Mapping the forest fire risk zones using artificial intelligence with risk factors data

Area of Science:

• Environmental Science
• Geography
• Artificial Intelligence

Background:

• Forest fire risk mapping commonly uses geographical information system (GIS) data.
• Forest fires result from complex interactions, including geographical, meteorological, and human-induced factors.
• Socio-economic conditions significantly influence human-induced factors contributing to forest fires.

Purpose of the Study:

• To develop a comprehensive forest fire risk zone map for Turkey.
• To integrate human-induced factors, natural elements, and temperature into risk assessment.
• To analyze regional variations in forest fire risk across Turkey.

Main Methods:

• Utilized a dataset comprising 11 human-induced factors, one natural factor, and temperature.
• Employed the k-means clustering algorithm, an artificial intelligence technique, for risk zone classification.
• Selected Turkey as the study area due to its diverse regional socio-economic variations.

Main Results:

• Categorized Turkish regional forest directorates into four risk clusters: extreme, high, moderate, and low.
• Generated a visual map illustrating these distinct forest fire risk zones.
• Identified western and southwestern coastal regions of Turkey as generally high-risk areas for forest fires.

Conclusions:

• Forest fire risk is not solely determined by geographical and meteorological factors.
• Human-induced factors, influenced by socio-economic conditions, are critical in fire occurrence.
• The study highlights a higher propensity for forest fires in Turkey's western and southwestern coastal areas compared to northern, central, and eastern regions.