Accepted Manuscript

Public Transport Accessibility Tools Matter: A case study of Gothenburg, Swe-

den

Carey Curtis, Erik Ellder, Jan Scheurer

PII: S2213-624X(17)30367-X
DOI: https://doi.org/10.1016/j.cstp.2018.12.003
Reference: CSTP 309

To appear in: Case Studies on Transport Policy

Received Date: 15 December 2017

Revised Date: 12 October 2018
Accepted Date: 10 December 2018

Please cite this article as: C. Curtis, E. Ellder, J. Scheurer, Public Transport Accessibility Tools Matter: A case study
of Gothenburg, Sweden, Case Studies on Transport Policy (2018), doi: https://doi.org/10.1016/j.cstp.2018.12.003

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Public Transport Accessibility Tools Matter: A case study of Gothenburg,
Sweden
Carey Curtis Curtin University, Western Australia & University of Gothenburg,
Sweden
Erik Ellder University of Gothenburg, Sweden
Jan Scheurer Curtin University, Western Australia

Public Transport Accessibility Tools Matter: A case study of Gothenburg,

Sweden

Abstract
Urban transport investment decision-making has relied on traditional modelling tools
that forecast travel demand based on existing travel patterns. This approach has also
underpinned decisions about future urban development. Latent travel demand is
poorly understood, and this is particularly important given policy aspirations for the
take-up of more sustainable transport modes such as public transport. Accessibility
tools can make an important contribution to future development and investment
decisions given this policy aspiration. In this paper, public transport supply is
assessed using an accessibility tool and then compared to travel patterns. By
assessing both public transport supply and travel demand through an accessibility
lens we are able to gain insight into the latent demand for public transport. An
innovative segmentation type approach is employed by comparing supply and
demand and from this a typology is proposed as a framework for a more holistic
public transport investment strategy. This takes into account public transport
accessibility, travel demand of trip makers and their socio-economic situation, set in
the context of the urban region.
Keywords: Public transport; Accessibility; Latent travel demand

Introduction
Policy decisions about investment in urban transport have been dominated by
decision support tools drawn from the traffic engineering and transport economics
fields1 (Hatzopoulou and Miller, 2009; Naess et al, 2012; Keblowski and Bassens,
2017). One assumption underlying these tools is that savings in travel time are
beneficial, such that high-speed transport modes (usually the private car) are the
favoured solution. These tools not only facilitate infrastructure decisions but also, as
a consequence, work to underpin urban development decisions. Their method relies
on predicting future travel demand based on existing demand (Owens, 1995). This

1
Such as the traditional four step model and cost benefit analysis. The four step model, in use since
the 1950s, is used to test scenarios at corridor and network levels for transport and land use
arrangements. While more recently including public transport modes, they are dominated by a focus
on car-based traffic and measure efficiency such as vehicle-kilometres-travelled, mode share and
travel time savings.

1
approach perpetuates car-oriented planning, at odds with a policy direction that aims
for a shift to sustainable modes of transport (Banister and Hickman, 2013).
Additionally these tools reveal little about latent demand for other modes. Clifton and
Moura (2017, p78) define latent demand as “the activities and travel that are desired
but unrealised because of constraints.” In public transport networks, such constraints
often concern insufficient spatial or temporal coverage of origins and destinations,
poor service levels (such as low frequencies or slow speeds) or a network
configuration characterised by connectivity gaps and time-consuming detours over
geographical desire lines. In this context, conventional decision support tools seek to
predict public transport patronage after network or service improvements based on
past demand patterns (thus favouring existing public transport markets), but are not
designed to inform public transport coverage goals (Walker, 2008) as a means to
break into new public transport markets. Where overall policy settings are informed
by sustainable transport goals that aim to expand the role of public transport in a
given settlement context, a consequential dissonance between policy goals and
practical tools emerges.
Whereas latent demand can manifest as induced demand when it results from an
improvement in transport supply, latent demand can also be an expression of
constraints other than infrastructure provision, such as socio-economic
circumstances among specific user groups (Hine and Grieco, 2003). Induced
demand, however, is often not considered in travel demand models for transport
projects (both road and public transport), leading to discrepancies between forecasts
and actual usage (Naess et al, 2012) and highlighting the difficulty for mainstream
transport decision support tools in tackling latent demand. There is a need to
differentiate between induced demand and latent demand. Clifton and Moura (2017)
highlight this in distinguishing between redistributed demand and generative demand.
They too note that forecasting tools are typically applied at a project level and mainly
for car traffic, rather than for an urban area and for other modes.
Compounding these problems, contemporary urban development strategies have
tended to be ‘solutionist’, based on top-down and superficial planning aspirations
(Freestone, 2012) - such as the ‘compact city’, ‘new urbanism’, and ‘transit oriented
development’ with little regard to detailed implementation. Instead, strategies need to
be formed from an analysis that identifies the gaps between existing travel patterns
and policy aspirations. But there has been a demise in such systematic spatial
planning practices. Promoted by early planners (‘survey-analysis-plan’ after Patrick
Geddes’ rational planning process (Davoudi, 2006)) they formed the main-stay of
planning strategy-making during the 1970s and 1980s. Alternative options for urban
development could be tested against an overarching vision (or aspiration) for a given
city (Albrechts, 2006). Instead the experience now is of an ad hoc, incremental
approach to urban development planning, heavily influenced by development
speculation at the urban fringe and relying on pragmatic and incremental add-ons to
existing urban fabric on a case-by-case basis. This has led to decision support tools
that work in the context of particular projects rather than for considering the city as a
whole (Healey, 2006; Hrelja, 2011).

