2''" Pali-Arri. Synrp. I,rrrr(lsli~les."'I COilKAIi, Kio (le Jrrrieiro, 1397

Rainfall and Senerchia Landslides, Southern

Precipitaciones y Deslizarnientos en Senerchia, Sur de Italia

M. Polemio Istituto di Geologia Applicata e Geotecnia. Facoltà di Ingegneria, Politecnico di Bari, Italy

ABSTRACT: The aim of this'work is to highlight the influente exerted by meteoric events on landslide triggering. Taking into account the main geomorphological and hydrogeological features af the mass movements, simple hydrologicai/statistical methods are suggested. Cumulative daily rainfalls over 1 to 180 days are studied to determine probability distribution functions. Hence the critica1 rainfall period is deterrnined for a specific landslide and the return period of the hydrological event associated with the landslide is quantified.

RESUMEN: E1 objetivo de este trabajo es poner en evidencia la influencia ejercida por 10s episodios meteoncos en e1 desencadenar la aparicion de deslizamientos. Teniendo en cuenta 10s grincipales rasgos geomorfologicos e hidrogeologicos de 10s movimientos de masas, se pueden sugerir métodos hidrogelogico-estadisticos simples. Las precipitaciones diarias acumuladas, que comprenden periodos de 1 a 180 dias, serb estudiadas con objeto de determinar las funciones de iiistribuci8n probabilistica. Asi, e1 periodo de precipitacion critico puede ser determinado para un de specifico caso de deslizamiento y del mismo modo, e1 periodo de "repeticion" de un episodio 1iidrogeologico asociado al deslizamiento puede ser también cuantificado.

I. JNTRODUCTION A nurnber of geological, geomorphological, hydrological and geotechnical features briefly 'ne study area is located in Southern Italy, summarised in this paper are examined in south-east of Naples (Figure 1). Phenomena of detail in the fina1 technical report drawn up slope instability and floods have long been a under the co-ordination of Prof. Cotecchia constant feature in this portion of the (EEC 1996). Apemines, with earthquakes and meteoric civents representing the main triggering factors. Landslides have ofien resulted in dreadful 2. GEOLOGICAL AND GEOMORPHO- economic and human losses. LOGICAL SETTING This work exarnines the rainfalillandslide relations for selected events in the upper The outcropping lithofacies can be divided into valley (Figure l), in the Senerchia area. The four types (Figure 1). purpose of the study is to define the impact of 1) The limestones and dolomitic limestones rainfall events on selected landslides. At (Trias-Cretaceous) which constitute the relief. present, the area is being investigated to 2) Clayey-marly , and clayey-marly- identifi the spatial trend of environmental arenaceous flysch, shales, marls, chert variables related to landslides; a site with an limestones, sandstones and varicoloured clays active landslide best suited for this research (Upper Cretaceous - Paleocene). The flysch project has been selected. shows a marked heterogeneity and anisotropy induced by a long geologica1 and tectonic essentially of clays and marls. Pelitic fractions history. Due to their peculiar mechanical and are found along with a granular component geotechnical behaviour, these successions are enriched by chaotic and non - homogeneous highly susceptible to macroscopic slope pebbles and calcareous blocks. deformation and mass movement in genera1 3) The detrital and breccia deposits of (Parise et al. 1997). The flysch is composed rockfall or scree (Quatemary) have a carbonate or piroclastic nature. with the carbonate rocks, and is threatened by 4) The alluvial deposits (Pleistocene- the adjoining SE1 and SE2 landslides in the Holocene) are present along the course of the south and an ancient landslide to the east. Its Sele River and its main tributaries. relative permeability is medium to high. The As a result of the Apennine tectonics -begun estimated thickness is of one hundred metres. iri the Middle - Upper Miocene and still active The flysch lies at the bonom. On the upslope throughout the Plio-Pleistocene-, the local side, it is hydrogeologically in contact with the rocks suffered large scale dislocations and main aquifer. The deposits make up a associated