Journal of Transport Geography 9 (2001) 229±242
www.elsevier.com/locate/jtrangeo
Location, economic potential and daily accessibility:
an analysis of the accessibility impact of the high-speed line
Madrid±Barcelona±French border q
Javier Gutierrez *
Departmento de Geografia Humana, Facultad de Geografia e Historia, Universidad Complutense de Madrid, 28040 Madrid, Spain
Abstract
This paper evaluates the accessibility impact of the future Madrid±Barcelona±French border high-speed line. Accessibility impact
of the new infrastructure is measured by means of three indicators: weighted average travel times, economic potential and daily
accessibility. These indicators respond to di€erent conceptualizations and o€er complementary information about the issue accessibility. The results are quite di€erent: very concentrated e€ects in the daily accessibility indicator, less concentrated in the
economic potential one and more dispersal in the location indicator. The sign (polarizing/balancing) of these e€ects depend on the
geographic scale: polarizing e€ects at the national level and balancing e€ects at both corridor and European levels are identi®ed. A
geographic information system (GIS) was used to carry out this study. Ó 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Accessibility; Railway transport; High-speed train; Spain; European Union; Geographic information systems (GISs)
1. Introduction
Historically, changes in transport infrastructures
have produced a progressive contraction of space, in the
sense that travel times have been shortened and transport costs reduced. Improvement of transportation
networks reduces interaction costs, increases the overall
competitiveness of the system and allows for more specialisation such that economies of scale and specialisation bene®ts can be utilised. Hence, we should expect
that production and other economic activities can be
carried out more eciently as the quality and capacity
of a region's tranportation networks increase (Forslund
and Johansson, 1995).
The high-speed train permits links between cities in
conditions hitherto unimaginable. Its competitivity in
the transport market is based on quality of service, access
times to the chief economic activity centres and capability to handle large passenger volumes. This throughput is vitally important in the European regions, where
q
This is a revised and enlarged version of a paper presented to the
ERSA Conference, Dublin, 1999. It is based on research supported by
the Spanish government (Ministerio de Fomento).
*
Tel.: +34-1-394-5949; fax: +34-91-394-5960.
E-mail address: [email protected] (J. Gutierrez).
large conurbations are separated by distances of several
hundred kilometres (Campenon, 1995). By shortening
travel times, it brings about changes in accessibility
conditions and hence in the relative location of places,
with obvious consequences for the development potential of the regions a€ected, since the spatial situation of
the EC regions within Europe and the accessibility of
large agglomerations and infrastructure facilities are
factors of attractiveness and development capabilities of
the regions (Lutter et al., 1992) and good connections in
the international infrastructure networks will be a critical
success factor in the distribution of economic activity in
Europe (Bruinsma and Rietveld, 1993).
The European high-speed train network is currently
at an early stage of development, with a limited number
of mainly unconnected lines, but a number of new links
are planned in order to complete an European highspeed train network (Commission of the European
Communities, 1990). One of these planned links is the
Madrid±Barcelona±French border line, which is currently under construction to standard gauge and will be
completed by the year 2005. Its antecedents are to be
found in the AVE Madrid±Seville, inaugurated in the
year 1992 and also built to standard gauge, so that both
lines will constitute a continuous structuring axis from
Seville to the French border (Seville±Madrid±Zaragoza±
Barcelona±French border). Madrid±Barcelona services
0966-6923/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 9 6 6 - 6 9 2 3 ( 0 1 ) 0 0 0 1 7 - 5
230
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
Table 1
Travel times by train between the main Spanish cities on the corridor
Routes
Travel time, 2005
Madrid±Barcelona
Madrid±Zaragoza
Zaragoza±Barcelona
Travel time saving
Without new line
With new line
Absolute
%
5 h 28 min
3 h 3 min
3 h 27 min
2 h 40 min
1 h 25 min
1 h 15 min
2 h 48 min
1 h 38 min
2 h 12 min
51.2
53.5
68.6
Table 2
Travel times by train between Barcelona±Madrid and other European cities
Routes
Travel time, 2005
Travel time saving
From
To
Without new line
With new line
Absolute
(%)
Barcelona
Marseilles
Milan
Lyon
Paris
Brussels
Frankfurt
5 h 21 min
10 h 39 min
6 h 01 min
7 h 31 min
8 h 51 min
10 h 36 min
3
8
4
5
7
8
1
1
1
1
1
1
h
h
h
h
h
h
50
50
50
50
50
50
min
min
min
min
min
min
34.27
17.21
30.47
24.39
20.71
17.29
Madrid
Marseilles
Milan
Lyon
Paris
Brussels
Frankfurt
10
16
11
10
11
13
6 h 11 min
11 h 29 min
6 h 51 min
8 h 21 min
9 h 41 min
10 h 56 min
4
4
4
1
1
1
h
h
h
h
h
h
38
38
38
39
39
39
min
min
min
min
min
min
42.84
28.75
40.35
16.50
14.51
12.64
h
h
h
h
h
h
49
07
29
00
22
03
min
min
min
min
min
min
will run at up to 350 km/h, so that travelling times between the main cities on the corridor will be drastically
reduced (Table 1) and time savings will be also important for some relevant European connections (Table 2).
