Earthquake Induced Landslides (EILs) occurrence and earthquake parameters: new empirical relationships developed using the updated CFTI historical dataset
Author(s)
Type
Conference paper
Language
English
Obiettivo Specifico
OST2 Deformazione e Hazard sismico e da maremoto
Status
Published
Date Issued
February 2023
Conference Location
Bologna
Abstract
The ground shaking generated by the arrival of seismic waves released during earthquakes, besides
generating damages to the urban environment and artificial infrastructures, can trigger
widespread environmental coseismic effects, overall defined as secondary Earthquake
Environmental Effects (EEE) in the Environmental Seismic Intensity scale (ESI; Michetti et al., 2007).
These natural effects include river ponding or diversion, liquefaction, ground compaction and
failures and, among all, landslides. If, due to geological and geomorphological reasons, the area hit
by the earthquake is susceptible to their occurrence, they take place almost constantly, regardless
of the kinematic characteristics of the seismogenic sources, and affect areas of variable size
according to the earthquake magnitude. Besides potentially increasing overall earthquake damage
immediately after the mainshock, they pose long lasting secondary hazards to earthquake affected
areas (e.g. Fan et al., 2019), driving also the long-term morphological evolution of the territory
(e.g. Wang et al., 2020).
Earthquake Induced Landslides (EILs) are among the most diffused environmental hazard
connected to the earthquake activity, and especially affect recent, high relief mountain landscapes
associated to the ongoing tectonic plate dynamics. There exist in fact reciprocal feedbacks
between active tectonics, which in the long-term creates the predisposing morphological and
geological factors, and landslide activity, which works for lowering the topographic gradients.
Italy sits on the plate boundary separating the slowly converging Eurasia and Nubia plates, which in
the long run created its mostly mountain territory, and is characterized by a relatively high seismic
hazard. The combination of the relatively frequent seismic release and the locally high landslide
susceptibility (Fig. 1), makes the Italian territory especially prone to EILs occurrence. This is testified by the numerous coseismic landslide inventories compiled after recent earthquakes (e.g. Guzzetti et al., 2009).
As a first step of our workflow, we revised the database of seismic induced environmental effects
connected to the CFTI5Med historical seismic catalogue (available at
https://storing.ingv.it/cfti/cfti5/; Guidoboni et al., 2018; 2019). To achieve this goal, we collected
new data points from new historical sources, revised the EILs already included in CFTI, and finally,
in a dedicated GIS environment, we tried to locate them on topographic maps using the coeval
descriptions. In some cases, it was possible to associate the historical EILs to landslides included
and described in the Italian Landslide Inventory (IFFI database; Trigila et al., 2007). Following this,
landslides were subdivided into three classes according to their positioning precision, i.e.: Class A if
they were geographically well located using a toponymal (AI if associated to one of landslides
included in IFFI); Class B, if they were generally located in an area (BI if associated to one of
landslides included in IFFI); Class C, if not localized. The revised CFTI database includes more than
1000 historical EILs, associated to 159 individual earthquakes or seismic sequences occurred
between 117 B.C.E. and 1997 (Fig. 2). With respect to the datasets used so far for deriving
empirical relationships (e.g. Livio and Ferrario, 2019; Tanyaş et al., 2017), these figures
dramatically increase the number of available data points for the Italian territory.
As a second step, the implemented database was used to develop new empirical attenuation
relationships between the EILs density and the distance from the epicenter as a function of the
earthquake magnitude. Then, we subdivided seismic events into three magnitude classes to
account for the different extent of the maximum area affected by EILs and released energy. The
three classes are as follows: Class 1, M < 5.5; Class 2, 5.5 ≤ M < 6.5, and Class 3, M ≥ 6.5. Based on
the locations of EILs and of epicenters, we set circular search areas with a 5 km large moving
window and calculated for each area the cumulative density. We found that the cumulative density
decreases with the distance following a power law relationship, with coefficients of determination
variable between 0.95 and 0.99 according to the magnitude class, and that the maximum distance
including 95th percentile of landslides increases with increasing magnitude range. In particular, the
95th percentile distance, i.e. the distance within which 95% of EILs are expected to occur, is 42 km
for earthquakes of Class 1, 66 km for those of Class 2, and 77 km for those of Class 3.
In addition, using the shakemaps of 38 out of 159 historical earthquakes of our dataset, we
developed an empirical relationship between EILs density and PGA values. The shakemaps of
historical earthquakes were recently published by the Istituto Nazionale di Geofisica e Vulcanologia
and are available at http://shakemap.ingv.it/shake4/. EILs density was computed by classifying PGA values into classes of 0.1 g range. We used the EILs location to sample the PGA value and derive a
power law relationship showing an increasing landslide density with increasing acceleration.
As a final step, these empirical relationships were used to relate each landslide belonging to the
Italian national inventory (IFFI) to individual segments of known seismogenic sources located in the
Italian territory or in nearby countries within a maximum distance corresponding to the 95th
percentile distance calculated from the expected magnitude of the seismogenic source (Fig. 2). We
used the composite seismogenic sources of the DISS database version 3.3.0 (DISS Working Group,
2021) hypothesizing that they may trigger the EILs activation during future earthquakes and, using
a floating hypocenter approach, i.e. shifting the hypocenters along the seismogenic sources with 4
km fixed steps, built preliminary EILs scenarios for each potential hypocenter location.
generating damages to the urban environment and artificial infrastructures, can trigger
widespread environmental coseismic effects, overall defined as secondary Earthquake
Environmental Effects (EEE) in the Environmental Seismic Intensity scale (ESI; Michetti et al., 2007).
