D. 1...SLAIN 2 Version 1.2 Document Control Sheet Version History Input by Consortium partners V1.0...
Transcript of D. 1...SLAIN 2 Version 1.2 Document Control Sheet Version History Input by Consortium partners V1.0...
SLAIN 1 V1.2
Grant Agreement Number: INEA/CEF/TRAN/M2018/179967
Project acronym: SLAIN
Project full title: Saving Lives Assessing and Improving TEN-T Road Network Safety
D. 1.0
Due delivery date: 15 January 2020 Actual delivery date: 6 March 2020
Organisation name of lead participant for this deliverable:
EuroRAP
D1.1, D1.2, D1.3, D1.4, D1.5 (Crash Rate) Risk Mapping
Co-financed by the Connecting Europe Facility of the European Union
SLAIN 2 Version 1.2
Document Control Sheet
Version History
Input by Consortium partners
V1.0 D1.1, D1.2, D1.3, D1.4, D1.5 input by Croatia, Greece, Spain, Italy
V1.1 Review by iRAP Technical Coordinator, Steve Lawson
V1.2 Review by EuroRAP Coordinator, Lina Konstantinopoulou
Legal Disclaimer The information in this document is provided “as is”, and no guarantee or warranty is given that the information is fit for any particular purpose. The above referenced consortium members shall have no liability for damages of any kind including without limitation direct, special, indirect, or consequential damages that may result from the use of these materials subject to any liability which is mandatory due to applicable law. © 2020 by SLAIN Consortium.
Acknowledgement The SLAIN beneficiaries are grateful to the many data providers in Croatia, Greece, Italy and Spain (including SLAIN partners at Dirección General de Tráfico and Generalitat de Catalunya) who have provided data for these maps. The maps were assembled by teams from Anas (Italy), FPZ at the University of Zagreb (Croatia), the RACC Foundation (Spain) and the Road Safety Institute Panos Mylonas (Greece). The report was coordinated and prepared by iRAP, with final editing and liaison with INEA by the project coordinator EuroRAP.
Contact points [email protected] [email protected] and [email protected]
Abbreviations and Acronyms
Acronym Abreviation
Anas Azienda Nazionale Autonoma delle Strade
RACC Real Automóvil Club de Cataluna
SLAIN Saving Lives Assessing and Improving Network Safety
TEN-T Trans-European Transport Network
SLAIN 3 Version 1.2
Table of Contents
1 Objectives .......................................................................................................................................8
1.1 SLAIN project objectives ......................................................................................................... 8
1.2 SLAIN Activity 1 ....................................................................................................................... 8
2 Methodology ..................................................................................................................................9
2.1 Task 1.1: Define the Core TEN-T network to be mapped and resources ................................ 9
2.2 Task 1.2: Allocate traffic data ............................................................................................... 10
2.3 Task 1.3: Disaggregate crashes and allocate to network for each section by type and severity 10
2.4 Task 1.4: Review .................................................................................................................... 11
2.5 Task 1.5: Compute and assess crash risk per kilometre travelled and density of crashes per kilometre .................................................................................................................................... 11
2.6 Task 1.6: Assemble required data and produce high-quality risk maps ............................... 11
3 Results ......................................................................................................................................... 13
3.1 Maps of individual and collective risk ................................................................................... 14
3.1.1 Croatia Comprehensive TEN-T – individual risk (fatal and serious crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020 ........................................................................ 14
3.1.2 Croatia Comprehensive TEN-T – collective risk (fatal and serious crashes per kilometre) with normalisation to Risk Bands 2020 .......................................................................................... 15
3.1.3 Croatia Comprehensive TEN-T – collective risk (fatal and serious crashes per kilometre) based on percentile distribution 40% (low risk), 25%, 20%, 10%, 5% (high risk) .......................... 16
3.1.4 Croatia Comprehensive TEN-T – deviation of risk from risk shown at 3.1.1 ..................... 17
3.1.5 Croatia Comprehensive TEN-T – potential for crash reduction ........................................ 18
3.1.6 Croatia Comprehensive TEN-T – crash costs per kilometre .............................................. 19
3.1.7 Greece Core TEN-T – individual risk (fatal crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020 .................................................................................................. 20
3.1.8 Greece Core TEN-T – collective risk (fatal crashes per kilometre) with normalisation to Risk Bands 2020 ..................................................................................................................................... 21
3.1.9 Italy, Anas network, Comprehensive TEN-T – individual risk (all injury crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020 ............................................................ 22
3.1.10 Italy, Anas network, Comprehensive TEN-T – collective risk (all injury crashes per kilometre) with normalisation to Risk Bands 2020 ........................................................................ 23
3.1.11 Spain state road network (including TEN-T) individual risk (fatal and serious crashes per billion vehicle kilometre) without normalisation to Risk Bands 2020 ........................................... 24
3.1.12 Spain Comprehensive TEN-T – individual risk (fatal and serious crashes per billion vehicle kilometre) without normalisation to Risk Bands 2020 ...................................................... 25
3.1.13 Spain Comprehensive TEN-T – individual risk (fatal and serious crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020 ............................................................ 26
3.1.14 Spain state road network (including TEN-T) collective risk (fatal and serious crashes per kilometre) without normalisation to Risk Bands 2020 ................................................................... 27
SLAIN 4 Version 1.2
3.1.15 Spain Comprehensive TEN-T – collective risk (fatal and serious crashes per kilometre) without normalisation to Risk Bands 2020 .................................................................................... 28
3.1.16 Spain Comprehensive TEN-T – collective risk (fatal and serious crashes per kilometre) with normalisation to Risk Bands 2020 .......................................................................................... 29
3.1.17 Catalonia (Spain) – primary roads network individual risk (fatal and serious crashes per billion vehicle kilometre) without normalisation to Risk Bands 2020 ........................................... 30
3.1.18 Catalonia (Spain) – individual risk Core TEN-T (fatal and serious crashes per billion vehicle kilometre) without normalisation to Risk Bands 2020 ...................................................... 31
3.1.19 Catalonia (Spain) – individual risk Core TEN-T (fatal and serious crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020 ............................................................ 32
3.1.20 Catalonia (Spain) – primary roads network collective risk (fatal and serious crashes per kilometre) without normalisation to Risk Bands 2020 ................................................................... 33
3.1.21 Catalonia (Spain) – Core TEN-T collective risk (fatal and serious crashes per kilometre) without normalisation to Risk Bands 2020 .................................................................................... 34
3.1.22 Catalonia (Spain) – individual risk Core TEN-T (fatal and serious crashes per kilometre) with normalisation to Risk Bands 2020 .......................................................................................... 35
4 Data and analysis by country ...................................................................................................... 36
4.1.1 Croatia ............................................................................................................................... 36
4.1.2 Greece ............................................................................................................................... 37
4.1.3 Italy .................................................................................................................................... 40
4.1.4 Spain .................................................................................................................................. 45
4.1.5 Catalonia (Spain) ................................................................................................................ 48
4.1.6 Spain – illustrative example of analysis without normalisation ........................................ 50
4.1.7 Catalonia (Spain) – illustrative example of analysis without normalisation ..................... 52
4.1.8 Catalonia TEN-T network ................................................................................................... 55
5 Discussion .................................................................................................................................... 57
5.1 Role and purpose .................................................................................................................. 57
5.2 Project learning – considerations for the RISM directive ..................................................... 57
5.2.1 Messages for road-users and operators ............................................................................ 57
5.2.2 Allocation of data, normalisation, ratios and thresholds .................................................. 58
5.2.3 Under-reporting ................................................................................................................ 61
5.2.4 Documentation .................................................................................................................. 61
5.2.5 Scope of mapping .............................................................................................................. 62
6 Conclusions ................................................................................................................................. 63
References .......................................................................................................................................... 64
Appendix 1 – Reporting of crashes and casualties ............................................................................. 65
Appendix 2 – Country meta-analysis .................................................................................................. 67
SLAIN 5 Version 1.2
List of Figures
Figure 1: Core TEN-T in the SLAIN Grant Proposal and Agreement ....................................................... 9 Figure 2: Individual risk bandings .......................................................................................................... 36 Figure 3: Collective risk bandings .......................................................................................................... 37 Figure 4: Frequency of road sections by risk band (individual risk) ...................................................... 38 Figure 5: Frequency of road sections by risk band (collective risk) ...................................................... 39 Figure 6: Distribution of sections number on Anas TEN-T CORE network ............................................ 41 Figure 7: Distribution of sections number on Anas TEN-T Comprehensive network............................ 42 Figure 8: Distribution of sections extension on Anas TEN-T CORE network ......................................... 42 Figure 9: Distribution of sections extension on Anas Comprehensive TEN-T network ......................... 42 Figure 10: Distribution of sections number on Anas TEN-T CORE network .......................................... 43 Figure 11: Distribution of sections number on Anas TEN-T Comprehensive network ......................... 43 Figure 12: Distribution of sections extension on Anas TEN-T CORE network ....................................... 43 Figure 13: Distribution of sections extension on Anas TEN-T network ................................................. 44 Figure 14: individual risk distribution per sections. Spain Comprehensive TEN-T network. ................ 45 Figure 15: individual risk distribution per km. Spain Comprehensive TEN-T network .......................... 46 Figure 16: collective risk distribution per section. Spain Comprehensive TEN-T network. .................. 47 Figure 17: collective risk distribution per km. Spain Comprehensive TEN-T network. ......................... 47 Figure 18: individual risk distribution per sections. Catalan Comprehensive TEN-T network. ............. 48 Figure 19: individual risk distribution per km. Catalan Comprehensive TEN-T network. ..................... 48 Figure 20: collective risk distribution per sections. Catalan Comprehensive TEN-T network. ............. 49 Figure 21: collective risk distribution per km. Catalan Comprehensive TEN-T network. ...................... 49 Figure 22: Individual risk distribution per sections. Spain state road network. .................................... 50 Figure 23: Individual risk distribution per km. Spain state road network ............................................. 51 Figure 24: distribution of collective risk- Spain ..................................................................................... 51 Figure 26: Individual risk distribution per sections. Catalan primary roads network. .......................... 53 Figure 27: Individual risk distribution per km. Catalan primary roads network. .................................. 53 Figure 28: Collective risk distribution per sections. Catalan primary roads network. .......................... 54 Figure 29: Collective risk distribution per km. Catalan primary roads network. .................................. 54 Figure 30: Individual risk distribution per sections. Catalan TEN-T network. ....................................... 55 Figure 31: Individual risk distribution per km. Catalan TEN-T network. ............................................... 55 Figure 32: Collective risk distribution per sections. Catalan TEN-T network. ....................................... 56 Figure 33: Collective risk distribution per km. Catalan TEN-T network ................................................ 56 Figure 34: Risk banding varies dependent upon calibration choice ..................................................... 60 Figure 35: Risk varies substantially when lower tiers of roads are also mapped ................................. 62