As a result, it is expected that the new line will have the
potential to improve signi®cantly the accessibility by rail
in several European cities and regions. There are plans
to extend the Spanish high-speed network beyond the
Madrid±Barcelona±French frontier, including the links
Madrid±Valencia/Alicante, Madrid±Valladolid, Zaragoza±Bilbao and Madrid±Lisbon.
This paper evaluates the accessibility impact (predicted e€ects) of the high-speed line Madrid±Barcelona±
French border. In order to re¯ect exclusively the e€ects
of the new line, two scenarios are considered: the year
2005 ``with'' and ``without'' the new line. Accessibility
impact is measured by means of di€erent indicators.
Inequality measures are used in order to verify whether
the line will reduce or increase existing accessibility
disparities between cities on di€erent spatial scales.
Selection of indicators and geographical scale are
very important issues by measuring accessibility changes. In fact, results can be di€erent depending on the
indicators used and on the geographical scale. In this
paper we are going to investigate the in¯uence of both
factors by analysing accessibility changes in the example
of this new high-speed line:
h
h
h
h
h
h
31
49
11
41
01
46
min
min
min
min
min
min
· Accessibility indicators. Accessibility impact of the
new infrastructure is measured by means of three indicators: weighted average travel times, economic potential and daily accessibility. 1 These indicators
respond to di€erent conceptualizations and o€er
complementary information about the issue of accessibility.
· Geographical scale. The new line will play an important role for both European and Spanish internal relations. From the European point of view, this line
will enable a peripheral space (the Iberian Peninsula)
to link up to the rest of Europe; from a national perspective, it will link the two main cities in Spain and
other urban agglomerations along an important corridor of the country. 2 Therefore the problem of scale
has to be carefully taken into account: we analyse
whether the new line will increase or reduce dispari-
1
A recent paper (Bruinsma and Rietveld, 1998) compares a number
of accessibility indicators at the European level. The authors use as
input results from other papers, with di€erent data and with di€erent
demarcations of the study area. Because of this, it is not clear to what
extent variations in the accessibility values depend properly on the
di€erent operationalizations and to what extent they depend on
di€erences in data and demarcation of the study area.
2
This is a situation similar to that which will result in Italy with the
building of the Rome±Milan line.
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
ties between cities on the international, national and
corridor scales.
In order to de®ne the study area, it should be borne in
mind that much of the territory of the European Union
(including cities such as London, Amsterdam or Brussels) will be less than 10 h away from Madrid or Barcelona once the new line is built and connected up to the
French TGV (Table 2), so that apart from day services
for short or medium distances, there is a possibility that
night trains could be put into service for longer links. 3
For this reason, it would seem reasonable to extend the
study to the whole of the territory of the European
Union, with the exception of the islands and Sweden,
Finland and Greece. For these areas, the bene®ts of the
new line can be considered irrelevant at the outset on
account precisely of their insularity (in the case of the
former) or their great distance from the new line (in the
case of the latter three countries), although they could
nevertheless distort the values of certain accessibility
indicators. Great Britain should logically be taken into
account due to its relative proximity and to the fact that
the Channel Tunnel has eliminated its erstwhile island
condition. Consequently, the countries which will be
taken into account are Spain, Portugal, France, Belgium, the Netherlands, Luxemburg, the United Kingdom, Germany, Denmark, Austria and Italy, to which
Switzerland must be added because of its geographical
proximity and economic connections.
This paper is divided into six parts, including this
section. Section 2 examines and discusses the three
accessibility indicators selected. Section 3 describes the
network building and accessibility calculation procedures. In Section 4 the accessibility impact of the new
line according to the indicators selected is discussed. In
Section 5 the inequality measures are used in order to
verify whether the line will reduce or increase existing
accessibility disparities. The research ®ndings are summarized in Section 6.
2. Accessibility indicators
231
tempt to encapsulate notions of the opportunities
available to people and ®rms to reach places where they
can carry out activities that are important to them
(Linneker and Spence, 1992). Usually, by accessibility,
we mean the ease with which activities can be reached
from a certain place and with a certain system of
transport (Morris et al., 1978). Since at the regional,
national or international levels, activities are particularly concentrated in the urban agglomerations and the
competitiveness of the high-speed train is based on
quality of service and access times to the chief cities, it
would seem logical to measure accessibility with respect
to such cities.
