The AVI Project:
A bibliographical and archive inventory of landslides and floods in Italy
A revised version of this document was published on Environmental Management, Volume 18, Number 4, July/August 1994, p623-633
Abstract
The AVI project was commissioned by the Minister of Civil Protection to the National Group for Prevention of Hydrogeologic Hazards to complete an inventory of areas historically affected by landslides and floods in Italy. More than 300 people, divided into 15 research teams and 2 support groups, worked for one year on the project. Twenty-two journals were systematically searched for the period 1918-90, 350,000 newspaper issues were screened and 39,953 articles were collected. About 150 experts on mass movement and floods were interviewed and 1,482 published and unpublished technical and scientific reports were reviewed. The results of the AVI project, in spite of the limitations, represent the most comprehensive archiving of mass movement and floods ever prepared in Italy. The type and quality of the information collected and the methodologies and techniques used to make the inventory are discussed. Possible applications and future developments are also presented.
Introduction
Landslides and floods are among the most destructive natural disasters. The yearly cost of mass movement worldwide can be estimated in billions of US dollars (Alexander 1989, Swanston and Schuster 1989, Schuster and Fleming 1986, Taylor and Brabb 1986). In Italy, regional landslide events have been extremely costly. The July 1987 catastrophic rainfall in the southern Alps caused damages estimated at 1.4 billion US dollars (Guzzetti and others 1992). Single landslide events were also extremely costly. The Vajont slide of 9 October 1963 claimed 1,917 lives and cost more than 6 million US dollars; the Ancona landslide of 13 December 1982 caused damages estimated at one billion US dollars; and the damages caused by the M. Zandilla rock avalanche of 28 July 1987 were estimated at 800 million US dollars (Catenacci 1992, Alexander 1989).
Flood damage is also expensive worldwide. Floods in Asian countries cost more than 5 billion US dollars annually a decade ago, and since then the figures have risen steadily (Sehmi 1989). In Italy in the last 50 years, catastrophic floods along the Po River in November 1951 and the Arno River in November 1966 left thousands of people homeless and cost more than 13 million US dollars.
Government and research institutions worldwide have been working for years to assess hydrogeologic hazard and risk, and to find effective remedial measures (Brabb and Harrod 1989, Starosolszky and Melder 1989). Small scale nationwide landslide maps have been prepared for several countries through bibliographical research (Brabb and Harrod 1989) or by interpretation of medium and small scale aerial photographs (Cardinali and others 1990). Data base information on historic landslides have also been produced. Alger and Brabb (1985) collected more than 6,500 scientific papers, technical reports, maps, and newspaper articles related to mass movement in the U.S. In Italy, Guida and others (1979) listed about 950 papers on landslides and hydrogeologic hazards published from 1900 to 1978. Recently, the International Geotechnical Societies UNESCO Working Party on mass movement has proposed a world inventory of historic landslides as a contribution to the International Decade of Natural Disaster Reduction (IDNDR) "to help the U.N. Education Scientific and Cultural Organization and UNDRO in understanding the world distribution of landslides" (UNESCO 1990).
Within the framework of the IDNDR, the AVI project (an Italian acronym for Aree Vulnerate Italiane, Areas Affected by Landslides or Floods in Italy) was commissioned by the Minister of Civil Protection of the National Group for Prevention of Hydrogeologic Hazards (GNDCI) of the Italian National Research Council (CNR) to compile an inventory of information on areas historically affected by landslides and floods in Italy. An earlier inventory of mass movement was made at the beginning of the century by R. Almagią in which he published two reports listing the occurrence of landslides in the northern and south-central Apennines during the period 1103-1908 (Almagią 1907, Almagią 1910). He analyzed 79 scientific and technical reports and identified 504 landslide events, 222 in the northern Apennines and 282 in central and southern Italy.
In 1957 and 1963, inventories of landslides were made by the Consiglio Superiore dei Lavori Pubblici, an office of the Italian Ministry for Public Works. The Minister asked the local offices of the Genio Civile (the Publik Work Office) to investigate the occurrence of landslides within their provinces and to provide information on the type of movement and the extent of damage. The information gathered by the local offices was summarized in a report of landslide events that described their characteristics and the extent of damage on national and regional levels. In the 1957 inventory, 1,987 mass movements were reported; the number increased to 2,685 in 1963. A comparison between the two inventories was presented and discussed (MM LL PP 1965).