List of Tables
Table 1: The distribution of individual and collective risks on TEN-T network ..................................... 37 Table 2: Average AADT and individual and collective risk values for different carriageway types ...... 37 Table 3: Frequency of road sections by risk band (individual risk) ....................................................... 38 Table 4: Frequency of road sections by risk band (collective risk) ........................................................ 39 Table 5: Distribution of network length by risk band (individual and collective risk) ........................... 40 Table 6: Anas TEN-T CORE network- Total km ...................................................................................... 40 Table 7: Anas TEN-T Comprehensive network-Total km ....................................................................... 40 Table 8: Anas TEN-T CORE network....................................................................................................... 41 Table 9: Anas TEN-T Comprehensive network ...................................................................................... 41 Table 10: Collective Risk ........................................................................................................................ 44 Table 11: Individual Risk ........................................................................................................................ 44 Table 12: Relative risk by road type ...................................................................................................... 45
SLAIN 6 Version 1.2
Table 13: Individual risk distribution. Spain Comprehensive TEN-T network. ...................................... 46 Table 14: collective risk distribution. Spain Comprehensive TEN-T network........................................ 47 Table 15: Individual risk distribution. Catalan Comprehensive TEN-T network. ................................... 49 Table 16: collective risk distribution. Catalan Comprehensive TEN-T network. ................................... 50 Table 17: Individual risk distribution. Spain state road network. ......................................................... 51 Table 18: Collective risk distribution. Spain state road network. ......................................................... 52 Table 19: Individual and collective risk rates per road type. Spain state road network. ...................... 52 Table 20: Individual risk distribution. Catalan primary roads network ................................................. 53 Table 21: Collective risk distribution. Catalan primary roads network. ................................................ 54 Table 22: Individual and collective risk rates per road type. Catalan primary roads network. ............ 55 Table 23: Individual risk distribution. Catalan TEN-T network. ............................................................. 55 Table 24: Collective risk distribution. Catalan TEN-T network .............................................................. 56
SLAIN 7 Version 1.2
Executive Summary
Crash Risk Maps have been produced for each of Croatia, Greece, Italy, Spain and the Catalonia region of Spain. In Italy, Croatia and Spain the maps are for the Comprehensive TEN-T and in Greece for the Core TEN-T. In Catalonia maps are presented for the TEN-T and also for primary roads. The maps presented show both individual and collective risk measured in terms of crashes per billion vehicle kilometre and crashes per kilometre respectively. In Croatia and Spain the measure is of fatal and serious crashes. Limitations in the availability and reliability of the data in Greece and in Italy due to under-reporting (Greece) and lack of differentiation between “serious” and “slight” crashes (Italy) mean that in Greece only fatal crashes have been mapped and in Italy an aggregation of all injury levels is presented. Results are being launched or released on websites, subject to local consultation and discussion, as appropriate. Differences in the nature and distribution of these crashes have been analysed and differences in the ratios of fatal to other severities of crash noted on different road types. Such differences have implications for the presentation of Crash Risk Maps and there is discussion of steps to be taken if a map for one country is to be presented alongside others for comparison. Comment is made on the use of maps for the assessment of risk by road-users and by road operators and the application of such maps in the Road Infrastructure Safety Management Directive 2019/1936/EC. Recommendations have been made for updating the EuroRAP guidelines on crash rate risk mapping.
SLAIN 8 Version 1.2
1 Objectives
1.1 SLAIN project objectives
The project’s Action fits in the EC’s 2010 Communication ‘Towards a European Road Safety Area’ and aims to contribute to the long-term goal for zero road deaths in 2050. With partners in the different countries, Project SLAIN is a transnational project aiming to extend the skills and knowledge base of partners in performing network-wide road assessment.
The main areas to be covered within the SLAIN project are
• Demonstration of a methodology of network-wide assessment
• Assessment of the Safety Performance Management of the TEN-T core road network and, if possible beyond, in 4 European countries: Croatia, Italy, Greece and Spain where we will perform road surveys (10,000 km of mapping)
• Proposals of section-specific, economically-viable crash countermeasures designed to raise infrastructure quality to achieve significant reductions in severe injuries and deaths
• Preparation of the readiness of Europe’s physical infrastructure for automation
The SLAIN consortium consists of eight core partners, coming from six EU member states, namely Greece, Italy, Spain, Croatia, UK and Belgium. The list of partners are EuroRAP - Project Coordinator, Anas, FPZ, RSI Panos Mylonas, RACC-ACASA, DGT Spain, SCT Spain, TES Spain (Catalonia), iRAP.
1.2 SLAIN Activity 1
The objective of Activity 1 is to produce maps showing crash risk as an overall part of network-wide road assessment for Croatia, Greece, Italy, Spain (and Catalonia). Subject to the availability of appropriate data, the objective of this task was to produce a Crash Risk Map of death and serious injury for each country illustrating both the individual risk and the collective risk for crashes for at least sections of the TEN-T Core Network in each of the four countries. This provides a preliminary and immediate basis for comparison of the safety of the networks being examined and is often used as the basis for further analysis. Such maps can be used to compare current performance and also track that over time. The relevant beneficiaries in each territory have been responsible for producing the map in that territory and for collecting the data from which they are formed. Crash Risk Maps are a convenient and relatively inexpensive means of portraying risk across a network and how that changes as one travels from one road section to the next. They relate the number of severe crashes to the amount of vehicle travel on each section (crashes per billion vehicle kilometre) or to the length of the section (crashes per kilometre) for given time periods. Mapped over time, the crash rates of individual road sections can track the performance of the road. The report on these Deliverables (D1.1-D1.5) is written both from the perspective of an analysis of the maps and data and from the perspective of assessing how useful such information is as a tool in network-wide road safety assessment. It is also written from the perspective of these Deliverables being a pilot exercise in Crash Risk Mappings, with refinement of the techniques and analysis being provided in a second round of mapping due in January 2021. That work will include assessment of changes over time and comparison between countries.
SLAIN 9 Version 1.2
2 Methodology
2.1 Task 1.1: Define the Core TEN-T network to be mapped and resources
The network for the Core TEN-T in each of Croatia, Greece, Italy, Spain (including a separate map for Catalonia) has been identified and mapped. In Italy the network is limited to that operated by Anas and in Italy, Croatia and Spain, includes the Comprehensive TEN-T. The methodology used is described in the RAP-RM-2-1 Risk Mapping Technical Specification in the methodology section of the iRAP website: http://resources.irap.org/Specifications/RAP-RM-2-1_Risk_Mapping_technical_specification.pdf.
RAP-RM-2-1 sets out the technical specification for the production of RAP Risk Mapping to a standardised format. It details how networks are constructed and the rationale for the selection of road sections and their related parameters in building a data set. RAP-RM-3-1 sets out the design and cartographic specification for the production of RAP Risk Mapping to a standardised format and will be considered for use in future productions of these maps. It too is stored on the iRAP website: http://resources.irap.org/Specifications/RAP-RM-3-1_Risk_Mapping_design_specification.pdf.
The mapping in the Grant Agreement is limited to the relevant Core TEN-T network (see Figure 1) although, as described above, at no additional cost to the project, in some circumstances it has been possible to provide mapping that includes other roads and to include roads included in the Comprehensive TEN-T.
Figure 1: Core TEN-T in the SLAIN Grant Proposal and Agreement
SLAIN 10 Version 1.2
2.2 Task 1.2: Allocate traffic data
This task involves collecting Annual Average Daily Traffic (AADT) volumes for each section from counts provided by the relevant road authorities in each of Croatia, Greece, Italy, Spain (and including a separate map of Catalonia). In Greece, for those Core TEN-T roads operated by concessionaires, data were collected by the relevant department of the Ministry. Traffic flows were assigned according to the collected data. The data were the latest currently available (i.e. for the most recent time periods, including years 2014-2019) and not previously collected for earlier studies. In Italy, the data were limited to the Core TEN-T network for which Anas has responsibility as road authority (see Figure 1) and supplemented by data included in the nearby Comprehensive TEN-T. Italy can act as an exemplar of other countries. In Italy traffic data on Anas road network are acquired and analysed with two different platforms that are implemented to facilitate interaction and mutual exchange of information:
• Automatic System for Traffic Statistical Survey;
• DSS – Decision Support System. The Automatic System for the Traffic Statistical Survey consists of over 1,150 counting sections distributed over the entire Anas road network, recording traffic volumes and conditions 24 hours a day, 365 days a year. Data acquired are checked, stored and processed by PANAMA platform (Anas monitoring and analysis platform). The average annual daily traffic (AADT), is elaborated for each counting section, and then transferred to the DSS, a mathematical traffic simulation model of national level, which enables estimates of AADT of light and heavy vehicles on the entire road network managed by Anas. This estimation procedure was used to obtain the average annual daily traffic (AADT) for the year 2015, 2016 and 2017 on the TEN-T network for which Anas has responsibility.
