Improving Landslide Susceptibility Assessment through Frequency Ratio and Classification Methods-Case Study of Valencia Region (Spain)

[EN] Landslide susceptibility maps are widely used in land management and urban planning to delimit potentially problematic areas. In this article we improve their reliability by acting on the frequency ratio method and map classification systems. For the frequency ratio method, we have worked with...

Descripción completa

Detalles Bibliográficos
Autores: Cantarino-Martí, Isidro|||0000-0002-9962-5734, Carrión Carmona, Miguel Ángel|||0000-0003-2998-3031, Martínez Ibáñez, Víctor|||0000-0003-4445-2896, Gielen, Eric|||0000-0002-4591-2914
Tipo de recurso: artículo
Fecha de publicación:2023
País:España
Institución:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:riunet.upv.es:10251/197134
Acceso en línea:https://riunet.upv.es/handle/10251/197134
Access Level:acceso abierto
Palabra clave:Frequency ratio method FR
Sensitivity and specificity
Geographical information systems GIS
Landslide susceptibility maps
Classification systems
ROC analysis
URBANISTICA Y ORDENACION DEL TERRITORIO
INGENIERIA DEL TERRENO
11.- Conseguir que las ciudades y los asentamientos humanos sean inclusivos, seguros, resilientes y sostenibles
Descripción
Sumario:[EN] Landslide susceptibility maps are widely used in land management and urban planning to delimit potentially problematic areas. In this article we improve their reliability by acting on the frequency ratio method and map classification systems. For the frequency ratio method, we have worked with continuous variables and established intervals grouped by probability according to the landslide inventory and based on the characteristics of the data rather than on standard divisions. For map classification systems, we have compared the efficacy of conventional classifications and those based on the concepts of sensitivity and specificity, with the specificity classifications being supported by the information offered by available comparative data. Both strategies make it possible to avoid subjective and repetitive procedures that are alien to the nature of the data being assessed. We present a case study in the 23,000 km2 Region of Valencia where a total of 48 different susceptibility maps were generated. We demonstrate that the methods applied in this study to calculate the frequency ratio provide an improvement in specificity in areas of high susceptibility while maintaining good sensitivity. In particular, the Area Under Curve (AUC) values increase from 0.67 for the conventional methods to 0.76 with the methods proposed in this work. This improvement is transferred to susceptibility mapping much more clearly when classifications that incorporate sensitivity, and especially specificity parameters, are used.