There is a wide variety of indicators to measure accessibility (see, for example, Vickerman, 1974; Pirie,
1979; Jones, 1981; Bruinsma and Rietveld, 1993;
Reggiani, 1998). This variety re¯ects the numerous approaches to the concept of accessibility. Most accessibility measurements combine the travel cost to, and
attractiveness of, di€erent destinations in a single indicator (Geertman and Ritsema van Eck, 1995). Travel
cost is a measure of the e€ect of friction of the distance
(the resistance to movement between two points). It can
be expressed in di€erent cost units, such as distance
(Keeble et al., 1988), travel time (Lutter et al., 1992;
Bruinsma and Rietveld, 1993; Dundon-Smith and Gibb,
1994; Geertman and Ritsema van Eck, 1995; Gutierrez
and Urbano, 1996) or generalised cost of transport
(Linneker and Spence, 1992; Spence and Linneker,
1994). Attractiveness of urban agglomerations depends
on their masses. Given the available data, di€erent indicators of capacity of attraction can be used, such as
population (Lutter et al., 1992; Bruinsma and Rietveld,
1993), employment (Linneker and Spence, 1992; Spence
and Linneker, 1994) or gross domestic product (Keeble
et al., 1988; Gutierrez and Urbano, 1996; Gutierrez
et al., 1996). In European studies travel time is normally
used as an indicator of friction of the distance and
population or gross domestic product as an indicator of
attractiveness.
Accessibility is not easy to de®ne in unambiguous and
quanti®able terms. In its most abstract form, accessibility involves a combination of two elements: location
on a surface relative to suitable destinations, and the
characteristics of the transport network or networks
linking points on that surface (Vickerman, 1974). Actually accessibility concepts are generally used to at-
2.1. Accessibility indicators selected
3
Studies carried out at the European level have clearly identi®ed the
favoured niches of the tree main passenger transport systems: under
300 km, the motorcar enjoys a virtual monopoly; above 1400 km, air
transport dominates; and in between the two, high-speed rail occupies
a niche which, at both extremities, is subjected to competition from
both the motorcar and the plane (Campenon, 1995).
2.1.1. Weighted average travel times (location indicator)
The weighted average travel time between each node
and all urban agglomerations is calculated taking as
weight the mass of the centres according to the following:
To carry out this study, we have selected three indicators which respond to di€erent conceptualizations and
o€er complementary information as to the problem of
changes in accessibility conditions brought about by a
new infrastructure.
232
Pn
jˆ1 …Tij Mj †
;
Li ˆ Pn
jˆ1 …Mj †
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
…1†
where Li is the accessibility (location) of node i, Tij is the
travel time by the minimal-time route through the network between node i and urban agglomeration j (in
min), and Mj is the mass (for example, gross domestic
product) of the destination urban agglomeration.
The mass of the urban agglomerations is used as
weight in order to value the importance of the minimal-time routes (Gutierrez and Urbano, 1996;
Gutierrez et al., 1996). The measure is not a gravitybased indicator (there is no distance-decay), so that,
unlike economic potential, it does not place the emphasis on short distances. Thus, for example, in the
economic potential model, the relationship Madrid±
Guadalajara could weigh more than the relationship
Madrid±Paris, for Guadalajara is much closer to
Madrid than Paris. But it seems quite clear that, from
the perspective of the European high-speed network,
the ®rst of these relationships is less relevant, since
the second is one of the key trans-European connections.
This average-distance-based indicator should be interpreted from the locational rather than the economic
point of view. But economic implications are obvious,
since the spatial situation of the EC regions within Europe is a factor of attractiveness and development capabilities of the regions (Lutter et al., 1992). This
measure expresses the relative location of each place and
the extent to which a new link modi®es this location by
reducing access times to the main urban agglomerations.
The results are very easily interpreted, for example: from
node A the average travel time to all centres is 400 min
in the scenario ``without the line'' and 360 min in the
scenario ``with the line'', which means a time saving of
40 min.
2.1.2. Economic potential
The economic potential is a gravity-based measure,
widely used in accessibility studies (see, for example,
Harris, 1954; Keeble et al., 1988; Linneker and Spence,
1992; Smith and Gibb, 1993; Spence and Linneker,
1994). It is a measure of the nearness or accessibility of a
given volume of economic activity to a particular point/
region and can be interpreted as the volume of economic
activity to which a region has access, after the cost/time
of covering the distance to that activity has been accounted for (Dundon-Smith and Gibb, 1994). According to this model, the level of opportunity (accessibility)
between a node i and a destination node j is positively
related to the mass of the destination and inversely
proportional to some power of the distance between
both nodes. Its classical mathematical expression is as
follows:
Pi ˆ
n
X
Mj
;
Tija
jˆ1
…2†
where Pi is the economic potential of node i, a is a parameter re¯ecting the rate of increase of the friction of
distance (distance decay) and the other terms are still
known.
In this paper (as in most accessibility studies) the
parameter value a used is 1. Using higher values than 1
means giving too much importance to relations over
short distances (which would not seem appropriate
when analysing the e€ects of a new infrastructure of a
national and international nature such as the line which
is the object of this study) and it also means increasing
the problem known as self-potential (see Frost and
Spence, 1995; Bruinsma and Rietveld, 1993).
When discussing the former indicator (weighted average travel times), it has been argued that in the evaluation of the impact of large transport infrastructures at
the European level it would seem reasonable to point
out the long distance e€ects. Yet from a merely economic point of view, there is no doubt that the economic
e€ects of a new infrastructure are inversely related to the
distance (there are many trips over short distances and
few trips over long distances), so that in this context it
would seem appropriate to use a gravity-based operationalization.
