Conference PaperPDF Available

Understanding Vulnerable Road User Crash Risk (On Auckland's High Risk Routes)

Authors:

Abstract and Figures

Like many large cities Auckland's busy and high-risk arterials carry the bulk of traffic flows and include a mix of active/vulnerable road users (cyclist / motorcyclist / pedestrian). Auckland Transport is adopting Vision Zero and understanding vulnerable road user crash risks on such routes can be difficult due to the wide variety of factors that lead to crashes and the diverse nature of such users, which include many types of impairments. To supplement the limited information available on crash risk from crash history, two new methods have been developed. The Crash Risk Assessment Framework (CRAF) extends the Austroads safe system risk framework (where risk is broken down into exposure, likelihood and severity) to include several new pedestrian and bicycle crash types. CRAF can be used to assess the risk of serious and fatal crashes for existing routes and improvement projects at intersections and short (50 to 200m) mid-blocks. The multiuser assessment framework (MUAF) is a route inspection method that records and rates (from low to extreme) safety issues through observing road user behaviour and looking at operational matters along with design and maintenance faults. It considers different types of active/vulnerable road users including those with visual, hearing, sensual and physical impairments due to age, disabilities and temporal impairments (e.g. alcohol). The CRAF method focuses attention on medium to higher severity issues and speed management, while the MUAF method typical, but not exclusively, identifies lower cost improvements and maintenance activities. This paper will present the use of these tools on Mt Albert and Carrington Roads to assess the crash risk on the existing route and a number of improvement options. The paper also presents the GIS mapping tool that has been developed to display the results of the analysis.
Content may be subject to copyright.
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 1
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
UNDERSTANDING VULNERABLE ROAD USER CRASH RISK
(ON AUCKLAND’S HIGH RISK ROUTES)
Dr Shane Turner, Principal Road Safety
PhD, BE (Hons), CPEng, CMEngNZ, IntPE(NZ)
Stantec New Zealand, 6 Hazeldean Road, Hazeldean Business Park, Christchurch
E-Mail: shane.turner@stantec.com
----------------------------------------------------------------------------------------------------------------------------------
Mike Smith, Principal Road Safety
NZCE (Civil), MET, CPEng, CMEngNZ
Stantec New Zealand, 6 Hazeldean Road, Hazeldean Business Park, Christchurch
E-Mail: mike.a.smith@stantec.com
Irene Tse, Road Safety Engineering Team Leader - Urban
BEng (Civil), CPEng, IntPE(NZ), CMEngNZ
Auckland Transport
E-Mail: irene.tse@at.govt.nz
Andrew Garratt, Principal Road Safety Engineer
IEng, FIHE
Auckland Transport
E-Mail: Andrew.garratt@at.govt.nz
ABSTRACT
Like many large cities Auckland’s busy and high-risk arterials carry the bulk of traffic flows and
include a mix of active/vulnerable road users (cyclist / motor-cyclist / pedestrian). Auckland
Transport is adopting Vision Zero and understanding vulnerable road user crash risks on such
routes can be difficult due to the wide variety of factors that lead to crashes and the diverse nature
of such users, which include many types of impairments. To supplement the limited information
available on crash risk from crash history, two new methods have been developed. The Crash
Risk Assessment Framework (CRAF) extends the Austroads safe system risk framework (where
risk is broken down into exposure, likelihood and severity) to include several new pedestrian and
bicycle crash types. CRAF can be used to assess the risk of serious and fatal crashes for existing
routes and improvement projects at intersections and short (50 to 200m) mid-blocks. The multi-
user assessment framework (MUAF) is a route inspection method that records and rates (from low
to extreme) safety issues through observing road user behaviour and looking at operational
matters along with design and maintenance faults. It considers different types of active/vulnerable
road users including those with visual, hearing, sensual and physical impairments due to age,
disabilities and temporal impairments (e.g. alcohol). The CRAF method focuses attention on
medium to higher severity issues and speed management, while the MUAF method typical, but not
exclusively, identifies lower cost improvements and maintenance activities. This paper will present
the use of these tools on Mt Albert and Carrington Roads to assess the crash risk on the existing
route and a number of improvement options. The paper also presents the GIS mapping tool that
has been developed to display the results of the analysis.
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 2
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
INTRODUCTION
Since 2014 Auckland’s road deaths and serious injuries has increased by more than five times
the rate of travel growth and triple the rate of the rest of NZ. In 2017, 64 people died and 749
were seriously injured on Auckland’s roads. A significant proportion (45%) of these deaths and
serious injuries involved active/vulnerable road users (pedestrians, pedal cyclists and motor-
cyclists/moped). In the urban environment a relatively high proportion of serious injuries and
deaths occur on urban arterials (80% in Auckland). Auckland Transport (AT) recognised the
safety challenge of urban arterials in Auckland. The study presented in this paper is one of
several that is looking at how to target and address the risk factors that lead to trauma crashes.
