Earthquakes play a significant role in triggering landslides. The seismic activity can destabilize slopes, resulting in various types of landslides that pose serious risks to both life and property. Gaining insight into this relationship is crucial for effective hazard assessment and developing mitigation strategies in areas prone to earthquakes. Additionally, recent seismic events have highlighted the extra dangers posed by landslides, which can cause extensive damage beyond the initial shaking. Understanding how earthquake-induced landslides (EILs) behave is vital for establishing links between earthquake parameters and landslide distribution, especially in Italy, a region known for its susceptibility to both earthquakes and landslides. To achieve this, a comprehensive dataset is necessary, one that captures the earthquake-induced impacts of both historical and contemporary seismic events. Identifying as many documented effects of landslides triggered by past earthquakes as possible is essential, as is establishing a reliable framework for assessing future seismic events. This will enable the continuous expansion of the dataset, yielding increasingly accurate empirical values reflective of EIL trends in Italy. In this study, we undertook a review and integration of historical EIL data available in the CFTI5Med database, identifying new landslides and enhancing the accuracy of existing records. We focused on examining historical sources—both newly discovered and previously archived—alongside recent scientific literature and technical reports. This effort allowed for a more precise localization of historical EILs and a clearer classification of landslide types, linking effects to landslides cataloged in the IFFI database. This research resulted in a new independent dataset, CFTI Landslides, encompassing over 1,000 recorded landslides. We utilized this database to develop empirical relationships illustrating the correlation between EIL density and distance from the epicenter, as influenced by earthquake magnitude. Additionally, we created a relationship based on Peak Ground Acceleration (PGA) using shake maps from historical earthquakes published by INGV. A novel method for identifying EILs for contemporary seismic events has also been established, leveraging satellite imagery analysis through the Normalized Difference Index (NDI) to detect ground variations linked to landslide impacts. This method, augmented by field observations, enables a comprehensive assessment of the epicentral area and the identification of seismic-induced effects. Ongoing data collection is important as it will enhance the dataset, refining empirical relationships. Indeed, although historical records are crucial for research, they have limitations in defining statistical trends concerning landslide dimensions and volumes. The thesis concludes with case studies illustrating the application of empirical relationships derived from this research. The first case applies our findings to the distribution of seismic-induced landslide density following the November 2022 seismic sequence that struck offshore Pesaro (central Italy) off the northern Marche coast, assisting INGV emergency response teams. The second case explores the potential for EILs in the Valmarecchia area, analyzing local seismic events in relation to their magnitude and proximity. Ultimately, this study has produced a new database on historical EILs, the CFTI Landslides database. This initiative has facilitated the development of empirical relationships that describe the phenomenon's trends in Italy. To enhance these findings, we propose a study methodology that integrates both indirect and direct observations. The use of NDI proves effective for investigating extensive areas where field surveys may be insufficient. The resulting dataset and empirical outcomes hold significant potential for diverse applications, particularly in emergency response and disaster management.
I terremoti sono i principali fattori scatenanti delle frane le quali,a loro volta,comportano ulteriori rischi,causando danni significativi che vanno ad aggiungersi a quelli dovuti allo scuotimento iniziale.Per questo motivo,è essenziale conoscere le frane sismo-indotte (EILs) del passato,per definire delle relazioni empiriche che consentano di mettere in relazione i parametri sismici con la distribuzione delle frane,non solo a livello globale ma soprattutto a scala locale come il territorio italiano.Per fare ciò è importante avere un dataset rappresentativo del problema e quindi che prenda in considerazione gli effetti sismo-indotti di terremoti del passato, sia lontano che recente. Risulta quindi fondamentale identificare un numero più elevato possibile di effetti di EILs documentati in fonti storiche affidabili,ma anche di effetti di terremoti recenti. Inoltre,è necessario definire un metodo di indagine che,grazie alle attuali tecnologie, permetta una rapida analisi delle aree epicentrali alla ricerca di fenomeni sismo-indotti, in previsione di terremoti futuri. Così sarà possibile ampliare in modo continuo il dataset per ottenere dei valori empirici sempre più rappresentativi dell’andamento del fenomeno su territorio italiano.In questo studio, abbiamo intrapreso una revisione e un’integrazione dei dati storici delle EILs disponibili nel database CFTI5Med, identificando nuove frane e migliorando l’accuratezza dei dati esistenti. Ci siamo concentrati sulla revisione di fonti storiche e letteratura scientifica recente. Questo sforzo ha consentito una localizzazione più precisa degli EILs storici,una classificazione dei tipi di frana,ed ove possibile un collegamento alle frane del database IFFI.Questa ricerca ha prodotto un nuovo set di dati indipendente, CFTIlandslides, che comprende oltre 1.000 frane. Il database ha permesso di sviluppare relazioni empiriche che illustrano la correlazione tra la densità dell'EILs e la distanza dall'epicentro per classi di magnitudo,ed una relazione basata sulla Peak Ground Acceleration (PGA) utilizzando le mappe delle scosse dei terremoti storici pubblicate dall'INGV.