Analysis of fire risk elements in the Wildland-Urban Interface. Case study: "La Ribera Alta", Valencia, Spain

Abstract

Forest fires are a real problem at the interface of vegetation and settlements with several areas being burnt annually along the Mediterranean coast. This thesis is focused on the La Ribera Alta (RA) region where a Spanish consultancy found a gap in knowledge in the location and assignment of fire risk levels for houses built up in or close to forest areas. Firstly it presents a test of automated extraction using an orthophoto for rooftops detection obtaining location of houses in vectorial format. This allows associating the level of requirement of Contingency Plan against Forest Fires (CPAFF) calculated in the second part of the project. In that part elements related to the fire risk analysis at both parish and house level are implemented in order to select which houses are in more urgent need of a CPAFF. A confusion matrix was used to evaluate the accuracy of the rooftops extraction, with the overall accuracy being equal to 85%. A poll was taken (interviewing a group of experts) in order to weight all parameters and obtain the final results of the fire risk analysis. This classifies the territory depending on the priority of actuation showing the houses that require a CPAFF, in grades of high, medium and low requirement. This classification is known as the Plan Against Fire Requirement Index (PAFRI). It was concluded that the Feature Analyst software is a powerful tool for rooftop extraction and the analysis allows detection of houses located in areas more likely to be at risk of burning, so companies as MEDI XXI ¿ GSA can offer the CPAFF to these areas.