GEOL. CROAT. 1 51/2 1 195 - 204 1 5 Figs. 1 3 Tabs. 1 1 ZAGREB 1998 Recommendations for Landslide Hazard and Risk Mapping in Croatia Snjezana MIHALIC Key words: Gcohazards, Landslides, Landslide hazard, Lands lide risk, Hazard and risk mapping, GIS. Abstract The preparation of landslide hazard and risk maps is required as a base for rational land lise planning and decision-mak ing, in landslide prone arcas. In this paper, an inventory was made of the currently available methods for landslide hazard and fisk zonation, in order 10 produce recommendations for the lise of specific methods in relation to lhe sca le of analysis . A hierarchical sct of activities aimed at obtaining landslide-related information for all levels of land use plan­ ning in the Republic of Croatia is conSlrllclcd. Thi s sel encompasses: the establishmen t of a national landslide inventory on a regional scale « 1: 100,(00), stati stical landslide hazard analysis of geological-mor­ pholog ical fac tors at the medium scale (I :25,000) and the geotcchni­ cal c haracteri sation of slope movements followed by landslide ri sk analysis on a dctailed scale (> 1 : 5 ,(00). 1. INTRODUCTION The term landslide denotes the movement of a mass of rock, debris or earth down a slope (CRUD EN, 199 1). Most of the terrain in hillside areas has been subjected to landslides at least once under the influence of a vari­ ety of causal factors . The number of slope instability is significantly increased due to urbanisation and develop­ ment in landslide-prone areas. I-Tcnce, landslides conti­ nue to be one of the most threatening and widespread geohazards. There is an increasing trend for geohazards to be recognised in planning legislation and gu idance, espe­ cially in the last decade (the International Decade for Natural Disaster Reduction). Many countries presently have specific planning or development policies aimed to reduce the losses due to natural hazards (SCHUS ­ TER, 1991; MciNNES, 1996). Methods vary from gui­ dance documentat ion alone, through mandatory build­ ing codes and finally to insurance or disaster relief schemes (STATIIAM et al. . 1995). Zoning and subd ivi­ sion ordinances are used to divert development into areas where the risks are less and to ensure that , where deve lopments are permitted, the appropriate engineer­ ing measures are incorporated, Faculty of Mining, Geology and Petroleum Engineering, University of Zagreb, Pierotlijcva 6, HR-J 0000 Zagreb, Croatia. Risk assessment is a prerequisite for all the methods of hazard prevention and mitigation. It consists of three steps: (1) hazard assessment - identification of past/pre­ sent landslides as well as the prediction of future occur­ rences; (2) vulnerability analysis - identification of the location and distribution of population, infrastructure and vital economic activities exposed to a potential or present land slide; and (3) calculation of the expected loss (risk) from the hazard and vulnerability. Hazard analysis requires a detailed knowledge of the geoenvi­ ronmental predisposition factors and initiation events that lead to landsliding. This lies within the domain of earth scientists. Vulnerability and risk evaluation also includes other disciplines, such as urban planning, social geography, economy, etc. The end result of haz­ ard and risk analysis should be presented in informative documents, usually in the form of various maps which display the spatial distribution of hazard and risk clas­ ses, These documents are used by decision-makers who have to define a general risk prevention policy, Many methods and techniques have been proposed for landslide hazard and risk mapping over the last 30 years (BRABB. 1984; I'IANSEN, 1987; MIHALIC. [996), Significant progress has been made by establish­ ing the basic definition of terms related to hazard and risk assessment (VARNES. 