È stato inoltre stabilito un nuovo metodo per identificare le EILs, sfruttando l’analisi delle immagini satellitari attraverso il Normalized Differences Index (NDI). Questo metodo, arricchito da osservazioni sul campo,consente una valutazione completa dell'area epicentrale per l'identificazione delle EILs per i terremoti più recenti e per quelli futuri. La raccolta dati così ottenuta andrà a popolare ulteriormente il set di dati,perfezionando le relazioni empiriche. Infatti,sebbene i dati storici siano cruciali per la ricerca, essi presentano limitazioni nel definire le tendenze statistiche relative alle dimensioni e ai volumi delle frane,a causa dei pochi dati ad oggi disponibili.La tesi si conclude con casi di studio che illustrano l'applicazione delle relazioni empiriche derivate da questa ricerca: l’utilizzo della distribuzione della densità di EILs rispetto all’epicentro per guidare sul terreno le squadre del gruppo operativo EMERGEO dell’INGV a seguito della sequenza sismica del novembre 2022 lungo la costa marchigiana pesarese; esplorare il potenziale di EILs nell'area della Valmarecchia,analizzando gli eventi sismici locali in relazione alla loro magnitudo e prossimità.In definitiva, questo studio ha prodotto un nuovo database sulle EILs storiche: il database CFTIlandslides.Questo risultato ha permesso lo sviluppo di relazioni empiriche che descrivono le tendenze del fenomeno in Italia.Per migliorare questi dati, proponiamo una metodologia di studio che integra osservazioni dirette e indirette.L’uso dell’NDI si rivela efficace per indagare aree estese dove i rilievi sul campo potrebbero essere limitati: il set di dati e i risultati empirici derivanti hanno un potenziale significativo per diverse applicazioni,in particolare nella risposta alle emergenze e nella gestione delle catastrofi.
The study of earthquake-induced landslides: from historical data to empirical relationships
ZEI, CATERINA
2025
Abstract
Earthquakes play a significant role in triggering landslides. The seismic activity can destabilize slopes, resulting in various types of landslides that pose serious risks to both life and property. Gaining insight into this relationship is crucial for effective hazard assessment and developing mitigation strategies in areas prone to earthquakes. Additionally, recent seismic events have highlighted the extra dangers posed by landslides, which can cause extensive damage beyond the initial shaking. Understanding how earthquake-induced landslides (EILs) behave is vital for establishing links between earthquake parameters and landslide distribution, especially in Italy, a region known for its susceptibility to both earthquakes and landslides. To achieve this, a comprehensive dataset is necessary, one that captures the earthquake-induced impacts of both historical and contemporary seismic events. Identifying as many documented effects of landslides triggered by past earthquakes as possible is essential, as is establishing a reliable framework for assessing future seismic events. This will enable the continuous expansion of the dataset, yielding increasingly accurate empirical values reflective of EIL trends in Italy. In this study, we undertook a review and integration of historical EIL data available in the CFTI5Med database, identifying new landslides and enhancing the accuracy of existing records. We focused on examining historical sources—both newly discovered and previously archived—alongside recent scientific literature and technical reports. This effort allowed for a more precise localization of historical EILs and a clearer classification of landslide types, linking effects to landslides cataloged in the IFFI database. This research resulted in a new independent dataset, CFTI Landslides, encompassing over 1,000 recorded landslides. We utilized this database to develop empirical relationships illustrating the correlation between EIL density and distance from the epicenter, as influenced by earthquake magnitude. Additionally, we created a relationship based on Peak Ground Acceleration (PGA) using shake maps from historical earthquakes published by INGV. A novel method for identifying EILs for contemporary seismic events has also been established, leveraging satellite imagery analysis through the Normalized Difference Index (NDI) to detect ground variations linked to landslide impacts. This method, augmented by field observations, enables a comprehensive assessment of the epicentral area and the identification of seismic-induced effects. Ongoing data collection is important as it will enhance the dataset, refining empirical relationships. Indeed, although historical records are crucial for research, they have limitations in defining statistical trends concerning landslide dimensions and volumes. The thesis concludes with case studies illustrating the application of empirical relationships derived from this research. The first case applies our findings to the distribution of seismic-induced landslide density following the November 2022 seismic sequence that struck offshore Pesaro (central Italy) off the northern Marche coast, assisting INGV emergency response teams. The second case explores the potential for EILs in the Valmarecchia area, analyzing local seismic events in relation to their magnitude and proximity. Ultimately, this study has produced a new database on historical EILs, the CFTI Landslides database. This initiative has facilitated the development of empirical relationships that describe the phenomenon's trends in Italy. To enhance these findings, we propose a study methodology that integrates both indirect and direct observations. The use of NDI proves effective for investigating extensive areas where field surveys may be insufficient. The resulting dataset and empirical outcomes hold significant potential for diverse applications, particularly in emergency response and disaster management.File | Dimensione | Formato | |
---|---|---|---|
PhD_Thesis_Zei.pdf
embargo fino al 04/04/2026
Dimensione
20.34 MB
Formato
Adobe PDF
|
20.34 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/218806
URN:NBN:IT:UNIFE-218806