2
Traditional transport models also tend to be employed for infrastructure projects
rather than strategic decision making (Hatzopoulou and Miller, 2009). There is also a
need for tools to support planning strategy making in a more deliberative way (te
Brömmelstroet and Bertolini, 2011). Contemporary accessibility tools are an
emerging decision support tool that, framed appropriately, can fill this vacuum as well
as addressing the desire for a policy shift towards sustainable transport modes.
Unlike traditional transport models, accessibility planning tools measure the ease of
access for a given population and the opportunities that are accessible in order to
meet an individuals’ daily needs. Accessibility tools can reflect the dynamic
interaction between transport-related factors and land use activity (traditional
transport models tend to assume land use remains static) (Papa et al, 2017; Hull et
al, 2012). Accessibility tools have been employed as a planning support tool (see
examples at http://www.accessibilityplanning.eu/accessibility-instruments-list/’.
The Spatial Network Accessibility for Multimodal Urban Transport Systems
(SNAMUTS) accessibility tool is one such tool that has offered a new potential for
strategic spatial planning. It has been applied in planning practice in order to assist in
deliberations on investment and urban development decisions. Of significance is that
the tool takes as its starting point the policy goal – in this case that inhabitants be
afforded accessibility by public transport, and that this be understood directly from the
perspective of the individual as traveller. This differs from conventional public
transport planning practice which tends to plan from the perspective of patronage
(Walker, 2008) and also focuses on parts of the network rather than the network as a
whole.
Unlike traditional four-step models, SNAMUTS is a supply side tool; it assesses the
supply of public transport rather than being a demand-led model. Up to the point of
this paper, SNAMUTS data has not been empirically linked to patterns of travel
demand generated by the supplied public transport network. The key benefit to
planners and decision makers of SNAMUTS (and other accessibility tools) has been
in its capacity to consider the city as a whole as well as the functioning of the network
as a whole, and to consider this in the context of accessibility of both the transport
network and the place (node). This has seen some initial confusion by some
transport modellers who are more familiar with demand modelling. They have not
seen this as an impediment, however, since they note that public transport supply
has been based on assessments of demand. However, prior to the advent of
transport-related ‘big data’ (whose utilisation in planning remains incidental and
unsystematic in most jurisdictions, demand figures used by practitioners
predominantly related to passenger counts or ticketing data, and were only compared
with car traffic flow data to ascertain ‘demand’. The aim in this paper is consider how
analysis using an accessibility lens may advance the understanding of the
relationship between public transport supply and travel demand and also consider
latent demand. A key contribution in this paper is that by assessing both the supply of
public transport and the travel demand, as evidenced in comprehensive travel
surveys, we are able to more precisely capture the latent demand for public transport
by reference to a travel behaviour framework (Clifton and Moura, 2017).

3
In this paper, after describing our research approach and the case study
characteristics, we start by measuring the public transport accessibility of the Greater
Gothenburg region from a supply perspective. We describe the accessibility tool we
employed to do this and explain the indicator we use to measure the level of service
that the public transport network provides. Next we report travel demand by the
different transport modes. Then we compare public transport supply with actual
travel demand, and go on to show how this relationship can be used to create a
typology of public transport supply and demand. We conclude the paper by
highlighting how this approach can be used to understand latent demand for public
transport, considering the mismatch between supply and demand.

Research Approach
Study context
A case study approach, using Greater Gothenburg, is employed to test this method.
The choice of case study was initially precipitated by data availability since the public
transport accessibility analysis was complete and, rarely, detailed travel diary data
was available making the assessment of travel demand possible, Importantly,
Greater Gothenburg2 provided an interesting case from a planning perspective given
its dual objective since the 1950s to serve the car as well as provide for public
transport accessibility.
The region has a well-developed public transport system comprising 11 full-time tram
lines in the core city, 21 regular and seven express bus lines meeting the SNAMUTS
minimum service standard (see below), one urban and several archipelago ferry
routes, and three commuter train lines which serve the wider region. Within central
Gothenburg, tram has the highest public transport market share representing 55 % of
all public transport trips, followed by bus (43 %) and boat (2 %) (Trafikkontoret 2017).
In the greater region, bus has the highest public transport market share (52 %),
followed by tram (39 %), train (7 %) and boat (2 %) (Västra Götalandsregionen
2018).
From a public transport accessibility perspective, Gothenburg (Figure 1) scores
highly compared with many other cities in the SNAMUTS sample;
www.snamuts.com). Figure 1 depicts the multimodal public transport system which is
differentiated by four modes: commuter rail to the outer reaches of Älvängen (north),
Alingsås (east) and Kungsbacka (south); a bus network; tram network; and ferries
serving the Gothenburg archipelago.
Gothenburg is densely urbanised compared to many Australian and U.S. cities, but of
lower density than many other European cities of similar population size. Gothenburg
is Sweden’s second largest city after Stockholm. From an urban structure
perspective, it has a strong, mixed-use and compact central city. Beyond the central
area, since the early 1950s new residential areas were planned in suburban locations
and aimed at creating stronger sub-centres, such as Mölndal and Frölunda in the

2
Greater Gothenburg, the functional region where most daily activities are carried out and where
SNAMUTS minimum service extends (see below), includes 14 municipalities: Ale, Göteborg, Öckerö,
Härryda, Kungälv, Lerum, Partille, Mölndal, Kungsbacka, Alingsås, Bollebygd, Lilla Edet, Stenungsund
and Tjörn