Catenacci (1992) published a comprehensive report on geological catastrophes (landslides, floods, earthquakes, volcanic eruptions and groundwater contaminations) that occurred in Italy during the period 1945-90. He identified landslides and floods in 4,568 municipalities (56% of total) that claimed 3,488 lives and cost the Italian government more than 33,000,000 million lire (20,000 million US $).
The AVI Project
On 23 June 1989, the Minister of Civil Protection requested the National Group for Prevention of Hydrogeologic Hazards to complete an inventory of historic hydrogeological events. The Minister's request followed several meetings of a government committee on natural and human-induced risks (the Commissione Grandi Rischi) that urged the Ministry of Civil Protection to complete an inventory aimed at defining the hydrogeologic risk for the entire nation.
In the early months of 1990, the Director of GNDCI asked a small group of experts to prepare a feasibility plan for the AVI project. The Scientific Committee of GNDCI and the Earth Science Committee of CNR approved the plan and the budget of 2.6 billion lire (1.6 million US $) on July 1990. On June 1991, CNR signed the research contracts with the external contractors and on 6 December 1991, in a meeting held at the Ministry of Civil Protection, the research activities began. The project ended a year later on 9 December 1992.
The working plan of the AVI project provided for a Central Coordination Unit (CCU), originally with CNR-IRPI in Cosenza and later moved to the CNR-IRPI in Perugia, whose task it was to organize, oversee and direct the activities of 14 regional and one national Research Team (RT), and of two CNR support and analysis groups. The CCU was the link between the Director and Scientific Committee of GNDCI, the CNR administrative offices, and the inspectors. Midway and at the end of the project, the work performed by each RT was tested. The CCU was also responsible for synthesizing the information collected on a regional basis by the RTs, and for producing synoptic maps and a final report at the end of the project (Cardinali and others 1993).
Each regional RT collected, organized and summarized the information on mass movement within one or two regions, and on floods within several hydrologic basins, covering approximately the same territory. The investigation was guided by a framework based on two levels (Table 1). The first level aimed at the widespread collection of information through interviews, systematic review of newspapers, and analysis of technical reports, scientific papers, monographs, and books. The information was stored in standard forms: news-forms (S0), for single newspaper articles; source-forms (S1), for data obtained from a single source, i.e., a single newspaper issue, a book, or a technical report; and finally, event-forms (S2), summarizing all the information available on a single hydrogeologic event. News (S0) found in the same newspaper issue and regarding a single event, were combined into a source-form (S1). Source-forms regarding a particular event were then merged to compile an event-form (S2) (Figure 1).
Event forms (S2) contain a variety of information and are quite complex. They consist of more than 140 fields, divided into 6 sections (Table 2). Such complex forms have disadvantages, such as the complexity of the related data base and a certain degree of dishomogeneity, but these forms are preferred to a simpler form with a limited number of fields because they account for all the possible types of data that could be collected for a hydrogeologic event. Despite all efforts, a limited number of peculiar or local data were overlooked. This information was not discarded but was written in the note-field at the end of the form.
The second level of investigation aimed at refining the knowledge of a limited number of particularly important events (8% of the total). For these events, new data were added to the event-form to produce a more detailed analytical-form (S3). Data were obtained mostly through interviews and the review of unpublished technical documents. At the end of the second level of investigation, the RTs prepared regional reports, listing the most damaging hydrogeologic events and summarizing the results of their research activity, the problems that were found and the limitations of the results.
Besides the 14 RTs acting at the regional level, the plans of the AVI project provided a RT working at the national level and two CNR support groups located in Perugia and in Cosenza. The tasks of the national RT were to review a weekly national newspaper (La Domenica del Corriere) as a possible backup for inconsistencies or flaws in the regional inventories, and to complete a research of historic events occurring before 1918, the year from which the systematic investigation of news articles began. The two CNR research teams developed the software used to enter and store the event-forms into the data base; summarized the information on maps at different scales; carried out experiments on possible applications of the data collected by the AVI project; and proposed methods and techniques to improve, maintain and distribute the information.
