ENVIRONMENTAL ASSESSMENT

Procedures for the Documentation of Historical Debris Flows:Application to the Chieppena Torrent (Italian Alps)

Lorenzo Marchi Æ Marco Cavalli

Received: 11 August 2006 / Accepted: 22 February 2007

� Springer Science+Business Media, LLC 2007

Abstract The reconstruction of triggering conditions,

geomorphic effects, and damage produced by historical

floods and debris flows significantly contributes to hazard

assessment, allowing improved risk mitigation measures to

be defined. Methods for the analysis of historical floods and

debris flows vary greatly according to the type and quality

of available data, which in turn are influenced by the time

the events occurred. For floods and debris flows occurring

in the Alps a few decades ago (between about 1950 and

1980), the documentation is usually better than for previous

periods but, unlike events of most recent years, quantitative

data are usually scanty and the description of the events

does not aim to identify processes according to current

terminology and classifications. The potential, and also the

limitations of historical information available for the

reconstruction of historical debris flows in the Alps have

been explored by analyzing a high-magnitude debris flow

that occurred on November 4, 1966 in the Chieppena

Torrent (northeastern Italy). Reconstruction of the event

was based on the use of written documentation, terrestrial

and aerial photographs, and geomorphological maps. The

analysis aimed to define the temporal development of

phenomena, recognizing the type of flow processes and

assessing some basic flow variables, such as volume,

channel-debris yield rate, erosion depth, total distance

traveled, and runout distance on the alluvial fan. The his-

torical development of torrent hydraulic works, both before

and after the debris flow of November 1966, was also

analyzed with regard to the technical solutions adopted and

their performance.

Keywords Debris flow � Historical documents �Alluvial fan � Channel erosion � Torrent Control Works �Alps

Introduction

Debris flows and flash floods in mountainous headwaters

are common in Italy, as well as in many other countries,

and cause the loss of lives and economic damage.

Knowledge of past floods and debris flows may have a

major role in hazard assessment and in the definition of

mitigation measures. Among the most important outcomes

of reconstruction of historical floods in headwater basins

are the following:

• identification of areas affected and damage caused;

• assessment of the frequency of events;

• recognition of the type of flow processes, with partic-

ular attention to discriminating water floods with

sediment transport from debris flows;

• evaluation of the effectiveness of mitigation measures.

The use of documentary evidence, both published and

unpublished, gives an important contribution to our knowl-

edge of historical floods and debris flows (Eisbacher

and Clague 1984, Tropeano 1989, Brazdil and others 1999,

Barnikel and Becht 2003, Barnikel 2004, Cœur and others

2002, Tropeano and Turconi 2004). Documents from

historical archives can be integrated with information from

geomorphological, sedimentological, and paleo-hydrologi-

cal studies.

The relative importance of data from historical docu-

ments compared to other methods for the reconstruction of

past floods and debris flows depends on the quality of

available information, which, in turn, is affected by several

L. Marchi (&) � M. Cavalli

CNR IRPI Padova, Corso Stati Uniti 4, 35127 Padova, Italy

e-mail: lorenzo.marchi@irpi.cnr.it

123

Environ Manage (2007) 40:493–503

DOI 10.1007/s00267-006-0288-5

factors, in particular by the time of occurrence. In many

cases, the documents concerning ancient events report just

the date of occurrence and, sometimes, summary infor-

mation on the damage caused, whereas detailed informa-

tion on temporal development and characteristics of the

events occurred in the last decades often exists, together

with some quantitative data.

Benito and others (2004) present a review of the meth-

ods that can be used, in an integrated way, in reconstructing

river floods in order to improve risk assessment. These

authors point out the extra value represented by documents

that allow the assessment of social and economic impact

caused by past floods, which had a much greater intensity

than events reported in the instrumental series.

From a space scale viewpoint, two approaches can be

distinguished: 1) the analysis of historical events on a re-

gional scale, where the vastness of the collection offers

samples that can be used for statistical elaborations, even if

single cases cannot be studied in depth; 2) the detailed

study of a single basin or set of adjacent basins. The

temporal scale of the study can depict either a single event

or the reconstruction of a complete historical series of

floods in a certain area (Table 1). Obviously, there are also

intermediate situations in which detailed studies on par-

ticularly interesting basins are accompanied by a regional-

scale analysis, or studies dedicated to a single flooding

event that are subsequently integrated with data, albeit a

summary, of previous events.

One of the earliest investigations on collection and anal-

ysis of archive data for the study of flash floods and debris

flows in the Italian Alps was carried out by Govi (1975). In

the following decades, this topic was considerably devel-

oped, leading to the setting up of a national database of

documents on floods and landslides (Guzzetti and others

1994, Guzzetti and Tonelli 2004). The reconstruction of time

series of flash floods and debris flows in single catchments

(class 3 in Table 1) and in wider regions (class 4) provides an

important integration to the detailed analysis of specific flow

events and to studies for hazard assessment. There are fewer

studies dedicated to the reconstruction of a particular event

in a single catchment (class 1) or in relatively limited areas.

Among studies that integrate different methods for the

reconstruction of a historical flood in a medium-sized area in

Italy, a study that is worthy of note is a recent monograph by

Esposito and others (2004) about the flood occurring in

October 1954 in a coastal area near Salerno (southern Italy).