AT have adopted a ‘vision zero’ policy that concludes “It can never be acceptable that people
are killed or seriously injured” on Auckland’s roads. It is acknowledged that to achieve vision
zero (i.e. to eliminate death and serious injury) there needs to be a paradigm shift in
responsibility from the ‘people using the roads’ to the ‘people designing and operating them’. A
key focus is on moving towards a safe (transport) system, which will require infrastructure
changes and speed management and a focus on active/vulnerable road users. It is also
important to 1) identify behaviours by road users that require education and enforcement to
change and to 2) take into account other outcomes desired on each route for these
vulnerable/active road users, like those assessed as part of a) non-motorised user audit
(Department of Transport, 2005) b) healthy streets review (Transport for London, 2017a and
2017b) and c) walkability review (Abley, 2011).
To target and address fatal and serious crash risk on urban arterials in Auckland a safe system
investigation process (SSIP), that incorporates the following steps, has been adopted by AT:
1. Route Prioritisation Process to identify high risk routes and the high-risk sections
2. Diagnosis of Safety Issues on routes using CRAF/MUAF and (historical) Crash Data
3. Development of Improvement Options that achieve or move towards a safe system
4. Development and use of new tools to support Funding Applications
5. Implementation of improvement options that maximise highest benefits
AT use a number of performance metrics to identify the high-risk urban arterials, including
Urban KiwiRap risk mapping (Step 1). This risk mapping is based on historical crash data. AT
have developed and agreed a list of high-risk urban arterials to target improvement works. This
list includes the Mt Albert Road and Carrington Road corridors discussed in this paper. This
paper outlines the diagnosis tools that were developed at a road segment level to identify safety
issues (Step 2) were developed along with a list of potential improvement options (Step 3) to
address the identified crash risk. This method has been applied to Mt Albert Road and
Carrington Road and a number of other high-risk urban arterials in Auckland.
CRASH DIAGNOSIS TOOLS
Three methods were used to diagnosis the safety issues faced by vulnerable road users. This
includes the traditional analysis of historical crash data. Historical crash data is useful for
prioritising high-risk routes and for understanding general crash trends along a corridor (e.g. high
number of pedal-cyclist crashes). Crash data is often not particularly useful for active road users at
a detail road element level (intersection and short mid-block lengths) due to the relatively low crash
numbers, especially the more severe crashes. Low crash numbers may also reflect a low usage of
the route by some active/vulnerable road user types (e.g. bicyclists), as high crash risks may put
off some users, resulting in a high latent demand. Using historical crash analysis method, it tends
to drive a reactive outcome rather than a proactive / safe system solution that can prevent fatal and
serious crashes from happening in the first place. Hence two additional crash diagnosis methods
(CRAF and MUAF) have been developed that help crash investigators understand the risk faced
by different road users as they use arterials roads, where crash data is not sufficient on its own.
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 3
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
Crash Risk Assessment Framework (CRAF)
This process is based on the Austroads’ Safe System Assessment Framework ( Austroads,
2016), which breaks crash risk into an assessment of exposure (volume of users), likelihood
(e.g. crossing facilities provided) and severity (operating speed). This framework looks at the
risk of DSi for key crash types on urban and rural roads. Crash types include, loss-of-control.
head-on, intersection, pedestrian, cyclist and motor-cyclist. For each crash type, the exposure,
likelihood and severity is scored from 0 to 4, with 0 being no risk (e.g. there are no pedestrians
allowed on the road) through to 4, high risk of a DSi. The score for exposure, likelihood and
severity are then multiplied together to produce an overall score for each crash type. A high
score (maximum of 64) indicates a very high risk. Many jurisdictions, especially in Victoria are
using this framework to assess routes and safety improvement projects.
CRAF includes a more detailed breakdown of pedestrian and cyclist safety, through adding
different pedestrian and bicycle crash categories (only one for each in Austroads Framework)
and by removing crash types that are not common in urban areas (e.g. head-on crashes). The
CRAF focuses on ten different crash types that can occur on urban arterials, including three
pedestrian, four cyclists and one each of pedestrians versus cyclists, motor-cyclists/mopeds
versus cars and intersection (motor-vehicle only). Rarely do all ten crash types features on a
road section. The ten crash types include (diagrams for each crash type are provided in
Appendix A):
1. Pedestrian crossing side road/access vs left and right turning vehicle (P1)
2. Pedestrians crossing main (study) road vs through vehicle uncommon (P2)
3. Pedestrian crossing side road/access vs through vehicle (at crossroad) (P3)
4. Pedestrians vs cyclist conflicts (on shared path or footpath) (P4)
5. Cyclists sideswiped by vehicle on a mid-block section (C1)
6. Cyclist vs right and left turning vehicle (emphasis on right turning) (C2)
7. Cyclists sideswiped by vehicle through intersection (often extra lanes) (C3)
8. Cyclist on main road vs through vehicle on side road (at crossroads) uncommon (C4)
9. Motorbike vs turning vehicle (at side road or access) (M1)
10. Intersection vehicle vs vehicle (4 wheel plus - turning in and out of side road and
access) (V)
The CRAF process is undertaken at a road element level along a specific corridor (it has been
developed for urban arterials). Typical road elements vary in length from 50 to 200m. This can
include major intersections (traffic signals & roundabouts), and road lengths that include;
railway level crossings, bridges, priority cross-roads (especially across three or more lanes),
one or more priority T-intersections, school or other vulnerable user crossings. Element
boundaries also occur at speed limit changes, change in number of lanes (including addition of
bus lanes and clearways), bus stop sections (especially when bus stops are provided on both
sides of road) and land-use changes (e.g. higher numbers of commercial/industrial accesses).