2.3 Task 1.3: Disaggregate crashes and allocate to network for each section by type and severity
The beneficiaries in Croatia, Greece, Italy and Spain have assembled an Excel spreadsheet, with the number of fatal and serious crashes for each road section disaggregated by the principal road user groups and the total. As an example, in Croatia FPZ collected all relevant data on traffic crashes for the TEN-T network road sections within the Republic of Croatia. The data about the number, severity, types and locations of road traffic crashes in the observed time period were collected from the available publications. This included the Road Traffic Safety Bulletin and data from official Ministry of Interior, Croatian Highways Ltd, Croatian Roads Ltd databases and additional data provided from remaining relevant highway concessionaires. The collected data on road traffic crashes were disaggregated by type and severity and then assigned to the corresponding road network sections. In Greece, all relevant data on road crashes within the Core TEN-T were collected from the Hellenic Statistical Authority. The data were the latest currently available (i.e. for the most recent time periods, including years 2014-2017). The collected data on road crashes were disaggregated by severity and then assigned to the corresponding road network sections. In Spain, the data related to the road crashes has been sourced by the Dirección General de Tráfico (Spanish road safety authority) and the data regarding average traffic from the Ministerio de Fomento. Regarding Catalonia, the data regarding road crashes has been sourced by the Servei Català de Trànsit (Catalan road safety authority) and the data related the average traffic is sourced by the owners of the different roads of the network analysed (Generalitat de Catalunya, Diputacions and Ministerio de Fomento). The iRAP/EuroRAP methodology sets an aspirational target of 20 fatal and serious crashes per network section but it has been noted in practice that in many circumstances this is impossible to achieve
SLAIN 11 Version 1.2
without extending the length of the network to a length that diminishes the ability to differentiate risk (and in particular to identify higher risk road lengths) or means that it is necessary to use data from much longer time periods. In Greece, but also in the other countries, it was noticeable that there is substantial under-reporting of crashes. In Greece, the number of reported serious crashes was actually lower than the number of fatal crashes whereas one would typically expect the ratio of fatal:serious to be in the region of 1:10. A summary of the potential extent of under-reporting is provided in Appendix 1. Due to the significant under-reporting of crashes in Greece, only fatal crashes were taken into consideration, in order to compute and assess crash risk per kilometre travelled and density of crashes per kilometre. In addition, the assessment period was extended to 4 years, instead of the standard 3-year period, as the minimum target of 20 fatal and serious crashes per network section could not be achieved. The relevant assessment period considers years 2014 to 2017. In Italy, it was not possible to disaggregate between serious and slight crashes and therefore the maps reflect the risk for all injuries. In Italy crash data are acquired by national institute of statistics and located on the national road network by ACI (Automobile Club Italia), which provided to Anas all crash data occurred on the TEN-T network. Currently, in Italy it is not possible to disaggregate between serious and slight crashes, so all the crash with injuries regardless of the severity of the consequences were considered in producing Crash Risk Maps. Where there were residual crashes not allocated to the road system, they were excluded from the spatial analysis in the mapping.
2.4 Task 1.4: Review
Data were reviewed for accuracy of allocation and for underreporting. It is well-known for example that some crash types are under-reported, notably low severity and pedestrian crashes. Comment has been passed on the relevant observations in the reporting of results.
2.5 Task 1.5: Compute and assess crash risk per kilometre travelled and density of crashes per kilometre
Calculations were based upon crashes divided by the amount of traffic using the road or by the number of crashes per kilometre and ranked using an excel file. The task computed the crash risks according to the standard procedures for RAP Risk Mapping Type I: Individual crash risk per vehicle km travelled and RAP Risk Mapping Type II: Crash density (Collective or Community risk). Crash risk per kilometre travelled (Type I Crash risk) is expressed as the number of fatal and serious crashes per billion vehicle kilometres travelled. This is the risk for individual road users of being involved in fatal or serious crash whilst using a specific road length. The determined crash risk rates were then allocated into five RAP risk bandings (low, low-medium, medium, medium-high and high risk categories) and the standard Type I and Type II Risk Maps were produced.
2.6 Task 1.6: Assemble required data and produce high-quality risk maps
This task was done using mapping shapefiles. In order to produce the EuroRAP Risk Maps, the data on road network geometry, road traffic crashes and road traffic volume data, extracted from the relevant databases were recorded in shapefile (.shp) format, compatible with the webGIS systems which was used for further data processing and calculation of crash risk and crash density rates. The resulting Crash Risk Maps were also stored in shapefile format in order to enable the fast and easy data transfer between stakeholders. Unless otherwise stated, the maps presented are normalised for comparison between countries
SLAIN 12 Version 1.2
using “Risk Bands 2020”. The rationale and methodology adopted is explained at section 8.2.1, from page 35 at: http://resources.irap.org/Specifications/RAP-RM-2-1_Risk_Mapping_technical_specification.pdf
SLAIN 13 Version 1.2
3 Results
One set of results is provided for each of Croatia, Greece, Italy and Spain plus Catalonia with maps of both individual and collective risk and accompanying description of each. Results are being launched or released on websites, subject to local consultation and discussion, as appropriate. Maps are presented in the style and design of the beneficiaries’ own organisation and in manner familiar to national stakeholders. In Croatia, three additional maps are presented – a map based on the percentage split of road sections (40% in the lowest risk band and then 25%, 20% 10% and, in the highest, 5%), one showing the “potential for crash reduction” (comparing risk with roads of a similar type) and another of crash costs. In Spain, additional maps are shown for risk on the state network and also for primary roads in Catalonia – as indicated, these maps use thresholds determined for national and local use and illustrate an application of the maps within this context, rather than for pan-European comparison.
SLAIN 14 Version 1.2
3.1 Maps of individual and collective risk
3.1.1 Croatia Comprehensive TEN-T – individual risk (fatal and serious crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020
SLAIN 15 Version 1.2
3.1.2 Croatia Comprehensive TEN-T – collective risk (fatal and serious crashes per kilometre) with normalisation to Risk Bands 2020
SLAIN 16 Version 1.2
3.1.3 Croatia Comprehensive TEN-T – collective risk (fatal and serious crashes per kilometre) based on percentile distribution 40% (low risk), 25%, 20%, 10%, 5% (high risk)
SLAIN 17 Version 1.2
3.1.4 Croatia Comprehensive TEN-T – deviation of risk from risk shown at 3.1.1
SLAIN 18 Version 1.2
3.1.5 Croatia Comprehensive TEN-T – potential for crash reduction
SLAIN 19 Version 1.2
3.1.6 Croatia Comprehensive TEN-T – crash costs per kilometre
SLAIN 20 Version 1.2
3.1.7 Greece Core TEN-T – individual risk (fatal crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020
SLAIN 21 Version 1.2
3.1.8 Greece Core TEN-T – collective risk (fatal crashes per kilometre) with normalisation to Risk Bands 2020
SLAIN 22 Version 1.2
3.1.9 Italy, Anas network, Comprehensive TEN-T – individual risk (all injury crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020
SLAIN 23 Version 1.2
3.1.10 Italy, Anas network, Comprehensive TEN-T – collective risk (all injury crashes per kilometre) with normalisation to Risk Bands 2020
SLAIN 24 Version 1.2
3.1.11 Spain state road network (including TEN-T) individual risk (fatal and serious crashes per billion vehicle kilometre) without normalisation to Risk Bands 2020
The Spanish state road network (Red de carreteras del estado) is the network managed by the Spanish Ministry for Public Works and Transport. It includes 1,388 road sections, that make up a total of 25,082km.
SLAIN 25 Version 1.2
3.1.12 Spain Comprehensive TEN-T – individual risk (fatal and serious crashes per billion vehicle kilometre) without normalisation to Risk Bands 2020
Source of TEN-T network roads: TENTEC portal: https://ec.europa.eu/transport/infrastructure/tentec/tentec-portal/site/maps_upload/annexes/annex1/Annex%20I%20-%20VOL%2017.pdf
SLAIN 26 Version 1.2
3.1.13 Spain Comprehensive TEN-T – individual risk (fatal and serious crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020
SLAIN 27 Version 1.2
3.1.14 Spain state road network (including TEN-T) collective risk (fatal and serious crashes per kilometre) without normalisation to Risk Bands 2020
SLAIN 28 Version 1.2
3.1.15 Spain Comprehensive TEN-T – collective risk (fatal and serious crashes per kilometre) without normalisation to Risk Bands 2020
SLAIN 29 Version 1.2
3.1.16 Spain Comprehensive TEN-T – collective risk (fatal and serious crashes per kilometre) with normalisation to Risk Bands 2020
SLAIN 30 Version 1.2
3.1.17 Catalonia (Spain) – primary roads network individual risk (fatal and serious crashes per billion vehicle kilometre) without normalisation to Risk Bands 2020
The Catalan primary roads network (xarxa bàsica) is a network of 6,300km of interurban roads owned by the State, the Generalitat de Catalunya (regional government) and the provincial councils.
TEN-T source: Generalitat de Catalunya.
http://territori.gencat.cat/es/03_infraestructures_i_mobilitat/carreteres/observatori_viari_de_catalunya_viacat/descripcio_xarxa_autopistes_vies_alta_capacitat/2_3_caracteritzacio_de_la_xarxa/2_3_3_xarxa_transeuropea/
SLAIN 31 Version 1.2
3.1.18 Catalonia (Spain) – individual risk Core TEN-T (fatal and serious crashes per billion vehicle kilometre) without normalisation to Risk Bands 2020
SLAIN 32 Version 1.2
3.1.19 Catalonia (Spain) – individual risk Core TEN-T (fatal and serious crashes per billion vehicle kilometre) with normalisation to Risk Bands 2020
SLAIN 33 Version 1.2
3.1.20 Catalonia (Spain) – primary roads network collective risk (fatal and serious crashes per kilometre) without normalisation to Risk Bands 2020
SLAIN 34 Version 1.2
3.1.21 Catalonia (Spain) – Core TEN-T collective risk (fatal and serious crashes per kilometre) without normalisation to Risk Bands 2020
SLAIN 35 Version 1.2
3.1.22 Catalonia (Spain) – individual risk Core TEN-T (fatal and serious crashes per kilometre) with normalisation to Risk Bands 2020
SLAIN 36 Version 1.2
4 Data and analysis by country
The following tables show histograms for the frequency distributions of low-high risk for each of the four countries for individual and collective risk.