1984). Van WESTEN's (1993) overview of the available methods is of great importance for improving the quality, as well as for achieving a uniform approach to landslide hazard map­ ping on an international level. On the other hand, exam­ ples of landslide risk zonations are st ill rare because of the difficulties in assessing the probability of landslid­ ing and of the vulnerability of elements at risk. Howev­ er, the evaluation of risk corresponds to a political, eco­ nomic and social necessity. Therefore, research of oper­ at ional risk evaluation methods is in progress (RAGO­ ZIN. 1994; REZIG et aI., 1996). The first fundamental step of hazard and risk assess­ ment is the identification and mapping of all landslide phenomena, i.e. compilation of the landslide inventory (FERNANDEZ et al.. 1996; ROSENBAUM & POPES­ CU, 1996). Collection and management of spatial data req uire the utilisation of geographic inform ation sys­ tem s (GIS). Furthermore, the cartographic landslide data bases should be va lid nation-wide in order to ena­ ble extrapolati on of acquired experience to areas with similar characteristics (LEROI, 1996). In many coun­ tries, the development of such databases started in the 196 1990's. The most important arc examples from France (LEROI, 1996), Germany (POSCHlNGER , 1994; KRAUTER et aI., 1996), the USA (BROWN, 1992) and Canada (CRUDEN, 1996). Moreover, there is a tendency towards establi shment of a World Landslide Inventory (BROWN et aI., 1992). To ensure the consis­ tcncy of data recording, the International Geotechnical Societies' UNESCO Working Party on World Land­ slidc Invcntory (WP/WLl), initiated in the 1988 at the 51h Internat ional Symposium on Landsl ides, is suggest­ ing a standard terminology for describing landslides (WP/WLl, 1993). GIS technology is also essential for assembling the hazard and risk models, as well as for an efficient and rapid information exchange between sci­ entists, engineers, policy makers, and all the people and institutions dealing with the landslide hazard. Before slarti ng any data collection, a number of interrelated things should be clearly defined, such as the aim of a study, the scale and degree of precision of the prescnted rcsults, and the available resources in the form of money and manpower. To achieve the optimi­ sation of costs and quality, the application of different data analysis methods at various scales is required. Accordingly, a concisc rcview of current methods of landslide hazard and risk assessment is presented in this paper. The review is aimed at a comparison of the me­ thods, and proposit ion of a logical set or activities relat­ ed to the preparation and im plementation of hazard and risk maps, in the field of land use planning in the Repu­ blic of Croatia. This set comprises all levels of urban planning, from national to the local scale. 2. LANDSLIDE HAZARD AND RISK The terminology concerning hazards and risk used in this paper conforms to the definitions proposed by VARNES (1984). Evaluation of various ri sk compo­ nents (hazard, vulnerability, cost) , and of the landslide risk as a whole, presupposes that answers are available 1'01' the questions as shown in Table 1 (LEROI, 1996). Accordingly, landslide research aimed at hazard and risk mapping comprises the aspects of landslides sum­ marised in the following paragraphs. Since the term of landsliding encompasses "all movement of a mass of rock, debris or earth down a Component Question Hazard 1) Which type of movement is involved? 2) Where are the potentially unstable areas? Geologia Croalic;) 51n slope" (CRUDEN, 1991), which types of movement arc present in the stud ied area should be defined. The types of movement arc essentially those defined by the Inter­ national Geotechnical Societies' UNESCO Working Party on World Landslide Inventory: fall, topple, slide, spread and flow (CRUDEN & VARNES, 1996). Location of the unstable areas should be determined by the engineeri ng geo log ical mapping of the land­ slides. The objective would bc to record identifi able landslide features and their dimensions (IAEG COM­ MISSION ON LANDSLIDES, 1990). Evaluation of the probabili ty or timing of the future occurrence is dependent on the probability of occur­ rence of the triggering factor. A trigger is an external stimulus, such as intense rainfall, that causes a near­ immediate response in the form of a landslide by rapid­ ly increasing the stresses or by reducing the strength of slope material (WIECZOREK, 1996). Hence, it is of primary importance to differentiate the conditions that