For each road element a score is calculated for each relevant crash category for 1) exposure, 2)
likelihood and 3) severity. The raw score ranges from 1 to 5 for exposure and 1 to 4 for
likelihood and severity depending on the level of risk. For exposure the risk is dependent on the
number of road users of each type that might collide. Depending on crash category this
includes pedestrians crossing the road, cyclists and motor-cyclists travelling along the road,
vehicles travelling along the road, pedestrians walking along the footpath, cyclists turning in and
out of side-roads and vehicles turning in and out of side-roads and accesses. The volume
bands used in the analysis are shown in Table 1. Where possible these are based on counts.
With counts not readily available for vulnerable users these counts are often estimated based
on short term observations of usage levels (hence some level of uncertainty exists with these).
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 4
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
1 2 3 4 5
Low Med-Low Med Med-High High
Traffic Volume up to15,000
15,000 to
25,000
25,000 to
30,000
30,000 to
40,000
Over 40,000
Side-road/Access Vol (per 100m) <100 100 to 300 300 to 3000
3000 to
15000
Over 15000
Motorcycle Volume (2 way) <50 50 to 100 100 to 200 200 to 300 Over 300
Over 500
Cycle Volume (2 way)
Over 1,000
<50
50 to 100
100 to 200
<100
100 to 200
Volume bands
Pedestrian Vol (Across & Along)
200 to 500
200 to 500
500 to 1,000
Table 1 Exposure (Daily User Volume) risk bands (ADTs)
The exposure rating is based on a combined scoring of the two movements that are in conflict
(e.g. crossing pedestrians versus through traffic volumes). When both are high then an
exposure of 5 is selected. For example, the score is 4 for a daily pedestrian crossing volume of
200 to 500 and daily vehicle volume of 25,000 to 30,000. This differs from the Austroads
framework which tends to focus on one of the user volumes. For example, the number of daily
pedestrians for pedestrian crashes.
For likelihood the risk band depends on the facilities provided along a section. The raw score
for likelihood is derived using the facilities specified in Tables 2 and Table 3. Further refinement
of these criteria is likely as experience with the tool occurs.
Crash Risk/Score
1
3
4
Pedestrian 1 (P1)
Intersection/Access
(Vehicle Turning)
Signalised
intersection with
turning arrows
Single lane
crossing/refuge
island
Two-lane crossing
together
Three-lane crossing
together
Pedestrian 2 (P2)
Mid Block/90 degree
movement
Signalised
crossings
Crossing
aid/zebra or
flush median
with regular
crossing aids on
2-lane road
Crossing aid or flush
median on four lane
road or centreline on
2-lanes
Four plus lane road
with no crossing
aids or centreline
Pedestrian 3 (P3)
Intersection (Vehicle
Straight)
Signalised
intersection with
turning arrows
Signalised
Intersection with
filtered turns
Uncontrolled across
2 traffic lanes
Uncontrolled across
three plus lanes
Pedestrian 4 (P4)
Ped versus cyclist off-
road cycle/shared path
Clear separation
of Cyclists and
Pedestrians
Wide shared
path (3m plus)
2 to 3m shared path
Narrow shared path
or cycling on
footpath
Table 2 Likelihood Categories for Pedestrians (P1 to P4)
Crash Risk/Score*
1
2
3
4
Cyclist 1 (C1) - Mid Block
side swipe
Separated
cycleway
cycle lane
No facility and wider
traffic lanes or no
parking
No facility and
narrow traffic lanes
near parking
Cyclist 2 (C2) -
Intersection/Access
(vehicle turning)
Separated
cycleway with
cycle signals
and signal with
arrows
cycle lane
No facility and wider
traffic lanes or no
parking, good
visibility
No facility and
narrow traffic lanes
near parking, poor
visibility
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 5
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
Cyclist 3 (C3) -
Intersection/Access
Approach side-swipe
Separated
cycleway
cycle lane
No facility and
narrow approach
lanes, on
access/minor
intersection
No facility and
narrow approach
lanes, on major
intersection
Cyclist 4 (C4) -
Intersection (vehicle
straight)
Traffic signals
Minor priority
crossroads
Major priority
controlled cross-
roads
Table 3 Likelihood Categories for Cyclists (C1 to C4)
For ‘motorcycle (M1)’ and ‘intersection (V1)crashes the scores are 1) signalised intersection
with turning arrows, 2) signalised intersection with filter turns, 3) non-signalised T-intersections
and accesses and 3) non-signalised crossroads and locations with major visibility restrictions.
