The increasing incidence of forest fires represents a major environmental and territorial challenge on a global and national scale, with significant impacts on ecosystems, population safety,and land management. In this context, analyzing forest fire susceptibilityis a fundamental tool for identifying the areas most prone to the phenomenon and supporting effective prevention and planning strategies. This thesis aims to develop a forest fire susceptibility map for the entire Marche Region, using the Weight of Evidence (WoE) statistical method. The choice of this study area is motivated by the lack, identified following a critical review of the scientific literature, of specific studies dedicated to mapping forest fire susceptibility in this regional context. The analysis is based on the integration of historical forest fire data with the main topographic and ecological predisposing factors, selected based on their recurrence in the literature and the availability of data at the regional scale. Specifically, we considered altitude, slope exposure, slope slope, and land cover, which were then processed and integrated into a GIS environment for the implementation of the WoE model. The results highlight how the spatial distribution of forest fires in the Marche region is not random, but is strongly influenced by specific combinations of territorial factors. The following classes: altimetry 500-1000 m, exposure southeast, southwest, and south, slope 15-30°, as well as certain land cover types, particularly broadleaf forests, coniferous forests, shrublands, and preforests, are most associated with high levels of susceptibility. Following the implementation of the WoE, these classes were found to have significant contrast values: 500-1000 m = 0.82; South East = 0.50, South West = 0.68, South = 0.79; 15-30° = 1.79; broadleaf forests = 0.27, coniferous forests = 2.30, shrublands and preforests = 2.61. The overlap of these classes means that 67.08% of the regional area is characterized by a medium susceptibility class to forest fires and 14.40% by a high susceptibility class; both classes highlight areas along the Umbria-Marche Apennines and generally throughout the entire regional area that require greater attention from the competent authorities and citizens. Subsequently, the model validation conducted using the Success Rate and Prediction Rate demonstrated the susceptibility map's good discriminatory ability with respect to historical fire data, confirming the reliability of the adopted approach. The validation highlighted that the spatial and temporal distribution of forest fires in the Marche region between the historical fire data for the period 2010-2023 and 2005, 2006, 2007, and 2009 showed no substantial differences, but rather excellent correspondence, confirming that the analysis performed using the WoE is reliable and robust. The resulting map could therefore represent a useful tool to support prevention, fuel management, and territorial planning activities, contributing to a more informed and sustainable management of forest fires in the Marche region.
La crescente incidenza degli incendi boschivi rappresenta una delle principali criticità ambientali e territoriali a scala globale e nazionale, con effetti rilevanti sugli ecosistemi, sulla sicurezza delle popolazioni e sulla gestione del territorio. In questo contesto, l’analisi della suscettibilità agli incendi boschivi costituisce uno strumento fondamentale per individuare le aree maggiormente predisposte al fenomeno e supportare efficaci strategie di prevenzione e pianificazione. La presente tesi ha come obiettivo la redazione di una mappa di suscettibilità agli incendi boschivi per l’intero territorio della Regione Marche, mediante l’applicazione del metodo statistico del Weight of Evidence (WoE). La scelta dell’area di studio è motivata dall’assenza, emersa a seguito di una revisione critica della letteratura scientifica, di studi specifici dedicati alla mappatura della suscettibilità agli incendi boschivi in questo contesto regionale. L’analisi si basa sull’integrazione dei dati storici degli incendi boschivi con i principali fattori predisponenti di natura topografica ed ecologica, selezionati in base alla loro ricorrenza nella letteratura e alla disponibilità dei dati a scala regionale. In particolare, sono stati considerati l’altitudine, l’esposizione dei versanti, la pendenza e la copertura del suolo, opportunamente elaborati e integrati in ambiente GIS per l’implementazione del modello WoE. I risultati evidenziano come la distribuzione spaziale degli incendi boschivi nelle Marche non sia casuale, ma fortemente condizionata da specifiche combinazioni di fattori territoriali. Le classi: altimetria 500-1000m, esposizione Sud Est, Sud Ovest e Sud, pendenza 15-30° nonché alcune tipologie di copertura del suolo, in particolare boschi di latifoglie, boschi di conifere e arbusteti e prebosco, risultano maggiormente associate a livelli elevati di suscettibilità. Tali classi a seguito dell’implementazione del WoE sono risultate infatti avere valori significativi di contrasto: 500-1000m = 0,82; Sud Est = 0,50, Sud Ovest = 0,68, Sud = 0,79; 15-30° = 1,79; boschi di latifoglie = 0,27, boschi di conifere = 2,30, arbusteti e prebosco = 2,61. La sovrapposizione di tali classi determina che il 67,08% dell’area regionale sia caratterizzata da una classe media di suscettibilità agli incendi boschivi e il 14,40% invece dalla classe alta di suscettibilità; entrambe le classi mettono in evidenza aree lungo l’appennino umbro-marchigiano e a livello generale sull’intera area regionale che necessitano di maggiore attenzione da parte delle Autorità Competenti e dei cittadini. Successivamente, la validazione del modello condotta attraverso il Success Rate e il Prediction Rate, ha dimostrato una buona capacità discriminante della mappa di suscettibilità rispetto ai dati storici degli incendi, confermando l’affidabilità dell’approccio adottato. La validazione ha infatti evidenziando come la distribuzione spaziale e temporale degli incendi boschivi nelle Marche tra i dati storici degli incendi compresi nell’intervallo 2010-2023 e 2005, 2006, 2007, 2009 non abbia differenze sostanziali, ma al contrario un’ottima corrispondenza confermando il fatto che l’analisi svolta attraverso il WoE risulta affidabile e robusta. La mappa prodotta potrebbe rappresentare pertanto uno strumento utile per il supporto alle attività di prevenzione, gestione del combustibile e pianificazione territoriale, contribuendo a una gestione più consapevole e sostenibile degli incendi boschivi nella Regione Marche.
VALUTAZIONE DELLA SUSCETTIBILITA’ AGLI INCENDI BOSCHIVI NEL TERRITORIO MARCHIGIANO: APPLICAZIONE DEL METODO WEIGHT OF EVIDENCE IN AMBIENTE GIS
ORTU, ALESSANDRO
2024/2025
Abstract
The increasing incidence of forest fires represents a major environmental and territorial challenge on a global and national scale, with significant impacts on ecosystems, population safety,and land management. In this context, analyzing forest fire susceptibilityis a fundamental tool for identifying the areas most prone to the phenomenon and supporting effective prevention and planning strategies. This thesis aims to develop a forest fire susceptibility map for the entire Marche Region, using the Weight of Evidence (WoE) statistical method. The choice of this study area is motivated by the lack, identified following a critical review of the scientific literature, of specific studies dedicated to mapping forest fire susceptibility in this regional context. The analysis is based on the integration of historical forest fire data with the main topographic and ecological predisposing factors, selected based on their recurrence in the literature and the availability of data at the regional scale. Specifically, we considered altitude, slope exposure, slope slope, and land cover, which were then processed and integrated into a GIS environment for the implementation of the WoE model. The results highlight how the spatial distribution of forest fires in the Marche region is not random, but is strongly influenced by specific combinations of territorial factors. The following classes: altimetry 500-1000 m, exposure southeast, southwest, and south, slope 15-30°, as well as certain land cover types, particularly broadleaf forests, coniferous forests, shrublands, and preforests, are most associated with high levels of susceptibility. Following the implementation of the WoE, these classes were found to have significant contrast values: 500-1000 m = 0.82; South East = 0.50, South West = 0.68, South = 0.79; 15-30° = 1.79; broadleaf forests = 0.27, coniferous forests = 2.30, shrublands and preforests = 2.61. The overlap of these classes means that 67.08% of the regional area is characterized by a medium susceptibility class to forest fires and 14.40% by a high susceptibility class; both classes highlight areas along the Umbria-Marche Apennines and generally throughout the entire regional area that require greater attention from the competent authorities and citizens. Subsequently, the model validation conducted using the Success Rate and Prediction Rate demonstrated the susceptibility map's good discriminatory ability with respect to historical fire data, confirming the reliability of the adopted approach. The validation highlighted that the spatial and temporal distribution of forest fires in the Marche region between the historical fire data for the period 2010-2023 and 2005, 2006, 2007, and 2009 showed no substantial differences, but rather excellent correspondence, confirming that the analysis performed using the WoE is reliable and robust. The resulting map could therefore represent a useful tool to support prevention, fuel management, and territorial planning activities, contributing to a more informed and sustainable management of forest fires in the Marche region.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi Magistrale - Valutazione della suscettibilità agli incendi boschivi nel territorio marchigiano_applicazione del metodo weight of evidence in ambiente GIS - Alessandro Ortu.pdf
accesso aperto
Descrizione: Tesi Magistrale - Alessandro Ortu
Dimensione
2.7 MB
Formato
Adobe PDF
|
2.7 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.12075/25269