caused s lope instability from the processes that trig­ gered the movement (POPESCU, 1994). Landsliding causes damage both in the areas of instability initiation and in the areas of transport and of the reception of thc movements. In order to be able to describe where the landslide is moving, it is necessary to investigate it's activity. The UNESCO WORKING PARTY ON WORLD LANDSLIDE INVENTORY (1993) suggests describing the landslidc activity ill terms relating to state, distribution and style of activity. The assessment of losses consists of analysis of the interactions between the phenomenon and goods, i. e. behaviour of structures and people that are exposed to landsliding. Hence, it is fundamental to determine thc level of intcnsity of a potential phenomenon. An impor­ tant characterist ic of the movement comprised in the intensity analysis is the rate of Jandsliding (lUGS WG/L, 1995). 3. METHODS OF LANDSLIDE HAZARD ZONATION All the methods proposed are founded upon a single principle "the past and present are keys to the future" which implies that slope-failures in the future will be more likely to occur under those condit ions which led 3) At which moment can the identified phenomenon be triggered? Vulnerability Cost 4) How far can the phenomenon be propagated? 5) What are the interactions with the environment, natural or modified by Man? 6) What is the cost of the resulting damage? Table [ Risk components with ques­ tions connected to landslide risk assessment. Mih:llic ; Rccommcndalions for Land~ l ide I-Ia;wrd and Risk Mapping in eroalia van WESTEN, 1996 LEROI, 1996 Heuristic approach Expert evaluati on Stati stical approach Statistical return analysis Deterministic approach Mechanical models Table 2 Methods of landslide hazard zonation. to post onci present instability (CARRARA et aI., 1995) . Application o f the above principle requires mapping both the landslides and a se t of geological -morphologi­ cal causal fac tors, and cstabl ishment of a hazard modcl. There arc three main approaches for the developing of hazard models : heuristic, statistica l and determini stic approach. Each o f them is based on different e lements as shown in Table 2. 3.1. HEU RISTIC APPROACH In heuristic methods the expert opinion of the engi­ neering geolog ist and/or geomorpholog ist is used to classify the hazard. Two types of heuristic analysis can be di stingui shed: geomorphic analysis and qualitative map combination. The gcomorph ic method is also known as the direct mopping method (HANSEN, 1987). It consists or geo­ morpholog ical and/or engineering geol ogical mapping through wh ich the surveyor identifies pas t and present landsl ides and makes assumptions on those sites where failures are likely to occur in the future. Direct hazard de tcrm ination is based on indi vidua l experi ence. The dec ision rules vary from place to place and are difficult to form ulate. In addition, the resulting documents gen­ erally are "paper" ones (KIEN HOLZ, 1978). To overcome the problem of the " hidden ru les" in di rect mappi ng, indirect mappi ng methods have been developed. Qualitat ive map comb ina ti on is based on a priori knowl edge o f the cau ses of landsliding in the in ves tigated area. Hence, instability factors arc ranked and we ighted according to their assumed or expected importance in caus ing a mass-movement. In this meth­ od the expert 's knowledge can be formali sed into rules, but the res ult essent ially depends on the experience of the surveyor. At present , maps obtained by th is method canno t readil y be evaluated in terms of reli ability or certainty. 3.2. STATISTICAL APPROAC H In the statis tical approach, causal factors arc dcfi ned a posteriori, through back analysis of historical events. Therefore, the ro le o f each factor (that led to landslides in the past) is de termined on the basis o f the observed re lations with th e past/present land slide di stribut ion. The sta tist ical approach can be applied foll owi ng differ­ ent techniques which essentia lly differ on the statistical procedure used: bivariate or multi variate analysis. 