1 INTRODUCTION

Understanding the mechanisms that control the fail-ure, post-failure and propagation stages of extremely rapid landslides of the flow-type is a fundamental is-sue for mitigating the high risk related to these phe-nomena. This is the case of an area of about 3,000 km2 of the Campania region where the hillslopes surrounding 212 municipal territories are covered by shallow deposits of pyroclastic soils recurrently af-fected by catastrophic landslides of the flow-type.

The huge related consequences, in terms of loss of life, are testified by a comprehensive catalogue of historical incident data concerning the landslide-caused fatalities recorded in the period 1640-2006. The last catastrophic event caused, on May 1998, 159 fatalities within four municipalities located at the toe of a carbonate massif (Pizzo d’Alvano), few kilometres far from the Vesuvius volcano.

After the May 1998 event, a National law (Piano Stralcio per l’Assetto Idrogeologico) was pro-claimed requiring the Authorities for zoning the ar-eas at landslide risk all over the Italian territory. This law is currently active and it regulates the land-use within the areas at landslide risk. These areas were recognised and zoned essentially through basic methods (Cascini 2002) that, according to Fell et al. (2008a), mainly furnish a preliminary level of risk zoning.

Considering the wide territory classified at high risk and the related constraints for the urban plan-

ning and development, more advanced risk zoning level are necessary at the present in order to better define the real extension of the areas at risk and to rank the risk inside these areas. However, the updat-ing of the current landslide risk zoning requires con-siderable technical efforts and the use of geotechni-cal models when an advanced level of zoning is pursued (Fell et al. 2008a). In such a case, under-standing the mechanisms controlling the different stages of the landslides of the flow type is the first necessary step for obtaining reliable zoning maps. However, the application of this approach to Cam-pania region is challenging due to both the large ar-eas affected by landslides, and the limited portions of the territory where geotechnical dataset is avail-able for landslide characterisation.

The paper is aimed to suggest an approach able to overcome these difficulties managing the dataset at different scales (i.e. small, medium and large scales, as defined by Fell et al. 2008a). Particularly, the histori-cal data concerning the fatal events occurred in the Campania region are firstly processed and, for the landslide risk, a ranking scale is defined, at small scale, to classify the geological settings affected by landsliding. Then, referring to the area with the highest landslide risk for life loss, the triggering factors and mechanisms are recognized at medium scale. Finally, focusing on the May 1998 event, which occurred in-side the most at risk area, the different triggering mechanisms are analyzed at large scale with the aid of sophisticated engineering models.

Detection of mechanisms for destructive landslides in Campania region – southern Italy

L. Cascini, S. Cuomo, S. Ferlisi & G. Sorbino Department of Civil Engineering, University of Salerno, Italy

ABSTRACT: The detection of the mechanisms that control the landslides of the flow-type is a fundamental issue in the risk analysis. This task is challenging when large areas are affected by landslides and related geo-technical datasets only refer to small portions of the analysed area. In such a case, a methodology based on a multi-scale approach can be profitably used. The present paper deals with the application of the above meth-odology to an area of the Campania region in southern Italy which is among the most at risk areas in Europe. Particularly, at small scale, preliminary insights on landslides mechanisms are obtained through the F-N curves. Then, referring to the medium scale, evaluations on the predisposing factors and triggering mecha-nisms are carried out through basic geological, geotechnical and hydrological analyses. Finally, the role played by the local geological and geotechnical features for the onset of different triggering mechanisms is assessed, at large scale, by using standard and advanced geotechnical analyses.

2 SMALL SCALE ANALYSIS The territory of the Campania region is 13,590 km2 large (Fig. 1a) and it includes 551 municipalities.

Owing to the complexity of its geological-structural setting, the region is prone to different types of landslides. However, those associated to the highest consequences, in terms of human loss, are concentrated in the area where the hillslopes are covered by pyroclastic soil deposits. In this area, these deposits mantle three main geo-environmental contexts (Fig. 1b) which differ for their geological, hydrogeological and geomorphological features, as well as for age and thickness of the pyroclastic de-posits.

The first main context (A) coincides with large sectors of the Campanian Apennine chain formed by Mesozoic carbonate rocks overlaid by thin layers (0.1 to 5.0 m) of pyroclastic soils. This context, in turn, can be distinguished into two different sub-contexts (A1 and A2) (Fig. 1b). In the first one (A1), pyroclastic soils were mainly originated from the Somma-Vesuvius volcanic activity and cover the hillslopes either as primary air-fall or re-worked de-posits (Rolandi 1997). The second one (A2), located in the northern part of the region, includes pyroclas-tic soils essentially derived from the Roccamonfina volcanic activity.

The second main context (B) corresponds to the Phlegraean district, including the city of Naples as well as Ischia and Procida islands. This context pre-sents a bedrock formed by volcanic tuffs and lavas of Late Pleistocene age overlaid by pyroclastic soil deposits having thicknesses up to several meters (Calcaterra et al. 2004).