The severity rating is based on estimated 85%ile speed of fastest moving vehicle type (normally
motor-car) with scores reflecting 1) less than 40km/h, 2) 40 to 50km/h, 3) 50 to 60km/h and 4)
60km/h plus.
The raw scores can be modified (i.e. increased) in two ways. The first modification is if a
number of additional risk factors are present. Risk factors include; high-turnover parking,
clearways, bus and HOV lanes, high levels of congestion, high proportion of trucks, poor vertical
and horizontal alignment and poor surface quality. The second modification is if there are one
or more high or extreme MUAF issues on a road element related to a crash type (discussed
below). The impact of the changes is to increase the range of scores for exposure, likelihood
and severity beyond 4 (the base score). The overall CRAF score is calculated by multiplying
the base or modified exposure score by the likelihood score and the severity score.
The overall process may include a CRAF assessment of the existing route, the identification of
high-risk sections of the route for different crash categories, prioritisation of high-risk sites to
investigate further, and can also be used to assess the likely effectiveness of various upgrade
options, at least whether mildly or highly positive or negative.
Multimodal User Assessment Framework (MUAF)
MUAF is a narrative- and evaluation-based assessment that combines the element of network
inspection (urban), road safety auditing, multi-modal user audit and knowledge of risks
associated with vulnerable road users. The purpose of the assessment is to collect
behavioural, interaction and safety elements associated with all road users, and especially
those elements that may lead to fatal and serious injuries.
The specific focus in an urban context are vulnerable road users, including pedestrians, pedal
and motor cyclists. Within these groups are highly vulnerable road users, such as children and
those with a disability. The behavioural elements of the assessment include desire lines,
conflicts (observed), lack of clear guidance/direction, user composition and interaction, and
driver behaviour.
The assessment is undertaken via a walk-over of the route in both directions, ideally at different
times of the day. The assessor observes road user behaviour and identified route deficiencies.
The data is recorded via a dicta-phone, with observations recorded in compliance with a field
guide. The field guide includes four key elements, 1) location (name of road and position along
it), 2) road user type (pedestrians, cyclists, both modes, classes of more vulnerable pedestrians
school, elderly and mobility and visually impaired), 3) Description of the issues/problem and 4)
risk level (six levels from low to extreme).
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 6
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
ASSESSMENT OF MT ALBERT AND CARRINGTON ROADS
Description of Routes
Mt Albert Road and Carrington Road (that abut each other) are two of a number of arterial roads
that criss-cross the Auckland Isthmus. The overall route extends from the Royal Oak
Roundabout (Manukau Road) in the south to Great North Road in the north (see Figure 1).
These road intersect with a number of other arterial roads including Pah Road (traditional route
to the Airport), Mt Eden Road, Dominion Road, Sandringham Road and New North Road. The
route generally runs parallel to the recently completed Waterview tunnel motorway (south-
western motorway).
There are a number of locations along the route that generate high numbers of vulnerable road
users and access turning movements, including big-box retail, commercial shopping areas (e.g.
Three Kings), schools and the Unitech (on Carrington Road). There is a growing bus patronage
on this route with the associated risks created by crossing pedestrians.
Figure 1 Mt Albert and Carrington Road Location Map (blue line)
CRAF and MUAF Analysis Existing Routes
The CRAF process included a walk-over to score each applicable crash type (P1-4, C1-4, M1, and
V1) for each sub-section along the Carrington Road and Mount Albert road corridors. Figure 2
shows each walking section. These eight walking sections where further divided into
approximately 10 road element sections each for the CRAF analysis. Sectioning is not required for
the MUAF assessment, as issues are recorded at a specific location (utilising the Mobile Road app
for smart phones).
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 7
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
Figure 2 Mt Albert and Carrington Road Sections
The sites scores were compiled in a spreadsheet tool for each section and crash type. Table 4
shows the highest risk locations for each crash type. Scores above 27 (exposure 3* likelihood
3* severity 3) are considered high risk, while score above 48 (exposure 4 * likelihood 4 *
severity 3) are considered extreme, and have often been modified due to observed MUAF risk
elements. Notably the highest risk scores along this route are associated with pedestrians
crossing the main road (P2) (many of the pedestrian crashes are of this type) and pedestrians
being hit while crossing sideroads and accesses (P1). The risk score of the latter is on average
lower due to lower speed of turning vehicles compared with through vehicles. However, there
are several places where the intersection/access alignment (e.g. Y-junctions and those with
large flares) does lead to higher route exit speeds and to lesser extent higher entry speeds.