Particularly severe events (for a historical flood from a large

river, Turitto (2004)) are sometimes the object of studies on a

regional scale (class 2).

This paper aims to illustrate the most important meth-

odological aspects concerning the reconstruction of a

debris flow on a single basin scale by taking into account

the Chieppena Torrent (northeastern Italy). The event

studied is the debris flow of November 4, 1966, which was

the most recent and one of the most serious disasters that

occurred in this basin. The collection of basic information

on past flooding events in this basin complements the

reconstruction of this event.

Methods

Bibliographic sources and historical documents and maps

were collected and analyzed to identify the frequency of

past floods and debris flows and their relevant character-

istics. The evidence of the 1966 event consisted of ac-

counts of witnesses reported in local newspapers,

monographs, scientific papers, and photographs. These

were used to reconstruct the temporal development of the

flow processes, their typology, and the determination of

some quantitative parameters. Some empirical equations

were used to analyze basic quantitative parameters of the

November 1966 debris flow.

Basin Studied

The Chieppena Torrent is a stream of the Eastern Italian

Alps and a tributary of the Brenta River (Fig. 1 and Ta-

ble 2). The Cinaga Torrent joins the Chieppena Torrent on

the alluvial fan, so that sediment dynamics in its basin are

independent from those of Chieppena Torrent basin.

The Chieppena Torrent follows a fault line, which sep-

arates magmatites and metamorphites in the north from

sedimentary rocks in the south (Fig. 2). The northern part

of the basin consists mostly of granite. In the central part of

the basin, metamorphic rocks (phyllites and cornubianites)

crop out. Dolomites and limestones occupy most of the

southern part of the basin where other sedimentary rocks

are also found, in particular the arenaceous complex of Val

Gardena sandstones and the Bellerophon Formation (an

alternation of silty marls or sandstones with gypsum nod-

ules). Quaternary deposits are widespread in the basin and

are the main source of sediments for the debris flows of the

Chieppena Torrent. They consist of moraines, partly re-

worked as fluvioglacial deposits, scree, and alluvial

deposits.

Table 1 Space and time scales in the historical analysis of torrent

hazards

Space scale

Single

catchment

Region or large

river basin

Time scale Single event 1 2

Multiple events—

Time series

3 4

494 Environ Manage (2007) 40:493–503

123

Coniferous forests cover large areas between 900 and

1600–1700 m a.s.l. Deciduous forests are located along the

middle and lower parts of the streams. Meadows and

farming areas are present near urban settlements. Alpine

grasslands and bare ground (tussock, scree, and outcrop-

ping rocks) occupy vast areas at the highest elevations in

the northern part of the basin. Agricultural areas (meadows,

orchards) mixed with urban settlements occupy most of the

alluvial fan. A railway and a state road cross the lower part

of the alluvial fan.

The study area is characterized by an alpine climate with

a Mediterranean influence. Precipitation is relatively

abundant throughout all the year, with maxima in May–

June and October–November. Mean annual precipitation

amounts to approximately 1250 mm.

Historical Events

Table 3 presents basic information on past events reported

or estimated from historical records. For most historical

events, available descriptions of the phenomena made it

possible to classify the flow processes as debris flows or

water floods. The historical events have been classified into

two classes of intensity (high and moderate) on the basis of

the description of the phenomena and reported damage.

The lack of detailed information does not permit assess-

ment of the intensity of the 1564 and 1655 events. It is

important to note that the occurrence of casualties is not

necessarily linked to the severity of an event. As an

example, occasional circumstances caused seven fatalities

Table 2 Principal morphometric parameters of the drainage basin

and alluvial fan

Drainage basin above the fan apex

Area (km2) 27.0

Average slope (%) 54

Average elevation (m) 1333

Maximum elevation (m) 2482

Minimum elevation (m) 459

Length of the main channel (km) 8.9

Average slope of the main channel (%) 18.6

Alluvial fan

Area (km2) 2.84

Average slope (%) 6.8

Average channel slope on the alluvial fan (%) 5.1

Fig. 2 Geological settings of the Chieppena Torrent basinFig. 1 Mapof the study area, showing source areas of the debris flowof

November 4, 1966, extent of deposits, and sites described in the text

Environ Manage (2007) 40:493–503 495

123

during the flood of September 24–25, 1924, for which

available documents depict a relatively moderate intensity.

Historical data show a rather low frequency of high-

intensity events in the Chieppena Torrent basin, with only

three events (August 1851, September 1882, and November

1966) in the last two centuries. The availability of data

regarding minor events in the 19th and 20th centuries should

be ascribed to greater attention for document processing and

better conservation of the documents themselves.

The Event of November 4, 1966

General Situation

The most recent flood in the Chieppena Torrent occurred

on November 4, 1966. The flood of November 1966 in

northeastern Italy was caused by widespread precipitation,

which coupled long duration with the persistence of high

intensity. The most intense phase of the rainstorm lasted

about 40 hours, from the morning of November 3 to the

evening of November 4 (Dorigo 1969). Tonini (1968)

suggested that the principal factor for the severe floods of

November 1966 was the occurrence of a large amount of

precipitation, corresponding to the average monthly total,

in a short time, affecting soils that were already in critical

saturation condition. A further non-negligible contribution,

although not decisive, came from the melting of the snow

cover at elevations above 850–1200 m.