4
south. This has comprised modernist development with separation of transport
modes and land use, and the planning for speed-induced accessibility through
improved car connectivity and some light rail links. The road network was expanded
rapidly, resulting in urban sprawl. Recent decades have seen a gradual shift
underway with new development plans advocating densification adopted, and major
public transport investments are underway. New high-density residential
development is focussed on the north side of the river bank, replacing older port
related and industrial activity.
From a transport perspective, the City of Gothenburg’s urban development strategy
has long seen a dual policy agenda of both catering for the car and providing for
public transport. Gothenburg, the home of Volvo, has also been a car-oriented city,
but public transport investment has continued in the post-war era. Rather than close
the tramways, like other cities throughout the world, city planners retained them and
added several higher-speed, reserved-track extensions to suburban growth areas.
The planning approach has been to limit car access to the city centre for many
decades, instead catering for public transport access; and to ensure that housing
areas within the City of Gothenburg are served by public transport. In the Greater
Gothenburg region, catering for the car is strongly evident, with wide, high-standard
highways traversing the region. More recently, there are also ambitious goals by the
regional authority to increase public transport mode shares (Göteborgsregionens
kommunalförbund 2009), with a focus on fast links to the regional towns (e.g.
Kungsbacka, Alingsås, Mölnlycke) and across Gothenburg city for commuting
(facilitated by the latest proposal for the Västlänken train tunnel under the central
city).
A congestion charge was introduced early 2013. The maximum daily cost is 60 SEK
(7.5 USD), with single charges ranges from 9 SEK to 22 SEK. This can be compared
to a one-way public transport ticket (29 SEK or 3.6 USD), and a monthly season
ticket in only Gothenburg city (610 SEK or 76.6 USD). While the early effect of the
congestion charge saw car commuting trips decrease (by 9 % in 2013 compared to
2012) and public transport commutes increase (by 24 %) (Börjesson & Kristoffersson
2015), by 2016 car trips were only 5-6 percent lower than before the congestion
charge (Swedish Transport Administration, 2017) but the number of public transport
trips continues to increase faster than that of car trips (for example, by 7 % in 2016
compared to 2015).
Measuring public transport accessibility (supply)
The SNAMUTS tool was employed to measure public transport accessibility for
Greater Gothenburg (data was captured in September 2014). This tool forms part of
a new generation of accessibility-focussed planning support tools (te Brömmelstroet
et al, 2014). SNAMUTS is designed to assess and quantify the performance of public
transport networks in their settlement context, generating information both at the
metropolitan and local scale. It has been used comparatively for a sample of 27
global cities and extensively for the elaboration of land use-transport future scenarios
predominantly in Perth and Melbourne

5
Spatial accessibility of public transport is assessed from a variety of angles using
separate indicators for each, and by visualising these both for the metropolitan area
as a whole, and for individual activity nodes (www.snamuts.com). In this analysis the
‘contour catchment’ indicator is utilised for the majority of the analysis. This indicator
measures the spatial reach (using a travel time contour) of public transport journeys
within a pre-defined time window of 30 minutes as a percentage of metropolitan
residents and jobs within this contour, thus combining measurements of network
performance (speed and configuration) with those of urban structure (density and
transit orientation of land uses).
The Greater Gothenburg region is defined by reference to the SNAMUTS minimum
service standard3 and from this the municipalities have been selected in order to
compare public transport and car use against public transport accessibility. The
analysis explores geographical detail through the identification of nodes and points
on the public transport network as clusters of land use activity. In the Gothenburg
region, 84 such activity nodes were identified and make up the spatial basis, or
matrix, for the analysis.
The number of residents and jobs allocated to each activity node is determined by
counting these figures as aggregates of the statistical areas that lie within walking
distance of the corresponding rail stations (800 metres) or tram/bus routes (400
metres). In Sweden, the smallest statistical areas for which this data is available are
known as SAMS zones (Small Areas for Market Statistics). SAMS are defined by
Statistics Sweden to represent residential neighbourhoods. Each SAMS zone within
the walking range of one or more activity nodes (occasionally, their catchments
overlap) is tagged with the names of these nodes. In total, 846 SAMS zones in
Greater Gothenburg were allocated.
Capturing travel behaviour/mode shares (demand)
Information concerning mode shares was retrieved from the Swedish National Travel
Survey (RVU) carried out 2011-20164. It comprises data on the everyday travel of a
large randomized sample of the Swedish population aged 6–84 years. Data were
collected using one-day trip diaries of all out-of-home movements during a randomly
assigned survey day. The diary was sent by mail a few days beforehand, then phone
interviews referring to prepared individual travel diaries were carried out. The overall
response rate was about 40%.
In the data, trips are defined and arranged in a hierarchy of main trips, “trip legs”, and
trip elements. A main trip starts and ends at a main trip location (i.e. the respondent’s
main dwelling, holiday house, workplace, school, or temporary overnight abode). A
main trip involves one or more trip legs, which emerge when the respondent performs

3
The minimum standard for inclusion of a route into the SNAMUTS analysis requires a service
frequency of at least 3 departures per hour per direction during the weekday inter-peak period (2
during the day on Saturdays and Sundays) for tram and bus lines, and 2 departures per hour per
direction during the weekday inter-peak period and a 7-days-a-week operation for rail and ferry lines.
This rationale is based on the need to compete with the car where potential public transport users
have car as an option.
4
Note that, overall, mode shares from NTS data before and after the introduction of the congestion
charge show no notable differences.