Finally, the third main context (C), essentially lo-cated in the north-west of Sarno and Picentini Mounts, includes thin pyroclastic soil deposits (< 2.0 m) mantling flysch and terrigenous bedrock.

For these three geo-environmental contexts, the availability of a comprehensive catalogue of incident data allows the analysis of both temporal and spatial distribution of fatal landslides occurred in the past (Cascini et al. 2008d).

Documentary sources essentially consist in liter-ary works, documents housed in public and private archives, historical books, etc. (Operative Unit 2.38 1998; Cascini et al. 2002; Cascini & Ferlisi 2003). The collected data were systematically integrated by Cascini et al. (2008d) with those of the Italian AVI Special Project archive (Guzzetti 2000, Guzzetti et al. 2005) which represents the 30% of the whole dataset. The latter includes 311 fatal landslide inci-dents occurred in the whole territory of the Cam-pania region; however, only the events (293) for which the number of related fatalities is exactly quantified have been taken into account for analysis purposes. These events refer to a time period span-ning from 1640 up to 2006.

Figure 1. a) The territory of the Campania region (southern It-aly); b) map of the areas in Campania region where pyroclastic soils cover different bedrocks (modified from Cascini et al. 2005).

Focusing on the spatial distribution of fatal land-slides within the defined geo-environmental con-texts, the dataset shows that the 34.1% of the events (100 out of 293) affected the context A; among these, 93 occurred in A1 and the remaining 7 in A2. As for the geo-environmental contexts B and C, 166 and 3 events were respectively recovered.

Referring to the temporal distribution of fatal landslides during the hydrological year, the dataset highlights that the events are essentially concen-trated at the beginning and during the rainy season (Cascini et al. 2008d).

As it concerns the consequences – in terms of life loss – associated to landslides of given frequency, historical data allow the estimation of the societal risk (Fell et al. 2008a). At small scale, this goal can be pursued by correlating, on a log-log scale, the an-nual frequency F of landslides causing N or more fa-talities versus the number N of fatalities (F-N curve). Figure 2 compares the F-N curve obtained for Cam-pania region with those derived for Italy and other

Figure 2. F-N curves related to fatal landslides in the territories of Campania region, Italy and other Countries (from Cascini et al. 2008d). Countries, such as Canada, China, Colombia, Hong Kong, Japan, Nepal and Norway (Evans 1997, Guzzetti 2000, Düzgün & Lacasse 2005).

It is worth noting that the F-N curve of the Cam-pania region is very close to that of the Italian terri-tory – this last obtained on the basis of an incident database spanning from 1410 to 1999 (Guzzetti 2000) – thus highlighting that Campania is one of the Italian regions with the highest societal risk ow-ing to landslides. Moreover, the Campania curve is located below the F-N curves of Japan and China only, while the data points corresponding to the highest numbers of fatalities (N ≥ 40) are well fitted by the F-N curve referring to Hong Kong. This re-sult could suggest a similar level of societal risk ow-ing to landslides among the Campania and Hong Kong territories (Cascini et al. 2008d).

Furthermore, the F-N curve of the Campania region shows two rollover points in correspondence of the highest and the lowest number of fatalities (Fig. 3). The first rollover point could be a consequence of different predominant triggering factors affecting events causing no more than 3 fatalities (which are essentially triggered by anthropogenic factors) and events causing 4 or more fatalities (wich are induced by rainfall). On the other hand, the second rollover point implies few events of high intensity exceeding a certain threshold in terms of caused fatalities (N = 247 in such a case) and can be probably related to

1.0E-04

1.0E-03

1.0E-02

1.0E-01

1.0E+00

1 10 100 1000N

F

Campania RegionContext A1Context A2Context BContext C

rollover

rollover

Figure 3. F-N curves related to fatal landslides in the Campania region and in different geo-environmental contexts (modified from Cascini et al. 2008d).

both rainfalls and other factors not well documented at the present; this second rollover point concides with the one showed by the F-N curve of the context A1.

Further interesting considerations arise by com-paring the F-N data points for fatal landslides occur-ring in the three geo-environmental contexts covered by pyroclastic soils (Fig. 1b). In particular, Figure 3 shows that context C is characterized by fatal land-slide events with the lowest values of both annual frequency and consequences.

The F-N plot of the territory where pyroclastic soils rest on volcanic tuff and lava deposits (context B) is characterized by a relative high slope due to the great number of incident data (166 out of 293) which refer to landslide events resulting in no more than 25 fatalities. Similarly, a high slope also char-acterizes the F-N plot of context A2 which indicates few data of fatal landslides, associated to no more than 3 fatalities, even though the annual frequency of events is very low.