Therefore, the interpretation of the results provided
by this indicator must be carried out from an economic
viewpoint: the indicator measures the economic potential of each place in each of the scenarios considered and
the changes in potential caused by the new infrastructure.
2.1.3. Daily accessibility indicator
This indicator consists of calculating the amount of
population or economic activity that can be reached
from a node within a certain travel time limit. The time
limit is usually established in 3 or 4 h, so that it is
possible to go and return within the day and carry out
an activity at the visit location (Lutter et al., 1992). The
limit of 4 h travel is considered as a critical cut-o€ point
since it represents the likely limit of comfortable day
return business trac, although the limit of 3 h is the
likely cut-o€ point for major transfers from air to rail
transport. A study by the Institute of Air Transport
suggests a loss of 30±50% of air trac for a 3-h rail
connection, 15±30% at for 4 h and less than 20% at more
than 6 h (see Vickerman, 1995). On account of lack of
space, in this paper the results will only be expressed in
number of inhabitants within the travel time limit of 4 h.
This measurement is particularly useful for calculating accessibility in business and tourist trips, for the
need to stay overnight in the destination city means an
important extra expense for both companies and indi-
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
viduals. In fact, the empirical evidence shows that new
high-speed lines produce an increase not only in the
number of travellers in the relations served by the line,
but also in the proportion of those who return within the
same day (Bonnafous, 1987).
In the context of the high-speed train, this indicator
provides basically the number of possible business
contacts (for business trips) and the market potential
(for tourist trips). It measures how much population can
be reached from a place (or can reach a place) in a
certain travel time limit and the changes in accessible
population brought about by a new infrastructure. The
results are of the following type: from city A, within a
travel time limit of 4 h, 10 million inhabitants can be
reached in the scenario ``without the line'' and 15 million
in the scenario ``with the line'', which means an increase
of 5 million inhabitants.
2.2. Comparison among indicators
By analysing the accessibility e€ects of a new link by
means of the three indicators selected it is necessary to
take into account various critical issues.
2.2.1. Distance (travel time) decay
The three indicators handle the distance (travel time)
in a di€erent way (Fig. 1 and Table 3):
233
· The daily accessibility indicator considers exclusively
the centres within a certain travel time limit. Within
this time limit there is no distance decay.
· The economic potential indicator takes into account
all relationships within the study area, but inversely
to the distance, so that short distance routes contribute very heavily but those over long distances very little (distance decay).
· The weighted average travel time indicator considers
all relationships within the area of study too, but
routes on short distances do not contribute more
than others in accessibility calculations, since there
is no distance decay.
Thus, for example, a reduction of the travel time between Madrid and Barcelona from 6 h 30 min to 2 h 30
min will be picked up by the three indicators as following:
· The absolute saving on travel time between Madrid
and Barcelona is equally large compared with that
between Lisbon and Barcelona, but according to
the economic potential indicator the accessibility improvement will be greater for the former relationship
than for the latter, because Madrid is closer to Barcelona than Lisbon: a longer distance (travel time) leads
to less weight for the particular relationship (Fig.
1(a)).
· The weighted average travel time indicator will
record the same accessibility improvement for the
Fig. 1. Travel time decay and study area limit (partially based on Bruinsma and Rietveld, 1998).
Table 3
Characteristics of the selected accessibility indicators
a
b
Indicator
Economic centres
considered
Distance decay
Weight according to size
of economic centres
Measure units:
results in
Weighted mean travel costs
Economic potential
Daily accessibility
All
All
Several
No
Yes
No
Yes
Yes
Yes
Travel costsa
Economic activityb
Population or economic
activityb
The lower the value attained ˆ more accessibility.
The higher the value attained ˆ more accessibility.
234
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
relationship Madrid±Barcelona than for the Lisbon±
Barcelona one (4 h), since there is no distance decay
in the operationalization (Fig. 1(b)).
· Finally, the daily accessibility indicator will record an
accessibility improvement for Madrid (because the
travel time Madrid±Barcelona is reduced from 6 h
30 min to only 2 h 30 min), but not for Lisbon (because the travel time Lisbon±Barcelona exceeds the
critical level of 4 h both before and after the link improvement) (Fig. 1(c)).
As a result, the selected indicators pick up the impact of
a new link in a di€erent way: it is expected that the effects would be very concentrated in the daily accessibility indicator, less concentrated in the economic
potential one and more dispersal in the location indicator.
2.2.2. Treatment of internal accessibility: the self-potential problem
The internal accessibility of a city has not a signi®cant
in¯uence on the calculation of indicators with no distance decay, but may have a substantial impact on the
®nal outcome in the case of gravity type models (Bruinsma and Rietveld, 1998). Thus, by analysing the e€ects
of a new high-speed link, potential values may depend to
a great extent on local accessibility, when local accessibility (for example, the relationship Paris±Paris) has
nothing to do with the high speed train network (Gutierrez et al., 1996).