197 In bivariate stat istical analysis each instabili ty fac tor map is combined with a landslide distribution map, and weighting values based on landslide densit ies are calcu­ laled (SIDDLE ct aI. , 1991; van WESTEN, 1993; YIN , 1994). Mul tivariate statistical analys is of the important fac­ lars re lated to landslide occurrence, g ive the re lative contribution of each of these fac tors to the total hazard within a defined land unit. For each sampling unit, the prcsence or absence of landslides is also de termined. The model is conceptually rai rl y simple, but large data se ts arc needed to obtain cnough cases to produce rcli ­ able resuits (CARRARA et aI. , 1995). 3.3. DETERMINIST IC APPROACH There are some examples of landslide hazard asses­ sment by calculating safe ty fac tors over large areas (van WESTEN, 1993; LEROl , 1996). The resulting safe ty fac tors are only indi cat ive and are used to tes t multipl e scenarios based on variable triggering hypo­ theses. The most frequently co nsidered are hydrau lic and seismic triggers. The ma in problem with these me­ thods lies in the choice o f the representative input para­ meters and the slope stability model. For the rational consideration of the natural va ri ab i­ li ty and uncertainty o f each input variable in slope sta­ bility analyses , a probabili s ti c approach is essenti al (HAMMOND et aI. , 1992; TER LIEN et aI. , 1995). The objecti ve is to obtain the probabili ty distribution of the fac to r of safe ty and hence probability o f failure. The most important limitation to the application of proba­ bilistic methods in landslide hazard assessment may be the lack o f statistical data on soil properties, pore water pressures and on loads (CHOWDHURY, 1984). 4. LA NDSLIDE RISK ASSESSMENT Landslide risk assessment requi res the understand­ ing, analysis and control o r damage which are the con­ sequences of thc interac ti on between slope movement s and exposed element s (property, people and various activ ities). However, due to the complexity of the phe­ nomena and partly to an absence o f conceptual know­ ledge of certain risk components, a unified approach 10 the problem docs not exist. As a result of technical and sociological advances several researchers and organi sa­ tions sta rt ed to develop a me th odo logy for landsl ide risk evaluati on in the last decade (FELL, 1994; RAGO­ ZIN, 1994; LEONE et a I. , 1996; LEROUEIL e t ai. , J 996). ANDERSON 's et al. ( 1996) proposa l of a risk­ based meth od for selecting alte rn atives for lands lide risk mitigation is presented as follows. The proposal is in teres ting because it comprises the whole procedure: the ident ification of risk, the estimation of risk, and the evaluat ion of risk through either avers ion or acceptance (Fig . I). Ri sk identification involves development o r 198 Geologia CroaliC:t 51/2 INITIATING SYSTEM OUTCOME EXPOSURE CONSEQUENCES EVENT RESPONSE Freeze-Thaw Action Debris slide Mobilize Debris Tim e of Day Rapid Snowmelt Debris flow Complete Failure Prox. to Failure Property Damage IDENTIFICATION Earthquake Rockslide Partial Failure WamingTimc Loss of Revenue (Examples) Extreme Precipitation Rockfall Discrete Boulders Time of Year Loss of Life ---- - - ------- EVENT SYSTEM OUTCOME EXPOSURE EXPECTED ~ RESPONSE r------- ~ f----. LOSSES PeE) p(FI E) P(O I F) P(L I 0) P(C I L) ESTIMATION i Remove Zidc Debris Warning 1ystems i ------------0 Weather Stabilize Slide Rockfall Barricade Modification Relocate Road I SELECT RI SK AVERSION I m",o" .~ - - --------------------------------- ACCEPTANCE RI SK ACCEPTANCE CRlTERION MET? -rYES Fig. I Framework model fo r ri sk-based method to mitigate lands li des (AN DERSON ct al" 1996). the risk model for the evaluation of an ex isti ng land­ s lid e risk. In orde r to achieve thi s it is necessary to recognise and list the va ri ous factors whi ch could con ­ tribute to the landslide failure risk, and then 10 organise these into logical event sequences. The model is organi­ sed in the foml of an event tree, which commences with events that can in itiate fai lure, and ends with the conse­ quences or a ra il ure (Fig. 2). In the later phase the risk model serves fo r evaluat ion of the effectiveness of pro­ posed rehab ilit at ion alt ernati ves . The second step involves ri sk es timation, i.c. ass