Finally, the fatal landslides which systematically involve the territories in context A1 are character-ized by an annual frequency lower than that of the fatal landslides in context B only by N ≤ 2; on the contrary, they show a higher annual frequency for N ≥ 3. Moreover, the (F, N) data points practically overlap those of the entire Campania Region for N ≥ 40. The latter, in turn, correspond to events associ-ated to the occurrence of landslides of the flow-type, as highlighted by the analysis of the incident data (Cascini et al. 2008d). This result testifies that peo-ple living in context A1 are the most exposed to the landslide risk.

In conclusion, the selection of the historical inci-dent data, on the basis of the three homogeneous geo-environmental contexts to which they refer, al-lows the differentiation of the levels of the landslide risk to which people living in Campania region is exposed. At small scale, being the soils involved in fatal landslides of analogous origin, these different levels of risk also outline the existence of different mechanisms controlling the failure and propagation stages of landslides of the flow-type, as discussed in the following sections.

3 ENVIRONMENT AND MECHANISMS The previous section highlights that the most catas-trophic fatal landslides systematically affect the con-text A1 (≈ 1,400 km2). However, among the re-corded events, significant differences can be observed referring to their annual frequency and consequences. This is outlined by comparing the F-N curve of the context A1 with those referring to two different sub-contexts (Fig. 4) for which the highest numbers of landslide-caused fatalities were recorded in the past.

km

Figure 4. Map of the sub-contexts individuated within the geo-environmental context A1.

The first refers to an area (≈ 130 km2) including territories of the Amalfi Coast, Salerno and Cava de’ Tirreni; the second one refers to the Pizzo d’Alvano massif (≈ 60 km2).

In particular, Figure 5 shows that the F-N curves differ in slope and annual frequency. This is due to the existence of different mechanisms controlling the landslides’ stage of movements within the two areas located in the same geo-environmental context A1. This statement seems to be confirmed focusing on the events characterized by major consequences and, particularly, on the season of their occurrence. To this aim, Figure 6 provides the monthly distribution of landslide of the flow-type and fatalities recorded in both the considered sub-contexts. It can be noted that: i) in the territory of Amalfi Coast-Salerno-Cava de’ Tirreni the fatal events mainly occur in Autumn (i.e. October) and in Winter (i.e. January) while the caused fatalities essentially concentrate in Autumn months of October and November; ii) in the Pizzo d’Alvano massif, on the contrary, the four recorded fatal events are distributed during the hydrological year, but fatalities concentrate in May (i.e. at the end of the rainy season).

Figure 5. F-N curves for landslides of the flow-type occurred, within the context A1 as well as in the two sub-contexts de-picted in Figure 4 (from Cascini et al. 2008d).

0

1

2

3

4

5

JUL AUG SEPT OCT NOV DEC JAN FEB MAR APR MAY JUN

Num

ber

of e

vent

s

Amalfi Coast - Salerno - Cava de' Tirreni

Pizzo d'Alvano massif

050

100150200250300350400450500550

JUL AUG SEPT OCT NOV DEC JAN FEB MAR APR MAY JUN

Num

ber

of fa

talit

ies

Amalfi Coast - Salerno - Cava de' Tirreni

Pizzo d'Alvano massif

Figure 6. Monthly distributions of fatal landslides of the flow-type occurred within two sub-contexts of A1 (a) and of related fatalities (b).

The differences outlined in Figures 5 and 6 can

be preliminarily explained referring to a medium scale (1:25,000) as defined by Fell et al. (2008a). In fact, this scale allows to consider some factors which in literature are indicated as relevant for this type of landslides such as: i) shape of the mountain basins (Pareschi et al. 2000), ii) slope morphology (Guadagno et al. 2005), iii) stratigraphy of pyroclas-tic soil deposits (Cascini et al. 2005), iv) rainfall pat-terns (Fiorillo & Wilson 2004), v) hydrogeological factors (Cascini et al. 2005) and vi) anthropogenic factors (Guadagno et al. 2005). Insights on the role played by these factors in controlling the landslides’ mechanisms can be provided referring to the most recent catastrophic events dated on 1954 and 1998.

The 25th October 1954 event affected several mu-nicipalities along the Amalfi Coast as well as the town of Salerno (Esposito et al. 2003) causing 325 fatalities and huge damages (Cascini et al. 2008d). During this event, both landslides of the flow-type (Hungr et al. 2001) and hyper-concentrated flows (Coussot & Meunier 1996) occurred as a conse-quence of the above mentioned factors. Particularly, this event was triggered by an extraordinary rainfall pattern estimated as equal to 504 mm in 8 hours (Tranfaglia & Braca 2004) and it affected an area characterized by funnel shaped mountain basins with a high order of drainage networks (Fig. 7). There-fore, during the rainstorm, superficial run-off and soil erosion played a major role for the attainment of slope instability conditions.