These natural effects include river ponding or diversion, liquefaction, ground compaction and
failures and, among all, landslides. If, due to geological and geomorphological reasons, the area hit
by the earthquake is susceptible to their occurrence, they take place almost constantly, regardless
of the kinematic characteristics of the seismogenic sources, and affect areas of variable size
according to the earthquake magnitude. Besides potentially increasing overall earthquake damage
immediately after the mainshock, they pose long lasting secondary hazards to earthquake affected
areas (e.g. Fan et al., 2019), driving also the long-term morphological evolution of the territory
(e.g. Wang et al., 2020).
Earthquake Induced Landslides (EILs) are among the most diffused environmental hazard
connected to the earthquake activity, and especially affect recent, high relief mountain landscapes
associated to the ongoing tectonic plate dynamics. There exist in fact reciprocal feedbacks
between active tectonics, which in the long-term creates the predisposing morphological and
geological factors, and landslide activity, which works for lowering the topographic gradients.
Italy sits on the plate boundary separating the slowly converging Eurasia and Nubia plates, which in
the long run created its mostly mountain territory, and is characterized by a relatively high seismic
hazard. The combination of the relatively frequent seismic release and the locally high landslide
susceptibility (Fig. 1), makes the Italian territory especially prone to EILs occurrence. This is testified by the numerous coseismic landslide inventories compiled after recent earthquakes (e.g. Guzzetti et al., 2009).
As a first step of our workflow, we revised the database of seismic induced environmental effects
connected to the CFTI5Med historical seismic catalogue (available at
https://storing.ingv.it/cfti/cfti5/; Guidoboni et al., 2018; 2019). To achieve this goal, we collected
new data points from new historical sources, revised the EILs already included in CFTI, and finally,
in a dedicated GIS environment, we tried to locate them on topographic maps using the coeval
descriptions. In some cases, it was possible to associate the historical EILs to landslides included
and described in the Italian Landslide Inventory (IFFI database; Trigila et al., 2007). Following this,
landslides were subdivided into three classes according to their positioning precision, i.e.: Class A if
they were geographically well located using a toponymal (AI if associated to one of landslides
included in IFFI); Class B, if they were generally located in an area (BI if associated to one of
landslides included in IFFI); Class C, if not localized. The revised CFTI database includes more than
1000 historical EILs, associated to 159 individual earthquakes or seismic sequences occurred
between 117 B.C.E. and 1997 (Fig. 2). With respect to the datasets used so far for deriving
empirical relationships (e.g. Livio and Ferrario, 2019; Tanyaş et al., 2017), these figures
dramatically increase the number of available data points for the Italian territory.
As a second step, the implemented database was used to develop new empirical attenuation
relationships between the EILs density and the distance from the epicenter as a function of the
earthquake magnitude. Then, we subdivided seismic events into three magnitude classes to
account for the different extent of the maximum area affected by EILs and released energy. The
three classes are as follows: Class 1, M < 5.5; Class 2, 5.5 ≤ M < 6.5, and Class 3, M ≥ 6.5. Based on
the locations of EILs and of epicenters, we set circular search areas with a 5 km large moving
window and calculated for each area the cumulative density. We found that the cumulative density
decreases with the distance following a power law relationship, with coefficients of determination
variable between 0.95 and 0.99 according to the magnitude class, and that the maximum distance
including 95th percentile of landslides increases with increasing magnitude range. In particular, the
95th percentile distance, i.e. the distance within which 95% of EILs are expected to occur, is 42 km
for earthquakes of Class 1, 66 km for those of Class 2, and 77 km for those of Class 3.
In addition, using the shakemaps of 38 out of 159 historical earthquakes of our dataset, we
developed an empirical relationship between EILs density and PGA values. The shakemaps of
historical earthquakes were recently published by the Istituto Nazionale di Geofisica e Vulcanologia
and are available at http://shakemap.ingv.it/shake4/. EILs density was computed by classifying PGA values into classes of 0.1 g range. We used the EILs location to sample the PGA value and derive a
power law relationship showing an increasing landslide density with increasing acceleration.
As a final step, these empirical relationships were used to relate each landslide belonging to the
Italian national inventory (IFFI) to individual segments of known seismogenic sources located in the
Italian territory or in nearby countries within a maximum distance corresponding to the 95th
percentile distance calculated from the expected magnitude of the seismogenic source (Fig. 2). We
used the composite seismogenic sources of the DISS database version 3.3.0 (DISS Working Group,
2021) hypothesizing that they may trigger the EILs activation during future earthquakes and, using
a floating hypocenter approach, i.e. shifting the hypocenters along the seismogenic sources with 4
km fixed steps, built preliminary EILs scenarios for each potential hypocenter location.
File(s)![]()
Name
EILS_Abstract.pdf
Size
3.41 MB
Format
Adobe PDF
Checksum (MD5)
The MD5 message-digest algorithm can be used to verify the integrity of the file that you have uploaded. You can calculate its value locally via tools that are generally available in each operative system like md5sum
319bfdeafacd120efec7b0f097525164
