Modelling of spatial prediction of fire ignition risk in the Antalya-Manavgat district

Abstract

The aim of this study was to present the fire ignition risk for Manavgat-Antalya District to enable the planning of firefighting sources in a more qualified way. From sites within the study area, where forest fires broke out or not during the past five years, we obtained geographical coordinates, climate data, topographical data and variables like bedrock, stand types, settlement areas, roads and power lines and prepared them with geographical information systems. For all variables we performed Wilcoxon rank-sum test, interspecific correlation analysis and logistic regression analysis and obtained 4 different models. When ROC analysis was applied to these models, model 4 was determined as the most significant model and therefore used to prepare the fire ignition risk map for the Manavgat-Antalya District. According to this map, ignition risk within the study area was highest in and around settlement areas where roads and power lines concentrate and Turkish red pine is distributed, but it was lowest afar of settlement areas without roads and where species apart from Turkish red pine are distributed. According to the results some suggestions were made.

Keywords

• Forest fire
• ignition risk
• fire risk
• fire risk map
• forest fıre occurrence probability