intensive deformations (Cotecchia et secondary aquifer. al. 1992). The present morpho-structural Wherever the flysch does not outcrop, it setting of the upper Sele River valley has been underlies the remaining lythofacies with which mainly shaped during the last Lower Pliocene - it is in contact. Its thickness reaches some Quaternary events (Parise et al. 1997). hundred metres. With respect to the carbonate The areas where carbonate rocks crop out hydrogeological unit, this complex acts as an are only marginally influenced by -mass impervious bounding, with a continuous movements; they include mostly rockfalls and permeabilty limit, for the carbonate rocks and topples limited to the steep slopes (mean slope the detrital deposits. The soils malung up this > 30") bordering the carbonate massif. The complex have been generally remoulded by a flysch areas are characterised by low-mediurn sequence of tectonic events and overlapping acclivity slopes (10t 12O), and widespread mass movements (Figure 1, Budetta 1983, creep and landslide phenomena (Agnesi et al. Agnesi et al. 1983). These mechanical actions 1983, Budetta 1983, Cotecchia et al. 1996). have disturbed the origina1 strutture of the soil and occasionally favoured some localized 2.1 Hydrogeological features improvements of the hydrogeological features. Lastly, some surface-alteration phenomena Tlie major hydrogeological features at the slope have participated in setting up particular scale are brieflv set forth. conditions favourable to groundwater flow. In the area surrounding Senerchia, on the The role of irnpervious limit played by this right side of the Sele River, the only detectable complex does not improve the stability of ihe hydrogeologic unit is that of Mts. slope Senerchia spreads upon. Some losses Polveracchio-Rainone (these mounts, which fiom the two hydrogeological water-bearing an: not included in Figure 1, area found on the lithotypes towards the area under study have to leiì of Magnone Mt.) (Celico & Civita 1976, be assumed, given: Celico et al. 1987). It is formed by carbonate - the occurrence of conspicuous springs of the rock. main aquifer, located along the slope where the These rocks exhibit the highest relative flysch outcrops (spring no. 3 on Figure 1, 360 permeability, owing to fiacturing and karst m a.s.l.), far fiom the permeability limit which phenomena. The carbonate rocks, some is found at 600 m a.s.1.; thousand meters thick, outcrop above the Iirnit - the precise match of the permeability limits which establishes a contact with the remaining around the secondary aquifer, near the scarps two significant hydrogeological lythofacies, the of landslides; detrital and breccia deposits and the flysch. In - the occurrence of some peremial old springs thc most depressed point of such a contact, the around the secondary aquifer, having an main sp~gsof the hydrogeological unit are estimated discharge which exceeds ten litres found (the Piceglia - Abbazzata springs, spring per second; no. 1 and 2 on Figure 1, about 570 Vs). - the after landsliding disappearance of some Moving towards the valley, the carbonate old springs. mass, divided in scales by the tectonic activity, underlies for some hundred metres. Some of these scales are uplifted and almost surfaced. 3. THE HYDROLOGICAL-STATISTICAL Tbus, the Pozzo S. Nicola spring is formed METHOD T0 STUDY THE MALL (spring no. 3 on Figure 1, about 180 Vs). INFLUENCE OVER LANDSLIDES The detrital and breccia deposits outcrop mcjstly in a limited quadrangular area, equa1 to As to the role played by precipitation on slope 0.5 square Km,which is in contact, to the west, instability, a lower limit can be easily established. These phenomena are actually Freeze and Herlam in 1969. These models had included among t& "stresses" comrnonly the merit of relying on processes and undergone by a slope. Therefore, researchers parameters directly observable on site. must proceed with extreme caution while Nowadays the unprecedented outburst of assuming slope instabili9 to be iriggered by numerica1 calculation allows to build complex precipitation, whenever these phenomena have hydrologic models for vast basins. However, a been concornitant with