In the same days, extreme rainfall caused widespread

flooding of many rivers in Tuscany, with the infamous

inundation of Florence.

Only one rain gauge (Bieno, Figure 1) was located in

the Chieppena basin. Rainfall recorded in Bieno within

Table 3 Historical floods and debris flows in the basin of Chieppena Torrent

Date Notes Type of flow

event

Estimated

intensity

Reference

1564 Not known Not known Eisbacher and Clague (1984);

Cerato (1999)

1649 Destruction of Gallina bridge

(ponte Gallina, Figure 1)

Not known High Eisbacher and Clague (1984);

Cerato (1999)

1655 Not known Not known Eisbacher and Clague (1984);

Cerato (1999)

August 18–19, 1748 Seven fatalities, damage to

agricultural areas and settlements

Not known High Cerato (1999)

August 30–31, 1757 Debris flow of the Cinaga Torrent,

destruction of houses and four

casualties in Samone

Debris flow High Eisbacher and Clague (1984)

1823 Flooding of the middle and lower part

of the alluvial fan

Water flood Moderate National Archive of Trento

1825 Flooding of the middle and lower part

of the alluvial fan

Water flood Moderate National Archive of Trento

1839 Flooding of the middle and lower part

of the alluvial fan

Water flood Moderate National Archive of Trento

1843 Flooding of the lower part of the

alluvial fan

Water flood Moderate National Archive of Trento

August 3, 1851 Strigno village damaged by

Chieppena and Cinaga and torrents,

deposition of large boulders in the

hamlet of Villa

Debris flow High Eisbacher and Clague (1984);

Cerato (1999)

September 17–18, 1882 The failure of a landslide dam on the

Rio Gallina triggered a major surge

that caused severe damage on the

alluvial fan

Debris flow High Cerato (1999)

September 24–25, 1924 Loss of lives (five on the Chieppena

Torrent, two on the Cinaga Torrent)

Water flood Moderate Local newspaper article reported

by Pedenzini (2001)

November 1, 1928 Erosion of channel bed and banks in

the lower part of the stream. Peak

discharge estimated to about 130

m3s–1

Water flood Moderate Archive of Autonomous Province

of Trento

496 Environ Manage (2007) 40:493–503

123

November 3–4 amounted to 213.7 mm, i.e., about 17% of

mean annual rainfall. In other rain gauges installed in the

vicinity, total rainfalls ranged from 177.2 to 217.2 mm. The

return period of 2 days’ rainfall (November 3–4) has been

estimated to be approximately 100 years.

Time Evolution and Effects Produced

The event that struck the Chieppena basin and alluvial fan

is quoted in various articles dealing with the effects of the

November 1966 floods in northeastern Italy. Some studies

cite the Chieppena Torrent as one of the streams in which

the most serious damage was produced, especially in

inhabited areas located on its alluvial fan (Dorigo 1969,

Castiglioni and others 1971, Croce and others 1971).

Moreover, an in-depth study by Venzo and Largaiolli

(1968) was specifically dedicated to the Chieppena basin.

These authors paid particular attention to the basin’s geo-

logical setting and mass wasting processes, which are

illustrated also by photographs and related captions. Their

monograph is accompanied by a geological map and a

geomorphological map, both at a 1:10,000 scale. Temporal

information available for the event is reported in Table 4.

Gorfer (1967) reports a particularly effective description

of the event: ‘‘… The second surge followed at 19:30

hours. In the streets, devastated and plunged in darkness,

people cried out that there was an earthquake. The second

‘‘wall of water’’ came rushing down, skimming past the

village of Strigno. The landslide broke out on the right

flank towards Villa whereas the water flowed down on the

left towards Agnedo.’’

The flood produced damage to roads, farming areas,

urban and production settlements on the alluvial fan, and

also caused the loss of three lives. In addition, most of the

torrent control works were destroyed.

Typology of Flow Processes

Most reports and papers, which mention the Chieppena

Torrent as one of the streams that produced severe damage

in the November 1966 flood, do not provide a classification

of the flow process that inundated the alluvial fan. In

particular, the basic distinction between debris flows and

water floods with sediment transport (Costa 1988, Pierson

2005) was not taken into account. However, some

descriptions of the event provide us with useful clues

leading to better classification. In particular, Gorfer (1967)

uses the term frana (landslide) to designate the flooding of

the hamlet of Villa, whereas the early occurrence of the

surge is described as a muraglia d’acqua (wall of water).

These two contrasting terms seem to be typical of debris

flows, which are processes intermediate between landslides

and sediment transport by water floods. The description by

Gorfer, reported in the previous paragraph, could be

interpreted as follows: the debris-flow front, particularly

rich in large boulders, moved towards Villa, whereas

Agnedo, located slightly downstream on the opposite side

of the channel (Fig. 1), was affected by more fluid mate-

rials, probably belonging to the debris-flow body following

the front, and to the recession phase.

It should be noted that at the time of the 1966 flood and

in the years immediately afterwards, most debris flows

were not recognized and analyzed as such in Italy. The

poor level of attention to and interest in debris flows and

their specific features had already been stressed by Casti-

glioni in 1971. This scientist, when quoting an Austrian

publication where ‘‘Muren-type phenomena’’ (i.e., debris

flows) were often described, says: ‘‘It might be due to

problems of terminology, since in Italian there is not a

specific word to describe this phenomenon, but I have the

feeling that it is not sufficiently understood or correctly

illustrated, although it is quite widespread also in our Alps

with characteristics that distinguish it both from purely

stream-related processes and mass wasting processes.’’