6
an errand outside of a main trip location (i.e. changes trip purpose). The trip legs are
divided into trip elements when the respondent changes travel mode. In this paper
the trip legs that started and ended up in any SAMS within the study area are used, a
total of 39,634 trips. A trip leg is characterized by a range of variables (e.g. travel
mode, distance, and time).
Average mode shares for all trips are calculated starting from the defined catchment
area of each SNAMUTS activity node. Where two or more nodal catchment areas
overlap geographically, the trips originating in the overlap zone have been counted
as a fraction and allocated evenly among the activity nodes concerned (eg. a trip that
originates in a SAMS zone that has been allocated to three overlapping SNAMUTS
nodal catchments is counted as one third of a trip for each activity node). The number
of resulting trips per nodal catchment ranges from 78 (Hammarkullen) to 905
(Frölunda).The average mode share for trips starting from SAMS not allocated to a
SNAMUTS activity node is also calculated (i.e. areas where some public transport
exists, but does not meet the defined minimum service standard). While aggregation
to SNAMUTS activity nodes masks individual variation and makes proper control for
individual co-founding factors problematic, it offers a useful approach to inform policy
makers about the whole transport system (consistent with the policy focus in this
paper). This relates back to the discussion in our introduction since this aggregated
approach provides for a view of the region as a whole as well as the functioning of
the network as a whole (a valid approach also confirmed by Ewing et al, 2017).
Analysis approach
An exploratory approach is employed with the primary focus being descriptive in
order to visualise and understand the relationship between public transport supply
and demand. We assert that this approach is useful in policy discussion and among
urban planners tasked with deliberating a strategic perspective on the effectiveness
of current and future investment and development strategies. By aggregating mode
shares to SNAMUTS activity nodes (as described above) it is possible to examine the
relationship between supply and demand. The analysis follows three steps:
In a first step (i), public transport accessibility is reported (using the SNAMUTS
Contour Catchment indicator) and travel behaviour separately employing descriptive
statistics.
The second step (ii) is an integrated analysis - examining how public transport
accessibility and revealed travel behaviour relate. This is examined by mapping the
relationship between public transport access and mode shares, as well as by
estimating correlations. In our experience, this step is likely to prove most useful for
deliberative discussion.
The third (iii), and final step, is an innovative segmentation-type approach. Here the
analysis digs deeper into the correlations between public transport accessibility and
travel behaviour, and provides explanations for the revealed patterns. Activity nodes
are characterized using a typology approach, considering factors such as PT
accessibility, mode share, socio-economic factors, land use development, centrality
and density. Statistics are used in combination with contextual and theoretical
knowledge of the region. Segmentation analysis has been utilised as a

7
methodological framework by other transport researchers in order to capture complex
relationships between multiple factors influencing travel behaviour. For example,
Ohnmacht, Götz and Schad (2009) found significant correlations between clustered
factors of demography, lifestyle orientations, transport attitudes and observed
patterns of leisure travel. Scheiner (2018) examines the context of mobility
biographies, residential self-selection and transport costs and discusses the nuanced
interactions between freedom and constraint to determine individual travel and
location choices.
We assert that the strength of the segmentation approach is that it can be readily
interpreted in a policy and planning context by those not familiar with Gothenburg. It
provides a clear building block for policy deliberation on future urban development
and investment strategies to sustain and increase public transport mode shares.
Importantly, by focussing on interrelations between public transport accessibility, built
form and travel patterns both along and away from dominant trend lines, the
approach enables a consideration of latent demand for public transport.

Findings

Public transport supply and travel demand

Public transport accessibility is measured using the SNAMUTS ‘Contour Catchment’
indicator as a way of representing the level of service by public transport at each
activity node (in this way we are able to differentiate between places in the city). The
indicator assesses how successfully the public transport supply translates into
practical travel opportunities for transport users. It measures how many residents and
jobs can be accessed within 30 minutes kerb-to-kerb travel time of a particular
activity node.5 Ideally, a region should have an average contour catchment of 50%
(equivalent to an even 30-minute radius of the settlement area from centre to
periphery), but in reality this is unlikely to occur except perhaps in a purpose-
designed model city for public transport accessibility. The average for all activity
nodes across the region is 25.6%, but there is considerable variation between activity
nodes with a range between 61% and 2% (Figure 1). A high level of service is
evident for the majority of nodes in the central city and inner suburbs, as these
locations allow for movement in a multitude of directions with short travel times. The
regional corridors north-east to Älvängen and east to Alingsås both perform poorly,
as they only allow movement in the radial direction with long travel times.

For the region as a whole, it is also possible to measure the coverage of the network
to describe the quantity of residents and employees with access to public transport
(at the SNAMUTS minimum service standard) and in this way cities can be compared
by their level of service. Our analysis of other global cities shows that Gothenburg
performs well on a global scale, and average in a European context, for network
coverage (Figure 2), with 69.4% of residents and jobs covered within walking
distance by a quality public transport system. That Gothenburg is a relatively small
city among our global sample helps to deliver this result.

5
Some assumptions are made: a) one transfer is allowed within 30 min, but only between services
that both run at least every 15 min; and b) a flat deduction applies for making the transfer, equivalent
to the actual average transfer time across the network (7.1 minutes in Gothenburg in 2014).

8
Figure 1: Gothenburg: Level of Service by Public Transport (using Contour
Catchments)

9
Figure 2: Gothenburg: Comparison of network coverage (left) and average
contour catchments (right) (percentage of metropolitan residents and jobs).

