Road safety and bicycle usage impacts of unbundling vehicular and cycle traffic in Dutch urban networks Paul Schepers Ministry of Infrastructure and the Environment SWOV Institute for Road Safety Research 22-01-2014 1
Contents 1. Introduction 2. Research questions 3. Hypotheses 4. Results 5. Conclusions and discussion 22-01-2014 2
Introduction: Crash modelling Basic crash prediction model (CPM): E(λ) = α V β e yixi = α V c β1 V m β2 R i Estimated using Negative Binomial Regression Modelling 22-01-2014 Effects of network level separation 3
Traditionally there is much research attention for infrastructure along distributor roads 22-01-2014 Effects of network level separation 4
Less attention for the impact of road networks and street hierarchies The Dutch hierarchical classification of roads: Through roads with a flow function Access roads to give access and for sojourning In between are distributors: Flow function on road sections Exchange of traffic at intersections 22-01-2014 Effects of network level separation 5
Examples Scale (0,5km=0,31mile) Distributor road 3-7-2012 How to make more cycling good for road safety? 6
Examples Scale (0,5km=0,31mile) Distributor road 3-7-2012 How to make more cycling good for road safety? 7
Research question What is the impact of separating bicycle traffic from the distributor road network, here called unbundling, on road safety and bicycle use Road safety: bicycle-motor vehicle crashes Bicycle use: bicycle mode share 22-01-2014 Effects of network level separation 8
Unbundling: separation by routes through traffic-calmed areas 22-01-2014 Effects of network level separation 9
Unbundling: grade-separated intersections, bicycle tunnels and bridges, to cross distributor roads 22-01-2014 Effects of network level separation 10
Operationalization and measurement Two measures: Routes through residential areas (share of the route) Grade-separated intersections to cross distributors per km As both appeared to be highly correlated they were combined in one unbundling measure using Principal Component Analysis Measurement: Reconstruction of routes using a cycling route planner for all 66 Dutch municipalities having over 50,000 inhabitants Trips drawn from the National Travel Survey (NTS) 22-01-2014 Effects of network level separation 11
Hypotheses Unbundling corresponds positively with cycling safety and bicycle usage Road safety: reduced exposure to high speed motor traffic Bicycle usage: improved competitiveness of cycling compared to driving in terms of trip length: Shorts cuts for cyclists where roads are closed to cars Standalone paths and grade-separated intersections Allowance of contraflow cycling on one-way streets NB this is an indirect effect via the length of car trips divided by those by bicycle (tested using SEM) 22-01-2014 Effects of network level separation 12
Descriptive statistics On average per municipality (2004-2009): 5 fatalities and 85 hospitalizations among cyclists Bicycle mode share: 34% for trips up to 7.5 km Below: 66 municipalities split in three equally large groups according to the degree of unbundling 100 90 80 70 60 50 40 30 20 10 0 low moderate high share (%) through residential areas grade-separated intersections / 100 km Degree of unbundling 22-01-2014 Effects of network level separation 13
Results cycling safety Less crashes than expected in municipalities with a higher degree of unbundling (based on traffic volumes, density, and age distribution) Ratio: Casualty number / predicted number of casualties 1.2 1.1 1 0.9 0.8 0.7 fatalities hospitalized casualties 0.6 0.5 low moderate high Degree of unbundling 22-01-2014 Effects of network level separation 14
Results bicycle use Unbundling measures tend to improve the competiveness of cycling (length of trips by car divided by those by bicycle), thereby slightly increasing bicycle mode share 22-01-2014 Effects of network level separation 15
Conclusions Unbundling: Increases cycling safety Slightly increases bicycle mode share due to shorter trips by bicycle compared to those by car 22-01-2014 Effects of network level separation 16
Discussion Other effects not studied here: Other crash types: also a positive effect Unbundling decreases cyclists exposure to exhaust fumes and noise See Jarjour et al, 2013. Cyclist route choice, traffic-related air pollution, and lung function: a scripted exposure study. Environmental Health, 12(1),1-14. Subjective (road) safety Depending on the types of measures implemented, it may worsen feelings of personal security, e.g. remote standalone paths or bicycle tunnels 22-01-2014 Effects of network level separation 17
Questions Paper available at: http://www.ejtir.tbm.tudelft.nl/issues/2013_03/pdf/2013_03_03.pdf 22-01-2014 Effects of network level separation 18
22-01-2014 19
Example: high level of unbundling (Houten) Bicycle tunnels Almost 50,000 inhabitants Two large residential areas Bicycle tunnels at crucial locations within the network High level of unbundling: almost any trip is made without travelling along or crossing distributors 22-01-2014 Effects of network level separation 20