In more recent years, the event of November 1966 on

the alluvial fan of the Chieppena Torrent has been classi-

fied as a debris flow by Eisbacher and Clague (1984),

D’Agostino and others (1996), Cerato (1999), and

Moscariello and others (2002). The sources available,

which essentially consist of descriptions of the event,

which have been reported in the papers quoted above and

related pictorial material, confirm this interpretation. In

particular, it is possible to state that the second surge,

which was characterized by the transport of huge amounts

of boulders, had the characteristics of a debris flow. The

debris flow that occurred in the Chieppena Torrent can be

Table 4 Time evolution of the event from accounts of witnesses collected by Gorfer (1967, 1977)

Approximate time Observed phenomenon Notes

2.15–2.30 PM Strigno village flooded by the Cinaga Torrent Sudden water flood; damage increased by stream culverting

3.30 PM First surge of the Chieppena Torrent Initiation caused by the failure of a channel blockage

(debris and log dam) near the confluence of Gallina

and Fierollo streams (Figure 1)

7.30 PM Second surge of the Chieppena Torrent Probable failure of a channel blockage

Environ Manage (2007) 40:493–503 497

123

ascribed to granular debris flow, characterized by abundant

coarse material and by a sandy matrix, with limited or

negligible percentage of the finest material (silt and clay).

The occurrence of a granular, cohesionless debris flow

agrees with the prevalence of granite rocks in the source

area (Moscariello and others 2002). In fact, although

channel erosion within the basin has also involved phyl-

lites, the deposits mostly consist of granite boulders and

cobbles. The presence of a sandy matrix greatly varied in

the deposits; matrix-free accumulation of boulders could be

due to the erosion of the sandy fraction in the recessional,

more liquid phase of the event. Some uncertainties exist

about the first pulse, whose deposits were reworked and

partly obliterated by the second surge. The first surge, less

rich in coarse material, might not have attained the con-

centration of a debris flow, and consisted of a hypercon-

centrated flow.

The limits of the deposits on the alluvial fan were set by

using two oblique aerial photographs of the alluvial fan

(views from north and from south), taken immediately after

the debris flow of 1966. The photographs were georefer-

enced and rectified by means of a polynomial transforma-

tion of control points identified on the photographs and on

a georeferenced topographic map. The map of the deposits

is shown in Figure 3. Two classes were distinguished:

deposits with abundant coarse-grained material and a wider

area flooded by water and finer sediments. The first class

corresponds to actual debris-flow deposition, encompassing

direct and indirect impact zones, according to the definition

by Kellerhals and Church (1990), whereas water flooding

with finer sediment, as well as the formation of erosion

tracks, can be related to the recession phase of the event

and to fluvial reworking of debris-flow deposits. Coarse-

grained deposits also display relevant thickness, as con-

firmed by some terrestrial photographs, but available

information does not permit quantitative assessment of the

depth of the deposits. An isolated area of coarse material

(indicated with ‘‘A’’ in Figure 3) can be ascribed to local

topographic conditions, including a railway embankment,

which favored the deposition of a debris-flow surge that

had left only scanty deposits in the upstream path.

Elements for Quantitative Analysis

Quantitative data about the November 1966 debris flow are

scarce. A first approximation of the solid volume deposited

was proposed by Cerato (1999): ‘‘over one million cubic

metres of debris on the alluvial fan where the villages of

Strigno and Villa Agnedo are located.’’ A slightly smaller

amount (950,000 m3) was reported by D’Agostino and

others (1996). These evaluations are necessarily approxi-

mate and do not consider the material conveyed as far as

the Brenta River and removed by the same watercourse.

Although with inevitable uncertainties, we can assess the

total volume of the material deposited on the fan as

approximately 106 m3. A huge boulder, with a size of 10 ·9 · 12 m, was transported for about 900 m and was

deposited a little upstream of the fan apex (Fig. 1).

Quantitative data from event documentation have been

compared with data on other debris flows in Alpine basins

and with the results of empirical equations. The total

deposit volume, assessed as 106 m3, was divided by the

total length of the hydrographic network affected by the

most intense erosions, corresponding to parts of the Gallina

and Fierollo streams and Chieppena Torrent as far as the

starting point of deposition on the alluvial fan. The iden-

tification of the parts of the hydrographic network to be

taken into account was carried out on the basis of the

geomorphological map by Venzo and Largaiolli (1968). It

was therefore possible to calculate the rate of yield of

debris, which is a particularly significant index in assessing

erosion due to debris flows. Figure 4 shows the cumulative

frequency distribution of debris-yield rate for unit channel

length, for a sample of more than 120 debris flows that

have occurred in northeastern Italy since the 19th century.

The value obtained for the Chieppena Torrent (120 m3 m–1)

lies among the highest historically recorded in this Alpine

Fig. 3 Map of debris-flow deposits on the alluvial fan

498 Environ Manage (2007) 40:493–503

123

region. This result is in agreement with high-intensity

erosion processes that affected the Chieppena Torrent on

November 4, 1966.

