A combined structural and seasonal approach to assess wildfire susceptibility and hazard in summertime

• Autor(a) principal: Bergonse, Rafaello
• Data de Publicação: 2021
• Outros Autores: Oliveira, Sandra, Gonçalves, Ana, Nunes, Sílvia, da Câmara, Carlos, Zêzere, José
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10451/46346
• topic: Wildfre susceptibility; Hazard assessment; Structural approach; Seasonal variables; Portugal

Wildfre susceptibility and hazard models based on drivers that change only on a multiyear timescale are considered of a structural nature. They ignore specifc short-term conditions in any year and period within the year, especially summer, when most wildfre damage occurs in southern Europe. We investigate whether the predictive capacity of structural wildfre susceptibility and hazard models can be improved by integrating a seasonal dimension, expressed by three variables with yearly to seasonal timescales: (1) a meteorological index rating fuel fammability at the onset of summer; (2) the scarcity of fuel associated with the burned areas of the previous year, and (3) the excessive abundance of fuel in especially fre-prone areas that have not been burned in the previous ten years. We describe a new methodology for combining the structural maps with the seasonal variables, producing year-specifc seasonal susceptibility and hazard maps. We then compare the structural and seasonal maps as to their capacity to predict burnt areas during the summer period in a set of eight independent years. The seasonal maps revealed a higher predictive capacity in 75% of the validation period, both for susceptibility and hazard, when only the highest class was considered. This percentage was reduced to 50% when the two highest classes were considered together. In some years, structural factors and other unconsidered variables probably exert a strong infuence over the spatial pattern of wildfre incidence. These fndings can complement existing structural data and improve the mapping tools used to defne wildfre prevention and mitigation actions.