Results
The most important results of the AVI project are summarized as follows:
The inventory was carried out by 15 research teams that systematically researched 22 regional and national newspapers for the period 1918-90. About 350,000 newspaper issues were screened and 39,953 articles were collected. Of these about 54.5% were related to landslides and floods, and the remaining 45.5% to other catastrophic events, such as snowstorms, snow-avalanches, sea storms, hail storms, salt intrusions, earthquakes, mining and tunnelling failures, etc. The RTs interviewed 150 university professors, CNR researchers, and other government and private experts on mass movement and floods, and analyzed 1,482 published and unpublished technical and scientific reports. The information collected from these sources was summarized in 18,849 source-forms (S1), 9,973 event-forms (S2) and 707 analytical-forms (S3).
Figure 2 summarizes the geographical distribution of data and shows the number of forms prepared for each region. Event forms and analytical forms are further subdivided according to their type. The information obtained from newspapers varies greatly from region to region because the journals chosen for the investigation varied in importance, circulation, and reliability. With only a few exceptions, for Umbria-Marche, Campania- Calabria, and Liguria-Lombardia regions, the number of event-forms was less than originally expected. The difference may be due to difficulties encountered by the RTs in finding enough information to fill an event- form (S2), or to specific problems experienced by the RTs within their regional or local socio-economical environments.
It is difficult to estimate the average cost of collecting the information. As a first estimate, if one takes the budget of each RT, divides it by the number of forms prepared, and then averages the results, the estimate is 63,000 lire (US $ 39) for a news-form, 100,000 lire (US $ 62) for a source-form, and 245,000 lire (US $ 153) for an event-form. The average cost of an analytical-form is estimated at about 2 million lire (US $ 1,250).
The regional reports
The 14 regional RTs, following a detailed outline provided by the CCU, produced regional reports summarizing all the findings, the methodological and technical limitations and the problems encountered during the research activity. These reports constitute one of the most important products of the AVI project. They provide a comprehensive summary of the activities carried out by each RT, and they allow for objective comparisons between different administrative regions as well as different morphological, geological, and socio- economical environments. They list all sources of information, and in particular the names and affiliations of the experts and professionals that were interviewed, the newspapers that were searched, and the published and unpublished reports, the scientific papers, and the books and monographs that were reviewed for the inventory. The reports also list, by date and by magnitude, the most damaging hydrogeologic events that occurred in each region.
The computer database
The RTs, using simple user-friendly software provided by the CCU, prepared regional data bases for landslides and floods. These data bases, a digital replica of the event-forms (S2), were successively merged into two national data bases, for mass movement and floods, containing about 30 Mbytes of data, that constitute the digital archive of the AVI project .
To handle such a large amount of information, the CCU designed a data-base managing-system with a powerful point-and-click graphical interface that allows for complex questions. The software, written for MS- Windows1 3.0 in the C language, fully supports DDE (Digital Dynamic Exchange), allowing for complex dynamic links with other commercial software such as Microsoft Word1 and Excel1 Macro-functions were written for this popular software that automatically draw histograms and charts, or that fill and print pre- defined, standard reports.
The data-base managing-system proved to be an efficient way of storing and retrieving the vast amount of information collected by the AVI project. It allows for complex analyses such as the definition of the geographical and economical extent of large events that caused damage in several regions, and the listing and statistical analysis of events by province, municipality, and basin boundary. The data-base was used to select the geographical coordinates of event and analytical forms used to prepare maps showing the distribution of the information.
Cartography
Four sets of maps were prepared, at different scales, to portray the location of the areas affected by
landslides or floods. These were:
An inventory of events occurring before 1918, the year from which the systematic review of newspapers began, was carried out by examining records in 8 universities, CNR, and public libraries. From 54 books and monographs, a total of 1,316 events, (1,007 landslides and 309 floods) were entered into a data base. The data base contains reference data, such as the title, date and page number; the availability of the document, as well as general information on the hydrogeologic event; the date of occurrence; and the locality, including the province.
The second product is a bibliography of technical reports, scientific papers, books, and monographs collected by the RTs during the investigation. The references are stored in GEOREF1 a bibliographical data- base.
Discussion
A comprehensive study of the vast amount of information collected by the AVI project is beyond the aims of this report. The following discussion involves the type and quality of the information, with emphasis on the advantages and limitations of the methodology, techniques used to make the inventory, and possible application to future developments.