Available documentation (i.e., a geomorphological map

by Venzo and Largaiolli (1968), and both aerial and ter-

restrial photographs) has enabled the runout distance of the

debris flow on the alluvial fan and total distance traveled

from the initiation point to the lowest point of deposition to

be approximately assessed. The analysis regards the second

surge, whose deposits are visible in the photographs taken

immediately after the event. Measured values have been

compared with the results of the following empirical

equations:

Lf ¼ 8:6 � V � tanJð Þ0:42 ð1Þ

Lf ¼ 15 � V1=3 ð2Þ

Lt ¼ 1:9 � V0:16 � H0:83e ð3Þ

Lt ¼ 30 � V � Heð Þ0:25 ð4Þ

where Lf is the runout distance on the alluvial fan (m), Lt is

the total distance traveled (m), J is the mean slope of the

transportation zone (�), V is the volume (m3), and He is the

elevation difference between the starting point and the

lowest point of deposition (m).

Eq. 1, reported by Ikeya (1989) for Japanese alluvial

fans, was rearranged by Bathurst and others (1997) in the

form presented above. Eqs. 2–4 were proposed by Ric-

kenmann (1999). The exponents of eqs. 2 and 4 satisfy

Froude scaling, whereas eq. 3 best fits the experimental

data (Rickenmann 1999). It has been assumed that the

volume deposited by the second surge amounted to

500,000 m3, i.e., 50% of the total event volume. Because

the starting point of the debris flow is not known for sure,

there is considerable uncertainty regarding the total path

run by the debris flow. Two hypotheses have been con-

sidered: debris flow initiation at the confluence of the

Gallina and Fierollo torrents, and from the upstream point

of the erosion areas mapped by Venzo and Largaiolli

(1968) on the Rio Fierollo (Fig. 1). The results, reported in

Table 5, show an underestimation of values for both vari-

ables analyzed.

In considering the outcome of the comparison between

the data observed and those calculated, it is important to

remember that such formulae produce only approximate

results. Nevertheless, the fact that empirical equations

produce values of Lf and Lt shorter than those arising from

the documentation of the event is not without meaning. A

possible explanation for this can be found in the relatively

low solid concentration which, on the basis of the available

descriptions, the debris flow seems to have had, as well as

in the low viscosity of the solid–liquid mix involved. Both

circumstances might have contributed to the high mobility

of the flowing mass. On the other hand, eqs. 1–4 lead to the

determination of values that reflect the comprehensive

conditions of the samples they are based upon and that

comprise diverse types of debris flows. Some simple

sensitivity tests (Fig. 5) have emphasized that the

assumption relative to the volume of the second surge does

not influence appreciably on the results: the values result-

ing from the application of eqs. 1–4 are lower than those

observed even by utilizing a volume double of the first

reference.

The values of channel erosion depth reported in papers

and monographs describing the event have been compared

with an empirical relationship proposed by Kronfellner-

Kraus (1984) for Austria and used also by Rickenmann and

Zimmermann (1993) for analyzing the debris flow that

occurred in Switzerland in 1987:

D ¼ 1:5þ 0:125 � S ð5Þ

Fig. 4 Debris yield rate for unit channel length in the Eastern Italian

Alps

Table 5 Runout distance on fan and total distance traveled

Lf observed (m) 2170

Lf eq. 1 (m) 1055

Lf eq. 2 (m) 1190

Lt Fierollo observed (m) 8935

Lt Fierollo eq. 3 (m) 6415

Lt Fierollo eq. 4 (m) 4900

Lt confluence Gallina—Fierollo observed (m) 7630

Lt confluence Gallina—Fierollo eq. 3 (m) 3860

Lt confluence Gallina—Fierollo eq. 4 (m) 4200

Environ Manage (2007) 40:493–503 499

123

where D is the depth of channel bed erosion (m), and S

is the channel slope (%).

Channel-bed incision in proximity of Ponte Gallina

(Fig. 1)wasmeasured at about 10m (ProvinciaAutonoma di

Trento 1991). More generally, Venzo (1968) reports ‘‘a

channel erosion of 6–7 m within a few hours’’ for the

Chieppena Torrent upstream of Bieno. Considering that the

channel-bed slope at Ponte Gallina is 28%, the value of 10 m

is higher than the value resulting from eq. 1. The same

outcome results by considering the erosion depth of 6–7 m

reported by Venzo (1968) in the Chieppena channel-bed

upstream of Bieno, which has an average slope of 21%. The

fact that the erosional depths observed are more pronounced

than the values resulting from a relationship based on a large

sample of high-intensity events further emphasizes the great

intensity of the erosional processes occurring along the

Chieppena Torrent during the event of November 1966. It

should be considered, however, that the depth of erosion

analyzed is the cumulative value of the two surges of the

November 1966 event. If we assume that each surge caused

50% of the total erosion observed, the resulting value would

be close to that obtained by means of eq. 5.

Torrent Control Works

Since the 19th century, severe damage produced by the

Chieppena Torrent to settlements, roads, and agricultural

areas has required mitigation measures.

Figure 6a presents a sketch, from an 1848 project,

showing stone masonry check dams being constructed a

little upstream of the alluvial fan. These dams incorporated

large boulders left in the channel bed by previous debris

flows. Other check dams had been designed for upstream

channel stretches. Some control works were carried out

after the 1882 flood, and a more systematic project was set

up in 1913. The implementation of the project was pre-

vented by World War I (the Chieppena Torrent was close to

the Front) but some interventions were carried out in the

1920s, following the guidelines of the 1913 project. A

comprehensive project aiming at complete control of torrent

erosion was prepared in 1927 and was subsequently

implemented. Works continued in the following decades,

until the 1960s. The check dams were built in stone ma-

sonry, using blocks taken from the channel bed (Fig. 6b).