heavy rainfalls problem remains to be overcome. Data 'are (Polemio 1993). insufficient to indicate reliable results (Polemio Precipitation may prove the only cause 1995 a). Therefore, it is hardly surprising that when rainfalls are rather exceptionally heavy. these models have otten been used in a inverse As a rule, the slope stability happens to be way with a view to deriving reliable value-lags already endangered by previous events of the hydrologic parameters of the (Polemio 1995 b). While investigating the investigated area, based on mode1 calibrations. relationships between precipitation and Despite simplifying assumptions, it is self- landslides, the behaviour of the temtory should evident that these models are rather be taken into account. Each temtory pursues a complicated. They require many parameters dynamic equilibrium under locally ordinary which are practically unavailable. Gaining a climatic stresses, provided they last long preliminary and thorough knowledge of the enough. As soon as extxaordinary hydrologic slopes to be simulated proves extremely costly events intervene, the equilibrium rnight be (Polemio 1995 b). Besides, difficulties arise in drastically changed (Govi et al. 1985). identifiing these parameters within a complex Therefore, the analysis of any precipitation- hydrogeological context, as the Senerchia induced irnpacts cannot disregard the vast array hillslope. of instability-triggering phenomena in the case in point or within the investigated area. In 3.1 The applied hydrologic-statistica1 rnethod particular, it must be generally agreed that a. single precipitation-landslide event, though Should the numerous expensive data required conspicuous, is but a tirne-limited event within to evaluate the above mentioned phenomena a long-lasting active evolution (Polemio 1993). not be available, valuable indications can still Determining the influence of rainfall upon be obtained by more strearnlined landslides requires a thorough approach that is hydrologicaYstatistica1 models. Usually, these possible only if an overall understanding of the are empincal or semi-empirical models of the soils forming the slope has been acquired. analysis of rainfall-landslide relation. Among Given the complexity of slope stability- them, the hydrologicaYstatistica1 models were related phenomena and the uneven variation devised as a means of investigating, through speed of the stability conditions induced, the study of the selected hydrological variable, simulations are required of both subsoil and to what extent the rainfall event believed to be surface phenomena. associated with a given landslide had been Wherever groundwater flow is involved, exceptional (Polemio 1993). attention should be focused on both the Experience has shown the way precipitation saturated and the unsaturated zones, and exerts an influence on the stability of some complete hydrological models of slopes must given slope types. Regardless of the soil nature, be constructed (Beven et al. 1987). Despite heavier and short-duration rainfall generally their complexity, however, such models do not trigger limited surface-movements (Govi et al. cover some phenomena that cannot be 1985), often associated to new slipping overlooked, such as the air reverse flow in the surfaces. By contrast, long-duration rainy unsaturated zone. periods reactivate pre-existing slipping A fu11 approach can only be derived from a surfaces, mostly located some metres below the deep knowledge of the slope soil behaviour. It soil surface. is rather difficult to gather data enabling to Consistently with the above scenario, reconstruct both underground and surface flow, Cascini and Versace (1986) indicate that for focusing on percolation in the saturated and deep-seated landslides, the only significant unsaturated areas. variables are the ones which take into account The feasibility of physically-based models rainfalls lasting more than one day. leading to run-off was first envisaged by Some new hydrologic variables need to be complex mixed iùnctions, such as the TCEV,is introduced afiesh, such as the so-called justified. cumulated or cumulative rainfall, that is the The method described was applied