4.1.1 Croatia
The TEN-T network in the Republic of Croatia is approximately 1,600km long and was divided into 145 individual road sections for the purpose of conducting the Road Risk Mapping methodology. The TEN-T network is mostly comprised of motorways, about 1,300km, and the rest are state roads, about 300km in length. The following two figures (Figure 1 and 2) show histograms of the frequency distribution of low-high risk rates on the Croatian TEN – T network for adjusted individual and collective risk bands. To assess the current risk rates on the Croatian network, data such as number of road segments, length of road sections, number of reported serious and fatal crashes, AADT and speed limit has been used. Data containing information about road sections which are part of the Croatian TEN–T road network was obtained from the Ministry of the Sea, Transport and Infrastructure while data regarding traffic crashes was obtained from the Ministry of the Interior. Data regarding Average Annual Daily Traffic was obtained from the official publication of traffic counting issued by Croatian Roads Ltd. In the assessment, 2020 risk bands were used, which were adjusted with a factor calculated specifically for the Republic of Croatia. In the case of Croatia, major part of the TEN – T road network is covered by highways (about 80%) while the remaining part (about 20%) is covered by state roads. By design, highways are the safest road category and that is the reason why Croatia has most of it road sections classified as “low” and “middle-low” risk band (Table 1).
Figure 2: Individual risk bandings
SLAIN 37 Version 1.2
Figure 3: Collective risk bandings
Table 1: The distribution of individual and collective risks on TEN-T network
Risk category Individual risk (road sections) [%]
Collective risk (road sections) [%]
Low 37% 40%
Low-Medium 48% 25%
Medium 7% 20%
Medium-High 5% 10%
High 3% 5%
Table 2: Average AADT and individual and collective risk values for different carriageway types
Carriageway type Individual risk (Risk rates)
Collective risk (Crash density)
AADT (veh/day)
Single carriageway
80.83 0.23 9960
Dual carriageway (motorways)
20.34 0.11 15783
The analysis of risk rates on different carriageway types (Table 2) shows that individual risk rates are 4 times higher on single carriageway roads than on dual carriageway roads (motorways), while crash density is 2 times higher. This is due to the fact that dual carriageway roads are inherently safer than single carriageway roads, while average AADT values are not that far apart, the difference is around 6,000 veh/day.
4.1.2 Greece
In Greece, the Core TEN-T network is approximately 1,700km long and is mostly operated by concessionaires. The major part of the Greek Core TEN-T network is currently comprised of motorways (about 95%), while the remaining part (about 5%) consists of single carriageways.
SLAIN 38 Version 1.2
The examined network was divided into 39 individual road sections, in order to meet all the required conditions and regulations (i.e. uniform traffic flow values, geometrical and functional characteristics) for the purpose of producing the maps for both individual and collective risk. During the assessment process, data such as number and length of road segments, number of fatal crashes and AADT was used. In addition, 2020 risk bands were used, which were adjusted with a scaling factor calculated specifically for the Core TEN-T network of Greece. Figures below show a frequency distribution of road sections of Greece by risk band (regarding both individual and collective risk, respectively).
Figure 4: Frequency of road sections by risk band (individual risk)
In terms of individual risk (Figure 3), it can be noted that approximately 60% of the road sections fall into the low or low-medium risk rate, while no segment is characterized by high risk. The relevant data from which Figure 3 was produced is shown in Table 3.
Table 3: Frequency of road sections by risk band (individual risk)
Risk category Individual risk Individual risk
(road sections) (road sections) [%]
Low 1 3%
Low-Medium 22 56%
Medium 9 23%
Medium-High 7 18%
High 0 0%
SLAIN 39 Version 1.2
Figure 5: Frequency of road sections by risk band (collective risk)
With regard to collective risk, it is observed that almost 70% of the Core TEN-T road sections fall into the low risk rate, while 95% of the total network belongs either to the low or low-medium risk rate. It is also remarkable that no road segment is characterized by medium-high or high risk. In addition, Table 4 below displays the data used for the production of Figure 4.
Table 4: Frequency of road sections by risk band (collective risk)
Risk category Collective risk Collective risk
(road sections) (road sections) [%]
Low 27 69%
Low-Medium 10 26%
Medium 2 5%
Medium-High 0 0%
High 0 0%
Typically, such crash rate distributions show a skewed distribution with many more road sections in the low end of the scale and relatively few at the top end. Those at the top end are likely to be those that attract priority attention for countermeasures. Particularly in Greece, from Figures 3 & 4, it is fully comprehensible that the majority of the road sections belong to the low or low-medium risk rate, while no segment is characterized as high risk. This kind of distribution is easily explained due to the fact that the majority of low or low-medium road segments belong to the motorway network of the country, indicating the crucial efforts that have been made towards road safety improvement during the latest years, with the completion of significant motorway axes. After all, motorways provide the highest standards of road safety to road-users. Finally, Table 5 displays the distribution of the network length by risk band, for both individual and collective risk.
SLAIN 40 Version 1.2
Table 5: Distribution of network length by risk band (individual and collective risk)
Risk category Individual risk Individual risk Collective Risk Collective Risk
(length) [km] (length) [%] (length) [km] (length) [%]
Low 43,00 2% 1298,00 75%
Low-Medium 1077,00 62% 351,50 20%
Medium 326,50 19% 74,00 5%
Medium-High 277,00 17% 0 0%
High 0 0% 0 0%
4.1.3 Italy
The TEN-T network in Italy is comprised of infrastructures managed by different operators including Anas S.p.A.. The total extension of the TEN-T network managed by Anas is about 4,200km, of which 790km of Core network, about 19% of total TEN-T. The Anas’ TEN-T road network is prevalent in the southern regions including the islands of Sicily and Sardinia (86% of Core and 63% of the total TEN-T) and follows the distribution shown below on the national territory (Table 6).
Table 6: Anas TEN-T CORE network- Total km
Area Total km %
Northern Italy 40 5%
Central Italy 68 9%
South Italy 682 86%
Table 7: Anas TEN-T Comprehensive network-Total km
Area Total km %
Northern Italy 397 9%
Central Italy 1140 27%
South Italy 2668 63%
The subdivision into sections was carried out in a step by step method. First step was based on traffic flow values and functional characteristics, then for sections not compliant with the “EuroRAP202 risk mapping technical specification” for the extent and number of crashes, adjacent sections with the same functional characteristics and comparable traffic volumes were aggregated. In the other cases, the sections identified were maintained even if characterised by a number of crashes over 3 years of less than 20 and with a length of less than 5km (single carriageway) or 10km (motorways and dual carriageways). A total of 233 sections have been identified on this network, of which 42 are on the Core network.
SLAIN 41 Version 1.2
Table 8: Anas TEN-T CORE network
Area N. sections %
Northern Italy 6 14%
Central Italy 8 19%
South Italy 28 67%
Table 9: Anas TEN-T Comprehensive network
Area N. sections %
Northern Italy 26 11%
Central Italy 77 33%
South Italy 130 56%
For each of the sections, crash data and traffic flows were then used to calculate the density of crashes (collective risk) and the risk rate (individual risk). Starting from the 2020 risk bands indicated in the EuroRAP methodological document, the thresholds for each band have been calculated using a scaling factor specific for Italy. This factor was identified with the aim of obtaining bands that could also be used for further risk representation that can be done on different parts of Anas road network. Therefore, the scaling factor has been calculated considering the total number of crashes detected and located on the whole Italian main network and not only on the basis of those falling on the TEN-T network. The following graphs show the distribution of the number of sections and the network extension in each of the risk bands, both considering the Core network and the entire Anas’ TEN-T network. As for the number of sections and with reference to collective risk, i.e. crash density, it can be observed that on the Core network the number of high risk sections is equal to that of low and Low-Medium risk ones, while for the whole TEN-T network the highest number of sections has a low risk level. High risk sections on the Core network represents more than 60% of the total number of high risk sections on the TEN-T network (Figures 5 and 6).
Figure 6: Distribution of sections number on Anas TEN-T CORE network
SLAIN 42 Version 1.2
Figure 7: Distribution of sections number on Anas TEN-T Comprehensive network
In terms of network extension (Figures 7 and 8), it can be noted that more than 80% of the network (both Core and the whole TEN-T) is characterized by low or medium/low risk. The extent of the high-risk network represents 13% of the Core and 4% of the entire TEN-T network.
Figure 8: Distribution of sections extension on Anas TEN-T CORE network
Figure 9: Distribution of sections extension on Anas Comprehensive TEN-T network
SLAIN 43 Version 1.2
With regard to individual risk, i.e. the ratio between the number of crashes and the billions of vehicles*km travelled, the distributions over the risk classes of both the number of sections and their extent are similar and shifted to the lower risk classes. In particular, most of the sections fall into the middle/low class (Figures 9 and 10) as well as most of the network extension (Figures 11 and 12).
Figure 10: Distribution of sections number on Anas TEN-T CORE network
Figure 11: Distribution of sections number on Anas TEN-T Comprehensive network
Figure 12: Distribution of sections extension on Anas TEN-T CORE network
SLAIN 44 Version 1.2
Figure 13: Distribution of sections extension on Anas TEN-T network
The distribution of both collective and individual risk over the National territory has similarities: most of the network with low crash density and low crash risk per kilometre travelled falls in the regions of southern Italy, while the most part of the highest risk class network is located in the central regions.