This factor is known as ``self-potential'' in gravity
models: the contribution of the potential of the individual city itself to the total potential of that city in
accessing to the network of places in the system. This
contribution may be very high for large cities, as the
function used lead to high weights for self-potential
(Frost and Spence, 1995). As a result, by analysing the
e€ects of a new link relative changes (in percentage) for
large cities may be very low, since the (®xed) self-potential are very high for such cities (see Table 5).
2.2.3. Demarcation of total area
A disadvantage of the no gravity indicators is that
they depend rather strongly on the demarcation of the
total area of study, because there is no distance decay, so
that such a demarcation should be chosen carefully
(Bruinsma and Rietveld, 1993). The demarcation of the
study area should be in correspondence with the distance within which the transport mode is competitive. If
the study area is too large, irrelevant relationships over
long distances may weigh heavily in the calculation of
the indicator; if the study area is too small, relevant
relationships are not considered (Fig. 1). It is also important to take into account political and economic
borders, since such borders produce a decrease in the
volume of the ¯ows (see Section 1).
2.2.4. Asymmetry
Network improvement may be picked up symmetrically or asymmetrically by accessibility indicators.
Asymmetry holds for the three selected indicators (since
they take into account the mass of the centres), so that a
reduction of the travel time in a link between a small city
and a large one leads to a larger increase in accessibility
for the former than for the latter. Thus, for example,
according to the indicators selected Zaragoza will bene®t more than Madrid from a new high-speed link between both cities, since Zaragoza will bene®t from a
better access to a large agglomeration as Madrid.
2.2.5. The ease of interpretation
Accessibility indicators must be understandable to
both the public and policy decision makers. Weighted
average travel time and daily accessibility indicators give
a value in meaningful units (minutes and population,
respectively) (Table 3), so that results are easily interpreted. This is the case, for example, when we say that a
new link produces a time saving of 20 min for city A
(weighted average travel time) or that 5 million inhabitants more can be reached from city A (daily accessibility). However, a potential value is not easily
interpreted in terms relating to geographical reality: an
increment of, say, 12,554 potential units for city A does
not tell us much (Geertman and Ritsema van Eck,
1995).
3. Modelling of networks and calculation of indicators
In order to calculate accessibility changes between the
scenarios 2005 without and 2005 with the new line, a
dense intermodal network was modelled in a geographic
information system (ARC/INFO). The main focus of
interest is logically the railway network, but the road
network is also included, for this latter enables access to
railway stations from places which have no station.
Data of the railway and road networks for the foreseen
situation in 2005 and predicted population and gross
domestic product (GDP) data for the main urban agglomerations in the same year were stored in the GIS.
When modelling the railway network, all lines at the
interregional and international levels were taken into
account. Stations at which long-distance trains stop
were considered as nodes. For each arc on the railway
network, the type of line, distance, speed and travel time
was registered. Distances and travel times for the existing lines were obtained from Cook's European Timetable (1996). For the horizon 2005, the changes foreseen
(see Fig. 2) were included by adding the new lines or
modifying the attributes (type of line and travel time) of
those lines to be upgraded. In order to calculate travel
times, we took into account a speed of 275 km/h for the
new lines and 200 km/h for those to be upgraded. In the
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
235
Fig. 2. The European high-speed network in the study area: scenario 2005.
case of the Madrid±Barcelona±French border line, the
travel times foreseen in the project were recorded.
Likewise, penalties times were recorded in turn tables
in order to simulate the change from road to rail mode
(60 min) and the change between the Iberian broad
gauge (conventional railways) and the European standard gauge (high-speed railways) in Spain (20 min). This
20 min penalty is not an estimated time, since it is the
time currently used in the operation of change in track
gauge. On the other side, penalty time for change from
road to rail could seem too high, but it also incorporates
the uncertainty linked to road trac.
A dense road network was also used in order to
guarantee access to stations from places which are not
directly served by the railway network. The nodes on
the road network were selected so that not only all
places with a signi®cant demand were included (in
accordance with their population), but also a number
of minor cities in order to cover the whole territory
homogenously, to guarantee the necessary accuracy in
mapping accessibility by using interpolation techniques. For each arc on the road network, the length,
estimated speed according to type of infrastructure and
travel time were also recorded. Travel time was obtained on the basis of the length of the arcs and estimated speeds according to the type of road: 120 km/h.
for motorways, 110 for expressways, 90 for interre-
gional roads and 70 for other roads. The new links
foreseen in the Trans-European Road Network Outline
Plan were also taken into account, assuming that such
links will be ®nished by the year 2005 (see Gutierrez
and Urbano, 1996).
In accessibility calculations are necessary not only
network data, but also population and/or GDP data of
the destination centres. The 88 urban agglomerations in
the study area (13 of them in Spain) with a population of
over the critical mass point of 300,000 inhabitants were
selected as destination centres. The population data of
these agglomerations come from the ocial statistics of
each country, whereas GDP was calculated on the
strength of the population of each agglomeration and
the GDP per capita of the region (NUT) in which each
agglomeration is located. This latter variable was obtained on the basis of REGIO data of the European
Union, which gives standardised data to this respect.