The three main sources of information used to complete the inventory were newspapers, interviews, and the review of technical and scientific documents. Various sources of information provided data of different quality and in different amounts. About 60% of the total information came from newspapers, 30% from technical and scientific documents, and the remaining 10% from interviews.
Of about 40,000 newspaper articles collected, slightly more than half (54.5%) were directly related to natural landsliding or flooding. The remaining articles discussed other natural geological and meteorological catastrophes (earthquakes, snowstorms, snow avalanches, sea storms, hailstorms, etc.), and a few human- induced events, such as mining or tunnelling failures.
As expected, journals emphasized large-magnitude events that occurred in urban areas, or that caused damage to well known or easily recognizable structures. They under-reported events of low magnitude or those that did not cause extensive or well-defined damage. This bias limits the definition of the full extent of the hazard in agricultural and rural areas. Newspapers also emphasized the reactivation and repetition of mass movement and floods.
The amount and reliability of information found in newspapers improved after World War II, and increased markedly during the 1950s, when several journals introduced regional and local chronicles. The review of newspaper articles provided quality data on the dates (but not the time) and the general triggering mechanism (rainfall, earthquake, etc.) of occurrences. The exact location was rarely reported and only for single, large events. Commonly, large areas were described where numerous landslides or bank overflows occurred. Economic estimates of the type and extent of damage were provided in a few articles, but geotechnical data, such as the type of movement and the kind and volume of material involved, were seldom reported.
Review of technical and scientific documents provided a wealth of high quality data for a small number of events. These documents described the geological, morphological and geotechnical characteristics of a single mass movement, and the hydrological and meteorological characteristics of particular flooding events. Most of them contain maps and drawings describing in detail the location and the geometrical characteristics of landslides and the extent of inundated areas. Scientific papers rarely describe the types and extent of damage; such information was more abundant in the technical reports and in the rare books or monographs written on particular, highly catastrophic events. Social and economic considerations were in general very scarce.
Interviews with experts in the fields of mass movement and floods provided general information on a limited number of hydrologic events. The interviews were useful in giving a comprehensive overview of the areas historically affected by landslides or floods, and in defining the awareness of the scientific and technical communities to the problem of hydrogeologic hazard in each region.
As already pointed out, the amount and quality of the information collected from the different sources varied largely between regions, and locally within the same region. The geographical dishomogeneity is related to geomorphologic and socio-economic differences as well as to the methodology applied and the techniques used to make the inventory. Discrepancies between regions reflect geomorphologic environments with a different predisposition to landsliding and flooding, as well as different socio-economical and institutional awareness of the problem.
Applications
Efforts were made to check the reliability of the data collected by the AVI project and to test possible applications. For the Umbria-Marche region, extending for about 20,000 km2 in East-Central Italy, a comparison was made between the AVI data base and small-scale landslide inventory maps available for the entire territory. The comparison was particularly meaningful because the two data sets, prepared using different techniques, aimed at the same general goal: the definition of the regional distribution of landslide hazard.
The small-scale landslide inventory maps were completed by CNR-IRPI in the late nineteen eighties as a part of a regional project aimed at recognition of landslide hazard and risk over large areas. Maps at 1:100,000 scale were produced through the interpretation of 1:33,000 scale, black and white aerial photographs made during flights in the 1950s. Limited field checks were performed (Guzzetti and Cardinali 1989a, 1989b, Antonini and others 1993).
This reconnaissance survey revealed about 27,000 landslides of different extent, type, age and degree of activity, covering a total of about 1,800 km2, 9% of the Umbria-Marche territory. Due to the reconnaissance techniques used to make the inventory, the scale of the maps and the date of the photographs, mass movement too small to be mapped at 1:100,000 scale was not shown on the maps. Landslides mapped in the inventory are the result of possibly more than 10,000 years of geomorphologic evolution of slopes.
For the same territory, a total of 1,065 landslide events (S2 forms) occurring in the last eighty years were inventoried by the AVI project. This represents a small fraction (less than 4%) of the amount of mass movement in the area.
Due to the different techniques used to complete the two inventories and their different time span, a direct comparison of the two products is difficult. Most of the major landslide events shown in the AVI data base are also reported in the small-scale reconnaissance inventory. Discrepancies occur because some events were too small to be mapped at 1:100,000 scale, too subtle to show on the aerial-photographs, or occurred after the date of the photographs used to carry out the photo-interpretation. Moreover data collected by the AVI project concentrate in and around cities and towns historically affected by landslides, in areas where the interpretation of aerial photographs is more difficult.