Table 10: Collective Risk (% of the extension by level of risk and geographical area)
NORTH CENTER SOUTH
L 7% 26% 67%
L-M 18% 20% 62%
M 9% 36% 55%
M-H 18% 31% 51%
H 4% 68% 28%
Table 11: Individual Risk (% of the extension by level of risk and geographical area)
NORTH CENTER SOUTH
L 4% 39% 57%
L-M 7% 25% 68%
M 19% 20% 60%
M-H 16% 45% 39%
H 9% 69% 22%
Almost all sections with crash density high risk have a lower risk level per travelled kilometre, the only exception on five sections and a total extension of 57km with a high risk level for both indicators. Four of this five sections have a very low extension and this aspect is supposed to affects the values of the two risk parameters. Finally, if the values of the risk indicators for the different types of infrastructure
SLAIN 45 Version 1.2
(motorway, dual carriageway and single carriageway) are analysed, the following average values are obtained for the whole TEN network (Table 12).
Table 12: Relative risk by road type
Collective risk
(Crash density)
Individual Risk
(Risk Rate)
MOTORWAY 7,45 141,57
DUAL CARRIAGEWAY 2,18 108,84
SINGLE CARRIAGEWAY 2,06 191,08
Motorways have collective risk levels more than three times higher than dual carriageways and single carriageways. Motorway infrastructures, however, are those that have AADT average value three times higher than those of single carriageway infrastructures. For this reason, when considering individual risk, single carriageway infrastructures are the one with the highest average value.
4.1.4 Spain
Analysis has been made of the Comprehensive TEN-T in Spain, comprising 507 road sections, that make up a total of 11,054km of the State Road Network managed by the Spanish Ministry for Public Works and Transport. The analysis considers the number of serious and fatal accidents of the last three years (2016-2018). For each of the sections, accident data and traffic flows were used to calculate the risk rate (individual risk) and the density of accidents (collective risk). The analysis referred here and in the following section for Catalonia uses Risk Bands 2020, normalising with the other countries participating in the SLAIN Crash Risk Mapping. The distribution of individual risk by both: on number of road section and extension in km is shown in Figure 13.
Figure 14: individual risk distribution per sections. Spain Comprehensive TEN-T network.
0
50
100
150
200
250
300
Low Low - medium Medium Medium - high High
Num
ber
of
road s
ections
Fatal crash rate per km
Adjusted individual risk band
SLAIN 46 Version 1.2
Figure 15: individual risk distribution per km. Spain Comprehensive TEN-T network
The data from which from these figures are produced is shown in Table 13.
Table 13: Individual risk distribution. Spain Comprehensive TEN-T network.
This data shows that the Spanish Spain EuroRAP Comprehensive TEN-T has a total of 868km with a ‘high’ or ‘medium high’ risk of having a serious or fatal accident, accounting for 8% of the total number of kilometres of the analysed network that is a reduction by 1.2 compared to the previous year. However, the band that clearly concentrates a higher percentage of kilometres is the low-medium band with a 63% of the total km of the TEN-T network. The distribution of collective risk by both: on number of road section and extension in km is shown in Figure 15.
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
Low Low - medium Medium Medium - high High
Num
ber
of
road k
ilom
ete
rs
Fatal crash rate per km
Adjusted individual risk band
Risk Km %km Sections %sections
Low 2.399 22% 132 26%
Low - medium 6.971 63% 272 54%
Medium 815 7% 48 9%
Medium - high 651 6% 40 8%
High 217 2% 14 3%
Total 11.054 100% 506 100%
SLAIN 47 Version 1.2
Figure 16: collective risk distribution per section. Spain Comprehensive TEN-T network.
Figure 17: collective risk distribution per km. Spain Comprehensive TEN-T network.
And the table from which these graphics are produced in Table 14.
Table 14: collective risk distribution. Spain Comprehensive TEN-T network.
This data shows that most of the sections and km have collective risk rate of fatal and serious accidents is the low band.
SLAIN 48 Version 1.2
4.1.5 Catalonia (Spain)
The Catalonia Comprehensive TEN-T network comprises 47 road sections, that make up a total of 1.078km of the State Road Network, the network that is managed by the Spanish Ministry for Public Works and Transport. The analysis considers the number of serious and fatal accidents of the last three years (2016, 2017 and 2018). The distribution of individual risk by both: on number of road section and extension in km is shown in the figures below.
Figure 18: individual risk distribution per sections. Catalan Comprehensive TEN-T network.
Figure 19: individual risk distribution per km. Catalan Comprehensive TEN-T network.
0
5
10
15
20
25
30
Low Low - medium Medium Medium - high High
Num
ber
of
road s
ections
Fatal crash rate per km
Adjusted individual risk band
0
100
200
300
400
500
600
700
Low Low - medium Medium Medium - high High
Num
be o
f ro
ad k
ilom
ete
rs
Fatal crash rate per km
Adjusted individual risk band
SLAIN 49 Version 1.2
Table 15: Individual risk distribution. Catalan Comprehensive TEN-T network.
In the road network analysed there are 70km with a ‘high’ or risk of having a serious or fatal accident, which account for 6 of the analysed network. However, the band that clearly concentrates a higher percentage of kilometres is the low-medium band with a 57% of the total km of the TEN-T network. The distribution of collective risk by both: on number of road section and extension in km is shown in the figures below.
Figure 20: collective risk distribution per sections. Catalan Comprehensive TEN-T network.
Figure 21: collective risk distribution per km. Catalan Comprehensive TEN-T network.
Risk Km %km Sections %sections
Low 379 35% 13 28%
Low - medium 611 57% 28 60%
Medium 20 2% 2 4%
Medium - high 24 2% 1 2%
High 43 4% 3 6%
Total 1.078 100% 47 100%
SLAIN 50 Version 1.2
Table 16: collective risk distribution. Catalan Comprehensive TEN-T network.
This data shows that most of the sections and km have collective risk rate of fatal and serious accidents is the low band.
4.1.6 Spain – illustrative example of analysis without normalisation
The analysis referred to here and in the following section for Catalonia does not normalise with the other countries participating in the SLAIN Crash Risk Mapping. It shows risk bands thresholds developed in Spain and in the Catalonia region of Spain to provide greater differentiation of risk based upon local needs and priorities. Separately from the analysis of the Comprehensive TEN-T, Spain analysed 1,388 road sections, that make up a total of 25,082km of the State Road Network, the network that is managed by the Spanish Ministry for Public Works and Transport. The analysed section are used by 52% of the road mobility in Spain; in other words, more than half of the kilometres covered by the vehicle fleet of the country. The analysis considers the number of serious and fatal crashes of the last three years (2016-2018).
For each of the sections, crash data and traffic flows were used to calculate the density of crashes (collective risk) and the risk rate (individual risk). The distribution of individual risk by both: on number of road section and extension in km is shown in Figure 21.
Figure 22: Individual risk distribution per sections. Spain state road network.
Concentration Km %km Sections %sections
Low 1.015 94% 42 89%
Low - medium 55 5% 4 9%
Medium 7 1% 1 2%
Medium - high 0 0% 0 0%
High 0 0% 0 0%
Total 1.078 100% 47 100%
SLAIN 51 Version 1.2
Figure 23: Individual risk distribution per km. Spain state road network
And the table from which from these tables are produced is shown in Table 17.
Table 17: Individual risk distribution. Spain state road network.
This data shows that the Spanish State Road Network (SRN) has a total of 2,753km with a ‘high’ or ‘medium high’ risk of having a serious or fatal crash, accounting for 11% of the total number of kilometres of the analysed network that is a reduction by 1.2 compared to the previous year. However the band that clearly concentrates a higher percentage of kilometres is the low band. The distribution of collective risk by both: on number of road section and extension in km is shown in Figure 23.
Figure 24: distribution of collective risk- Spain
Risk Km % Km Sections % sections
Low 9.827 39% 574 41%
Low - medium 9.053 36% 457 33%
Medium 3.450 14% 193 14%
Medium - high 2.072 8% 114 8%
High 680 3% 50 4%
Total 25.082 100% 1.388 100%
SLAIN 52 Version 1.2
Figure 25: Collective risk distribution per km. Spain state road network.
And the table from which these graphics are produced in Table 18. Table 18: Collective risk distribution. Spain state road network.
This data shows that most of the sections and km have collective risk rate of fatal and serious crashes in the 0.1-0.2 band. Very few km and sections on the 0.5-1 are recorded. Besides an analysis of the individual and the collective risk has been developed by type of road: single or dual carriage way. Please note that that the normal classification developed in Spain is Autovías, Autopistas and convencionales. For homogenization with the SLAIN project Autovías and Autopistas have merged into the dual carriageway category being the convencionales the single carriageways. There is not a direct twin category for motorways. This analysis is shown in Table 19.
Table 19: Individual and collective risk rates per road type. Spain state road network.
The analysis shows that the risk on the single carriageways is 3.7 times higher than in dual carriageways.
4.1.7 Catalonia (Spain) – illustrative example of analysis without normalisation
The EuroRAP study has been published every year in Catalonia since 2002, and it analyses more than 6,300 km of interurban roads owned by the State, the Autonomous Community and the Provincial Councils. The road network analysed by EuroRAP accounts for 53% of the total network and 90% of the Catalan road mobility. The analysis considers the number of serious and fatal crashes of the last three years (2016-2018). For each of the sections, crash data and traffic flows were used to calculate the density of crashes (collective risk) and the risk rate (individual risk). Starting from the 2020 risk bands indicated in the
Concentration Km % Sections %
Without FSA 4.063 16% 345 25%
< 0.1 17.420 69% 782 56%
0.1 - 0.2 2.690 11% 174 13%
0.2 - 0.5 875 3% 80 6%
0.5 - 1 34 0% 7 1%
Total 25.082 100% 1.388 100%
Carriageway typeIndividual risk
(Risk rates)
Collective risc
(Crash density)
AADT
(veh/day)
Single carriageway 26,80 0,05 5.840
Dual carriageway (autovías +autopistas) 7,27 0,08 27.883
SLAIN 53 Version 1.2
EuroRAP methodological document, the thresholds for each band have been calculated using a scaling factor. The distribution of individual risk by both: on number of road section and extension in km is shown in the figures below.