Turning to travel behaviour, Table 1 shows the mode shares for all trips starting from
within the Greater Gothenburg region. Motorized travel is the dominant transport
mode. Overall for this region, 14.6 % of the trips are made with public transport and
just over half of all trips are by car. Active modes of travel (walking and cycling) are
just under one-third of all trips.
Table 1. Mode shares for the entire region.
n %
Walk 9299 23,5%
Bike 2202 5,5%
Motorized 22079 55,7%
Public transport 5792 14,6%
Other 262 0,7%

10
Total 39634 100,0%

Figure 3 shows travel mode share by SNAMUTS activity nodes (ie. those nodes
which are covered by public transport meeting or exceeding the defined minimum
service standard). There is a clear variation in mode share by activity centre, which is
assessed in more detail below. The public transport mode share across the region
(including areas below minimum service) is 14.6% of all trips. In the areas where the
public transport supply falls below the minimum service the mode share by public
transport is much lower, at 5.7% of all trips.
Figure 3: Travel mode share by activity nodes that are served by the SNAMUTS
minimum service standard

11
12
The relationship between public transport accessibility and travel demand
The analysis primarily utilises the SNAMUTS ‘contour catchment’ indicator shown
above as a proxy for level of service. For completeness, however, the strength of the
relationship between supply and demand for five of the SNAMUTS indicators is also
assessed (Table 2). For all indicators, with the exception of ‘nodal connectivity’,
strong correlations between supply and demand are evident.
Table 2. Pearson correlations Public transport mode share/SNAMUTS
indicators.
SNAMUTS indicators# % Public Transport share
Closeness Centrality -,672**
Degree Centrality -,569**
Contour Catchments ,697**
Betweenness Centrality ,605**
Nodal Connectivity ,406**
** Correlation is significant at the 0.01 level (2-tailed). #Full details of the construct of each indicator can be
found in Curtis and Scheurer, 2016 and at www.snamuts.com

The first indicators, ‘closeness centrality’ and ‘degree centrality’, measure the
accessibility properties of the transport network alone, the first indicator measuring
the ease of movement within the network, derived from the travel impediment
exercised by travel times and service frequencies and the second indicator
measuring the number of transfers that have to be made on public transport journeys.
It is plausible that as these transport accessibility indicators improve, the public
transport mode share increases. Closeness and degree centrality are measured as
resistance to movement, with lower values indicating lower resistance and better
accessibility performance. Hence, the correlation between these measures and
public transport mode share is negative.
The remaining indicators add the land use dimension to these base indicators.
‘Contour catchments’ measure the spatial reach of journeys within a travel time
contour of 30 minutes, expressed in the percentage of metropolitan residents and
jobs within that contour. In this paper the relationship between this supply indicator
and travel demand had been examined in much greater geographic detail, using this
indicator as a proxy for level of service. It shows the strongest correlation with the
share of trips made with public transport. ‘Betweenness centrality’ measures the
extent to which specific locations or routes on the network act as conduits for
movement. ‘Nodal connectivity’ measures the extent to which the public transport
network facilitates user flexibility in navigating the system with minimal need to
consult timetables or maps, calculated from the number of routes that converge in
each node, and their level of service. This last indicator has the weakest relationship
with travel demand, indicating that the position of the network’s major transfer hubs
does not always coincide with areas designed or configured to attract a high public
transport mode share. This raises some important issues for the relationship between
the design of the public transport network and the design of the surrounding urban
environment (currently being explored in depth in a sister research project - Curtis et
al, 2017).

13
Level of service and latent demand
Given the high level of service achieved with the supplied public transport network in
Gothenburg, the central question of the relationship with travel demand is now
examined in more detail in order to trigger an insight into latent demand. The
following three figures portray this relationship in different ways in order to unpack
and aid the analysis.

Figure 4 compares public transport mode share and ‘contour catchments’ for all
nodes. The columns are ordered by mode share (highest to lowest) for all nodes, and
the SNAMUTS indicator ‘contour catchments’ is shown as a line. It is evident that
there is a clear relationship between supply and demand, the higher the mode share
by public transport the better the level of service. This is reinforced in further in Figure
5 which plots each node and shows the deviation from the regression line. This is
examined in more detail in the next section.
Finally, Figure 6 considers public transport supply-demand on a geographic base.
This map highlights that there is a clear geography of demand that relates to supply
and to urban geography. Within the City of Gothenburg at the core of the network,
walking-distance geographical coverage by public transport at the SNAMUTS
minimum standard is almost universal and of a good level of service particularly in
the inner neighbourhoods and along tram/light rail corridors, leaving out only major
industrial and port areas, open spaces and some lower-density suburbs near the
coast and on the islands. In the wider Gothenburg region outside the core city, only
larger towns (e.g. Kungälv, Alingsås and Kungsbacka) and the station catchments on
the commuter rail lines are included (though at a relatively low level of service) while
sizeable areas of scattered peri-urban settlements as well as some regional towns
are not covered.

14
 0%
5%
10%
15%
20%
25%
30%
40%
45%
50%
55%
60%

35%
Bäckebol
Kållered
Ytterby
Landvetter
Kungälv
Billdal
Sörredsvägen
Lindome
Alingsås
Peppareds Torg
Kullavik
Radiomotet
Älvängen
Torslanda
Kungsbacka
Lerum
Amhult
Nödinge
Gråbo
Vallhamra
Mölnlycke
Partille Centrum
Bohus
Kärra
Lilla Varholmen
Mölndal
Saltholmen
Hjalmar Brantingsplat
Kålltorp
Kungssten
Tuve
Skogome
Brunnsbotorget
Partille
Rannebergen
Selma Lagerlöfs Torg
Tynnered
Lana

PT share
Kallebäck
Krokslätt
Mariaplan
Rosendal
Hjällbo
Frölunda
Majorna
Eriksberg
Skeppsbron
Gamlestaden
Wieselgrensplatsen
Lövgärdet
Bjurslätts Torg
Contour catchment
Stigbergstorget
Eketrägatan
Linnéplatsen
Munkebäckstorget
Hammarkullen
Marklandsgatan
Johanneberg
Östra Sjukhuset
Länsmanstorget
Figure 4: Public transport mode share compared to ‘contour catchment’ level of service

