Landslide susceptibility assessment using different rainfall event based landslide inventories: advantages and limitations

• Author(s): Oliveira, Sérgio C. ; Zêzere, José Luís ; Garcia, Ricardo A. C. ; Pereira, Susana ; Vaz, Teresa ; Melo, Raquel
• Date: 2024
• Persistent ID: https://hdl.handle.net/10216/158987
• Origin: Repositório Aberto da Universidade do Porto
• Document Type: Journal article
• Language: English

Description

The present work aims to evaluate potential sources of uncertainty associated with rainfalltriggered event-based landslide inventories within the framework of landslide susceptibility assessment. Therefore, this study addresses the following questions: (i) How representative is an event-based landslide inventory map of the total landslide activity and distribution in a study area?; (ii) How reliable is an event-based landslide susceptibility map?; (iii) How appropriate is an event-based landslide inventory map for independently validating a landslide susceptibility map? To address these questions, two independent and contrasting rainfall event-based landslide inventories were used, together with a historical landslide inventory, to assess landslide susceptibility for diferent types of landslides in a study area located north of Lisbon, Portugal. The results revealed the following fndings: (i) contrasting rainfall critical conditions for failure can trigger similar landslide types, although they may vary in size and be spatially constrained by diferent predisposing conditions, particularly lithology and soil type; (ii) landslide susceptibility models using event-based landslide inventories are not reliable in the study area, regardless of the landslide inventory map used for training and validation; and (iii) complementary sources of uncertainty results from using incomplete historical landslide inventories to assess landslide susceptibility and non-totally independent landslide inventories for modeling validation. The present study enhances the understanding of regional landslide susceptibility patterns based on contrasting rainfall-trigger conditions, providing valuable information to minimize exposure; to design regional landslide early warning systems for specifc rainfall-trigger landslide events; and to improve the response and preparedness of civil protection services.