Figure 26: Individual risk distribution per sections. Catalan primary roads network.
Figure 27: Individual risk distribution per km. Catalan primary roads network.
Table 20: Individual risk distribution. Catalan primary roads network
In the road network analysed by EuroRAP (which accounts for 53% of the total network and 90% of the Catalan road mobility), there are 1,005km with a ‘high’ or risk of having a serious or fatal crash, which account for 16% of the analysed network. The distribution of collective risk by both: on number of road section and extension in km is shown in the figures below.
Risk Km %km Sections %km
Low 1.011 16% 75 18%
Low - medium 2.019 32% 132 31%
Medium 1.397 22% 96 22%
Medium - high 920 14% 64 15%
High 1.005 16% 60 14%
Total 6.352 100% 427 100%
SLAIN 54 Version 1.2
Figure 28: Collective risk distribution per sections. Catalan primary roads network.
Figure 29: Collective risk distribution per km. Catalan primary roads network.
Table 21: Collective risk distribution. Catalan primary roads network.
On the collective risk distribution, the volume of km on the “Without FSA” (“Without a fatal or serious crash”) band is lower than in the individual risk band since usually roads big have capacity like motorways has very low risk due to its high volume of traffic but equally for that, crashes with FSA are produced in this type of roads. Therefore, there is lower concentration of the km on the <0.1 collective risk band. An analysis of the individual and the collective risk has also been developed by type of road: single or dual carriage way. Please note that that, as explained in the Spain section, the normal classification developed in Catalonia is Autovías, Autopistas and convencionales. For homogenization with the SLAIN study Autovías and Autopistas have merged into the dual carriageway category being the convencionales the single carriageways. There is not a direct twin category for motorways. This analysis for the total analysed network is shown in Table 22.
Concentration Km %km Sections %Section
Without FSA 490 8% 52 12%
< 0.1 3.417 54% 183 43%
0.1 - 0.2 1.679 26% 119 28%
0.2 - 0.5 693 11% 66 15%
0.5 - 1 73 1% 7 2%
Total 6.352 100% 427 100%
SLAIN 55 Version 1.2
Table 22: Individual and collective risk rates per road type. Catalan primary roads network.
The analysis shows that the risk on the single carriageways is 4.2 times higher than in dual carriageways.
4.1.8 Catalonia TEN-T network
Regarding the TEN-T network in Catalonia has an extension of 1,078km that represent the 6% the total network analysed using he EURORAP methodology. The same methodology of analysis developed with the total network analysed by the EURORAP methodology has been done for the TEN-T network in Catalonia. The distribution of individual risk by both, on number of road sections and extension in km is shown in the figures below.
Figure 30: Individual risk distribution per sections. Catalan TEN-T network.
Figure 31: Individual risk distribution per km. Catalan TEN-T network.
Table 23: Individual risk distribution. Catalan TEN-T network.
Carriageway type
Individual risk
(Risk rates)
Collective risc
(Crash density)
AADT
(veh/day)
Single carriageway 39,89 0,11 8.165
Dual carriageway (motorways) 9,61 0,17 40.343
SLAIN 56 Version 1.2
Comparing Table 16 with Table 23, it is clear that there is far less medium-high and high risk category road on the Catalan TEN-T than on the primary network.
Figure 32: Collective risk distribution per sections. Catalan TEN-T network.
Figure 33: Collective risk distribution per km. Catalan TEN-T network
Table 24: Collective risk distribution. Catalan TEN-T network
Regarding the collective risk, the TEN-T network obtained better results than the total network analysed. However, the observed trend of lowering the km on the without on individual risk is intensified, since a high percentage of the TEN-T network are high capacity roads.
Risk Km %km Sections %sections
Low 411 38% 15 32%
Low - medium 579 54% 26 55%
Medium 44 4% 3 6%
Medium - high 17 2% 1 2%
High 26 2% 2 4%
Total 1.078 100% 47 100%
Concentration km %km Sections %section
Without FSA 40 4% 2 4%
< 0.1 682 63% 27 57%
0.1 - 0.2 225 21% 9 19%
0.2 - 0.5 96 9% 6 13%
0.5 - 1 34 3% 3 6%
Total 1.078 100% 47 100%
SLAIN 57 Version 1.2
5 Discussion
5.1 Role and purpose
The Crash Risk Mapping technical specification gives a good description of the role and purpose of this tool:
• “Risk Maps are statistically designed to support national road safety strategies and add an extra layer of information alongside existing approaches. As such, RAP Risk Mapping typically covers roads outside of towns and cities, where deaths and serious injuries are concentrated. Using an international and common basis of measurement that can be used to assess priorities, Risk Mapping identifies the safest and most dangerous road sections within a region or country. Comparisons between countries enable benchmarking and progress to be tracked. Knowing where risk has been reduced and the measures that have worked are essential in building best practice and knowledge transfer.
• “Although Risk Mapping shows that some sections carry higher risk than others, it does not necessarily mean that road authorities will and should regard these as the highest priority for improvement. Authorities rank roads for safety improvement, taking account of both the number of crashes likely to be saved through improvements and the cost of implementation. Discussion with authorities and police has shown that these bodies review high risk roads, comparing the output with road sections flagged by their own internal processes and in seeking to develop practical measures to reduce the risk to road users.
• “While not all roads can be managed to the same risk level, emphasis should be on keeping risk within acceptable boundaries. The RAP route structure is based on simple rules aimed at keeping as coherent a design as possible within any road section, while at the same time extending the section length far enough to give sufficient crash numbers to minimise year-on-year variation. Crash and traffic flow data are assigned to each section, typically compiled into three year periods to give a robust estimate of risk. The assessment period can be extended where crash numbers are low.
• “Risk Mapping, by its very nature, relies on the use of historic crash and traffic flow data. As such, when published, some routes may already have had road safety improvements. Others may be more difficult to change and on these roads it is particularly important for road users to be aware that they face higher risks than they might expect. Crash Risk Mapping should therefore be updated at regular intervals to ensure that they represent the most up-to-date picture.”
5.2 Project learning – considerations for the RISM directive
This application of the technique in Croatia, Greece, Italy and Spain has identified several items of note that are relevant to the application of the technique as part of EC/2019/1936.
5.2.1 Messages for road-users and operators
Crash Risk Rate Maps showing individual risk are primarily useful for showing the risk to an individual road user of being involved in a fatal or serious crash as the road user moves from one section to the next. Often that risk may be to the vehicle occupant but, and notably when passing through urban areas, the risk may include colliding with, and causing injury to, pedestrians or cyclists. The involvement of these road users on the TEN-T will be the subject of further study. The technical specification document captures this well: “The public are often most interested in their risk on the road as individual users. The simplest way to represent this is in terms of crash risk in relation to exposure. Rates per vehicle kilometre travelled
SLAIN 58 Version 1.2
can show the likelihood of a particular type of road-user (e.g. car driver, motorcyclist, lorry driver, pedestrian or cyclist), on average, being involved in a road crash. “An essential focus of RAP is to improve recognition among road-users that they must share in the responsibility for a safe road system. In producing maps aimed at individual risk, it is therefore important to counter the common assumption that their purpose is to inform the road-user of how best to modify the route taken to minimise their likelihood of being involved in a crash. This is especially true where mapping covers only higher-tier road networks, since it is known that roads off the main road network typically have higher crash rates. “The main purpose of mapping individual risk is to:
• inform road-users of how and where their behaviour needs to be modified to minimise risk and, in doing so, to help them to understand the role of road infrastructure in determining the risks they face. It is hoped that, over time, this will aid clearer recognition of the influence of road design on risk and how it can vary on different types of road;
• illustrate more generally how high-level infrastructure variables, such as carriageway type and road standard, influence risk.”
A Crash Rate Risk map showing crash density is designed to be useful to the road operator, showing where road sections have a high number of crashes and therefore may be used to target remedial action. Again, the technical specification explains: “Collective (or ‘community’) risk is used by road providers to reflect more broadly how the total risk to all road users is distributed across a network. This information is crucial in determining how to spend available budgets effectively. “At the simplest level collective Crash Risk Maps show the density, or total number, of crashes on a road over a given length. However, rates expressed in this way are largely influenced by the number of vehicles using a particular road section or link, given the positive correlation between fatal and serious crashes with traffic flow. “To assess how best to improve collective risk, it is important to understand not just the present level of risk, but also the extent to which a lower level can be achieved on a particular road at reasonable cost. By showing how much in total it is worth investing at one site compared with another, collective Crash Risk Maps enable the road provider to consider not just the safety quality of a network that should be planned for in the future, but also the level of funding required. “An alternative insight into safety performance is provided by crash rates related to road type averages. These demonstrate road sections with higher or lower risk after the expected variability between different road groups (i.e. motorways, dual carriageways, single carriageways, mixed carriageways) is taken into account. Benchmarking in this way involves highlighting roads that should be targeted, exploring why they fall short of the average safety standards for their road type, and assessing whether it is appropriate to apply countermeasures known to be effective on roads with similar design and usage characteristics. “Information provided in collective Crash Risk Maps can also be used as the basis for considering investment decisions, providing authorities and policy-makers with a valuable tool for estimating the total number of crashes that could potentially be avoided if safety on a road were improved. Used with cost information, this can indicate locations where the largest return on investment can be expected.”