Completely different features can be recognized for the landslides occurred on 5th May 1998 inside the Pizzo d’Alvano massif which caused 159 fatali-ties within the four municipalities located at the toe of the massif.

a)

b)

0 10 km

0 1 km

Figure 7. An example of a mountain basin affected by slope in-stability phenomena during the October 1954 event (Reginna Maior, Lattari Mounts).

0 2 4 M l

0 1 km

0 10 km

0 0.1 km

Figure 8. An example of a mountain basin affected by slope in-stability phenomena during the May 1998 event (Tuostolo ba-sin, Pizzo d’Alvano massif ).

In this case, landslides of the flow-type were trig-

gered in mountain basins characterized by elongated shapes and low order of drainage networks (Fig. 8). Furthermore, slope failure conditions were attained due to an intense but not extraordinary rainfall pat-tern evaluated as not exceeding 240 mm in 48 hours (Rossi & Chirico 1998, Cascini et al. 2005). These evidences testify that in addition to rainfall other factors were relevant for the failure onset. Particu-larly, for these landslides a major role was played by: i) the cumulated rainfall infiltration inside the pyroclastic deposits, ii) hydrogeological factors such as springs from bedrock (Cascini et al., 2005), iii) slope morphology factors such as bedrock scarps, and iv) anthropogenic factors such as trackways and slope cuts (Guadagno et al. 2005).

Due to the variety of triggering factors these landslides were surely triggered by different mecha-nisms (Cascini et al. 2008b) whose detection can be pursued at large scale.

4 LOCAL CONDITION AND MECHANISMS The role played by local conditions on triggering mechanisms can be well depicted referring to the last catastrophic event occurred on May 1998 along

the Pizzo d’Alvano slopes. Descriptions of both the landslides and consequences are available in the sci-entific literature (Cascini 2004, Guadagno et al. 2005, Fiorillo & Wilson 2004) as well as contribu-tions dealing with the landslides failure (Cascini et al. 2005, 2008b) and propagation (Pastor et al. 2008) stages.

Particularly, Cascini et al. (2005, 2008b) analyse the main geological and geomorphological features of both the bedrock and pyroclastic deposits and provide a framework for landforms and processes involving the slopes. Thanks to this framework, six different triggering mechanisms (Fig. 9) have been detected on the basis of the predisposing and trigger-ing factors as well as for the corresponding landslide source areas.

As Figure 9 shows, the triggering mechanism named M1 occurred inside colluvial hollows associ-ated to zero order basins affected by convergent sub-superficial groundwater circulation inside the pyro-clastic deposits and by temporary springs from the bedrock. The triggering mechanism M2 occurred in-side triangular shaped areas with an upper crown only a few meters wide, downslope enlarging. The land-slides source areas were located in the upper

1 2 3 4

M4 M5 M6

M1 M2 M3

M1 M2 M3

M4 M5 M6

Figure 9. Triggering mechanisms (M1-M6) for the May 1998 landslides of the flow-type: 1) bedrock, 2) pyroclastic deposit, 3) track, 4) spring from bedrock (modified from Cascini et al. 2008b).

open slopes and associated to outcropping or buried bedrock scarps. The triggering factor was the impact caused by small landslides occurred at the top of bedrock scarps or springs from karst conduits. The triggering mechanism M3 showed compound plan-forms that, based on in-situ evidence, can be related to overlapping and/or laterally enlarging local slope instabilities strictly connected to man-made tracks. The mechanism M4 was related to heavy superficial water supplies at the head of valleys and the land-slide source areas were shaped as grapes and they involved limited thicknesses of pyroclastic deposits with a “V” shaped transversal profiles. The mecha-nism M5 occurred along open slopes with a convex longitudinal profile where, due to natural or anthro-pogenic factors, deep gullies are present inside the pyroclastic deposits. Finally, the mechanism M6 was related to the existence of slope angle breaks (natu-ral or produced by anthropogenic activities) and they generally involved limited volumes within the pyro-clastic deposits.

In order to analyse at both massif and slope scales the above mechanisms, in-situ and laboratory inves-tigations were systematically carried out by several Authors. The in-situ investigations (i.e. pits, seismic prospections and soil suction measurements col-lected over the period 1998 – 2004) provided de-tailed stratigraphy of the pyroclastic deposits at slope and massif scales, as well as soil suction re-gime (Fig. 10) during the hydrological year all over the Pizzo d’Alvano massif, inside and near the land-

01020304050607080

daily

rai

nfal

l (m

m)

a)

05

101520253035404550

suct

ion

(kPa

)

from 0.2 to 0.4m from 0.4 to 0.6mfrom 0.6 to 0.8m from 0.8 to 1.0m

05

101520253035404550

Mar

-00

May

-00

Jul-0

0

Sep-

00

Nov

-00

Jan-

01

Mar

-01

May

-01

Jul-0

1

Sep-

01

Nov

-01

Jan-

02

Mar

-02

May

-02

suct

ion

(kPa

)

from 1.0 to 1.2m from 1.2 to 1.4mfrom 1.4 to 1.6m from 1.6 to 1.8md)

c)

Figure 10. a) daily rainfall; b) monthly average suction at depths lower than 1.0 m; c) monthly average suction at depths ranging from 1.0 m to 1.8 m (modified from Sorbino 2005).

