Citation

Novkovic I, Markovic GB, Lukic D, Dragićević S, Milosevic M, Djurdjic S, Samardzic I, Lezaic T, Tadic M. Sensors (Basel) 2021; 21(19): e6520.

Copyright

(Copyright © 2021, MDPI: Multidisciplinary Digital Publishing Institute)

DOI

10.3390/s21196520

PMID

34640837

Abstract

The territory of the Republic of Serbia is vulnerable to various natural disasters, among which forest fires stand out. In relation with climate changes, the number of forest fires in Serbia has been increasing from year to year. Protected natural areas are especially endangered by wildfires. For Nature Park Golija, as the second largest in Serbia, with an area of 75,183 ha, and with MaB Reserve Golija-Studenica on part of its territory (53,804 ha), more attention should be paid in terms of forest fire mitigation. GIS and multi-criteria decision analysis are indispensable when it comes to spatial analysis for the purpose of natural disaster risk management. Index-based and fuzzy AHP methods were used, together with TOPSIS method for forest fire susceptibility zonation. Very high and high forest fire susceptibility zone were recorded on 26.85% (Forest Fire Susceptibility Index) and 25.75% (fuzzy AHP). The additional support for forest fire prevention is realized through an additional Internet of Thing (IoT)-based sensor network that enables the continuous collection of local meteorological and environmental data, which enables low-cost and reliable real-time fire risk assessment and detection and the improved long-term and short-term forest fire susceptibility assessment. Obtained results can be applied for adequate forest fire risk management, improvement of the monitoring, and early warning systems in the Republic of Serbia, but are also important for relevant authorities at national, regional, and local level, which will be able to coordinate and intervene in a case of emergency events.

Language: en

Keywords

GIS; fire outbreak occurrence; forest fire susceptibility; fuzzy analytic hierarchy process (fuzzy AHP); IoT sensor networks; random forest (RF); remote sensing; technique for order of preference by similarity to ideal solution (TOPSIS)