The filtration process induced by the excavation of the drainage tunnel was analyzed under the assumption of homogeneous isotropic and un-deformable medium where quiet aquifer is present.The analysis of the phenomenon was conducted with simplified models whose solutions are given in closed form.The results depend on the initial position of the free surface relative to the tunnel axis, the depth of the impermeable substrate, the permeability, and the porosity of the medium [7,8].Was considered the filtration process under varying motion conditions induced by the tunnel excavation due to the change of piezometric height.Therefore we want to estimate the filtering flow rate q(t) in a tunnel section of 1 m and the area affected by the lowering of the groundwater R(t).We have considered a rectilinear tunnel with a horizontal axis under construction in a ground water space, originally horizontal and located at a distance H from the cent of the tunnel.The soil is assimilated to a homogeneous, isotropic, non-deformable medium with porosity n and permeability k.The fluid follows Darcy's law.As a result of drainage, the free surface of the water table lowers, assuming variable positions over time.The following assumptions were made: (i) all-draining tunnel contour (ii) the tunnel was built instantaneously (no forward speed was considered); (iii) the surface filtering contribution e is neglected.The filtration problem is depicted in Figure 5 and follows the model of Polubarinova and Kochina [9]: where f is the potential of filtration velocity in space.

Figure 5. Filtration problem
For this model, the functions R(t) and q(t) are known for horizontal direction of motion and for a finite region of space [10].With reference to the diagram in Figure 3 and for the above assumptions, the variation of ground-water level over time is expressed by the following equation: where b0 is the position of tunnel respect x axis.The equation for R(t) and q(t) are: Therefore, the total drained flow rate Q(t) can be estimated as: where   () are the tunnel length during the time (for the assumption (ii) this term takes the value of 1000 m).The calculations were performed under a hypothesis of following parameters: H=50 m, b0 = 50 m, k = 10 -5 m/s and n = 0.30.The results were given in terms of the ground water profile as a function of time, and the filter flow rate was estimated for a generic tunnel section of length par to 1 m and it is showed in Figures 6 and 7.

Discussion and conclusion
The case study concerns a proposed intervention to mitigate the Lettomanoppello landslide through the construction of a drainage tunnel.A geotechnical reconstruction of the complex subsurface (based on boreholes) and hydraulic quantities such as permeability and porosity was carried out.An analytical method was adopted to predict groundwater subsidence and water inflows into the tunnel.The results show that the tunnel has an effect on groundwater and in particular area of subsidence extends down to the landslide slope.In fact, in the long term (t = 1000 days), Figure 6 shows a subsidence amplitude of about 800 m in the two half-spaces separated by the tunnel and a minimum total filter flow rate Q of 0.5 m 3 /s is therefore estimated showed in Figure 7.These estimates might be valid under the assumption of no aquifer recharge, but we know that the flow coming from the mountainous area of Majella (ESE direction in Figure 5) ensures continuous inflow.This effect can be understood as the cutting of the water supply to the debris layer involved by the landslide directly in the rock substratum.However, deep drainage should be accompanied by a regularisation of surface water circulation to cut and to limit the infiltration from shallow groundwater located in debris layer.The proposed methodology should be considered as a reference for more indepth calculations also carried out in order to study the consolidation processes whose consequences that may occur with considerable delay after the completion of the work.
The drainage waters will be conveyed to pipe towards a hydraulic turbine in the lower part of the slope to produce electric energy.The expected minimum power of the system can be estimated as  = 7 and thus approximately equal to 300 KW (considering a total piped flow of 0.5 m 3 /s and a hydraulic load of 80 m). Figure 8 show a framework of the hydroelectric system.This expedient would allow the economic recovery of the cost of the tunnel by transforming geotechnical stabilisation from a 'losing' work into a virtuous and profitable process [11].

Figure 1 .
Figure 1.Geographic framework of Lettomanoppello studied area.In the red line the presumed boundaries of landslides and the sections of study (A-A' and B-B') .

Figure 2 .
Figure 2. Geological maps asset with the position of borehole

Figure 7 .
Figure 7. Filtering flow rate in a tunnel section of 1 m over time.

Table 1 .
Surveys along the tunnel route.