0102030405060708090

100

0 10 20 30 40 50 60 70 80 90 100effective vertical stress σ'v (kPa)

Shea

r stre

ngth

τf (

kPa)

12789Class "A"

Class "B"

a)

3,4,5

9) 2a φ' = 41°8) 2b φ' = 40°7) 1c φ' = 37.5°6) 1d φ' = 36°5) 1f φ' = 35°4) 1e φ' = 35°3) 2c φ' = 35°2) 1e φ' = 34°1) 3a φ' = 30°

6

100Effective vertical stress σ’v (kPa)

Shea

rst

reng

thτ f

(kPa

)

0.001 0.01 0.1 1 10 100Grain size (mm)

100

80

60

40

20

0Perc

enta

geby

wei

ght(

%)

a)

d)

Figure 11. Geotechnical properties of pyroclastic soils: a) grain size distributions, b) saturated shear strength envelopes, c-d) peak shear strength of ashy A and B soils versus net vertical stress for some range of saturation degree (Srf) (from Bilotta et al. 2005).

slide source areas (Cascini & Sorbino 2002, 2004, Cascini 2004, Sorbino 2005).

As it concerns the laboratory investigations, they provided the physical and mechanical properties of the pyroclastic soils in both saturated and unsatu-rated conditions (Sorbino & Foresta 2002, Bilotta et al. 2005, 2006, 2008) (Fig. 11). Based on the avail-able dataset, Sorbino et al. (2007) dealt with the modelling of the triggering areas at massif scale by using sophisticated models. Particularly, they used three different physically based models developed in a GIS framework: SHALSTAB (Montgomery & Dietrich 1994), TRIGRS (Baum et al. 2002) and TRIGRS-unsaturated (Savage et al. 2004).

Fig. 12a shows the results obtained by the three models in terms of two percentage indexes (Succes Index – SI and Error Index – EI) that quantify the “Success” and the “Error” of each model in simulat-

a)

b)

c)

a)

b)

c)

d)

Figure 12. Results obtained for the entire area of Pizzo d’Alvano massif with a) the SHALSTAB, TRIGRS and TRIGRS-unsaturated codes and b) with the TRIGRS codes for the different triggering mechanisms (from Sorbino et al. 2007).

ing the observed source areas of the May 1998 land-slides all over the Pizzo d’Alvano massif. The obtained results highlight that the TRIGRS transient models pro-vide, for the same Success Index values, Error Index values lower than those obtained by the steady-state SHALSTAB model that furnishes a systematic overes-timation of the observed source areas. Moreover, the TRIGRS-unsaturated is more suitable than TRIGRS due to its capability to model transient infiltration proc-ess in the partially saturated pyroclastic cover.

Referring to the most reliable models (TRIGRS), Figure 12b shows the SI and EI indexes obtained for the mechanisms M1-M6. As it can be seen, TRIGRS models do not provide the same SI and EI values for the source areas affected by the above mechanisms. These results do confirm that the mechanisms M1-M6 are different because they are associated to different boundary conditions, strictly related to the local geo-logical and geotechnical features of the landslides’ source areas. Furthermore, the highest values of SI are systematically provided for the mechanism M4, which is strongly affected by the convergence of fluxes and the slope angle both properly taken into account by TRIGRS models. On the contrary, lower SI values are obtained for the remaining triggering mechanisms, which are mainly associated to local boundary conditions not considered by the models.

The role played by the local geological and geo-technical features on the different triggering mecha-nisms can be clearly evidenced by either classical or advanced geotechnical analyses.

Scheme 8

Scheme 2Scheme 1

Class B ashy soils

Pumice soils

Class A ashy soils

Outlet

H

α α

α

H = 4.5 m, α = 35°

H

H

Computed slip surfaces0 5m 10m

Figure 13. Mobilised volumes for different stratigraphy (modi-fied from Cascini et al. 2005).

For instance, referring to the mechanism M1, limit equilibrium method and FEM coupled and un-coupled stress-strain analyses (Cascini et al. 2003, 2005, Cuomo 2006a, 2006b) highlight that rainfall infiltration and springs from bedrock are key factors for the onset of this mechanism both at slope and massif scales (Fig. 13). Analyses also show that the local stratigraphical conditions of the pyroclastic de-posits are responsible for the amount of the mobi-lised volumes (Fig. 13).