to the summation of precipitations occurred over a variables, cumulative daily rains Ch,: given number of consecutive days, prior to the considered day. The identification of the most suitable variable for the correlation between rainfalls and landslides results fiom essays run on inore than one cumulative rainfall. The where n stands for l to 180 consecutive days, j suggested method (Polemio 1993) leads to the is the serial number of the day on which identification of the cntical duration. This is measurements were taken over the observation the iiumber of consecutive days during which period and Hj is the daily rainfall of "j" day. total rainfalls may prove significant for The highest values CHMAX,, were extracted, landslides, under given climatic and for each year y, fiom the generated series of geomorphologic conditions. data. The observation period is 65 years long. The proposed hydrologic-statistica1 model By means of these models, the exceptional investigates the exceptional character of the character of landslide-related rainfall events meteoric event which is likely to be associated can be explained in tem of recurrence to landsliding, by exploring the assumed intervals or return period, T. As a result, one maximum values of the selected hydrologic can determine the statistical cyclicity by which variable. As to surface-instability phenomena, a slope has been subjected to hydrologic the time-lag used to define the hydrologic conditions similar to those related to a variable, expressed in te- of precipitation landslide. intensity, is generally very short (a few days, at It is interesting to consider that oniy this the most). Longer periods, ranging between type of approach pennits an evaluation of 180 and 360 days are covered by means of rainfall significante for historical landslide; the cumulated rainfalls, in case of major landslide only condition is to know daily rainfall data. reactivations. Available historical data about previous However, studies of statistical hydrology reactivations of the landslide body enable to suggest the use of two parent functions or better appreciate the cause-effect relationship probability dishibution functions (Polemio (Polemio & Sdao 1996 a, b and d). However, 1993 and references therein). The GEV once the influente of precipitation has been (Generalized Extreme Value) fùnction is ascertained, the calculation of the return period univocally defined by the three parameters may prove an useful prevention tool, al1 the scale, position and shape, with the well-known more if it is associated to the detection of Gumbel fiinction being one particular case. In warning- "hydrologic thresholds". Moreover, some particular hydrological situations whenever stabilization actions are forewen, (Cotecchia et al. 1995) it is necessary to utilize this method allows to assess the utility of a parent fùnction defined by means of a two- worb aimed at limiting the effects of component model. These are four-parameter precipitation on slope stability. functions such as TCEV (Two Component On the other hand, where a large area Extreme Value), obtained by superposing two periodically investigated shows a severe functions applied to the same set of data. susceptibility to landslides, often involving In theory, the larger the number of urban areas, a regionalized hydrological pararneters in the parent distributions, the more approach may be envisaged. This method could closely will the investigated phenomenon be then prove a good choice (Parise et al. 1997, represented. On the other hand, due to the Polemio & Sdao 1996 b and e). limited number of available data, the fùnctions Within the fiamework of evolution become more and more inaccurate and processes, the assessment of the exceptional ineffective as the number of parameters is character of rainfalls resulting in landslides increased. The use of the three-parameter GEV may support the geomorphological and function has by now become a very coriimon evolutionary study @'Ecclesiis et al. 1991, practice, but eficacy in those cases where the Cotecchia et al. 1995), as in this research. separation condition (Matalas et al. 1975) The hydrologic and statistical results occurs is