MAPPING ROAD CRASH SEVERITY: GIS-BASED NETWORK SCREENING OF STATE ROADS IN VOJVODINA, SERBIA

Effective road safety management requires transparent methods for identifying locations where limited resources can achieve the greatest benefit. This study develops a map-centred, severity-weighted GIS framework for screening state-road sections and critical microlocations in the Autonomous Province of Vojvodina, Serbia. The analysis covered 34,212 police-recorded crashes during 2019–2023, of which 18,723 crashes located on state roads were assigned to 589 predefined sections. Crash consequences were evaluated using the Severity-Weighted Casualty Index (SWCI), with weights of 1, 13 and 99 assigned to slightly injured, seriously injured and fatally injured persons, respectively. Geocoded crashes, section-level scores and critical intersections were processed and mapped in ArcGIS Pro. The maps revealed a network-constrained distribution of crashes and geographically selective concentrations of severe outcomes in western and central Bačka, the Novi Sad–Temerin area, approaches to Zrenjanin and the Pančevo region. The ten highest-ranked sections recorded SWCI values between 610 and 923, with a mean of 757.9 and a median of 773.5. The highest value was recorded on the IB-10 section of the Pančevo–Alibunar corridor. At the microlocation level, the leading site recorded an SWCI of 222 and exceeded the second-ranked site by 69.5%, while the four highest-ranked sites accounted for 68.8% of the total SWCI among the ten selected microlocations. The differences between section- and microlocation-level patterns demonstrate the value of combining corridor and site-specific screening. Although the SWCI represents observed casualty burden rather than exposure-normalised risk, the proposed framework provides a reproducible basis for prioritising field inspections, engineering diagnosis and subsequent road safety interventions.