To carry out a more meaningful comparison an isopleth map (Schmid and MacCannell 1955, DeGraff 1985), showing the distribution of landslide density, was prepared counting the percentage of area affected by landslides within a circular moving window of about 1 km2 (Figure 3a). In landslide-prone areas, landslide density varies form 0.01 (1 ha per km2) up to 1.0, where the whole area is affected by mass movement. Seventy per cent of data collected by the AVI project fall in landslide-prone areas. Of these about 42% fall in the 1-10% density class, 52% in the 11-60% density class, and about 5% in the 60-100% density class.
The comparison of the general distribution of landslide-prone areas (Figure 3a) shows that only the small-scale inventory, prepared through the systematic reconnaissance of aerial photographs, provides a reasonable estimate of the density of landslides in the Umbria-Marche region. Data collected by the AVI project (Figure 3b) concentrate in and around cities and towns historically affected by landslides, or in limited areas for which detailed geomorphologic studies have been carried out by CNR, universities and other institutions. Data in the AVI data base are not homogeneously distributed and suffer from bias due to sources of information and the methodology used to make the inventory. The AVI data base attempts to show the regional distribution of active mass movement, but inherent in the data base is the dishomogeneity in the spatial distribution of the information and in particular the bias towards rural and agricultural areas.
Although the AVI data base cannot be used to describe the distribution of landslides in any given area, it contains useful information pertaining to recurrence and the magnitude of meteorological events that have triggered landsliding and flooding. This improves the capability of assessing landslide hazard, defined as the probability of occurrence of mass movement of a given magnitude, over a given period of time (Varnes and others 1984). Most landslide hazard assessments do not take time into account, therefore, the possibility of defining recurrence intervals for meteorological events that may trigger mass movement becomes important (Carrara and others, 1991; 1992). An experiment, derived from the AVI inventory, was carried out in the upper section of the Tiber River basin in central Italy to relate occurrence of mass movement and flooding with the hydrological characteristics of the triggering meteorological events. Mean daily-discharge values at the Ponte Nuovo gauging station (Servizio Idrografico Nazionale, 1925-1980), where the basin has an area of 4,147 km2 and a lag time of 18 hours (Corradini and others, 1986), were digitized for the period 1925-1941 and 1951- 1980. Discharge values were charted in 5-year intervals, and 532 meteorological events exceeding 100 m3sec-1 of mean daily-discharge were identified. For each event an estimate of the approximate flood volume was also computed.
A meteorological event was broadly defined as a series of consecutive days (from 1 to 20 days) having mean daily discharge exceeding 100 m3sec-1. The highest mean discharge value was chosen as representative of the whole series, and was used to assign the event to the appropriate discharge class. Multiple flooding events having more than one peak discharge value were treated as single events. The flood volume was estimated by adding all the daily discharge values (in m1), multiplied by the number of seconds in a day.
Figure 4 shows the relationships between the hydrological characteristics (mean daily- discharge and total flood volume) of the meteorological events and the occurrence of landslides or floods. At the Ponte Nuovo gauging station, of all the events that exceeded 400 and 700 m3sec-1, 50% and 90%, respectively, triggered landslides or caused flooding somewhere in the basin. Virtually all events with a total flood volume greater than 250 million m3 triggered mass movement and caused flooding. For medium to small magnitude events, the occurrence of landsliding can be described by the estimated flood volume, an indirect measure of the hydrogeologic behavior of the entire basin and of its slopes. Mean daily discharge, a measure of flood level, indicates the probability of flooding.
These considerations can be used to define the probability that a meteorologic event of a given magnitute will trigger a hydrogeologic catastrophe. However, it will not be possible to define the exact time or location of the expected mass movement or flood. The data of the AVI inventory may be used to estimate the recurrence interval of catastrophic events. Information on landslides or floods were available for about 56% of the 532 events considered for a period of 36 years. Only about 5% of the meteorologic events triggered numerous landslides and/or caused extensive flooding in the basin.