Cycle crash risk (as measured by CRAF) is typically lower than the risk observed for
pedestrians due to the current low level of cyclists on these routes. While there were a number
of cycle crashes on Carrington Road, cycle lanes are provided, making this section safer than
parts of Mt Albert Road which have clearways and often narrow kerbside lanes for cyclists.
Interestingly most of the cyclists that were observed on Mt Albert Road were cycling on the
footpath, no doubt because of the high perceived crash risk of riding on the road. Refinement of
the two new methods should consider the risks associated with such behaviour.
Pedestrian
Movement
Corridor
Section
Risk
Score
Cyclist
Movement
Corridor
Sect
ion
Risk
Score
Vehicle Movement
Corridor
Section
Risk
Score
P1 - Vehicle
Turning
Mt Albert A
9
72
C1 - Mid
block SS
Mt Albert A
9
64
M1 - Intersection
Mt Albert A
8
60
Mt Albert A
1
60
Mt Albert A
3
48
Mt Albert A
10
48
Mt Albert A
3
60
Mt Albert A
5
48
Mt Albert A
7
48
Mt Albert B
5
54
Mt Albert A
6
48
Mt Albert A
6
48
Mt Albert D
1
36
Mt Albert A
7
48
Mt Albert A
5
48
Mt Albert F
3
36
Mt Albert A
8
48
Mt Albert A
3
48
Mt Albert D
3
30
Mt Albert A
10
48
Mt Albert A
2
36
Mt Albert C
4
30
Carrington B
7
32
Mt Albert A
1
36
Mt Albert A
6
30
Carrington B
9
32
Carrington B
10
32
Mt Albert B
7
30
C2 Vehicle
Turning
Mt Albert A
9
48
V1-Intersection
Mt Albert A
8
48
P2 - Vehicle
Straight
Carrington B
7
112
Mt Albert A
5
48
Mt Albert A
2
45
Mt Albert A
3
100
Mt Albert F
1
45
Mt Albert A
1
45
Carrington B
9
84
Mt Albert A
8
36
Mt Albert A
10
40
Mt Albert A
1
75
Mt Albert A
6
36
Mt Albert A
7
40
Mt Albert A
2
75
Mt Albert A
2
36
Mt Albert A
6
40
Mt Albert B
5
72
Mt Albert A
1
36
Mt Albert A
5
40
Mt Albert F
5
60
Mt Albert F
3
30
Mt Albert A
3
40
Mt Albert F
6
60
C3 -
Intersection
Mt Albert A
9
64
Carrington B
10
48
Mt Albert F
4
45
Mt Albert A
1
54
Carrington B
9
36
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 8
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
Mt Albert A
5
40
Side Swipe
Mt Albert A
10
48
Carrington B
8
36
Mt Albert A
7
40
Mt Albert A
8
48
Carrington B
7
36
Mt Albert A
8
40
Mt Albert A
7
48
P3 - Mid
block
Mt Albert D
1
36
Mt Albert A
6
48
Mt Albert A
5
48
Mt Albert A
2
36
Mt Albert A
3
36
Carrington A
5
32
Mt Albert D
8
30
C4 -
Intersection
Vehicle
Straight
Carrington B
14
36
Mt Albert B
5
30
Mt Albert E
8
27
Mt Albert D
1
24
Table 4 High CRAF Risk Road Sections
Mt Albert - Section A features in a number of the high-risk (sub-) sections. This is one of the
highest volume road sections and includes peak period clearways, but through most of its length
effectively operates as 4-lanes due to low kerbside parking demand.
The MUAF analysis highlighted a large number of reoccurring and some site-specific issue along
both routes. This included the following:
1. Bus stop locations and pedestrian desire lines with a lack of crossing opportunities and
facilities.
2. Pedestrians (including vulnerable) using flush medians to cross on two to four lane roads.
3. Poor visibility for pedestrian at mid-block crossing (desire lines).
4. Turning traffic into side streets and accessways causing various risks for pedestrians.
5. Cyclist hazards from A/C overlays resulting in high lips at edge of seal alongside kerb
channel.
6. Very narrow road side parking widths resulting in larger vehicles protruding into the cycle
lanes.
7. Need for effective cycle lane markings through intersections.
Table 5 shows the number of low, medium-low, medium, medium-high, high and extreme MUAF
risks identified along both routes (total length of 8km) for each road user type. The 109 high and
extreme risks were considered in the CRAF scoring. On average 45 issues were recorded per
kilometre.