As it concerns the mechanism M2 (Fig. 14), the performed analyses (Cuomo 2006; Cascini et al. 2008a) show that the most important factors are rep-resented by karst springs at the base of bedrock

uw>0

uw<0

uw>0

uw<0

uw<0

scheme 1 scheme 2

01020

300

1020

30

0102030

scheme 3

ashy Bpumiceashy Aslip surface

pore waterpressures (kPa)

M2a

M2b

Initial conditionsF=1.322

Load applicationF=1.531

Load removalF=0.989

a)

ashy Bslip surface

a)

load

b)

Figure 14. Failure stage induced by a) karst springs from bed-rock, b) impact phenomena in undrained conditions (modified from Cascini et al. 2008a).

a)

b)

Spring from bedrock

a) scheme 1FS=1.10

M3ab) scheme 2FS=0.89

ashy Bashy A

track track

pondingponding

Figure 15. Limit equilibrium analyses for zones close to bends of tracks (from Cascini et al. 2008c). scarps and/or the impact in undrained conditions of failed soil masses on stable deposits. Also for this mechanism, the amount of the mobilized volumes is strongly affected by the stratigraphical conditions.

Finally, referring to the mechanism M3, geotech-nical analyses (Cascini et al. 2008c) stress the fun-damental role played by some anthropogenic factors such as mountain tracks and cut slopes (Fig. 15). Mountain tracks are responsible for conveying large amount of superficial rainwater in singular points along the slopes, so inducing severe hydraulic pond-ing conditions at the ground surface. Cut slopes act as an important aggravating factor for slope stability conditions by inducing geometrical discontinuities in the slope morphology.

5 CONCLUDING REMARKS The risk analysis for landslides of the flow-type is a difficult task especially when they affect large areas. In this case, the detection and understanding of the landslide mechanisms – which is a fundamental step for landslide characterisation in risk analysis – can be adequately pursued adopting a methodology based on a multi-scale approach. As contribution to this topic, the paper deals with an area of southern Italy which is among the most at risk areas in Europe.

For this area, it is firstly outlined that preliminary insights can be obtained at small scale through the F-N curves, which allow distinguishing different portions of the area where landslides are character-ised by different intensity and mechanisms.

At the medium scale, the increasing of knowledge is obtained by using geological, geotechnical and hydrological approaches that provide useful insights on the predisposing factors and, consequently, on the detection of different triggering mechanisms.

Finally, the role played by the local geological and geotechnical features on the development of dif-ferent triggering mechanisms is properly assessed at large scale by using classical or advanced geotech-nical analyses.

Independently from the scale of analysis, the pa-per shows that knowledge improvement and updat-ing of landslide risk zoning, from a preliminary to an advanced level, can be pursued only thanks to the availability of a comprehensive dataset for the area of concern. To this aim, politicians and technicians should devote significant efforts to improve both the gathering of these data and their analysis according to proper procedures (Fell et al. 2008a, 2008b).

6 REFERENCES Baum, R.L., Savage, W.Z. & Godt, J.W. 2002. TRIGRS - a

FORTRAN program for transient rainfall infiltration and grid-based regional slope-stability analysis. U. S. Geologi-cal Survey Open-File Report 02-0424.

Bilotta, E., Cascini, L., Foresta, V. & Sorbino, G. 2005. Geo-technical characterization of pyroclastic soils involved in huge flowslides. Geotechnical and Geological Engineer-ing, 23: 365-402.

Bilotta, E., Foresta, V. & Migliaro, G. 2006. Suction controlled laboratory tests on undisturbed pyroclastic soil: stiffnesses and volumetric deformations. Proc. International Confer-ence on Unsaturated Soils, 2-6 April, Carefree, Arizona (USA), 1, 849-860.

Bilotta, E., Foresta, V. & Migliaro, G. 2008. The influence of suction on stiffness, shear strength, viscosity and collapse of some volcanic ashy soils. Proc. European Conference on Unsaturated Soils, 2-4 July, Durham (UK), 349-354.

Calcaterra, D., de Riso, R., Evangelista, A., Nicotera, M.V., Santo, A. & Scotto Di Santolo, A., 2004. Slope instabilities in the pyroclastic deposits of the Carbonate Apennine and the phlegrean district (Campania, Italy). In: L. Picarelli, ed. Proc. of the International Workshop on “Occurrence and Mechanisms of Flow-like Landslides in Natural Slopes and Earthfills”, 14-16 May 2003, Sorrento, Italy. Bologna: Pà-tron Editore, 61-75.

Cascini L. 2002. Il rischio da frana in aree urbane dell’Appennino Centro-meridionale. Proc. of the XXI Ital-ian Geotechnical Congress. L’Aquila, 11-13 September 2002. Bologna: Patron Editore, 127 -134.

Cascini L., Cuomo S., Pastor, M. & Fernández-Merodo J.A. 2008a. Geomechanical modelling of triggering mechanisms for rainfall-induced triangular shallow landslides of the flow-type. Proc. of the iEMSs Fourth Biennial Meeting: In-ternational Congress on Environmental Modelling and Software (iEMSs 2008). 7-10 July 2008, Barcelona, Spain, 1516-1523.