igning the probabi li ties and consequences to the occurrence of each fail ure mode. If these risks arc unacceptable, lhe assessment proceeds to the third step - ri sk avers ion. This involves the formulation and evaluat ion of remedial act ion (reha­ bilitat ion) alternatives. The final step in the risk assess­ ment process is the decision 0 11 what degree of safety is acceptable. A crucial stage fo r a good understanding of slope movements and the risk associated with them, is the characterisation of movement through fac tors havi ng a mechanical significance. It requ ires establishing a re la- tionship betwccn the charac teristics of a g iven mOve­ me nt , ex istence of defini te predi spos iti on fac to rs, occurrence of tr iggering or aggrava ti ng fac to rs , ex is­ tence of de fin ite revea li ng faC Ia l'S, and of the conse­ qucnccs of the movement. For this purpose, VAUNAT et al. ( 1994) arc developing geotechnical characterisa­ ti on o r s lope movements, taki ng int o accou nt slope movement type, involved materia l and movemenl sta­ ges. Such a characterisation constitutes an essential step fo r the development of expert systems on s lope engi­ neering, for the selection of numeri cal models fo r the simulati on of spec ifi c aspects of slope behav iour, as well as ror the design of remed ial measures for stabi lis­ ing a slope. 5. SCALE-RELATED RECOMMENDATIONS Not all the methods of landslide hazard zonati on are equally applicable at each scale of analys is, because of the d ifTerence in requi red input data and degree of pre­ cision o r the obtained results. Table 3 provides an over­ view of the var ious methods o r landslide hazard analy- INITIATING EVENT SYSTEM RESPONSE OUTCOME EXPOSURE CONSEQUENCES Rapid Snowmelt... Extreme Precipitation Freeze-Thaw Action ... Rockslide RockfalL .. Partial Failure Com iete failure ... Discrete Boulders ... Off-season Ni hI... Off-season Da . Tourist Season Night... Tourist Season Da Road Closed Car I-lils Debris Debris Hits Car Debris Kills Motorist Fi g. 2 Hypothetical event tree branch for evaluating outcome probability for landslide ri sk assessment (ANDERSON et aI., 1996). Mihalic: Recommcndalions for I_'l.nds lidc Hazard and Risk Mapping in Cro'<-~~-~' - - MEDIUM SCALE t :25,000 Statistical hazard analysis of geological-morphological factors HAZARD MAP I ELEMENTS AT RrSK / r ~ _ _ _ _____ _ _ _ _ _______________ _ _ _ ____ _ , , DETAILE D SCALE : > 1:5000 , Geotechnical characterisations of slope , movement & Risk analysis , i LANDSLIDES I ) RISK l\'1A P ~ ~ l , , , , , : , , , , , , , - ------ -- - - - - - --- - ---------~ Geologia Croalicl 5 In Fig. 3 Hierarchical mo ­ del for landslide ha­ za rd and risk map­ ping. slides in Croatia could also serve as input for the World Landslide Inventory. of the Basic Engineering Geological Map of the Rcpub­ lic of Croatia scale I: 100,000. Figure 4 is a cartogra­ phic representation of a landslide inventory in the cen­ tral part of the Zagreb section. The Croatian Institute of Geology is currently deve­ loping a lands lide inventory as part of the preparation Fig. 4 Landslide inventory map of the central part of Zagreb section (Engi ­ neering Geo ­ logi cal Data Base - In stitu ­ te of Gcology, Zagreb). Mih:llic: Recommendations for Lmdslidc Hazard and Rhk M:lpping in Croatia 2) At the medium scale (I : 25,000) sta ti stical hazard analys is of geolog ical-morphological causal factors is requ ired. The detail on the hazard map should be such, that adjacent slopes of the same lithology are evaluated separately, and may obta in different hazard scores, depending on other characterist ics, such as slope angle and slope segments. This map should represent a base for rational land-usc planning, in order to locale devel­ opmen ts on stable ground. The fie ld of app li cation would be the physical planning at the municipal and ci ty level. Accordingly. a landslide hazard map for the terri tory of Zagreb C it y should also become a compo­ nent of the Physical Plan or Zagreb City, by replacing the existing Map of lithologica l classification and the slope stability categori sa tion of the Mt. Med ved niea hill s id es. On the basis of a landslide hazard map, the legislation restric ti ng development in the areas most susceptible to landslides could be enacted. En largement or the lands lide hazard map to the I: 10 ,000 scale could a lso serve as the bas is for con ­ struction of the Physica l Devclopmen t Master Plan of Zagreb City. Overlaying of the hazard map with thc map wh ich displays clements a t ri sk could indicate the level of risk. In the areas where risk is low, landslide hazard analysis will su ffi cc. 201 Fig. 5 Aerial photograph of Kostanjck land slide ar­ ea lakcn in 1985 with inscribed landslide bou­ ndary (STAN I<: & NO­ NVEtLLER ,1996). 