Vårväderstorget
Axel Dahlströms Torg
Gårdsten
Guldheden
Redbergsplatsen
Korsvägen
Högsbohöjd
Vasaplatsen
Liseberg
Lindholmen
Svingeln
Chalmers
Sahlgrenska
Angered Centrum
Järntorget
Kungsportsavenyen
Hagakyrkan
Grönsakstorget
Bergsjön
Bellevue
Kortedala
Centralstationen
Brunnsparken
Lilla Bommen
15
Figure 5: Scatter dot of public transport mode share and ‘contour catchment’ and typology development

16
Figure 6: Geographic display of public transport supply and demand

17
Typology of public transport supply and demand
Figure 5 plots the relationship between public transport supply and demand
discussed above. While there is an evident dominant correlation of demand for
public transport with public transport supply, there are also two conspicuous clusters
of activity nodes that diverge from this trend. Exploring both on-trend and off-trend
patterns of supply and usage helps a clearer picture of latent (potential) demand to
emerge. In this way the accessibility lens provides valuable insights for policy
discussion around future investment and development. Table 3 depicts this analysis
as a typology for Gothenburg.
Seven distinct types are apparent, each with different public transport supply
investment considerations. In some areas opportunities to improve supply can be
considered (addressing latent demand), and in other areas any substantial
investment in supply must be evaluated against the rate at which demand can be
harnessed.
The group ‘Captive riders’ depicts those who are located in areas (in this case the
suburbs of the city) where public transport accessibility is medium to low, but the
demand for public transport is high. Typically, these areas have well-serviced radial
tram lines into central Gothenburg with a travel time of 30 minutes or less, but poor
or non-existent public transport links in non-radial directions. This group is
characterised by low household incomes and resultant low levels of car ownership.
In Gothenburg’s case, the example suburbs comprise medium density housing which
is in public ownership, and limited local employment opportunities. Investment
strategies may include improving the public transport service (frequency boosts
particularly on the non-radial lines, and greater network connectivity), and to improve
accessibility to employment and tertiary education opportunities. In addition, urban
development opportunities for increasing land use diversity should be considered,
but there is a risk that public transport demand is weakened if these areas are
gentrified and more people can afford cars. At the core are issues of social justice
and transport inequalities, which require policy priority since many people cannot opt
out of public transport use. Therefore it is important to invest in improved public
transport in these areas (for example by improving non-radial links).
The ‘inner city’ group already benefits from high accessibility by public transport for
which there is high demand. This group comprise a mixed socio-demographic
spectrum (a sizeable portion of whom access the area as employees or visitors) in a
high-density, mixed-use locality. It is important that high accessibility levels are
maintained as this not only serves this particular group, but is also essential for the
wider region. One important consideration in this locality is that in some areas
demand for active modes of travel is also high (e.g. Linnéplatsen), competing with
public transport for local journeys and during particular seasons. Conversely, a
sizeable number of trips are also made by car (e.g. Skeppsbron, Svingeln).
Therefore it is important to maintain, and ensure that the urban environment is walk
and cycle friendly as well as making it less accessible for cars (perhaps through
financial penalties).

18
Latent demand is a key question for the ‘middle suburbanites’ group. Here public
transport accessibility is at the mid to high-level and demand matches this. In some
localities car use is quite high. Where this is the case, efforts to further improve
public transport supply and, not least, offer alternatives to the car, including better
network connectivity for walking and cycling, may prove effective. The use of land
use policies to raise densities to create areas more like the inner city areas should
also be considered, although local resistance to such urban change will need to be
managed.
There is also latent demand in many of the areas termed “outer suburbanites”.
These areas are located at relatively long distance from the inner city and the public
transport supply ranges from low to medium. Some areas are served with commuter
trains and some by buses. Car use is generally high. Many of these areas have
lower than expected public transport demand. TOD-style developments are possibly
successful in these areas, but the location of these areas near or within rural or
semi-rural surroundings means that public transport supply is unlikely to be
competitive to the car outside the main (mostly radial) settlement corridors.
Another case for consideration is where accessibility by public transport (supply) is
high but demand is weak. Key questions are raised about the extent to which further
investment should be made in the absence of either accompanying travel demand
strategies, or where it is clear that investment improves the network at localities
beyond the immediate area. Here, particular attention must be paid to the ‘Car
dominant’ group. This group is characterised by high levels of car usage at trip
origins (high to medium income residents at the urban periphery), as well as (and
perhaps more importantly) at trip destinations. In Gothenburg’s case this applies
particularly to some industrial areas and large car-oriented shopping centres located
along the main expressway network – see for example Bäckebol and Hjalmar
Brantingsplatsen. In these examples there is a need to differentiate between the
residential population (which is likely low) and destination-based travel (which likely
dominates). The strategy proposed to boost public transport use there is less
focused on improving public transport supply (which is already high in some cases,
but to redevelop these localities with housing into more coherent urban hubs of high
activity intensity, where the car is accommodated but public transport accessibility
and local walkable neighbourhoods take priority. Gothenburg’s waterfront
developments offer a potential template, since these once shared a similar land
use/mobility profile.
The ‘High income dwellers’ and the ‘Country Living’ groups share low public
transport accessibility (supply) and low demand. Both groups are located at the
periphery of the city, the former in low density single family housing, the latter in very
low density single family housing. They tend to fall into higher socio-demographic
groups (greater variation in the ‘country living’ group, but this group has high car
ownership levels). It is questionable whether there is any latent demand for public
transport among these groups. One way forward could be to establish public
transport hubs strategically in these areas and provide incentive parking to improve
opportunities for multimodal trips. These could also be synergised with the potential
TOD-developments suggested in the “suburbanites” category. However, this

19
approach would need to be coupled with a strategy to deter car-based travel by
making it harder/more expensive to drive to the inner city areas to avoid additional
heavy car use developments far from daily activity opportunities.