The debris flow of November 4, 1966 almost completely

destroyed the control works in the Chieppena Torrent. The

failure of the containment works implemented before the

event of November 1966 can be ascribed both to inade-

quacy of design and to poor construction quality. Indeed,

the check dams had not been specifically designed to cope

with debris flows. Moreover, stone masonry check dams

could not provide sufficient resistance to the dynamic im-

pact of the flow.

New control works were carried out after the November

1966 event according to design criteria and management

policies prevailing in Italy at that time; these consisted

essentially of bulky concrete dams, with a design channel

slope equal to zero (Fig. 6c). Seventy check dams were

built: the aim of these works was to reduce channel slope in

order to stabilize channel bed and banks, thus preventing

formation and propagation of debris flows. The construc-

tion of a number of large, bulky check dams may appear to

be in contrast with recent trends in torrent management,

which produce control works with less environmental im-

Fig. 5 Sensitivity analysis of

the equations for the runout

distance on the fan and the total

distance traveled. a Runout

distance on fan. b Total travel

length—debris-flow initiation in

Rio Fierollo. c Total travel

length—debris-flow initiation at

the confluence of Rio Gallina

and Rio Fierollo

500 Environ Manage (2007) 40:493–503

123

pact, but they are justified by the need to contain high-

intensity torrent erosion that had caused major damage to

settlements on the alluvial fan. In recent years, an open

check dam has been added to the traditional check dams.

This was built immediately downstream of the confluence

of the Gallina and Fierollo torrents, and aims to prevent

downstream propagation of debris flows that could be

generated in the upper part of these channels (Fig. 6d).

The compact structure of the check dams built after the

November 1966 debris flow ensures much higher resistance

to dynamic pressures than the works destroyed by the 1966

debris flow. Up to now, the works implemented have

contributed effectively to the control of erosion and insta-

bility phenomena along the channel and on the adjacent

side slopes. We should stress, however, that these works

have not yet had to withstand events comparable in

intensity to the November 1966 debris flow.

Discussion and Conclusions

The reconstruction of the November 1966 debris flow in

the basin of the Chieppena Torrent offers two consider-

ations: both the particular case studied and methodological

considerations useful for reconstructing other historical

floods in stream basins of the Alps.

With regard to the first consideration, the reconstruction

of the time evolution of the event was coupled with the

recognition of the type of flow processes occurring in the

basin studied. The characteristics of the second surge lead

to its classification as a granular debris flow, whereas there

is some uncertainty about the first pulse, which could have

consisted of a hyperconcentrated flow. Data available for

characterizing the phenomenon quantitatively depict an

event in which the mobilization of huge amounts of solid

material was matched by a remarkable mobility of the

flowing mass.

A further comment regards the frequency of past events,

assessed on the basis of historical documentation. The

frequency of high-intensity flow events in the Chieppena

Torrent basin is rather low. This seems to depict a rela-

tively stable system, but one that is prone to large debris

flows when particular meteorological conditions trigger

widespread landslides and intense erosion on the channel

bed and banks. The mere availability of loose material

subject to mobilization is not a limiting factor for the

occurrence of debris flows in the Chieppena Torrent. Very

erodible Quaternary deposits are widespread in the basin,

so that the potential debris supply to the channel network

can be deemed unlimited and long recharge times between

two subsequent events are not necessary. With reference to

the classification of debris-flow prone basins into weath-

ering-limited and transport-limited, proposed by Bovis and

Jakob (1999), the large availability of erodible material

would cause the Chieppena Torrent to be classified as a

transport-limited basin. However, the relatively high sta-

bility of basin slopes and minor tributaries, which fails only

in response to high-intensity meteorological events, causes

large debris flows to be less frequent than is commonly

observed in Alpine basins with an unlimited sediment

supply.

The study of the November 1966 event in the Chieppena

Torrent has allowed assessment of the potential and the

limitations of information commonly available for the

reconstruction of debris flows occurring in past decades in

the Italian Alps (Table 6). We can therefore consider that

Fig. 6 Different types of check

dams outlining the historical

development of control works in

the Chieppena Torrent. a A

stone masonry check dam from

an 1848 project (National

Archive of Trento). b A stone

masonry check dam built in the

1930s (archive of Autonomous

Province of Trento). c Concrete

check dams built after the debris

flow of November 1966

(archive of Autonomous

Province of Trento). d An open

check dam built in the early

1990s

Environ Manage (2007) 40:493–503 501

123

the devised procedure, based on collection and analysis of

published and unpublished documents and aerial and ter-

restrial photographs, could be applied to the reconstruction

of other floods and debris flows occurring in past decades.

For floods and debris flows that took place in previous

times (approximately before the 1950s), available infor-

mation is often limited. For older events, aerial photo-

graphs are not available and terrestrial photographs are

usually scarce. As for the most important flooding events,

archive documentation may be sufficient for a reliable

reconstruction of the main characteristics.

