Introduction
Run-off-road crashes are among the most dangerous accidents on Austrian roads, with injury risks up to seven times higher than in other crash types. These crashes usually happen when drivers lose control and collide with obstacles such as trees, signposts, or gantries. A recent master’s thesis at Z_GIS presents a new approach to pinpoint hazardous roadside sections and strengthen road safety strategies.
Why Roadside Safety Matters
Run-off-road (ROR) crashes remain one of the most pressing safety challenges on Austria’s motorways. Research from a Z_GIS master’s thesis shows that the risk of a ROR crash (0.242) is almost seven times higher than the average risk across all crash types (0.034). Figure 1 shows a comparison of the risk values associated with different crash types. This highlights the urgent need for effective countermeasures – such as forgiving roadsides and self-explanatory road designs – both of which are already emphasised in ASFINAG’s Road Safety Programme 2030.
Figure 1: Distribution of risk of injury crashes on Austrian roads by crash type
The Limitations of Current Assessment Tools
Traditional roadside risk assessment methods – including the Roadside Hazard Scale, Roadside Hazard Rating, and Reliability Index – often fall short. Because they rely on subjective evaluations of road features, their results can be inconsistent and fail to explain the underlying causes of ROR crashes. This gap motivated the development of a more data-driven and transparent framework.
Figure 2: Top seven risk factors for assessing roadside risk
Key Roadside Risk Factors
A literature review combined with crash analysis revealed seven critical roadside risk factors that influence both the likelihood and severity of ROR crashes (see Figure 2). These include:
- Side slopes that allow drivers to regain control (see Figure 3).
- Clear zone width, providing space for safe stopping (see Figure 4).
- Hard shoulder width, where shoulders wider than 3 m significantly reduce crash probability.
- Embankment height, where guardrails become essential once drop-offs exceed 3 m.
- Hazardous objects, loacted adjacent to the carriageway, such as traffic signs and trees (see Figure 5).
When considered together, these factors provide a much more comprehensive picture of roadside safety than existing methods (see Figure 6).
Figure 3: GIS overlay of slope levels and orthoimage
Figure 4: Calculation of the Clear Zone Width for motorway with 3m shoulders width
Figure 5: Trees detected after implementing a pre-trained Deep Learning Model
Figure 6: Distribution of roadside risk factors affecting crash probability
Introducing the RISK Framework
To address the shortcomings of current approaches, a new RISK framework was developed. Based on the principle of equivalent risk, it calculates roadside risk as the product of probability and consequence of failure. This method makes it possible to compare different roadside configurations on a single scale and to identify the role of individual variables – offering valuable insights for engineers and planners.
Who Should Collect the Data?
A major challenge in roadside safety assessment lies in data collection. National Road Agencies often lack the capacity to perform regular, detailed surveys. Here, National Statistical Institutes (NSIs) could step in: with their expertise in systematic and standardised data collection, NSIs are well positioned to support reliable roadside monitoring.
Looking Ahead
Expert discussions with ASFINAG revealed that risk scores alone are not enough. To truly enhance roadside safety, risk assessment must be combined with detailed information on road geometry and design parameters. The RISK framework meets this need by offering both explanatory depth and practical guidance. In the coming months, these results will be discussed further with ASFINAG to explore their integration into future safety assessments on Austria’s motorway network.
About the Author
Christian Stefan at the graduation ceremony after finishing the UNIGIS Master of Science studies (© University of Salzburg)
Christian Stefan studied Spatial Planning and Regional Development at TU Wien before completing a master’s in Geoinformatics in Salzburg. He has worked as a project manager at the Austrian Road Safety Board (KFV) and the Austrian Institute of Technology (AIT), and trained as a Risk Analyst at ETH Zurich. Since 2023, he has been employed as a Geoinformatics Specialist at Statistics Austria.
LinkedIn Christian Stefan: www.linkedin.com/in/christian-stefan-84520730
Link to UNIGIS Master Thesis: unigis.at/files/Mastertheses/Full/106851.pdf
