Information Technology Reference
In-Depth Information
4 Simulations of the Curti Debris Flow of May 1998
The Curti landslide was selected among the whole population of slope movements
triggered in Campania by prolonged rainfalls on May 1998 [18]. On that occasion,
hundreds of small soil slips originated at Pizzo d'Alvano, in the volcaniclastic mantle,
on the uppermost portions of the slopes, and transformed into fast-moving debris
flows. These generally involved the entire depth of available detrital cover, eroding it
down to the bedrock, and increased their initial volume by scraping off material along
the path. Landslides hit the urbanised areas at the base of the massif (villages of
Sarno, Siano, Bracigliano, Quindici), killing 161 people and leaving more than 1000
others homeless.
In particular, the selected landslide started as a soil slip at 783 m a.s.l. (right above
a minor outcrop of bedrock). The sliding mass of volcaniclastic terrain, of about 100
m
3
, rapidly transformed into a fast flowing mixture of mud, debris and water, running
down the slope along a smooth pre-existing channel. After encountering a notable
bedrock outcrop (about 75 m high), the flow enlarged and entered the main channel,
triggering some “secondary” debris slides on both flanks (cf. “a”, “b”, “c” and “d” in
Fig.1). At about 200 m a.s.l., the phenomenon subdivided into two distinct flows
which,
CA
. 460 m down slope, merged again and the hit the urban area of Curti (at
115 m a.s.l.), killing two people. The total length of the Curti debris flow is greater
than 1750 m. According to field evidence, along both the upper and middle reach of
the path, the detrital cover was completely eroded by the flowing mass. As a
consequence, the debris flow reached the base of the massif with an estimated volume
of about 81000 m
3
.
The thickness of the mantle of detrital cover (available for the erosion along the
path), as evaluated through detailed field surveying, was given as input matrix to the
model. Note that “complete” erosion of the detrital cover was allowed along the path
of the debris flow - as suggested by field observations. The extent of this matrix is
about 2.7 times greater than the real landslide.
At present, values of global parameters have been assigned on the basis of a trial-
and-error procedure, starting with physically reasonable values and gradually
modifying them until satisfactory results obtained. Initial values have been chosen by
considering both the cell size, and their expected influence on the rheology of the
phenomenon. By simply changing these values, simulation after simulation, and by
comparing the results with the map of the real event, the set of provisional “optimal”
values for the particular case of study has then been obtained.
A quantitative evaluation of the performed simulations can be made, in a GIS
environment, by comparing the extent of the “real” landslide (as mapped through field
surveying and air photo-interpretation) and the simulated ones. The area A
b
affected
by both the simulated and the real debris flow was computed, as well as sectors which
pertain to only one case (A
r
= real and not simulated, A
s
= simulated and not real; note
that the error A
r
is more serious than A
s
).
In Fig.1, the best simulation obtained with S3hex release of SCIDDICA is shown.
The ratios of A
b
, A
r
and A
s
to the extent of the real debris flow, are 73%, 27% and
21%. Simulation may be considered successful in terms of extent of debris path, if
quality of available input data is considered; furthermore erosion and deposit obtained
by simulation are satisfying, where a comparison with real event is possible.