For the year 2005 predictions were made for both
population (based on former growth tendencies of agglomerations) and gross domestic product (taking into
account the population prediction and assuming an
annual increase of 2% in the GPD) of the selected urban
agglomerations.
Obviously there are some uncertainties in predictions for population, GDP and road and rail networks changes, but they do not in¯uence heavily the
236
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
Fig. 3. Changes in accessibility: location indicator.
results (changes in accessibility due to the new link),
since population, GDP and road and rail networks
are the same in both scenarios: 2005 without the line
and 2005 with the line (the only di€erence between
them is the new link). Thus, by comparing the
without and with accessibility values, we are only
picking out the e€ects of the reduction in travel time
due to the future Madrid±Barcelona±French border
high-speed line.
For each of the scenarios (2005 with the new line''
and 2005 without the new line), minimal-time routes
between each node and each destination centre were
calculated, bearing in mind arc and node impedances.
When the node of origin has a station, the travel time
between this node and the destination city is equal to the
sum of the travel times of the arcs on the railway network plus, where necessary, penalty times for change in
track gauge (20 min). When the node of origin has no
station, access time by road to the nearest station was
added and a penalty for the change from road to rail
mode (60 min).
In order to calculate accessibility values it is necessary
to take also into account internal relationships. Average
travel times for these internal relationships were estimated in this paper according to the following formula:
T ˆ 15 log…P 10†;
…3†
where T is the average internal travel time and P is the
population (in millions) living within this city.
This is a logarithmic function because the average
travel time of internal relationships tends to increase as
the size of the city increases, but not in a linear fashion
(Barber, 1986). The highest value in the study area
(Paris±Paris) is some 26 min.
Once the access times were obtained for all the
relationships considered, the accessibility values of the
4000 nodes in the network 4 were calculated applying
the corresponding formula (see Section 2) for each of
the scenarios. Once changes in accessibility values for
each node were computed, relative data (changes in
percentage) were interpolated and mapped, so that it
is possible to compare how di€erent indicators capture
accessibility changes (see Figs. 3±5). On the other side,
Tables 4±6 show both absolute and relative improvements for those cities that record more signi®cant
changes.
4
This high number of nodes is necessary to facilitate isoaccessibility
mapping, so that it is possible to obtain representative results of
accessibility conditions in the whole of the study area and not
exclusively in the chief cities.
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
Fig. 4. Changes in accessibility: potential indicator.
Fig. 5. Changes in accessibility: daily accessibility indicator.
237
238
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
4. Impact of the new line on accessibility
4.1. Weighted average travel times
The new line will bring about a reduction of 25 min
(about 5%) in the average travel times between the selected agglomerations (those in the study area with more
than 300,000 inhabitants). Logically, the greatest bene®ts were recorded in the Iberian Peninsula, for the new
line o€ers its cities better access with each other and with
cities in the rest of Europe; on the other hand, the improvements recorded in the rest of Europe are scant, for
the new line only enables its cities to improve their
connection with the cities in the Iberian Peninsula, these
latter having relatively little weight with respect to the
total group of cities in the study area (Fig. 3 and Table
4).
The greatest bene®ts logically occur in cities with a
station on the new line, but the e€ects on the Mediterranean corridor are also important, as likewise on the
corridor of the high-speed Madrid±Seville line. The agTable 4
Weighted mean times (in min) in the without and with scenarios for
selected cities (only cities showing a percentage di€erence above 2.5%
are included)
Cities
Alicante
Amsterdam
Barcelona
Brussels
Cordoba
Frankfurt
Geneva
Granada
Grenoble
La Coru~
na
Lisbon
London
Lyon
Madrid
Malaga
Marseilles
Milan
Murcia
Naples
Oporto
Paris
Rome
Seville
Turin
Toulon
Toulouse
Valencia
Valladolid
Zaragoza
Scenarios
Di€erence
Without
new line
With
new line
Absolute
(%)
805.6
321.8
613.6
276.9
897.2
300.1
380.6
1020.8
413.7
983.9
1227.8
337
337.1
742.3
1031.3
398.3
529.4
832.1
767.4
1121.7
269.8
703.3
941.8
540.4
443.4
507.8
741.5
709.2
729.2
696.5
313.1
504.4
268.1
749.0
289.8
359.7
951.3
392.7
957.3
1196.1
328.3
316.3
640.8
883.1
377.3
508.4
722.9
746.4
1095.0
261.1
682.3
793.6
519.4
422.5
493.6
631.6
677.5
568.5
109.1
8.7
109.2
8.7
148.2
10.2
20.8
69.5
20.9
26.6
31.7
8.7
20.8
101.5
148.2
21.0
21.0
109.1
21.0
26.7
8.8
21.0
148.2
21.0
21.0
14.1
110.0
31.7
160.6
13.5
2.7
17.8
3.1
16.5
3.4
5.4
6.8
5.0
2.7
2.5
2.6
6.1
13.6
14.3
5.2
3.9
13.1
2.7
2.3
3.2
2.9
15.7
3.8
4.7
2.7
14.8
4.4
22.0
glomeration which obtains the highest time saving (160
min, which is equivalent to 22.0%) is Zaragoza. It
should be borne in mind that Barcelona (with a 17.8%
improvement) uses the least amount of kilometres on the
new line in order to reach most European cities (situated
on the other side of the border) and that Madrid (with a
13.7% time saving) gains a few minutes in its relations
with the ``blue banana'' cities (Paris, London, Brussels,
Amsterdam) due to the fact that this connection before
the construction of the new line is made via the Basque
Country and after via Barcelona with a considerable
detour.