Final Remarks
The AVI inventory constitutes the most comprehensive data base on hydrogeologic events ever prepared in Italy. More than 300 people, in 15 regional and national research teams, worked for one year making an inventory of more than 9,300 landslides and floods. In spite of the limitations, the AVI inventory represents an important source of information for future regional hazard and risk assessment. The Minister's ordinance made explicit reference to the fact that the AVI project was intended to be the first step toward defining and mapping hydrogeologic risk in Italy.
A detailed analysis of the vast amount of information collected will certainly require much effort and a considerable amount of time. The study of large hydrogeologic events that triggered landslides and caused flooding over several regions, as well as the analysis of the hydrogeologic conditions in the different physiographic and political regions, will be of help in the assessment of small-scale hydrogeologic hazard. Detailed projects aimed at defining hydrogeologic risk can benefit from the information on areas repeatedly affected by catastrophic events, as well as on the type and extent of damage to structures vulnerable to various types and magnitude of events.
Currently, these projects include attempts to improve and complete the data base, to investigate techniques and methodologies for periodic upgrading, and to produce flooding and landsliding maps. Reports summarizing, mainly in graphs, the extent of the hydrogeologic problem within each political region are also being prepared for national and local civil protection authorities.
Literature Cited
Alexander, D. 1989. Urban landslides. Progress in Physical Geography. 13: 157-191.
Alger C.S. and E.E. Brabb. 1985. Bibliography of United States landslide maps and reports: U.S. Geological Survey Open-File Report 85-585, 119 p.
Almagią R. 1907. Studi geografici sopra le frane d'Italia. Memorie della Societą Geografica Italiana 13.
Almagią R. 1910. Studi geografici sopra le frane d'Italia. Memorie della Societą Geografica Italiana 14.
Antonini G., M. Cardinali, F. Guzzetti, P. Reichenbach, and A. Sorrentino. 1993. Carta Inventario dei Fenomeni Franosi della Regione Marche ed aree limitrofe.GNDCI publication 580 Map at 1:100,000 scale. In press.
Brabb E.E. 1984. Innovative approaches to landslide hazard and risk mapping. Proceedings IV ISL, 1:307- 323.
Brabb E.E. and B.L. Harrod (eds.). 1989. Landslides: extent and economic significance. Proceedings 28th International Geological Congress: Symposium on Landslides, Washington D.C., 17 July 1989, 385 p.
Cardinali M., F. Guzzetti, and E.E. Brabb. 1990. Preliminary map showing landslide deposits and related features in New Mexico. Maps at 1:500,000 scale, U.S. Geological Survey Open File Report 90- 293.
Cardinali M., F. Guzzetti, and P. Reichenbach. 1993. Progetto AVI - Relazione di sintesi. Consiglio Nazionale delle Ricerche - Gruppo Nazionale per la Difesa dalle Catastrofi Idrogeologiche, 39 p.
Carrara A., M. Cardinali, R. Detti, F. Guzzetti, V. Pasqui, and P. Reichenbach. 1991. GIS techniques and statistical models in evaluating landslide hazard. Earth Surface Processes and Landforms 20:427-445.
Carrara A., M. Cardinali, and F. Guzzetti. 1992. Uncertainty in assessing landslide hazard and risk. ITC Journal 1992(2):172-183.
Catenacci V. 1992. Geological and geoenvironmental failure from the post-war to 1990, Italy. Memorie Descrittive della Carta Geologica d'Italia 47:301 p.
Corradini C., F. Melone, L. Ubertini, and V. P. Singh. 1986. Geomorphic approach to synthesis of direct runoff hydrograph from the Upper Tiber River basin, Italy. Pages 57-79 in V.K. Gupta and others (eds.). Scale problems in hydrology, D. Reidel Publishing Co. 57-79.
Schmid F.C., and E. H. MacCannell. 1955. Basic problems, techniques, and theory of isopleth mapping. American Statistical Association Journal 50:220-239.
DeGraff J. V. 1985. Using isopleth maps of landslide deposits as a tool in timber sale planning. Bulletin Association of Engineering Geology 22:445-453.
Guida M., G. Iaccarino, G. Metcalf, and A. Vallario. 1979. Bibliografia delle frane dal 1900 al 1978. C.N.R., IT ISSN 0085-2309. Rome. 4:362 p.
