Wildland urban interface assessment and prediction in relation to land use and land cover changes: the Portuguese case study

Abstract

The Wildland-Urban Interface (WUI) refers to the zone of transition between wildlands and human development. Its spatio-temporal evolution is strongly influenced by land use/land cover changes (LULCC). In Mediterranean countries, WUI represents the most prone area for human-caused wildfires, and is, at present, a central focus of wildfire policy. Traditionally, WUI maps rest on the measurement of the distances among human settlements and wildlands. This methodology is effective to produce detailed maps and to give precise indications, but it is rigid and needs to pre-define fixed parameters. To overcome this uncertainty, we introduce a LULCC simulation model based on Machine Learning approach, allowing to produce prediction maps for the land cover classes related to the WUI. Finally, Fuzzy Set Theory is used to develop continuous non-categorical maps expressing the possibility of being part of this interface. The proposed methodology has been applied to the case study of Portugal, and considers the land covers referred to different epochs (1990, 2000, and 2012). Both a classical geospatial analyses and an innovative simulation approach were adopted: as result we obtained traditional maps of the WUI for the different investigated periods (“hard” boundaries), and a fuzzy WUI map (“soft” prediction) for the year 2030. The framework proposed in this paper correctly identifies the areas belonging to this interface, providing useful information for wildfires-prevention policies.