Cascini, L. & Ferlisi, S., 2003. Occurrence and consequences of flowslides: a case study. In: L. Picarelli (ed.) Proc. of the International Conference on “Fast Slope Movements – Prediction and Prevention for Risk Mitigation”, 11-13 May 2003 Napoli, Italy. Bologna: Pàtron Editore, Vol. I, 85-92.

Cascini, L. & Sorbino, G. 2002. Soil suction measurement over large areas: a case study. Proc. 3rd International Confer-ence on Unsaturated Soils, Recife (Brasil). Rotterdam: Balkema, vol. 2, 829-834.

Cascini, L. & Sorbino, G. 2004. The contribution of soil suc-tion measurements to the analysis of flowslide triggering.” Invited Lecture, In: L. Picarelli, ed. Proc. of the Interna-tional Workshop on “Occurrence and Mechanisms of Flow-like Landslides in Natural Slopes and Earthfills”, 14-16 May 2003, Sorrento, Italy. Bologna: Pàtron Editore, 77-86.

Cascini, L. 2004. The flowslides of May 1998 in the Campania region, Italy: the scientific emergency management. Italian Geotechnical Journal, 2: 11-44.

Cascini, L., Cuomo, S. & Guida, D. 2008b. Typical source ar-eas of May 1998 flow-like mass movements in the Cam-pania region, Southern Italy. Engineering Geology, 96: 107-125.

Cascini, L., Cuomo, S. & Pastor, M. 2008c. The role played by mountain tracks on rainfall-induced shallow landslides: a case study. Proc. of the iEMSs Fourth Biennial Meeting: International Congress on Environmental Modelling and Software (iEMSs 2008). 7-10 July 2008, Barcelona, Spain, 1484- 1491.

Cascini, L., Cuomo, S. & Sorbino, G. 2005. Flow-like mass movements in pyroclastic soils: remarks on the modelling of triggering mechanisms. Italian Geotechnical Journal 4: 11-31.

Cascini, L., Ferlisi S. & Vitolo, E., 2008d. Individual and so-cietal risk owing to landslides in the Campania region (southern Italy). Georisk, 2(3): 125-140.

Cascini, L., Ferlisi, S. & Tagliafierro, G., 2002. Il contributo delle indagini storiche nella definizione del rischio da fra-na: un caso di studio. Proc. of the XXI Italian Geotechnical Congress. 11-14 September 2002 L’Aquila, Italy. Bologna: Pàtron Editore, 135-142.

Cascini, L., Sorbino, G. & Cuomo, S. 2003. Modelling of flowslides triggering in pyroclastic soils. Proc. Int. Confer-ence on “Fast Slope Movements “Prediction and Preven-tion for Risk Mitigation”, Napoli, Patron Ed. 1, 93-100.

Coussot, P. & Meunier, M. 1996. Recognition, classification and mechanical description of debris flows. Earth-Science Review, 40, 209-227.

Cuomo, S. 2006a. Geomechanical modelling of triggering mechanisms for flow-like mass movements in pyroclastic soils. PhD Thesis, University of Salerno, Italy, 274 pp.

Cuomo, S. 2006b. Geomechanical modelling of triggering mechanisms for flow-like mass movements in pyroclastic soils. In: Vlasta Szavits-Nossan (ed.), Proc. of XVII Euro-pean Young Geotechnical Engineers’ Conference, 20-22 Luglio 2006 Zagreb.

Düzgün, H.S.B. & Lacasse, S., 2005. Vulnerability and accept-able risk in integrated risk assessment framework. In: O. Hungr, R. Fell, R. Couture, E. Eberhardt (eds.). Proc. of the International Conference on “Landslide Risk Manage-ment”. Vancouver, Canada 31 May-3 June 2005. London: Taylor and Francis, 505-515.

Esposito, E., Porfido, S. & Violante, C. 2003. Reconstruction and recurrence of flood-induced geological effects: the Vietri sul Mare case history (Amalfi Coast, Southern Italy). In: L. Picarelli (ed.) Proc. of the International Conference on “Fast Slope Movements – Prediction and Prevention for Risk Mitigation”, 11-13 May 2003 Napoli, Italy. Bologna: Pàtron Editore, Vol. 1, 169-172.

Evans, S.G., 1997. Fatal landslides and landslide risk in Can-ada. In: D.M. Cruden, R. Fell (eds.). Landslide risk assess-ment. Rotterdam: Balkema, 185-196.