rather poor. In such cases using more mferred, irrespective of the soil nature and the Table 1 - Probability grades of cumulative rainfall related to landslide. History case: landslide narne and the critical duration in brackets. problem geornetry, have always proven on slope stability rather than on trigger consistent with the slope hydrogeological thresholds. To this particular purpose a conditions and the characteristics of the soils classification of the influence of rains on slope involved. In every investigated case, given the stability and of rain-induced landslide hazard is grain-size fractions, the corresponding proposed here (Table l). hydraulic conductivity and the thickness of the landslide body, the duration of a critical infiltration episode has been estirnated - as is 4. INVESTIGATED LANDSLIDES the case in the Agrigento landslide (Cotecchia et al. 1995). The duration was always found to The SEI landslide was mobilized following the be almost equal to the critical duration or the 1980 earthquake (M = 6.8): the rnudslide duration of the rnost hydrologically and rneasures about 2500 m in length (Table 2); the statistically significant cumulative rainfall. landslide is likely to be a reactivation though The aforementioned positive sides of this no historical data are available (Cotecchia 198 1 method are discussed in detail in the and 1984, Cotecchia et al. 1986, Maugeri et al. bibliography (Polemio 1993, Cotecchia et al. 1982). Despite the 1930 and 1962 earthquakes 1995, Polemio & Sdao 1996 a-e). (9s10 MSK degree), the 1955 aerial photos do Unlike the work done on the rainfall/flood not show any well-defined landslide body; only correlations, the rainfallllandslide research in the area between 274 and 199 m a.s.1. deals with a subject -the slope- which, contrary landslide activity is recognizable (Maugeri et to catchrnent areas, changes its own al. 1982, Budetta 1983). It is a complex characteristic morphological, arid ofien landslide, started as a translational earth slide geotechnical, features with succeeding in the upper portion and evolved to flow in the landslide events. This implies theoretically medium-lower one. The rnain scarp is located that, when consiructing a trigger rnodel, one in the proximity of the contact between the should consider from a statistical viewpoint the flysch and carbonate or detrital deposits. time lapsed since the last movement or The initial landsliding coincided with the morphological change and the scale thereof. collapse of the aqueduct which crossed the Integrated studies should be undertaken on landslide body. statistical hydrology and evolving morphology The SE2 landslide represents the (D'Ecclesiis et al. 1991, Cotecchia et al. 1995). enlargernent of a subsidiary slide situated in the Attention should be focused on the middle-upper portion of the left flank of the influence exerted by some hydrological events SEI landslide (Table 2, Photo 1). The reactivation of a lateral portion of the SE1 rnudslide was observed on December 29, 1993, with a strong retrogressive evolution. The SE2 accurnulation zone is superimposed on or joined to the SE1 landslide body. This landslide is continously surveyed from 1995 and it is still active today (Polemio 1996). The SE2 landslide ground surface inclination ranges from 10 to 15O, without significant changes due to landslidiag (Wasowski 1994); the main scarp is very Table 2 - Investigated landslides and average features (m). Photo 2 - The SE4 landslide foot.

Here again, the initial landsliding coincided with the collapse of the aqueduct-which ran longitudinally through the landslide. The SE3 and SE4 landslides occurred in Sierra Piano and Bosco, respectively, are very Photo 1 - Panoramic view of the SE1 and SE2 landslides. recent. The Author was thus able to carry out surface surveys without resorting to topographic surveys and aerial views. Needless steep, its slope continuously changes up to 50". to say, the detected geometrica1 and The landslide body shows partial excursive geomorphological features are but preliminary accelerations mainly due to upslope feeding. In ones. fact some secondary rotational slides, falls and Some points of similarity do