The cities served by the high-speed line Madrid±Seville will bene®t not only from the new line, but also
from the suppression of penalty times (due to change in
track gauge) that before had to be carried out in Madrid. They will have a percentage improvement around
15%, somewhat higher even than that of Madrid, above
all because the need to change trains in Madrid will be
eliminated. The cities of the Mediterranean corridor will
also bene®t considerably, for they will take advantage of
the Tarragona±French border stretch in their relations
with most of the European urban agglomerations. The
average time saving of such cities as Valencia, Alicante
and Murcia is around 109 min, which is equivalent to
13±14%. In the north, northwest and west of the Iberian
peninsula, the bene®ts will be less. This is the case, for
example, of Valladolid (4.5% improvement), La Coru~
na
(2.7%), and Bilbao (2.0%), which will use the new line
for their links with cities located in the east (Barcelona,
Marseilles, Milan), but for which the new line will not
o€er anything as regards their links with the ``blue banana'' cities.
Over the border, the greatest bene®ts in absolute
values (21 min) are located on the natural prolongation
of the line eastwards (south of France, Italy). In other
directions the e€ects of the new line become weaker: 10
min for most German cities, 8.7 for Paris and British
cities and Benelux, and only 4 for Bordeaux and Nantes.
If time savings are measured in percentages, these will
logically be lower for remoter cities within each of the
forementioned directions. Thus, travelling eastwards,
Marseilles (5.3%) will give an improvement higher than
that of Naples (2.7%); the same thing occurs travelling
northeast with Lyon (6.2%) as against Copenhagen
(1.3%) or northwards with Paris (3.2%) as against Edinburgh (1.5%). However, in any case, the lowest values
correspond to south-west France (0.9% in Bordeaux),
where the new line hardly brings any bene®t at all.
4.2. Economic potential
The average variation in the economic potential of
the selected urban agglomerations (it only increases by
1.45%) is much less than the one which corresponds to
the location indicator (travel times are reduced by 5%).
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
239
Table 5
Potential indicator: potential values (million ECUS of 1995) in the without and with scenarios for selected cities (only cities showing a percentage
di€erence above 1% are included)
Cities
Alicante
Barcelona
Bilbao
Cordoba
Geneva
Granada
Grenoble
La Coru~
na
Lisbon
Lyon
Madrid
Malaga
Marseilles
Milan
Murcia
Oporto
Seville
Turin
Toulon
Toulouse
Valencia
Valladolid
Zaragoza
Scenarios
Di€erence
Without new line
With new line
Absolute
(%)
3892.2
6388.0
4639.2
3529.5
8290.3
2727.7
6886.3
2723.2
2393.0
10187.5
7149.3
2918.3
8270.8
6949.7
3628.7
2227.2
3480.4
6228.2
6506.8
5874.9
4635.2
4774.4
4156.8
4304.3
7472.1
4733.9
4049.7
8427.5
2880.3
7043.3
2793.5
2443.5
10366.9
7752.5
3278.4
8489.1
7019.8
4010.0
2278.6
3935.2
6307.4
6676.5
6085.9
5161.7
4999.3
5731.6
412.1
1084.0
94.7
520.2
137.2
152.6
157.1
70.3
50.5
179.3
603.2
360.2
218.3
70.1
381.3
51.3
454.8
79.2
169.7
211.0
526.5
224.8
1574.8
10.5
16.9
2.0
14.7
1.6
5.5
2.2
2.5
2.1
1.7
8.4
12.3
2.6
1.0
10.5
2.3
13.0
1.2
2.6
3.5
11.3
4.7
37.8
This is due to the fact that the potential indicator is a
gravity-based measure, so that most European cities
located far from the new line undergo very little variation in their potential values. In fact, changes in accessibility are concentrated on cities most directly a€ected
by the new line to a greater extent than the former indicator (Fig. 4 and Table 5). The city that most bene®ts
from the new line is Zaragoza (37%), which is located
very close in travel time to two large agglomerations
such as Madrid and Barcelona and greatly improves its
links with cities located beyond the French border.
Bene®ts are less for Barcelona (16%) and for Madrid
(8%).
In relative values, Madrid (8%) grows even less than
Cordoba (14%), Seville (13%) and Malaga (12%). This
apparently anomalous situation is due not only to the
fact that it is no longer necessary to change trains in
Madrid, but also that self-potential in Madrid (see Table
5) represents a very large part of its total potential. 5
However, if we bear in mind the absolute values of
di€erences, we observe that due to the distance decay,
Madrid's potential has a greater increase (603 million
5
Self-potential (%)
Self-potential of all the cities is equal in absolute terms in both
situations, ``with'' and ``without'' the new line, so that cities with high
self-potential values tend to record low percentage improvements (see
Section 2.2). Big cities, such as Paris, London and Madrid, tend to
have high self-potential values (Table 5).