3) At the large scale (> I: 5,000) the information of lands lidc ri sk is required. To achieve the expected deg­ ree of prec ision it is necessary to undertake comple­ mentary invest igation , fo llowed by geotechnical char­ acterisation of slope movement s, and thu s the risk assessment. The applicati on of ri sk maps li es in thc construction or urban plans and of the detailed physical deve lopment plans for the areas characterised by high risks. An example o f the area where the lands lide ri sk assessment could be undert aken is the large Kostanjek lands lide on the western o ut skirts of Zagreb, on the southern slopes of the M1. Medvednica (Fig, 5), The first movement of thi s active landslide occurred in 1963. It is est imated that a sl iding mass of some 32x 10(, 111 3 is involved, with a maximum depth of 90 m. The d isp laeemell1s on the su rface are 3-6 III (STANIC & NONVEILLER, 1996), The necessity for landslide ri sk assessment is conditioned by the presence of numerous houses in the area, as well as by the high level of explo­ ration including continuou s ground-displacement mea­ surements. 202 6. CONCLUSIONS Predictive models of landslide hazard and risk assessment constitute a major research field which may well take advantage of the potential of new technologi­ cal advancements - GIS-driven data acquisition, mani­ pulation and analysis. Consequently, the development of the methods of prod ucing landslide hazard and risk maps is still in progress, and no uniform approach is as yet accepted. By evaluation of the methodological approaches to landslide hazard zonation practices, the stati sti cal ana­ lysis or geological-morphological causal factors is sug­ gested, aimed at the prediction of the spatial probability of landslides (i.c. where j~lilurcs are most likely to oc­ cur). This method allows production of lands lide hazard maps at the scale of 1 :25,000 at an acceptable cost. Due to the variety of geological situations, the dive­ rsity of materials, the complexity of acting mechanisms and the variability of controll ing parameters, the indica­ tion of the tempora l probability of landslid ing (i.e. when failures are likely to occur) can only bc obtained by risk analysis at the detailed scale (> I : 5.000). Hence, of crucial importance for risk analysis is V AUNAT's et al. (1994) geotechnical characterisation of slope move­ ments. To bc of value, in term s of the eval uation and presentation of landslide mi tigation alternatives, risk analysis shou ld encompass risk identification, estima­ tion, aversion and acceptance, as proposed by ANDER­ SON et al. (1996). The priority areas for the construction of risk maps arc to be delimitcd on the basis of the hazard maps. The areas that arc to be covered by hazard maps should be determined on the basis of data fro m a national land­ slide inventory. The development of a genera l methodology for landslide hazard and risk mapping would require defini­ tion of the conceptual models, and extraction of simpli­ fi ed ope rational models from the conceptual model s. The choice of the models should also serve as a gu ide for the development of appropriate data bases, taking into account that the availability of adequate data (both in quantity and quality) is cruc ial issue enabling the task to be accomplished. Acknowledgements This paper bene fit ed from constructive rev iews by Prof. Dr. Mihael RIBIC IC, Prof. Dr. Hansj6rg OELTZ· SCHNER and Prof. Dr. Edmund KRAUTER. I wish to especially thank Prof. Dr. H. OEL TZSCHNER and col­ leagues from Bayerisches Geologischcs Landesam t for having introduced me into the GEORISK - landslide information system. Geologia Croarica 5 [/2 7. REFERENCES ANDERSON. L.R .