exist between earthflows, located along and near the crown, the SE3 and SE2 landslides as to the particle periodically load the body upslope, thus size of the soils involved and the motion feeding the flow; the landslide is composite thereof. However, there is an essential (Cruden & Vames 1994). The maximum difference. The evolution of the landslide is not displacement rate was 1.33 mlday (72 m after retrogressive. It was favoured by the failure of 54 days) and was observed at the main scarp the rock overlying the landslide crown. bottom. Therefore, the SE3 landslide would appear The SE2 landslide represents a geologica1 more similar to the SE1 landslide, though the hazard for some houses and infrastructures latter is by far smaller. It might result from the located upslope in the surrounding area. reactivation of a slightly bigger landslide interested by not-structural stabilization works. The failure of the overlying rocks threaten both - the road and the aqueduct connected to some E 20 - large tapped springs. Moving downward, the 5 200 -'2 5 mudslide flowed past a gravity wall of massive

.J l0 2 stones destroying the underlying local road and -; 100 the telephone lines. In the vicinity of another O x severely damaged road, not far from the foot of o O JAN MAR MAY JUL SEP I*)Y the landslide, the motion was al1 the more complex. 1 m2 -3 -0-L The SE4 landslide differs from the Figure 2 - Hydrological regimen. 1) surplus remaining ones, as it affected a detrital water, 2) rainfall, 3) rea1 evapotranspiration, 4) oalcareous body. The motion, enhanced by temperature. four mentioned landslides occurred in November and December. While studying the nine series of CHMAX,, of Senerchia, the separation condition was not found. It follows that the probability distribution functions of curnulated minfalls were calculated using the GEV hction (Figure 3). Note that the curves resulting from the standardisation by means of CHMAX, ,, - the mean of annual maximum values - are sirnilar. If a single mean curve was assumed for T<20, the maximurn deviation would equa1 12% of the mean, at the most. This circurnstance, common to al1 the sites investigated so far enables to set up simple waming tools. Once the probability distribution functions were known, the retum period of al1 curnulated rainfalls 60 days before and 30 days after the landslide occurrence was calculated. Each Roduced variala Figure 3 - Probability distribution functions of rainfall-landslide event was then assigned the Senerchia cumulative rainfall. CHnj = corresponding probability grade (Table 1). The cumiilative rainfall value, CHMAX,,, - cntical duration of the SE1 and SE2 landslides mean yearly maxirnum of cumulative rainfall. ulas not deterrnined, given the negligible role played by rainfalls. failures occurred above the crown lacking the The analysis confirms that the SE1 landslide main scarp, was. mostly translational uphill and was activated during ordinary rainfalls debns-flow-like downhill. The landslide (Cotecchia & Nuzzo 1986). destroyed an important aqueduct and a local Furthermore, in contrast to the hypothesis road which ran transversally through it and it put forward by other Authors (Wasoslu 1994), almost wiped out a house built next to the foot. the reactivation of the SE2 landslide cannot be The motion lasted less than 24 hours (Photo 2). associated with rainfalls. The concomitant rain Except for the SE1 landslide, Senerchia was a contributing factor rather than a landslides are known to have been triggered by determining factor, acting on a pre-existing concomitant heavy rains. This thesis expressed precarious stability. Lacking the seismic action by the local population and some scientists is involved in the SE1 case, the landslide is based on the notion that landslides occurred accounted for the evolution of the Sele Valley during rainy periods. We will show how sides - bnefly discussed in the conclusions - inaccurate this view is. still undergoing a complex dynamic process. The SE3 and SE4 landslides occurred during a relevant rainy penod - in late 1996 - 4.1 Rainfall associated with landslides when the yearly rainfall rate was 25% higher than the mean values. There is