Risk Category
Pedal Cyclist
Pedestrian
Both Users
Vulnerable
TOTAL
Extreme
4
7
1
15
27
High
17
11
7
47
82
Medium-high
25
17
8
49
99
Medium
23
12
9
30
74
Medium-low
19
13
5
22
59
Low
3
10
1
5
19
TOTAL
91
70
31
168
360
Total per KM
11
9
4
21
45
Table 5 MUAF Issues by Road User Type
Option Development
The high-risk locations and routes identified in the CRAF and MUAF assessments were used,
along with experienced gained while walking the routes, to select sections of the route for option
development. This following six sites/sections were identified:
1. Segar/Unitec Access 2 (Section 9 of Carrington B)
2. Farm Road Unitec Access (Section 7 of Carrington B)
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 9
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
3. Mt Albert Road/Alberton Avenue Intersection (Sections 4, 5 & 6 in Mt Albert F)
4. Three Kings Shopping Centre (Section 5 of Mt Albert B)
5. Hillsborough Road Intersection (Section 9 of Mt Albert A)
6. Royal Oak Road to Pah Road (Sections 1, 2 & 3 of Mt Albert A)
While options were developed for all the locations this section only provides details on locations 1,
2 and 4. While only two sections (Sections 7 and 9) were identified as high risk on Carrington
Road it was decided after reviewing the MUAF issues further to develop improvement options
for the entire corridor outside the Unitec campus. Many of the issues in these two sections also
occurred at other locations along the route, and hence a route treatment is preferable for
consistency, and to ensure that there is not crash / risk potential migration.
The key issues along this corridor include high numbers of vulnerable users, particularly
pedestrians and cyclists accessing the Unitec, school, bus stops, and several other facilities in
the area. A major concern along Carrington Road is the lack of high-quality crossing facilities
for pedestrians, which was observed to lead to high risk behaviour by pedestrians crossing
Carrington Road (see Figure 3).
Figure 3: Pedestrians crossing Carrington
Rd
Figure 4: Downhill vehicles northbound on
Carrington Rd
Another issue, that affects the severity of crashes on this corridor, is the high speed of vehicles
including northbound vehicles travelling downhill (this can be seen in Figure 4). The wide-open
feel of this corridor, due to the set-back of land-use, also contributes to the higher vehicle
speeds, particularly outside of peak periods.
There are numerous 3-arm intersections and Unitec accesses with stop or give-way control onto
Carrington Rd. These are considered high risk due to the relatively high right and left turn
vehicle movements combined with crossing pedestrians and cyclists. In this instance there is
no separation of movement leading to the risk of conflicts with serious consequences. This
results in a high crash risk potential for cyclists and pedestrians travelling along Carrington
Road as well as crossing from the side roads and accesses.
Cycle lanes on this part of Carrington Road (Unitec area) are disrupted by bus stops along the
corridor, putting cyclists at risk of side swipe or rear end crashes when buses are stopped or
approaching/leaving their stops. Figure 5 shows the worst example on this on the corridor,
where two bus stops are opposite one another. In this case drivers are forced onto part of the
flush median when buses are parked at the bus stops which puts crossing pedestrians at risk
and create a pinch-point for cyclists. There are several other locations where there is a single
bus stop that requires drivers to use the flush median in one direction to overtake buses putting
them in conflict with pedestrians and cyclists.
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 10
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
Figure 5: Bus stops and cycle lanes on Carrington Rd (sourced from google maps)
Two upgrade options were developed for this corridor from Section 12 (at the northern end of
Carrington Road) through to Section 7. The first option was higher costs and included an off-
road cycle path on the western side of Carrington Road and upgrading two new signalised
intersections (at the Unitec accesses) and signalisation (Toucan) of the zebra crossing and new
mid-block crossings at bus stops. Option 2 also has two new signalised T-intersections, retains
the zebra crossing and retains on-road cycle lanes.
Section 4 along Mt Albert Road (at the Three Kings Shops) is a complicated section of road with
multiple access points to surrounding side-roads and parking, as shown in Figure 6 below. This
is a complex section of road for cyclists, pedestrians and motor-cyclists to negotiate; both
crossing the main road and side-road/accesses. This shows up in the relatively high CRAF
scores calculated for these modes. The MUAF also highlighted the poor condition of the
pedestrian facilities in this section, along with high risk movement of users crossing at
inappropriate locations.
Figure 6: Three Kings Shopping area on Mt Albert Road (sourced from google maps)
Several options were initially developed for this location. Option 1 (see Figure 7) included
closing of the intersections of Dornwell Road and Hayr Road with Mt Albert Road, simplifying
the road layout. It also shows changes to the entry / exits into the Three Kings Shopping Centre
to reduce the number of turning movement locations onto Mt Albert Road. In addition, the
provision of an on-road cycle lane and indented parking on Mt Albert Road. Option 2 was
similar to option 1 but included a single lane entry and exit into Dronwell and Hayr Roads.
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 11
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
Figure 7 - Schematic of Option 1 for Three Kings Shops
CRAF Analysis - Options
The CRAF scores where then reassessed for the options. Table 6 shows the changes in risk score
that occurred for the main pedestrians and cycle crash types. The largest reduction observed is in
the risk of crashes involving pedestrians crossing Carrington Road. The formation of better cycle
facilities also reduces the risk of side-swipe crashes. The risk of pedestrian being hit by turning
vehicles at the two high risk Unitec accesses has also been reduced through introducing the traffic
signals, allowing separation \ protection of the turning and crossing movements.