The use of 1-D and 2-D models can provide important

contributions to the quantitative analysis of historical

debris flows, allowing reconstruction of flow depth and

velocity and simulations of flooded areas. On the other

hand, empirical evidence of large-magnitude historical

events can represent a challenging test for application and

validation of numerical models. The application of some

empirical equations has been deemed adequate for this

study, which aims to document a historical debris flow and

to outline its basic quantitative aspects. Numerical models,

which couple higher resolution with remarkable data

requirement, could represent a further development in the

analysis of this event.

In this study, a reconstruction of flood discharge by

means of rainfall-runoff models has not been carried out

because the flood event of November 1966 in the Chie-

ppena Torrent was substantially influenced by the failure of

temporary obstructions of the channel. An analysis of

rainfall-runoff transformation would have given a limited

contribution to the reconstruction of the timing and inten-

sity of the surges that produced major geomorphic effects

along the channel and on the alluvial fan.

In the Province of Trento, as well as in other Alpine

areas, torrent control works have been extensively carried

out since the 19th century. The study of the historical

evolution of such works is thus an important part of studies

intended to improve our knowledge of mountain territories.

The Chieppena Torrent offers a good example for this kind

of analysis, especially because of the great severity of past

debris flows, which also outlined the inadequacy of some

traditional control works, as demonstrated by the failure of

the check dams designed in the 1920s.

Acknowledgments This study was funded by the Provincia

Autonoma di Trento–Servizio Bacini montani (contract no. 1543

CONV) and the European Commission (Sixth Framework Pro-

gramme, HYDRATE Project, contract no. GOCE-037024). The

authors thank M. Cerato for the useful discussion on historical

records of flood events, and F. Tagliavini for the information on

the geolithological setting of the basin. The staffs from the His-

torical Archive of the Autonomous Province of Trento and State

Archive are thanked for their collaboration in the collection of

historical documents. We also thank R.L. Baum, F. Guzzetti, an

anonymous reviewer, and the Editor-in-Chief, V.H. Dale, for

valuable comments on the manuscript.

References

Barnikel F (2004) The value of historical documents for hazard zone

mapping. Nat Hazards Earth Syst Sci 4:599–613

Barnikel F, Becht M (2003) A historical analysis of hazardous events

in the Alps—the case of Hindelang (Bavaria, Germany). Nat

Hazards Earth Syst Sci 3:625–635

Bathurst JC, Burton A, Ward TJ (1997) Debris flow run-out and

landslide sediment delivery model tests. J Hydraulic Eng

123:410–419

Benito G, Lang M, Barriendos M, Llasat MC, Frances F, Ouarda T,

Thorndycraft VR, Enzel T, Bardossy A, Cœur D, Bobee B

(2004) Use of systematic, paleoflood and historical data for the

improvement of flood risk estimation. Review of scientific

methods. Nat Hazards 31:623–643

Bovis MJ, Jakob M (1999) The role of debris supply conditions in

predicting debris flow activity. Earth Surf Landforms Processes

24:1039–1054

Table 6 Summary of activities and results in the reconstruction of a historical debris-flow: the experience from the study of the Chieppena

Torrent

Activity Results achieved Drawbacks and limitations

Collection and analysis of data from historical archive Assessment of the frequency

of the events

Data on small-intensity events are seldom reported

for the most ancient periods

The information is essentially qualitative

Analysis of bibliographic sources, gray literature and

unpublished documents regarding the event studied

Reconstruction of the time

evolution of the event

Limited information regarding the upper parts of

the basin (debris-flow initiation zones)

Recognition of the type of

flow processes

The information is prevailingly qualitative

Assessment of damage

Analysis of aerial photographs Mapping of the areas

flooded

Difficulty in assessing the thickness of the deposits

Application of empirical equations Quantitative assessment of

flow variables

Intrinsic approximation of available equations

502 Environ Manage (2007) 40:493–503

123

Brazdil R, Glaser R, Pfister C, Dobrovolny P, Antoine J-M,

Barriendos M, Camuffo D, Deutsch M, Enzi S, Guidoboni E,

Kotyza O, Sanchez Rodrigo F (1999) Flood events of selected

European Rivers in the sixteenth century. Climatic Change

43:239–285

Castiglioni GB (1971) Natural disasters in the Alps. In: Atti del XXI

Congresso Geografico Italiano, Verbania, Italy, pp 7–37 (in

Italian)

Castiglioni GB, Pellegrini GB, Nardin M, Ferrari L, Mattana U (1971)

Notes on the map of the flood event of November 1966 in

Venetia and Trentino - Alto Adige: Morphological effects and

inundations. In: Atti del XXI Congresso Geografico Italiano,

Verbania, Italy, pp 269–290 (in Italian)

Cerato M (1999) Contributes for a history of hydraulic works and soil

bioengineering in Valsugana. Provincia Autonoma di Trento,

Azienda Speciale di Sistemazione Montana, Trento, Italy, 109 pp

(in Italian)

Cœur D, Lang M, Paquier A (2002) Flood knowledge: History,

hydraulics and hydrology. Houille Blanche 4-4:61–66 (in

French)

Costa JE (1988) Rheologic, geomorphic, and sedimentologic

differentiation of water floods, hyperconcentrated flows, and

debris flows. In: Baker VR, Kochel RC, Patton PC (eds.)