• BOWLES. D.S., PACK, R.T. & KEATON, LR. (1996): A risk-based method for landslide mitigation.- In: SENNESET. K. (ed.): Landslidcs. Proc. 7th Int. Symp. Landslides, Trond­ heim 1996 .• I. 135-140. Trondheim. BRABB, E.E. (1984): Innovative approachcs to land­ slide hazard and risk mapping.- Proc. 4th Int. Symp. Landslides. Toronto 1984 .. 1,307·323. Toronto. BROWN, W.M., III (1992): Information for disaster reduction: The National Landslide Information Cen­ ter. US Geological Survey.· In: BELLM, D.H. (ed.): Landslides. Proc. 6th Int. Symp. Landslides. Christ­ church 1992.,2.891·892, Christchurch. BROWN. W.M .• III. CRUDEN. D.M. & DENISON. J.S. (1992): The Directory of the World Landslide Inventory.· U.S. Geological Survey Open-Filc Rcp­ ort 92-427. CARRARA, A .• CARDINALI. M. , GUZZETTI. F. & REICHENBACH. P. (1995): GIS technology in mapping landsl ide hazard. - In : CARRARA, A. & GUZZETTI. F. (cds): Geographical Information Systems in Assessing Natural Hazards. Kluwer Academic Publishers. Dordrecht. 135- 176. CHOWDHURY. R.N. (1984) : Recent developmcnts in landslide studies : probabilistic methods. State-o[­ the-art report - Session VII (a).- Proc. 4th Int. Symp. Landslides. Toronto 1984 .• 1.209-228, Toronto. CRUDEN. D.M. (1991): A simple definition of a land­ slide.- Bull. of the Int. Assoc. of Eng. Geol., 43, 27- 29. CRUDEN, D.M. (1996): An inventory of lands lides in Alberta. Canada.- In: SENNESET, K. (ecl.): Land· slides. Proc. 7th Int. Symp. Landslides. Trondhcim 1996 .• 3.1877-1882. Trondhcim. CRUDEN. D.M. & VARNES, D.J. (1996): Land slide types and processes. - In: TURNER. A.K. & SCHU· STER, R.L. (eds): Landslides: Invcstigation and Mitigation. National Academy Press, Washington, 36-75. FELL. R. (1994): Landslide risk assessment and accept­ able risk.- Can. Gcotcch. Journal. 31, 261-272. FERNANDEZ, T .• IRIGARAY, C. & CHACON, J. (1996): Inventory and analysis of landslide dctenni· nant factors in Los Guajares Mountains, Granada (Southern Spain).- In: SENNESET. K. (ecl.): Land­ slides. Proe. 7th lnt. Symp. Landslides, Tronclheim 1996 .• 3.1891-1896. Trondheim. HAMMOND, c.J., PRELLWITZ. R.W. & MILLER, S.M. (1992): Landsl ide hazard assessment using Monte Carlo simu lation. - In: BELL. D.H. (cd.): Lands lides. Proc. 6th Int. Symp. Landslides, Christ­ church 1992 .• 2. 959-964, Christchurch. Mihalic: Recommendations for Landslide Ilazard and Risk Mapping in Croatia HANSEN, A. (1987): Landslide hazard analysis.- In: BRUNS DEN, D. & PRIOR, D.B. (cds.): Slope In­ stability. John Wiley and Sons, London, 523-602. IAEG COMMISSION ON LANDSLIDES (1990): Sug­ gested nomenclature for landslides.- Bull. of the Int. Assoc. of Eng. Geel., 41, 13-16. lUGS WG/L (1995): A suggested method for describ­ ing the rate of movement of a landslide.- Bull. of the Int. Assoc. of Eng. Geol., 52, 75-78. KIENHOLZ, H. (1978): Maps of gcomorfology and natural hazards of Grindewald, Switzerland, scale 1:10.000.- Arctic and Alpine Research, 10,2,169- 184. KRAUTER, E., LIPPOMANN, R, MOSER, M., MOL­ LER, B. & PRINZ, H. (1996): Kinematical-geotech­ nical aspects of landslides in Germany.- In: SEN­ NESET, K. (cd.): Landslides. Proe. 7th Int. Symp. Landslides ,'Trondheim 1996., 1, 251-256, Trond­ hcim. LEONE, F., ASTE, JP & LEROI, E. (1996): Vulnera­ bility assessment of elements exposed to mass­ movcmnet: Working toward a better risk percep­ tion. - In: SENNESET, K. (cd.): Landslides. Proe. 7th Int. Sym p. Landslides, Trondheim 1996., I, 263 -270, Trondheim. LEROI, E. (1996): Landslide hazard - Risk maps at dif­ ferent scales: Objectives, tools and developments.­ In: SENNESET, K. (ed.): Landslides. Proc. 7th Int. Symp. Land slides, Trondheim 1996., I, 35-52, Trondheim. LEROUEIL, S, LOCAT, J., VAUNAT, J. , PICAREL­ LI, L., LEE, H. & FAURE, R. (1996): Geotechnical characterization of slope movements.- Tn: SENNE­ SET, K. (cd.): Landslides. Proc. 7th Int. Symp. Landslides, Trondhcim 1996., I, 53-74, Trondheim. MciNNES, R.G. (1996): A review of coastal landslide management on the Is le of Wight, UK.