only one monthly temperature and The SE3 landslide was affected by rainfalls, rainfall minimurn and maximum during the the role of which is secondary though not mean year of Senerchia (600 m a.s.1.) (Figure negligible. The scheduled studies, including 2). Precipitation is influenced by the vicinity of on-site surveys and hydrogeological the Tyrrhenian Sea and the mountain. The determinations, will enable to check the mean annual rainfall is 1600 mm, the mean physical reliability of the estimated critical annual temperature is 12.8 "C. According to duration (Table 1). The low value of the critical the Thornthwaite-Mather method (1957), duration, given the particle size and the depth evapotranspiration is 570 mm, so 1030 mm are of the landslide body, is justified by the pre- available for infiltration and runoff. Hence, existing strong remoulding of the landslide. hydrogeological conditions are mostly prone to The SE4 activation is partially ascribable to landsliding between November and May. The the action of rainfalls 30 days before the event. As to the recent landslides, the collapse of Beven, K., Calver, A. & Moms, E.M. rocks above the crown, which can be hardly (1987). The institute of hydrology distributed dated, contributed to the instability of the site. modell:inst. of Hydrol. report n. 98, 1-33. Budetta, P. (1983). Geologia e frane dell'alta valle del F. Sele. Mern. e Note Ist. 5 CONCLUSIONS Geol. Appl., vol. 16, 53 pp, Napoli, Italy. Caschi, L. & Versace, P. (1986). Eventi This study highlighted the extreme pluvicpebici e movimenti fianosi. AGI, XVI complexity and diversity of the phenomena ihat Convegno Nazionale di Geotecnica, Bologna, govem the effects of rainfall upon slope Italy. stability. The study showed that conventipnal Celico, P. & Civita, V. (1976). Sulla statistica1 hydrology methods, based on flood tettonica del massiccio del Cervialto investigations and then applied to landslides, (Carnpania) e le implicazioni idrogeologiche ad are misleading. Whatever trigger mode1 is esso connesse. Boll. Soc. Nat. in Napoli, 85, applied to the rainfall/landslide study, one Napoli, Italy. cannot ignore recent slope changes. In the case Celico, P., Guadagno, F. M., Luise, G., in point, the progressive evolution of the Sele Tescione, M. & Vallario, A. (1987). valley sides proved crucial. It followed recent Idrogeologia del monte Polveracchio-Monte tectonic events and resulted in a sequence of Rainone (Monti Picentini, ). Mern. gravity phenomena. These instability Soc. Geol. It., 37,341-362, Rome, Italy. phenomena pursue new conditions established Cotecchia, V. (1981). Considerazioni sui by the geomorphological, seisrnic and climatic problemi geomorfologici, idrogeologici e status which are still far fiom equilibriurn. The geotecnici evidenziatisi nel temtono colpito single instability episodes have to be regarded dal sisma campano-lucano del 23 novembre as macroscopic and sudden signs of a slow, 1980 e possibilità di intervento del Progetto relentless evolution slightly affected by Finalizzato Conservazione del Suolo del CNR. rainfalls. Rend. Soc. Geol. It., 4,73-102, Rome, Italy. The continuous hydrogeological monitoring Cotecchia, V. (1984). Note sui fenomeni di started in 1995 on the SE3 landslide is likely to instabilità del temtorio e sulla loro clarify the role played by groundwater flow rappresentazione con particolare riguardo agli within the carbonate aquifers that make up the eventi sismici. Ricerche e Studi Formez, uphill portion of the investigated slope. Lineamenti di geologia regionale e tecnica, RS 37,207-282, Italy. Cotecchia, V,, (1986) . Ground ACKNOWLEDGEMENTS deformations and slope instability produced by the earthquake of 23 November 1980 in This research referred to the "Landslide Campania and . Geol. Appl. evolution controlled by clirnatic factors in a Idrogeol., 21 (5), 31-100, Bari, Italy. seismic area. Prediction methods and warning Cotecchia, V., Del Prete, M. & Tafuni, N. criteria " project, fmanced by EEC (1986). Effects of earthquake of 23th Environment Programme (Unit: Climatology November 1980 on pre-existing landslides in and Natura1 Hazards), The author is grateful to the Senerchia area (Southern