Section
6
7
8
9
10
11
12
Score
Score
Score
Score
Score
Score
Score
P1 - Vehicle
Turning
Existing
6
24
12
24
16
4
6
Option 1
4
12
6
12
12
4
3
Option 2
4
12
6
12
12
4
6
P2 - Vehicle
Straight
Existing
48
112
32
84
24
24
24
Option 1
16
16
8
6
12
12
8
Option 2
16
16
8
12
12
12
16
C1 - Mid block
SS
Existing
16
32
16
32
16
16
24
Option 1
4
4
4
4
6
12
6
Option 2
8
8
12
8
12
12
12
C2 Vehicle
Turning
Existing
4
12
4
32
8
4
4
Option 1
2
6
2
4
8
4
4
Option 2
4
12
4
8
8
4
4
C3 - Side Swipe
Existing
16
16
16
16
16
16
16
Option 1
8
4
4
4
8
12
6
Option 2
8
8
6
8
12
12
12
Table 6: CRAF Scores for Upgrades on the Unitec Corridor (Carrington Road)
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 12
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
MAP-BASED ANALYSIS TOOL
An interactive application has been developed to visualise the CRAF and MUAF results
captured. The premise behind the structure of the application is to provide a single common
platform view of the CRAF and MUAF information, potentially from multiple contributors, in a
standardised and easily understood format.
This application is web based and is built using ESRI technology. The information held within
the application is secure, with access gained via single factor authentication. Access to this
information can be granted to different users and/or organisations, either in its entirety or partly,
dependent upon requirements.
The user has the ability to search and view the captured information via an intuitive interface
that allows for the data to be displayed in a variety of manners including thematic ranges, charts
and heat maps. All the visualisation aspects of the application are zoom and scale dependant,
which enables the user to focus on discrete sections of the network and drill down to greater
levels of detail as required.
All features contained within the application retain the original source data tables. This data can
be queried, with the resultant views showing either partial or full field detail. This is particularly
useful when working with the MUAF information, where the information held within can often be
voluminous.
Figure 7 and 8 show the section of Mt Albert Road from the Frost Road to Hillsborough Road, at
the higher level and at a more detailed level (around Hillsborough Road). The crash location
(and severity), MUAF issues (heat maps) and the CRAF scores by crash type are shown.
Figure 7 shows that the CRAF scores are highest at 1) Mt Albert Road A section 9, which is
the Hillsborough intersection and 2) Mt Albert B section 5 which is the Three Kings Shopping
centre. Both have a cluster of vulnerable road user crashes (squares) and MUAF risks (circles).
Behind each point there is meta-data. This includes the full description of the MUAF issues, the
crash coding details and the CRAF component scores (for the exposure, likelihood and
severity). The heat maps in this case show the concentration of high and extreme MUAF issues
(but these can be changed to crashes and CRAF scores also).
Figure 7 - Crash risk data for Mt Albert Road (CRAF scores below)
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 13
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
Figure 8 Detailed Data at Mt Albert/Hillsborough Intersection with Meta-data
A spreadsheet tool has also been developed that estimates the crash risk for pedestrians and
cyclists before and after improvements, based on estimated user volumes and facilities provided.
It utilises crash prediction models and crash reduction factors from the Crash Estimation
Compendium of the economic evaluation manual and other research sources. We have forwarded
the outputs of this new tool to the NZ Transport Agency for their comments and inputs.
SUMMARY
This paper outlines two new methods (CRAF and MUAF) that have been developed to enable road
safety professionals to assess the pedestrian, pedal cyclists and motor-cyclists risk (with specific
focus on the first two) along high risk arterials. The two risk assessment methods have been
applied to the Mt Albert and Carrington Road corridors. The outcome from the analysis along with
other useful information (crashes and traffic volumes) have been presented in a web-tool. For the
high-risk areas identified in the (existing route) assessment a number of improvement options have
been developed. The options have then been assessed using the CRAF method, with the
resulting reduction in the CRAF score indicating the likely reduction in crash risk.
REFERENCES
Abley, S and Turner, S (2011), “Predicting Walkability” NZTA Research Report 452
Austroads (2016) ‘Safe System Assessment Framework, Publication No.AP-R509-16, Sydney,
Australia
Department for Transport (DfT) (2005) Non-Motorised User Audits
http://www.ukroads.org/webfiles/HD4205.pdf, United Kingdom.
Transport for London. (2017a). Healthy Streets for London; Prioritising walking, cycling and public
transport to create a healthy city. London, United Kingdom: Mayor of London
Transport for London. (2017b). Guide to the Healthy Streets Indicators; Delivering the Healthy
Streets Approach. London, United Kingdom: Mayor of London
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 14
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
APPENDIX A Crash Category Diagrams (P1 to V)
Pedestrian Movements (pedestrians are blue lines, cyclists green and motorists red)
P1 Intersection (Vehicle Turning)
P2 Mid-block/90 degree movement
P3 Intersections (vehicle straight through)
P4 Pedestrian and Cyclists
Cycle Movements
C1 Mid-Block Side Swipe
C2 Intersection/Access (vehicles turning)
Understanding Vulnerable Road User Crash Risk Turner, S, Tse, I, Smith, M, Garratt, A Page 15
Transportation Group 2019 Conference, Te Papa, 3-6 March 2019
C3 Intersection/Access Approach side-swipe
C4 Intersection (vehicle straight)
Motorcycle Movements (motor-cyclists are light blue lines)
... Many pedestrians engage in spatial violations while crossing to save time and reduce the walking distance (8). Nevertheless, pedestrians' decisions on whether to violate or not vary significantly depending on many factors. ...