- In: SEN­ NESET, K. (ed.): Landslides. Proc. 7th Int. Symp. Landslides, Trondhcim 1996., 1,301-307, Trond­ heim. MIHALIC, S. (1996): Zoniranje po opasnosti od klizan­ ja - hazard i rizik (Lands lide hazard and risk zona­ tion).- Unpublished M.Sc. Thesis, Faculty of Min­ ing , Geology and Petroleum Engineering, Universi­ ty of Zagreb, 70 p. MIHALIC, S. & STANIC, B. (1995): Pokreti na padi­ nama - hazard i rizik (Slope movements - hazard and risk).- In: VLAHOVIC, I., VELIC, 1., & SPAR­ ICA, M. (eds): Proc. 1st Croatian Geological Con­ gress, Opatija 1995., 2, 367-372, Zagreb. POPESCU, M.E. (1994): A suggested method for reporting landslide causes.- Bull. of the In!. Assoc. of Eng. Geel., 50, 7 I -74. 203 POSCHINGER, A.V. (1994): Recording and investiga­ tion of mass movements in the Bavarian Alps: The GEORISK-projeet.- Proc. 7th Int. IAEG Congress, Lisbon 1994.,3,4527-4531, Lisbon. RAGOZIN, A.L. (1994): Basic principles of natural hazard risk assessment and management.- Proe. 7th Int. IAEG Congress, Lisbon 1994.,3, 1277-1286, Lisbon. REZIG, S., FAVRE, J.L. & LEROI, E. (1996): The probabilistic evaluation of landslide risk. - In: SEN­ NESET, K. (cd.): Landslides. Proc. 7th Int. Symp. Landslides, Trondheim 1996., I, 351-356, Trond­ heim. ROSENBAUM, M.S. & POPESCU, M.E. (1996): Using a geographical information system to record and assess landslide-related risks in Romania. - In: SENNESET, K. (cd.): Landslides. Proc. 7th Int. Symp. Landslides, Trondheim 1996., 1,363-370, Trondheim. SCHUSTER, R.L. ( 1991): Landslide hazard manage­ ment - experience in the United States.- Slope sta­ bility engincering: developments and applications. Proe. Int. Conf. Slope Stability, Isle of Wight 1991., 253-263, Isle of Wight. SIDDLE, H.J., JONES, D.B. & PAYNE, H.R. (1991): Development of a methodology for landslip poten­ tialmapping in the Rhondda Valley.- Slope stability engineering; developments and applications. Proc. Int. Conf. Slope Stability, Isle of Wight 1991. ,253- 263, Isle of Wight. SOETERS, R. & van WESTEN, c.l. (1996): Slope instability: Recognition, analysis and zonation.- In: TURNER, A.K. & SCHUSTER, R.L. (eds.): Land­ slides: Investigation and Mitigation. Nat ional Acad­ emy Press, Washington, 129-177. STANIC, B. & MIHALIC, S. ( 1995). Zoniranje po opa­ snosti ad klizanja (Landsl ide hazard zoning).- Tn: MA V AR, R. (ed.): Geotechnical engineering in cit ies. Proc. 2nd Conf. HDMTTI , Varaidin 1995., I , 467-475, Varaidin. STANIC, B. & NONVEILLER, E. (1996). The Kosta­ njek landslide ill Zagreb.- Engineering geology, 42, 269-283. STATHAM, I., LANGER, M.F.B. & BOUCKOVA­ LAS , G. (1995): The identification and monitoring of geohazards.- Interplay bctwcen Geotechnical Engineering And Engineering Geology. Proc. XI Europ. ConI'. Soil Mechanics & Foundation Engi­ neering, Copenhagen 1995.,9,77-104, Copenha­ gen. TERLlEN, M.T.J., van WESTEN, C.J. & van ASCI-I, T.W.J. (1995): Deterministic modelling in GIS­ based landslide hazard assessment.- In: CARRARA, A. & GUZZETTI, F. (eds.): Geographical Informa- 204 lion Systems in Assessing Natural Hazards. Kluwer Academ ic Publi shers, 57-77, Dordreeht. UN ESCO WORKING PARTY ON WORLD LAND­ SLIDE INVENTORY ( 1993): A Suggested Mcthod for Describing thc Activity of a Lands lide.- Bull. of the Int. Assoc. of Eng. Gcol. , 47, 53-57. van WESTEN, C.J. (1993): GISSIZ. Application of Geographic Information Systems to landslide haz­ ard zo nation. - ITC Publi cation No. IS, Enschede, 245 p. VARNES, D.J. (1984): Landslide hazard zonation: a review of principles and praclicc.- Natural Hazards, 3, UNESCO, Pari s, 63 p. VAUNAT, J., LEROUEIL, S. & FAURE, R.M. (1994): Slope movements: A geotec hn ical perspectivc.- Geologia Croatica 51/2 Proc. 7th Int. IAEG Congress, Lisbon 1994. , 3, 1637-1646, Lisbon. WIECZOREK, G.F. (1996): Landslide tri ggering meehani sms.- In: TUR NER, A.K. & SCHUSTER, R.L. (eds.): Landslides: Investigation and Mitiga ­ tion. National Academy Press, Washington, 76-90. WP/WLI (1993): A suggested method for describing the activity of a landslide.- Bull. of the Int. Assoc. of Eng. Geol., 47, 53-57. YIN, K.L. ([994): A computer-assisted mapping of landslide hazard evaluation. - Proe. 7th Int. IAEG Congrcss, Lisbon 1994.,6,4495-4499, Lisbon. Manuscript received June 4, 1998. Revi sed manuscript accepted November 23, 1998.