 |
| Titel |
Large slope instabilities in Northern Chile and Southern Peru |
| VerfasserIn |
Giovanni B. Crosta, Reginald L. Hermanns, Elena Valbuzzi, Paolo Frattini, Andrea Valagussa |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250100342
|
| Publikation (Nr.) |
 EGU/EGU2014-16279.pdf |
|
|
|
|
|
| Zusammenfassung |
| Deep canyon incision into Tertiary paleosurfaces and large slope instabilities along the
canyon flanks characterize the landscape of western slope of the Andes of northern Chile and
South Peru. This area belongs to the Coastal Escarpment and Precordillera and is formed by
coarse-grained clastic and volcanoclastic formations. The area is characterized by
intense seismicity and long-term hyperaridity (Atacama Desert). Landslides along the
canyon flanks affect volumes generally up to 1 km3 and locally evolved in large rock
avalanches. We prepared a landslide inventory covering an area of about 30,000 km2,
extending from Iquique (Chile) to the South and Tacna (Peru) to the North. A total of
606 landslides have been mapped in the area by use of satellite images and direct
field surveys, prevalently including large phenomena. The landslides range from 1
10-3 km2 to 464 km2 (Lluta landslide). The total landslide area, inclusive of the
landslide scarp and of the deposit, amounts to about 2,130 km2 (about 7% of the
area).
The mega landslides can be classified as large block slides that can evolve in large rock
avalanches (e.g. Minimini landslide). Their initiation seems to be strongly associated to the
presence of secondary faults and large fractures transversal to the slope. These landslides
show evidence suggesting a re-incision by the main canyon network. This seems particularly
true for the Lluta collapse where the main “landslide” mass is masked or deleted
by the successive erosion. Other landslides have been mapped along the Coastal
Escarpment and some of the major tectonic escarpments with an E-W trend. We
examined area-frequency distributions of landslides by developing logarithmically
binned, non-cumulative size frequency distributions that report frequency density as a
function of landslide planar area A. The size frequency distribution presents a strong
undersampling for smaller landslides, due to the extremely old age of the inventory. For
landslides larger than 2 000 m2, the distribution exhibits a power-law behaviour with
scaling exponent, β, equal to -2.24. For comparison, we analysed the power-law
behaviour of other earthquake-induced landslide inventories, obtaining similar results,
although the geological and seismic conditions may have been very different (Buller,
New Zealand, β = -2.42; Iningahua, New Zealand, β = -2.53; Northridge, USA, β
= -2.39; Chi-Chi, Taiwan, β = -2.30; Wenchuan Earthquake, China, β = -2.19).
Volume estimates and slope stability modelling have been completed to characterize
the phenomena and the possible triggering mechanisms. For volume estimate, we
reconstructed the pre-failure surface for tens of landslides, in order to characterize
the area-volume relationship. By using this relationship, we assigned a volume
to all landslides of the inventory. The study area is subject to a high seismicity
associated to earthquakes of different type: interplate (superficial and intermediate
depth), subduction zone earthquakes, and earthquake along the Coastal Escarpment.
By analysing the frequency size relationships for earthquake-induced landslides
from literature, it is possible to observe that the higher the earthquake Magnitude,
the higher the frequency density curve. To quantify this observation, we used the
power-law relationships derived for each inventory to calculate the frequency density
associated to selected areas, and we plotted these frequencies as a function of the
magnitude of the respective earthquakes. By fitting these values, we derived the
expected Magnitude required to generate the landslide distribution of the study area. In
conclusion, we argue that the evolution of these landslides is controlled by: deep valley
incision, canyon walls undercutting and lateral migration of the river controlled by
valley flank instabilities, the Presence of weak lithologies and weak basal layers, the
river incision debuttressing the slope toe and especially brings to daylighting the
weak basal layers observed at some landslide sites, the possible deep groundwater
flow above the deep impermeable formations and clay layers, the movement along
sub-horizontal basal shear zones which can be locally extruded at the slope toe, the river
damming because of the strong lateral components of displacement and successive
reincision by the river with dam failure, the possible sequence of reactivations by
reincision of the deposit, and the occurrence of high magnitude (8-9) earthquakes. |
|
|
|
|
|
|