Geospatial Patterns and Drivers of Forest Fire Occurrence in Romania

Abstract Timely and accurate spatial explicit forest fire risk assessment and mapping is essential for forest fire prevention and suppression preparedness, firefighting resources allocation, and efficient multi-level fire management policies. This paper describes the application and validation of an...
Ausführliche Beschreibung

Abstract Timely and accurate spatial explicit forest fire risk assessment and mapping is essential for forest fire prevention and suppression preparedness, firefighting resources allocation, and efficient multi-level fire management policies. This paper describes the application and validation of an approach for forest fire risk analysis and fire risk zoning over Romania in order to identify areas at national scale where fires are most likely to occur and to threat existing values, resources and assets. A modeling approach based on logistic regression using historical fire observations has been developed based on environmental, socio-economic and demographic data availability at national scale. Ignitions were positively related to south-western slopes and occurred mostly in fuel type of xerophyte oaks as well as in areas of heterogeneous (natural/agricultural) landscape. In addition to the human variables the pattern of ignitions was also significantly related to slope and temperature of the driest quarter. The risk zones produced by the multiple logistic regression model presented satisfactory accuracy when compared with historical fire perimeters extracted from MODIS imagery. The findings of this study could be used by fire managers to implement prevention measures at forest areas with high fire risk. Furthermore, attention should be given to areas with high fire ignition probability, where the vulnerability and potential impact is higher. Ausführliche Beschreibung