5.2.2 Allocation of data, normalisation, ratios and thresholds
In Croatia, FPZ has raised concerns about the accuracy of the allocation of the crashes to the road sections. FPZ is working with the authorities to improve this accuracy. It is likely that, to a greater or lesser extent, accuracy of location coding will be an issue in all countries and indeed it is known that Police and road authorities routinely make efforts to refine data collection process. Similar problems were identified in Italy and may also be present in other countries. In Spain and in the Catalonia region of Spain, there has been comparison of the data with and without
SLAIN 59 Version 1.2
the normalisation to Risk Bands 2020. Without this normalisation, the maps show a wider spread in the risk distribution, providing therefore greater differentiation and therefore potentially a more useful analysis tool. In Greece, where the mapping has been on the basis of fatal crashes, the need for statistically robust crash numbers has meant that the analysis has been limited to 39 road sections and this in turn limits the usefulness of the tool as a differentiator of risk in that country. Crash Risk Map thresholds are determined by the ratio of fatal & serious crashes to fatal crashes for the network being examined. The study has shown that, where this ratio varies widely between road types (e.g. Core TEN-T compared with other Comprehensive roads), there may be an argument for providing separate Crash Risk Map thresholds for these different categories of roads. If this is not done and only one ratio is provided for the entire network, then it may over- or understate the risk for some roads near the threshold boundaries. Potentially, this is an issue mainly in Italy where the crash severity ratio differs substantially for Core TEN-T compared with other Comprehensive roads. This is explained further below. In Italy, Crash Risk Mapping was completed for the Core and Comprehensive TEN-T network. Given the differences in the fatal:serious crash ratio on different road types, various approaches to calculating the calibration factor(s) to be used were considered:
• Single factor based on all Italian main road network whether or not specific crash location was recorded
• Single factor based on all Italian main road network with specific crash location recorded
• Single factor based on the TEN-T only
• Different factors for Core and Comprehensive TEN-T
• Different factors for motorways, non-motorway dual carriageways and non-motorway single carriageways
The resulting calibration factors in each of these cases were approximately as follows:
• 31
• 28
• 30 Core: 41 Comprehensive: 25 Motorway: 40 Non-motorway dual carriageway: 31 Non-motorway single carriageway: 16 The effect on the risk bandings of using the different calibration factors was investigated further: the following two figures (Figure 40), for example, show the risk banding distributions of roads on the Core TEN-T network using approaches 2 and 4, i.e. the two most extreme of the calibration factors above for the Core TEN-T.
SLAIN 60 Version 1.2
Figure 34: Risk banding varies dependent upon calibration choice
It can be seen that the use of different risk factors for Core and Comprehensive networks resulted, in this case, in fewer roads in the higher risk bands and more roads in the lower risk bands. Specifically, the use of a higher calibration factor moves roads away from higher risk bands and towards lower risk bands, and vice versa. This highlights that different Crash Risk Maps are not necessarily comparable: if the same roads were mapped using two different calibration factors, the bandings for some roads could be different on the two maps. Crash Risk Maps from different countries may therefore not be comparable too. If a single factor is to be used when calculating the risk bands for a particular road network, as is traditional, the choice of factor might depend on the comparability desired: in this particular case in Italy, there was a desire to use a single factor and to ensure that the results on the TEN-T were comparable with those on the Italian main road network as a whole. Therefore, approach 2 was selected and the calibration factor of 28 was used on all roads. However, given the different fatal:serious ratios on different roads, this approach could result in the Core TEN-T’s risk bands being worse than they might otherwise be, and the Comprehensive TEN-T’s risk bands being better than they might otherwise be. In the longer term, therefore, a more sophisticated approach taking road type into account would be desirable. Using different calibration factors for Core and Comprehensive TEN-T would ensure that the decision to include the Comprehensive TEN-T in the analysis has no effect on the results on the Core TEN-T. While this would be preferable to using a single calibration factor, though, it remains far from ideal given the possible variability in Core TEN-T and Comprehensive TEN-T networks in different countries. For example, in Britain, the Core TEN-T has a mix of motorway and non-motorway, and the typical fatal:serious crash ratio is different on motorway from the typical one on non-motorway; the use of a single calibration factor across the Core TEN-T could subsequently result in some roads being mis-banded again. Therefore, using different calibration factors depending on road type (i.e. motorway, non-motorway dual or non-motorway single) would be preferable to using different factors depending on which TEN-T network a road is on. (There is no need for calibration to be done separately for both TEN-T network and road types: the
SLAIN 61 Version 1.2
different fatal:serious ratios calculated for Core and Comprehensive TEN-T simply reflect the different proportions of each road type on each of them. For example, in Italy, the Core network is almost the same as the motorway network, so the fatal:serious ratio for the Core network is inevitably very similar to that of the motorway network; the Comprehensive network, on the other hand, is mainly a mix of non-motorway dual carriageway and single carriageway, so the ratio for the Comprehensive network is inevitably part way between those of dual and single carriageways.) Using this approach would mean that Crash Risk Mapping results were more comparable between different countries. For example, the results for motorways in Italy could be compared meaningfully with those for motorways in Spain, because both sets of results would be calibrated against those of motorways in the country concerned, without the calibration being distorted by roads of other types which are included in the analysis. Some inconsistency would remain, of course, depending on the results displayed; for example, comparing the Core TEN-T in Great Britain with the Core TEN-T in Italy could be misleading given the different proportion of the Core TEN-T which is motorway in both countries; this lack of comparability is unavoidable given the differences in the road networks in each country. Further investigation into the road characteristics that are most likely to affect the fatal:serious crash ratio is required before a clear way forward can be recommended that would maximise the comparability of different Crash Risk Maps. The difference in fatal:serious ratio on motorways, non-motorway dual carriageways, and single carriageways has been considered above, but British Crash Risk Mapping is currently calibrating urban roads separately from rural roads and motorways. (The recent addition of urban roads to the annual British risk mapping exercise was what first drew attention to this issue: many rural roads and motorways moved into lower risk bandings at that point despite a lack of improvement in their crash rates.) The next steps will consider how fatal:serious crash ratios differ by road type, rurality and other characteristics that are already gathered as part of the standard Crash Risk Mapping protocol, in order to recommend an improved approach to calibration that will ensure different Crash Risk Maps are as comparable as possible.
5.2.3 Under-reporting
There is clearly a problem with the collection of serious road crash data in Greece. The number of serious crashes is recorded as less than the number of fatal crashes. In earlier work during the period 2002-2006 EuroRAP called for data in Italy to be improved – both the locational coding of crash data and the increased availability of traffic volume data. In the intervening 10-15 year period, data appear to have improved in Italy (with, for example, greater availability of traffic data), but it is clear that in Greece and in other countries too, there must be improvements. There are linkages between SLAIN Activity 1 and Activity 2. The EuroRAP/iRAP star rating protocol was initially devised to interpret and understand the differences in risk shown in Crash Risk Maps. Parts of the TEN-T network in Croatia and in Italy will be subject to Star Rating and comparison of the data will show how and why the risk varies over these networks. Note also that work in Spain as part of the project seeks to compare results from Crash Risk Mapping and Star Rating and provide a uniform procedure for such comparisons. These issues are addressed more fully in Appendix 1.
5.2.4 Documentation
Whilst addressing procedures and protocols, it should be noted that that the EuroRAP description of the Crash Risk Mapping procedure was written originally for an assessment of the British Motorway, trunk and primary road network. The guidelines suggest that an aspirational target for 20 crashes per section when mapping. There is a balance to be struck between the number of years of data necessary to collect this total or indeed the length of the road section required. This target is not often achieved,
SLAIN 62 Version 1.2
even in the UK where data collection is good. The guidance should be revised so that countries can work towards achieving the balance of:
• a statistically reliable crash total for the section
• a time span that will not have seen substantial road changes over time, and
• a road length that can be addressed in its entirety with improvements or an improvement package.
Further, it is proposed that the guidance be revisited to incorporate some of the salient material in the calculation of thresholds for Crash Risk Mapping described above.
5.2.5 Scope of mapping
Not all the maps that have been produced show substantial variation in risk to road users as they move from one road section to the next and some show relatively low individual risk. This highlights the importance of the amendments to the RISM directive extending its scope to the primary roads. It is often useful to go below the Core TEN-T to obtain results that are useful both in policy-setting and in determining targets for engineering casualty reduction. As this illustration from the UK (Figure 41) shows, individual risk is relatively low (green) on the major motorways (TEN-T) but increases sharply on the more minor roads nearby (red and black).
Figure 35: Risk varies substantially when lower tiers of roads are also mapped
An important issue to address is the suitability of the data for comparison between countries and indeed the appropriateness of showing different national networks on one map, thereby implying that the same colour on one network shows the same risk as on another. Among the factors to be considered in making that call are:
• The accuracy of location coding of the crash data
• The likely under-reporting of both fatal and serious crashes
• Whether the procedures in setting the thresholds have been the same between countries, whether for example the maps are portraying similar tiers of road hierarchy and whether the fatal: serious ratios set for each are similar
SLAIN 63 Version 1.2
6 Conclusions
Crash Risk Rate maps have been produced of the relevant Core TEN-T network in accordance with the deliverables required in Activity 1, namely D1.1, D1.2, D1.3, D1.4 and D1.5, for Croatia, Greece, Italy and Spain. Detailed description of the data has been provided. In Croatia and Italy, comparisons have been made in mapping and analysis between the Core TEN-T and other roads in the Comprehensive TEN-T. In Spain, the safety of the Comprehensive TEN-T can be compared with the safety of all national roads. Regional roads of importance in Catalonia can also be compared. In Greece the analysis is of the Core TEN-T. Results are being launched or released on websites, subject to local consultation and discussion, as appropriate.
• Applications of the data have been demonstrated in the analysis. The maps:
• show the risk to individual road-users of being involved in crashes as they move from one road section to another.
• provide guidance for operators on where there has been a concentration of crashes.
• can be used (as illustrated in Croatia) to show where crash rates deviate from the norm expected for roads of a particular type.