Article
Full-text available
The objective of this study is to understand the impact of a variety of factors on the frequency and severity of pedestrian-vehicle collisions that involve pedestrian spatial violations at mid-blocks. To that end, the historical collision records of the City of Hamilton between 2010 and 2017 were obtained, and collisions that had occurred at mid-blocks were filtered out. A Bayesian structural equation modeling (SEM) framework was developed to investigate the impact of a wide range of factors on such collisions. First, a classical SEM was developed to group the different factors into sets of latent variables. Four latent variables were defined, including location amenities and attractions, pedestrian/road network characteristics, exposure parameters, and location/collision-specific factors. The Bayesian SEM was then implemented to investigate the relationship between the latent variables and collisions. The results showed that the amenities and attractions of a location (e.g., parks, schools, bike-share stations, and bus stops) were the most influential factor on the frequency of collisions that involve spatial violation, followed by pedestrian network characteristics. Pedestrian network characteristics and location/collision-specific factors were found to be the most influential factors on the severity of collisions. The location of bike-share stations, pedestrian network connectivity, exposure to walkers, and the number of lanes were the four observed variables that explained the highest percentage of the variance in each latent group, respectively. The results of this study should assist engineers and planners to develop better design concepts to mitigate collisions that are caused by pedestrian spatial violations in urban areas.
... Several studies (Hamed, 2001;Yannis et al., 2013;Lue and Miller, 2019) evaluated the safety of pedestrians from an operational perspective, including the effect of traffic conditions and road environment (i.e., gap acceptance, vehicle speed, waiting times). Others (Keall, 1995;Lassarre et al., 2007;Lam et al., 2014;Turner and Smith, 2019) measured it through the exposure to crash risk such as time spent walking, the number of traffic lanes and volume to be crossed. However, fewer studies attempted to investigate the motivational factors of this behaviour. ...
Article
Full-text available
Pedestrians crossing roads at unprotected mid-block sections is a common behaviour associated with traffic accidents. It is a calculated risk that pedestrians take based on prevailing traffic conditions and their motivation. However, there is limited understanding of these factors. This paper investigates the motivational factors associated with pedestrians' risky crossing behaviour at unprotected, urban mid-block road sections. An on-site survey is conducted at four different locations in Auckland, New Zealand. It includes questions related to the constructs of the Theory of Planned Behaviour, habit and their relationships considering the effects of gender. Motivational factors are analysed using factor analysis and structural equation modelling. Results show that pedestrians' intention to cross a road at mid-block sections is mainly driven by habit and attitude. Some pedestrians , however, internalise the belief that risky crossing behaviour is an acceptable act in society from friends and important referents which is mediated through habit. Women's decisions are highly influenced by their attitude while men's' risky behaviour is influenced by their friends' perceptions. Crossing at mid-block sections is also perceived as a necessary risk worth taking, which is mentally linked to convenience gain, including saving travel time and reducing walking distances. The paper offers some insights into pedestrians' motivation to cross at mid-block. Findings are expected to assist in developing effective measures to reform the social acceptance of such behaviours and compliment engineering practices to reduce traffic accidents at unprotected mid-block sections.
Predicting Walkability
  • Abley
  • S Turner
Abley, S and Turner, S (2011), "Predicting Walkability" NZTA Research Report 452
Healthy Streets for London; Prioritising walking, cycling and public transport to create a healthy city
  • Austroads
Austroads (2016) 'Safe System Assessment Framework', Publication No.AP-R509-16, Sydney, Australia Department for Transport (DfT) (2005) "Non-Motorised User Audits" http://www.ukroads.org/webfiles/HD4205.pdf, United Kingdom. Transport for London. (2017a). Healthy Streets for London; Prioritising walking, cycling and public transport to create a healthy city. London, United Kingdom: Mayor of London Transport for London. (2017b). Guide to the Healthy Streets Indicators; Delivering the Healthy Streets Approach. London, United Kingdom: Mayor of London
Safe System Assessment Framework', Publication No.AP-R509-16, Sydney, Australia Department for Transport (DfT
  • Austroads
Austroads (2016) 'Safe System Assessment Framework', Publication No.AP-R509-16, Sydney, Australia Department for Transport (DfT) (2005) "Non-Motorised User Audits" http://www.ukroads.org/webfiles/HD4205.pdf, United Kingdom.