20
Table 2: Typology of public transport supply and demand.
DEMAND (LEVEL DWELLERS AND EXAMPLES OF
SUPPLY (LEVEL OF
OF PT MODE CENTRALITY DENSITY LAND USE TRIP-MAKERS AREAS IN
PT ACCESS)
SHARE) SOCIO-ECONOMY GOTHENBURG

Public housing, Low-income, low car

“CAPTIVE RIDERS” Low to medium High Outer city Medium Bergsjön, Angered
tenement buildings ownership

Centralstationen,
“INNER CITY” High High Central High Mixed Mixed
Grönsakstorget
Medium to high
External retail, Radiomotet,
“CAR DOMINANT” Low to high Low Most semi-periphery Low income, high car
industrial Bäckebol
ownership
“MIDDLE Johanneberg,
Medium to high Medium Semi-central Medium to high Housing, mixed Mixed
SUBURBANITES” Redbergsplatsen

“HIGH INCOME Semi-periphery/ High income, high

Low Low Low Single family housing Billdal, Kullavik
DWELLERS” peripheral car ownership
Medium to high
“OUTER Semi-periphery/ Housing, somewhat
Low to medium Low to medium Low to medium income, high car Partille, Mölnlycke
SUBURBANITES” peripheral mixed
ownership
Medium to high
Outside minimum
“COUNTRY LIVING” Low Low Peripheral Low Single family housing income, high car
service
ownership

21
Drawing on this approach a typology matrix is proposed (Figure 7) as a framework
for policy deliberation regarding public transport investment. This typology serves to
highlight where latent demand may exist, how it can be mobilised and also where
investment may be counterproductive without complimentary travel demand
management strategies to resolve car dependency (such as car parking charges in
areas of high public transport supply; revisiting the congestion charge pricing relative
to public transport price).
Figure 7: A typology framework for the relationship of public transport supply
and usage, including suggestions for policy directions

Conclusions and Policy Recommendations

The aim of this paper was to advance the methodology of understanding the
relationship between public transport supply and travel demand. In the first instance,
this approach allowed testing of the capacity of the supply-based accessibility tool for
tracing revealed patterns of public transport usage, and to identify where supply-
based and demand-based network performance measures match or differ. A
secondary aim of the paper was to consider the role of latent demand, and so review
how the accessibility tool may capture this potential and assist policy makers in
decisions about where to invest in better public transport infrastructure or services,
where to target urban development change, and which population groups or
residential/work locations to target for travel behaviour change programs.

22
By viewing travel demand through an accessibility lens, utilising an accessibility tool,
it is clear that public transport accessibility indeed matters. Strong correlations were
found between public transport supply and demand, particularly on measures that
capture the performance characteristics of the land use-transport system: travel
speeds, service frequencies and the density both of the public transport network and
the land use context. By contrast, the correlation of the network configuration around
hubs and transfer opportunities or constraints is more weakly correlated with usage
patterns, at least in the case study city of Gothenburg. This finding suggests that
investment in fast and frequent public transport services and an urban form that
supports these through urban density and diversity pays off in higher levels of transit
usage - build it and they will come (but this adage is not universal as we discuss
below). The findings also validate accessibility measures as effective support
instruments in understanding patterns of public transport demand, a vital quest for
those tasked with efficiently allocating finite resources to the development and
operation of public transport systems.
At a local scale, however, there are deviations and outliers from the trend
relationship between supply and demand which are of high importance in a planning
policy perspective. Even the limited range of activity centres in a medium-sized city
like Gothenburg constitutes a rich evidence base that provides clear building blocks
for policy deliberation on future development and investment strategies. By taking a
broader conceptual view of latent demand based on public transport accessibility
and supply, public transport mode share and demand, socio-economic composition
of travellers, land use and development trends and centrality/density, we are able to
develop a typology for understanding unmet accessibility needs. Seven typologies
for public transport-connected activity nodes in Gothenburg are identified: inner city,
captive riders, middle suburbanites, outer suburbanites, high-income areas, country
living areas and car-dominated locations. It provides a basis for deliberation as to
where to prioritise future investment in strengthening the public transport network,
and where to promote (or hinder) future urban development.
Overall, this dual consideration of supply and demand adds a significant dimension
to decision support that is not possible by using conventional planning tools, which
inherently struggle with the identification of latent demand. The potential to run and
compare concurrent and integrated future scenarios for public transport network
development and land use trends using SNAMUTS and other accessibility tools,
provides planners the opportunity to fundamentally rethink the relative roles of
different transport modes in a given city, for example under far-reaching assumptions
of energy policy transitions or changed planning priorities favouring non-car modes
for urban mobility. A better understanding of the relationship between the supply-
based elements of such scenarios and their patterns of impact on demand can help
planners to better anticipate and explore barriers and opportunities for policy
changes.
This analysis clearly shows that public transport investments cannot be made in
isolation, instead they must be considered in relation to the wider urban/regional
context, as well as to trip makers and their socio-economic situation, and also to the
land use system and the built environment. It is not a case of only improving

23
transport accessibility, although we clearly show that this is very important and
necessary. It is acknowledged that there is not a ‘one-size-fits-all’ situation, but the
typology proposed can provide a framework for policy deliberation regarding a more
holistic public transport investment strategy. It has been demonstrated that
SNAMUTS, and other new generation accessibility tools, in combination with other
data (including travel demand) can play a key role in planning for sustainable
transport and urban development.