Flood geomorphology. John Wiley & Sons, New York, pp

113–122

Croce D, Nodari P, Pellegrini GB, Tessari F (1971) Effects of the

November 1966 flood on inhabited areas of northeastern Italy.

In: Atti del XXI Congresso Geografico Italiano, Verbania, Italy

pp 291–301 (in Italian)

D’Agostino V, Cerato M, Coali R (1996) Extreme events of sediment

transport in the eastern Trentino torrents. In: Proceedings of

the International Symposium Interpraevent 1996, Garmisch-

Partenkirchen, Germany. 1:377–386 (in Italian)

Dorigo L (1969) The river basins of Venetia. In: L’evento alluvionale

del novembre 1966. Commissione interministeriale per lo studio

della sistemazione idraulica e della difesa del suolo. Istituto

Poligrafico dello Stato, Rome, Italy, pp 133–165 (in Italian)

Eisbacher GH, Clague JJ (1984) Destructive mass movements in high

mountains: Hazard and management. Geological Survey of

Canada, Paper 84–16, Ministry of Supply and Services, Ottawa,

Canada, 230 pp

Esposito E, Porfido S, Violante C (2004) The flood of October 1954 in

Vietri sul Mare (Salerno). Scenario and effects of a catastrophic

flood in a rocky coast area. Consiglio Nazionale delle Ricerche

GNDCI, IAMC, Naples, Italy, 381 pp (in Italian)

Gorfer A (1967) The general flood of November 4, 1966. Natura

Alpina 18:5–42 (in Italian)

Gorfer A (1977) The valleys of Trentino – eastern Trentino. Manfrini

Editori, Calliano, Italy, 1118 pp (in Italian)

Govi M (1975) Mapping, documentation and interpretation of floods

and mudflows in the Italian Alps. In: Proceedings of the

International Symposyium Interpraevent 1975, Innsbruck,

Austria, 2:209–233 (in French)

Guzzetti F, Cardinali M, Reichenbach P (1994) The AVI Project: A

bibliographical and archive inventory of landslides and floods in

Italy. Envir Manage 18:623–633

Guzzetti F, Tonelli G (2004) Information system on hydrological and

geomorphological catastrophes in Italy (SICI): A tool for

managing landslide and flood hazards. Natural Hazards Earth

Syst Sci 4:213–232

Ikeya H (1989) Debris flow and its countermeasures in Japan. Bull Int

Assoc Eng Geol 40:15–33

Kellerhals R, Church M (1990) Hazard management on fans, with

examples from British Columbia. In Rachocki AH, Church M

(eds.) Alluvial fans: a field approach. John Wiley & Sons, New

York pp 335–354

Kronfellner-Kraus G (1984) Extreme sedimentations and gullying of

torrents. In: Proceedings of the International Symposium

Interpraevent 1984, Villach, Austria, 2:109–118 (in German)

Moscariello A, Marchi L, Maraga F, Mortara G (2002) Alluvial fans

in the Italian Alps: Sedimentary facies and processes. In: Martini

P, Baker VR, Garzon G (eds.) Flood & megaflood processes and

deposits—recent and ancient examples. Blackwell Science,

Oxford, UK, pp 141–166

Pedenzini A (2001) Water from the sky and from ground – Strigno,

November 4, 1966. Il Foglio di Strigno, no. 3, 110 pp (in Italian)

Pierson TC (2005) Distinguishing between debris flows and floods

from field evidence in small watersheds. U.S. Geological Survey

Fact Sheet 2004–3142, 4 pp

Provincia Autonoma di Trento (1991) For the protection of the

territory. November 4, Trento, Italy, 237 pp (in Italian)

Rickenmann D (1999) Empirical relationships for debris flows.

Natural Hazards 19:47–77

Rickenmann D, Zimmermann M (1993) The 1987 debris flows in

Switzerland: Documentation and analysis. Geomorphology

8:175–189

Tonini D, (1968) The floods of November 1966 in Venetia and Friuli-

Venezia Giulia. In: Atti del Convegno sul Tema: Le scienze

della natura di fronte agli eventi idrogeologici, Rome, Italy, pp

91–146 (in Italian)

Tropeano D (1989) Floods and landslide in the Bormida River basin.

Retrospective study. Quaderni di Studio e Documentazione,

no. 10, Associazione Mineraria Subalpina, Torino, Italy, 155 pp

(in Italian)

Tropeano D, Turconi L (2004) Using historical documents for

landslide, debris flow and stream flood prevention: Applications

in Northern Italy. Natural Hazards 31:663–679

Turitto O (2004) Extreme Po River flood in November 1951 and

inundation of Polesine territory (Northern Italy). Houille

Blanche 5:64–70 (in French)

Venzo GA (1968) Observations on the relationships between

geological settings and characteristics of the flood of

November 1966 in Trentino. In: Atti del Convegno sul Tema:

Le scienze della natura di fronte agli eventi idrogeologici.

Accademia Nazionale dei Lincei, Roma, Italy, pp 147–160 (in

Italian)

Venzo GA, Largaiolli T (1968) The basin of Chieppena Torrent

(Trentino) - Stratigraphy - Tectonics – Geomorphology and

Hydrology - Landslides. Memorie del Museo Tridentino di

Scienze Naturali 17:1–103, includes two maps at 1:10000 scale

(in Italian)

Environ Manage (2007) 40:493–503 503

123