Guzzetti F., and M. Cardinali. 1990. Landslide inventory map of the Umbria region, Central Italy. 6th ICFL - ALPS 90, Milan, Italy, September 12, 1990. 273-284
Guzzetti F., and M. Cardinali. 1989. Carta Inventario dei Fenomeni Franosi della Regione dell'Umbria ed aree limitrofe. G.N.D.C.I. pub. n. 204. Map at 1:100,000 scale.
Guzzetti F., G. Crosta, M. Marchetti, and P. Reichenbach. 1992. Debris flows triggered by the July, 17-19, 1987 storm in the Valtellina area Northern Italy. Proceedings Interpraevent 1992 2:193-204.
Hansen, A. 1984. Landslide hazard analysis. Pages 523-602 in D. Brunsden and D.B. Prior (eds.). Slope Stability, John Wiley & Sons, New York.
MM LL PP 1965. I Movimenti Franosi in Italia. Consiglio Superiore dei Lavori Pubblici, Direzione Generale dell'A.N.A.S., Servizio Tecnico. 176 p.
Radbruch-Hall D.H., R.B. Colton, W.E. Davis, B.A. Skipp, I. Lucchitta, and D.J. Varnes. 1976. Preliminary landslide overview map of the conterminous United States. U.S. Geological Survey M.F.S. Map 771, map at 1:7,500,000 scale.
Reichenbach P., W. Acevedo, R.K. Mark, and R.J. Pike. 1992. Landforms of Italy. CNR-GNDCI, special project AVI, map at 1:1,200,000 scale.
Reichenbach P., W. Acevedo, R.K. Mark, and R.J. Pike. 1992. A new landform map of Italy in computer- shaded relief. Geodesia Bulletin. 52:21-44.
Schuster R.L., and R.W. Fleming. 1986. Economic losses and fatalities due to landslides. Bulletin Association of Engineering Geology 23:11-28.
Sehmi N.S. 1989. The hydrology of disastrous floods in Asia - an overview. Pages 106-122 in Starosolszky O. and Melder O.M. (eds.). Hydrology of Disasters, WMO.
Servizio Idrografico Nazionale 1925-1980. Annual hyrographic reports.
Starosolszky O., and Melder O.M. (eds.).1989. Hydrology of Disasters, WMO, Proceedings Technical Conference, Geneva, November 1988. 319 p.
Swanston D.N., and R.L. Schuster. 1989. Long-term landslide hazard mitigation programs: structure and experience from other countries. Bulletin Association of Engineering Geology 26:109- 133.
Taylor F., and E.E. Brabb. 1986. Map showing landslides in California that have caused fatalities or at least $ 1,000,000 in damages from 1906 to 1984. U.S. Geological Survey Miscellaneous Field Studies Map MF-1867.
UNESCO International Geotechnical Societies' Working Party on World Landslide Inventory. 1990. A suggested method for reporting a landslide. Bulletin International Association of Engineering Geology 41:5-12.
Varnes D.J. and Commission on Landslides and Other Mass movement IAEG. 1984. Landslide hazard
zonation: a review of principles and practice. Unesco Press, Paris. 63 p.
Figure and Table captions Not available here, sorry!
Figure 1. Scheme for the compilation of an event form (S2) from different sources of information.
Figure 2. Map showing the geographical distribution of data collected by the research teams.
The number of news (S0), source (S1), event (S2) and analytical (S3) form is given. For event
and analytical forms the number of landslides and floods in each region is given.
Figure 3. Comparison of the general distribution of landslides in the Umbria-Marche region
(Central Italy). Map A: Isopleth map showing the density of landslide area. Map B:
Geographical distribution of landslides inventoried by the AVI project. 1. Landslide density
less than 1%. 2. Landslide density between 1 and 10%; 3. Landslide density between 11 and
30%; 4. Landslide density between 31 and 60%; 5. Landslide density between 61 and 100%;
6. Landslide location.
Figure 4. Relationships between the hydrological characteristics of 532 meteorological events
and the occurrence of landslides and/or flood in the Upper Tiber River basin in the period
1925-1980. Horizontal heavy lines: thresholds of mean daily discharge. Vertical heavy line:
threshold of estimated flood volume. Dashed-dotted and dashed lines separates hydro-
meteorological conditions that can cause, respectively with high and low probability,
landslides and/or floods.
TABLE 1. Logical framework of the AVI project.
TABLE 2. Outline of an event-form (S2).
he AVI project are summarized as follows: