Wildfires are considered a devastating threat to natural ecosystems and human lives. Wildfires are complex phenomena, which depend on diverse factors including climatological factors, fuel availability and anthropogenic disturbance in many ways. The main objective of this study is (a) to identify the spatial autocorrelation of historical fire occurrences and (b) to explore the multilateral relationship between fire occurrence as the dependent variable and a set of independent variables reflecting human population patterns, vegetation types and climatological factors. In order to identify the spatial distribution of fire events, a cube hot spot analysis technique was applied to identify them within a span of 30 years (between 1984 and 2014). Thereafter, a Spatial Logistic Regression (SLR) was designed and applied to explore the potential relationship between the two sets of variables. Our findings reveal that fires relate distinctly to vegetation, and that, despite it being accidental, the presence of such vegetation together with other spatially continuous phenomena, lead to the clustering of fires on risk areas. Furthermore, that any socio-economical factor can be hardly conclusive. Our conclusions draw attention to the fact that fire protection policies, which have undergone diametrical changes over time, must be custom-made and characterised by assent on what actually is a hazard. Further discussions and recommendations drawn based on our results are presented.
|Title of host publication||Regional Intelligence : Spatial Analysis and Anthropogenic Regional Challenges in the Digital Age|
|Number of pages||26|
|Publication date||16 Mar 2020|
|Publication status||Published - 16 Mar 2020|