Italy). Geol. Appl. the project coordinator, Prof. V. Cotecchia, for Idrogeol., 21, 177-198, Bari, Italy. the positive resetlrch expenence. Cotecchia, V. & Nuzzo, G. (1986). Hydrological study of the upper valleys of the Sele and Ofanto rivers struck by the November REFERENCES 23 1980 earthquake. Geol. Appl. e Idrogeol., 2 1 (4), 65-97, Bari, Italy. Agnesi, V., Carrara, A., ~acal'uso,T., Cotecchia. V., Parise, M., Polemio, M., Monteleone, S., Pipitone, G. & Somso Valvo, Trizzino, R. & Wasowski, J. (1996). Primi M: (1983). Elementi tipologici e morfologici risultati della ricerca sulla Frana di Acquara- dei fenomeni $i instabiliti dei versanti indotti Vadoncello (Senerchia-AV). Conv. La dai sisma del 1980 (Alta valle del Sele). Geol. stabilita' del suolo in Italia, Accademia Appl. e Idrogeol.; 18 (l), 309-341, BaFi, Italy. Nazionale dei Lincei, Rome, Italy. Cotecchia, V., Salvemini, A. & Ventrella, Avigliano (Basilicata). Geol. Appl. e Idrogeol. N.A. (1 992). Interpretazione degli 3 1, Bari, Italy, in press. abbassamenti territoriali indotti dal terremoto Polemio, M. & Sdao, F. (1996 e). Critica1 del 23 Novembre 1980 e correlazioni con i rainfall and landslides in Basilicata region danni osservati su talune strutture (Southem Italy). 30" Int. Geologica1 Congress, ingegneristiche dell'area epicentrale Irpina. Beijng; Cina. Riv. It. Geot., 2, Italy. Polemio, M. (1993). Le precipitazioni Cruden, D.M. & Varnes, D.J. (1994). meteoriche nei fenomeni di instabilità dei Landlsides type and processes. In Landslides: versanti in ambiente mediterraneo. Medit, n. investigation and mitigation. Transportation 4/93, Bologna, Italy. Research Board. National Academy of Polemio, M. (1995 a). Gli effetti delle Sciences. precipitazioni atmosferiche sulle frane. D'Ecclesiis, G., Grassi, D., Merenda, L., Relazione. Visita di studio e ricerca alla City Polemio, M. & Sdao, F. (1991). Evoluzione University di Londra, 1- 13. geomorfologica di un'area suburbana di Polemio, M. (1995 b). Precipitazioni Castronuovo S.Andrea (PZ) ed incidenza delle atmosferiche e frane. Dimensione Ricerca, piogge su alcuni movimenti di massa. Geol. Area di Ricerca CNR, Bari, Italy . Appl. e Idrogeol., 26, Bari, Italy. Polemio, M. (1996). Gli strumenti del EEC (1996). Landslide evolution controlled monitoraggio idrogeologico continuo della by clirnatic factors in a seismic area. Prediction frana Acquara-Vadoncello di Senerchia. Rapp. methods and warning critena. Project Tecn. Int. CNR-CERIST, 41, 1- 17, Bari, Italy. coordinator and italian responsible scientist: Thomthwaite, C.W. & Mather, J.R. (1 957). Prof. Cotecchia, V.; italian research team: Instructions and tables for computing potential Cotecchia, V., Cotecchia, F., Parise, M., evapotraspiration and the water balance. Polemio ,M., Trizzino, R. & Wasowski, J.. Drexel Inst. of limat., 10,323 pp., Centerton. Govi, M., Mortara, G. & Sorzara, P. F. Wasowski J. (1994). Preliminary analysis of (1985). Eventi idrologici e frane. Geol. Appl. e the December 1993 Acquara-Vadoncello Idrogeol., 20 (2), Bari, Italy. landslide. Campania region. Rapp. Tecn. Int. Matalas, M.C., Slack, J.R. & Wallis, J.R. CNR-CERIST,40, 1-13, Bari, Italy. (1975). Regioni1 Skew in Search of a Parent. Water Resour. Res., 11, No. 6: Maugeri, M., Motta, E. & Sorriso Valvo, M. (1982). The Senerchia landslide triggered by the November 1980 earthquake. Prk. N Congr. MG,8, 139-149. Parise, M., Polemio, M. & Wasowski, J. (1996). Rainfall and Landslides in the Upper Valleys of Sele and Ofanto Rivers. Proc. Int. Symp. Engineering Geology and the Environment, IAEG-IAH, Athens, in press. Polemio, M. & Sdao, F. (1996 a). Le piogge e il rischio di frana in alcune aree urbane della Basilicata. Conv. La stabilita' del suolo in Italia, Accademia Nazionale dei Lincei, Roma. Polernio, M. & Sdao, F. (1996 b). Eventi di pioggia e fasi di attività di una hanei pressi di Calciano in Basilicata. Atti V Conv. Naz. dei Giovani Ricercatori in Geologia Applicata, 1011996, Cagliari. Polemio, M. & Sdao, F. (1996 C). Landslide risk and critica1 rainfalls in Southern Italy. Proc. 6th Int. Symp. on Landslides, 2,847-852. Polemio, M. & Sdao, F. (1996 d). Rischio da frana indotto dagli eventi di pioggia in