Fell R., Corominas J., Bonnard Ch., Cascini L., Leroi E. & Savage W.Z. on behalf of the JTC-1 Joint Technical Com-mittee on Landslides and Engineered Slopes. 2008a. Guide-lines for landslide susceptibility, hazard and risk zoning for land use planning. Engineering Geology, 102: 85-98. doi:10.1016/j.enggeo.2008.03.022

Fell R., Corominas J., Bonnard Ch., Cascini L., Leroi E. & Savage W.Z. on behalf of the JTC-1 Joint Technical Com-mittee on Landslides and Engineered Slopes. 2008b. Guidelines for landslide susceptibility, hazard and risk zon-ing for land-use planning. Commentary. Engineering Geol-ogy, 102: 99-111. doi:10.1016/j.enggeo.2008.03.014

Fiorillo, F. & Wilson, R.C. 2004. Rainfall induced debris flows in pyroclastic deposits, Campania (Southern Italy). Engi-neering Geology (75): 263-289.

Guadagno, F.M. & Revellino, P. 2005. Debris avalanches and debris flows of the Campania Region (southern Italy). In Jakob M. and Hungr O. (eds.) “Debris-flow Hazard and Related Phenomena”, Springer, 489-518.

Guadagno, F.M., Forte, R., Revellino, P., Fiorillo, F. & Fo-careta M. 2005. Some aspects of the initiation of debris avalanches in the Campania Region: the role of morpho-logical slope discontinuities and the development of failure. Geomorphology, 66: 237-254.

Guzzetti, F., 2000. Landslide fatalities and the evaluation of landslide risk in Italy. Engineering Geology, 58: 89-107.

Guzzetti, F., Stark, C.P. & Salvati, P., 2005. Evaluation of flood and landslide risk to the population in Italy. Environ-mental Management, 36(1): 15-36.

Hungr, O., Evans, S.G., Bovis, M.J. & Hutchinson J.N. 2001. A review of the classification of landslides of the flow type. Environ. & Eng. Geosci., VII (3): 221-238.

Montgomery, D.R. & Dietrich, W.E. 1994. A physically based model for the topographic control on shallow landsliding. Water Resources Research, 30, 1153-1171.

Operative Unit 2.38. 1998,. Ricerca storica sulle colate di fan-go in terreni piroclastici della Campania. G.N.D.C.I.-C.N.R. Report. University of Salerno, Italy.

Pareschi, M.T., Favalli, M., Giannini, F., Sulpizio, R., Zanchet-ta, G. & Santacroce, R. 2000. May 5, 1998, debris flow in circum-Vesuvian areas (southern Italy): Insights for hazard assessment. Geology, 28(7): 639-642.

Pastor, M., Haddad, B., Sorbino, G., Cuomo, S., Drempetic V. 2008. A depth-integrated, coupled SPH model for flow-like landslides and related phenomena. International Journal for Numerical and Analytical Methods in Geomechanics, 33(2): 143-172.

Rolandi, G., 1997. The eruptive history of Somma-Vesuvius. In: Cortini M., De Vivo B., eds. Volcanism and Archae-ology in Mediterranean Area, Research Signpost, Trivan-drum, India. 77-78.

Rossi, F. & Chirico, G.B. 1998. Definizione delle soglie plu-viometriche d’allarme. G.N.D.C.I. – Dept. of Civ. Eng., University of Salerno, Italy, (in Italian).

Savage, W.Z., Godt, J.W. & Baum, R.L. 2004. Modeling time-dependent areal slope stability. In: Lacerda, W.A., Erlich, M., Fontoura, S.A.B., Sayao, A.S.F., (eds.). Landslides – Evaluation and Stabilization. Proceedings of the 9th Inter-national Symposium on Landslides. Rotterdam: Balkema, 1, 23-36.

Sorbino, G. & Foresta, V. 2002. Unsaturated hydraulic charac-teristics of pyroclastic soils. Proc. 3rd International Con-ference on Unsaturated Soils, Recife (Brasil). Rotterdam: Balkema, 1, 405-410.

Sorbino, G. 2005. Numerical modelling of soil suction meas-urements in pyroclastic soils. Int. Symp. “Advanced ex-perimental unsaturated soil mechanics”, Tarantino, Ro-mero & Cui (eds), Taylor & Francis Group, London, 541-547.

Sorbino, G., Sica, C., Cascini, L. & Cuomo, S. 2007. On the forecasting of flowslides triggering areas using physically based models. In: V.R. Schuster, R.L. Schuster, A.K. Turner (eds.). Proc. of the 1st North American Landslide Conference (Vail, Colorado). AEG Publication n. 23. ISBN 978-0-975-4295-3-2.

Tranfaglia, G. & Braca, G. 2004. Analisi idrologica e meteoro-logica dell’evento alluvionale del 25-25 Ottobre 1954: con-fronto con le serie storiche e valutazione del periodo di ri-torno di eventi analoghi. In: Il Nubifragio dell’Ottobre 1954 a Vietri sul mare. Costa d’Amalfi, Salerno. Pubbli-cazione GNDCI n. 2870: 295-348 (in Italian).