Acknowledgements
The research leading to these results has received funding from the People
Programme (Marie Curie Actions) of the European Union's Seventh Framework
Programme (FP7/2007-2013) under REA grant agreement n° 608743.

References
Albrechts, L. 2006. Shifts in strategic spatial planning? Some evidence from Europe
and Australia, Environment and Planning A 38, 1149 - 1170
Banister, D., Hickman, R. 2013. Transport futures: Thinking the unthinkable.
Transport Policy 29, 293 – 293.
Börjesson, M., & Kristoffersson, I. (2015). “The Gothenburg congestion charge.
Effects, design and politics”. Transportation Research Part A: Policy and Practice, 75
134-146. doi: 10.1016/j.tra.2015.03.011.
Clifton, K., Moura, F. 2017. Conceptual framework for understanding latent demand:
Accounting for unrealized activities and travel. Transportation Research Record
2668, 78 – 83.
Curtis, Stjernborg, Petterson, (forthcoming) The duality of Swedish transport policy:
examining Stockholm, Gothenburg and the state.
Davoudi, S. 2006. Evidence-based planning disP – The Planning Review 42:165, 4 -
24
Ewing, R., Hamidi, S., Tian, G., Proffitt, D., Tonin, S., Fregolent, L (2017) Testing
Newman and Kenworthy’s Theory of Density and Automobile Dependence. Journal
of Planning Education and Research
Freestone, R. 2012. Futures thinking in planning education and research. Journal for
Education in the Built Environment 7(1), 8 – 13.
Göteborgsregionens Kommunalförbund (2009) K2020 Målbild för kollektivtrafiken i
Göteborgsregionen [K2020 Long term vision for the Public Transport System in the
Gothenburg Region]. Gothenburg, Sweden: Göteborgsregionens kommunalförbund.
Hatzopoulou, M., Miller, E.J. 2009. Transport policy evaluation in metropolitan areas:
The role of modelling in decision-making. Transportation Research Part A 43, 323 –
338

24
Healey, P. 2006. Transforming governance: Challenges of institutional adaptation
and a new politics of space. European Planning Studies 14 (3), 299 - 320
Hine J, Grieco M (2003) Scatters and clusters in time and space: Implications for
delivering integrated and inclusive transport. Transport Policy, Vol 10, pp 299-306
Hrelja, R. 2011. The tyranny of small decisions. Unsustainable cities and local day-
to-day transport planning. Planning Theory and Practice 12 (4), 511 – 534
Hull, A., Silva, C., Bertolini, L., 2012. Accessibility Instruments for Planning Practice.
Porto, COST Office
Keblowski, W., Bassens, D. 2017. All transport problems are essentially
mathematical: The uneven resonance of academic transport and mobility knowledge
in Brussels. Urban Geography
Naess, P., Nicolaisen, M., Strand, A. 2012. Traffic forecasts ignoring induced
demand: a shaky fundament for cost-benefit analysis. European Journal Transport
Infrastructure Research 12 (3) 291 - 309
Ohnmacht T, Götz K, Schad H (2009) Leisure mobility styles in Swiss conurbations:
construction and empirical analysis. Transportation, Vol 36, pp 243-265
Owens, S. 1995. From ‘predict and provide’ to ‘predict and prevent’?: Pricing and
planning in transport policy. Transport Policy 2 (1), 43 – 49.
Papa, E., Coppola, P., Angiello, G., Carpentieri, G. 2017. The learning process of
accessibility instrument developers: Testing the tools in planning practice.
Transportation Research Part A 104, 108 -120
Scheiner J (2018) Transport costs seen through the lens of residential self-selection
and mobility biographies. Transport Policy, Vol 65, pp 126-136
Stone J, Scheurer J (2014) Building Confidence in the Future of Melbourne’s Public
Transport Systems. Chapter 4.1 in Whitzman C, Gleeson B, Sheko A (2014, Eds)
Melbourne: What Next? Research Monograph No 1, Melbourne Sustainable Society
Institute, University of Melbourne
Swedish Transport Administration (2017) Västsvenska paketet. Uppföljning av bil-
och kollektivtrafik. Redovisning av 2016. [The West Swedish Agreement. Follow-up
of car and public transport. Report of 2016.] Available at:
http://www.trafikverket.se/contentassets/2109d7678fa74783b431d595e23d9ab5/rap
porter/uppfoljning-av-bil-och-kollektivtrafik-2016_170306-vp.pdf
Trafikkontoret. 2017. Trafik- och Resandeutveckling 2017. [Development of trafic
and travelling 2017]. Göteborg, Sweden: Göteborgs stad. Available at:
http://www2.trafikkontoret.goteborg.se/resourcelibrary/Trafikochresandeutveckling20
17.pdf
te Brömmelstroet M, Silva C, Bertolini L (2014, eds) Assessing Usability of
Accessibility Instruments. Amsterdam, Netherlands: COST.

25
te Brömmelstroet, M., Bertolini, L. 2011. The role of transport-related models in
urban planning practice. Transport reviews Vol 31 (2), 139 -143
Walker, J. 2008, Purpose-driven public transport: creating a clear conversation about
public transport goals. Journal of Transport Geography 16 (6), 436-422
Västra Götalandsregionen. 2018. Fakta om trafik & transporter. [Facts about trafic
and transport]. Available at: https://www.vgregion.se/faktavg

26
Highlights
 Accessibility tools can make an important contribution to future urban
development and investment decisions.
 By assessing both public transport supply and travel demand through an
accessibility lens an insight into the latent demand for public transport can be
gained.
 An innovative segmentation type approach is employed by comparing supply
and demand for public transport.
 A typology is proposed as a framework for a more holistic public transport
investment strategy.

27