• have regional applications (such as in Catalonia) in showing how to compare risk on the major road network with risk on adjacent roads
• Data deficiencies involving the reliability and definition of data in Greece and Italy respectively have meant that in Greece only fatal crashes have been mapped and in Italy aggregated “all injury” crashes have been mapped.
Further work will be undertaken in 2020 to resolve a methodological issue that has been identified. Dependent upon the ratio of fatal: serious crashes on different types of road, the calibration of the risk band varies and hence the allocation of road sections to the risk groups. A map of only one type of road (such as the TEN-T) using calibration of risk band thresholds derived from more than one type of road with differing fatal: serious ratios could show a different distribution of risk (for example, sometimes more green sections) than had that calibration been derived from only one type of road. This finding has been illustrated with data from Italy in the present study. It therefore cannot be assumed that one Crash Risk Map is directly comparable with every other without further assessment of the fatal: serious ratio for each road type and without the additional step in computation to calibrate a factor for each road type.
SLAIN 64 Version 1.2
References
Lawson S (2017) Analysis of changes in the rate of severe crashes for typical road infrastructure investments. Report to the European Bank for Reconstruction and Development. Road safety framework procurement (Reference 41196) consultancy services, December https://www.irap.org/2018/01/analysis-of-changes-in-the-rate-of-severe-crashes-for-typical-road-infrastructure-investments/
SLAIN 65 Version 1.2
Appendix 1 – Reporting of crashes and casualties1
Bull and Roberts (1973)2 were among the early reporters on casualty under-reporting and showed how this may vary by casualty severity and road-user group and others by location and road class. In many High Income Countries there is near-complete reporting of road fatalities but serious casualties may be substantially under-reported (by a factor of around 2-3) and slight casualties by a similar rate or more. Great Britain is an interesting High Income Country example. For example, more than 2 million damage-only crashes are estimated to have occurred but are not recorded in official statistics in Great Britain. The reported ratio of fatal:serious:slight casualties is around 1:10:100. The “true” ratio is likely to be closer to around 1:40:350 (see Table A1).
Reported
Approx.
ratio to
deaths --
Reported
Unreported
Potential
actual total
Approx. ratio
to deaths –
using
potential
“actual totals”
Total casualties
About 186,000
--
About 500,000 unreported
Total casualties 660-880k with 730k as central estimate [1]
Fatal About 1,700-2,000 (A)
1 1% not in records? (B)
About 1,700-2,000 ((A) + (B)
1
Serious 22,000 (C) 10 57,000 (D) About 80,000 (C) + (D)
40
Slight 162,000 (E) 100 471,000 (F) “More than 600,000” (E) + (F)
350
Damage only
--
--
--
About 2.3 million [2]
--
Table A1: Under-reporting of road casualties in Great Britain in a recent year (rounded data) ([1] see page 5 – estimates of unreported serious and slight. [2] See page 3 – total damage-only Great Britain) https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/9275/rrcgb2011-02.pdf
Even in the Netherlands, another road safety leader, it is recognised that not all fatalities may be reported to all data systems. Stipdonk ((2004) has argued that that there are road fatalities missing in all files. As this number in the example he provides is small (5) and there is no information about them, these fatalities are omitted in the overall estimation of reporting (Bos and Derriks (2009)). The example serves to illustrate why some sources are more complete than others and the different reporting mechanisms.
1 derived from Lawson (2017)
2 Bull JP and Roberts BJ (1973) Road accident statistics – a comparison of police and hospital information. Accident Analysis & Prevention
5(1), 45-53
SLAIN 66 Version 1.2
Figure 1: Databases and recording of fatal road casualties (from Bos and Derriks 2009)
http://www.itf-oecd.org/sites/default/files/docs/4-bos.pdf
WHO (2015) has compared road reported fatality data with that estimated as closer to the true number (from other national data sources such as morbidity statistics) and presents the data shown in Table A2. The degree of under-reporting is unlikely to be uniform and on the inter-urban roads that are predominantly of interest here, will vary between not only between country, but by location (eg proximity to villages), urban/rural, injury severity, type of road-user involved and dependent upon factors such as the level of police presence on the network. In many countries a simple check of the ratio of reported pedestrian casualties to all casualties, or the ratio of serious casualties to deaths, may be a good indicator of the quality of the crash data and of the reporting. Implied low numbers of pedestrian or serious casualties may cause concern. One implication of this variation is that simple comparison of fatality rates between individual countries may be of limited benefit. It is likely to be more useful to compare roads of different kinds in individual countries by comparing ratios of crash and injury rates.
Table A2: Deaths and reporting3 in Croatia, Italy, Greece and Spain [1]
Reported WHO estimate Reporting Deaths 2013
deaths 2016 deaths 2016 level [3] per 100,000
Croatia 307 340 90% 8.1
Italy 3283 [2] 3333 98% 5.6
Greece 824 1026 80% 9.2
Spain 1810 1922 94% 4.1
3 Notes to Table A2 Source: [1] 2016 data from Global status report on road safety 2018; [2] www.istat.it; [3] "Reported"/"WHO estimate" –
including definition discrepancies and under-reporting
SLAIN 67 Version 1.2
Appendix 2 – Country meta-analysis
Croatia
Item Units / Description Data
Network description
Current year 2016
Data sources Crash data Ministry of the Interior of the Republic of Croatia
Traffic data Ministry of the Interior of the Republic of Croatia
Data period 1 Year 2014
total fatal 22
total serious 84
total serious and fatal 106
Data period 2 Year 2015
total fatal 25
total serious 188
total serious and fatal 213
Data period 3 Year 2016
total fatal 47
total serious 183
total serious and fatal 230
Data period 4 Year 2017
total fatal 33
total serious 174
total serious and fatal 207
Data period 5 Year 2018
total fatal 44
total serious 211
total serious and fatal 255
Data period all 5 2014 to 2018
total fatal 171
total serious 840
total serious and fatal 1011
Scaling factor F&S / F 5.912280702
Aditional Parameters 1 billion 1,000,000,000
SLAIN 68 Version 1.2
Italy (Anas network)
Item Units / Description Data
Network description TEN-T network
Current year 2019
Data sources Crash data ISTAT/ACI database
Traffic data Anas Road Transportation Model
Data period 1 year 2015
total fatal 127
total serious and slight 3509
total serious, slight and fatal 3636
Data period 2 year 2016
total fatal 116
total serious and slight 3590
total serious, slight and fatal 3706
Data period 3 year 2017
total fatal 126
total serious and slight 3442
total serious, slight and fatal 3568
Data period all 2015 to 2017
total fatal 369
total serious 10541
total serious and fatal 10910
Scaling factor F&S&SL / F 27,77
Risk Bands Band Collective Individual
Standard Low 0 0
Low - medium 0,08 1,2
Medium 0,16 4,9
Medium - high 0,24 8,4
High 0,32 14,2
Adjusted Low 0 0
Low - medium 2,22 33,32
Medium 4,44 136,08
Medium - high 6,66 233,27
High 8,89 394,34
SLAIN 69 Version 1.2
Greece
The data collected in this tab provides a summary of the Risk mapping data
Item Units / Description Data
Network description TENT Core Road Network of Greece
Current year 2019
Data sources Crash data Hellenic Statistical Authority
Traffic data National Transport Plan of Greece
Data period 1 year 2014
total fatal 56
total serious 25
total serious and fatal 81
Data period 2 year 2015
total fatal 44
total serious 40
total serious and fatal 84
Data period 3 year 2016
total fatal 38
total serious 14
total serious and fatal 52
Data period 4 year 2017
total fatal 38
total serious 19
total serious and fatal 57
Data period all 2014 to 2017total fatal 176
total serious 98
total serious and fatal 274
Scaling factor F&S / F 1,556818182
Risk Bands Band Collective Individual
Standard Low 0 0Low - medium 0,08 1,2
Medium 0,16 4,9
Medium - high 0,24 8,4
High 0,32 14,2
Adjusted Low 0 0Low - medium 0,04 1,868
Medium 0,08 7,628
Medium - high 0,12 13,077
High 0,17 22,107
SLAIN 70 Version 1.2
Spain
Item Units / Description Data
Network description
Current year
Data sources Crash data Dirección general de tráfico
Traffic data Ministerio de Fomento
Data period 1 year 2016
total fatal 357
total serious 1056
total serious and fatal 1413
Data period 2 year 2017
total fatal 351
total serious 961
total serious and fatal 1312
Data period 3 year 2018
total fatal 423
total serious 875
total serious and fatal 1298
Data period all 2016 to 2018
total fatal 1131
total serious 2892
total serious and fatal 4023
Scaling factor F&S / F 3.56
Risk Bands Band Collective Individual
Standard Low 0.00 0.00
Low - medium 0.08 1.20
Medium 0.16 4.90
Medium - high 0.24 8.40
High 0.32 14.20
Adjusted Low 0.00 0.00
Low - medium 0.28 4.27
Medium 0.57 17.43
Medium - high 0.85 29.88
High 1.14 50.51
SLAIN 71 Version 1.2
Catalonia (Spain)
Item Units / Description Data
Network description
Current year
Data sources Crash data Servei Català de trànsit
Traffic dataMinisterio de Fomento de España, Generalitat
de Catalunya, Diputacions
Data period 1 year 2016
total fatal 109
total serious 582
total serious and fatal 691
Data period 2 year 2017
total fatal 128
total serious 520
total serious and fatal 648
Data period 3 year 2018
total fatal 140
total serious 523
total serious and fatal 663
Data period all 2016 to 2018
total fatal 377
total serious 1625
total serious and fatal 2002
Scaling factor F&S / F 5.31
Risk Bands Band Collective Individual
Standard Low 0 0
Low - medium 0.08 1.2
Medium 0.16 4.9
Medium - high 0.24 8.4
High 0.32 14.2
Adjusted Low 0.00 0.00
Low - medium 0.42 6.37
Medium 0.85 26.02
Medium - high 1.27 44.61
High 1.70 75.41