10.4
36.6
18.9
7.3
18.3
11.2
11.4
12.6
22.7
21.2
52.4
14.9
23.5
28.0
9.5
7.6
16.9
22.5
10.1
21.0
21.9
7.9
12.4
units) than those of Cordoba (520), Seville (454) or
Malaga (360).
We should also point out the bene®ts (around 10%)
which the line brings to cities in the southern stretch of
the Spanish sector of the Mediterranean arc, such as
Valencia, Alicante and Murcia, which obtain better
connections not only with Barcelona, but also with
many of the European cities located beyond the
French border. In north and northwest Spain the
changes are less. Thus, for example, Bilbao and La
Coru~
na only record increases of 2.0% and 2.5%, respectively.
Outside Spain, improvements in percentages are
somewhat reduced, more so when their distance in time
from the new line is greater. Both in Portugal (Lisbon,
Oporto) and in the south of France (Toulouse, Marseilles, Toulon) potential increases higher than 2% are
recorded. In French regions even further from Spain
and in the north of Italy, the potential variations are
even around 1% and 2%. But in the rest of the study
area, the changes are almost irrelevant (below 1% and
even below 0.5% in the farthest regions).
4.3. Daily accessibility
With the building of the new line the average accessible population within the 4-h limit for the selected
urban agglomerations rises from 20.7 to 21.1 million
240
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
Table 6
Daily accessibility indicator: reachable population (thousands of inhabitants) in a limit of 4 h in the without and with scenarios for selected cities (only cities showing a percentage di€erence above 0% are
included)
Cities
Barcelona
Cordoba
Madrid
Marseille
Murcia
Seville
Toulouse
Valencia
Valladolid
Zaragoza
Scenarios
Table 7
Changes on the coecient of variation for selected accessibility
indicators: urban agglomerations in the study area
Indicators
Di€erence
Without
new line
With
new line
Absolute
(%)
5597
7442
9945
16,992
6789
7093
5022
10,402
8488
9690
13,375
8038
13,209
20,256
10,053
8038
8286
10,998
12,814
12,451
7778
596
3264
3264
3264
945
3264
596
4326
2761
138.9
8
32.8
19.2
48
13.3
64.9
5.7
50.9
28.4
inhabitants, which means an increase of 1.64%. The
bene®ts of the new line are very concentrated according
to this indicator (Fig. 5 and Table 6). They are located
not only along the Madrid±Barcelona corridor, but also
in other parts of Spain and in southern France. It should
not be forgotten that Barcelona is one of the chief metropolises of southern Europe and that it will be accessible in under 4 h from most of the south of France.
Logically, most European cities will record no bene®t
according to this indicator, which only re¯ects daily
accessibility.
The increase in accessible population is particularly
important in Barcelona, 7.7 million inhabitants (139%),
which fundamentally corresponds to the population of
Madrid, Valladolid, Marseilles and Toulouse. Zaragoza
gains 2.7 million (28.4%) through improvements in its
links with Valencia, Seville and Valladolid. The increase
in 3.2. million for Madrid (32.8%), Murcia (48%),
Toulouse (64.9%) and Marseilles (19.2%) are due to the
fact that with the new line they can reach Barcelona in
under 4 h.
5. Increasing or reducing accessibility inequalities?
Finally, there is the question of whether the new line
will contribute to increasing inequalities (in terms of
accessibility) among European cities, or whether it will
on the contrary favour a reduction in same. Increase or
reduction in disparities among cities can be measured by
the coecient of variation. The changes on the coecient of variation of the three indicators used clearly
point in the same direction: the new line will reduce
existing disparities by 1.87% in travel times, by 1.37% in
economic potential and by 2.30% in daily accessibility
(Table 7). This evolution was predictable in accordance
with the analysis of the spatial distribution of the e€ects
Location
Economic potential
Daily accessibility
Scenario, 2005
Without the
new line
With the
new line
Di€erence
44.30
41.90
71.97
42.43
40.53
69.67
)1.87
)1.37
)2.30
Table 8
Changes on the coecient of variation for selected accessibility
indicators: Spanish cities
Indicators
Location
Economic potential
Daily accessibility
Scenario, 2005
Without the
new line
With the
new line
Di€erence
28.20
46.36
63.70
33.15
49.75
67.25
4.95
3.39
3.55
of the new line, which basically favoured the Iberian
Peninsula (when a transport infrastructure mainly favours a peripheral space, it is obvious that it lessens the
centre-periphery disparities).
However, if we change the scale and look merely at
the impact of the new line on Spanish internal links with
cities with a population of over 75,000 inhabitants, the
results are quite di€erent (Gutierrez and Jaro, 1999)
(Table 8). In all the selected indicators, the new line
brings about an increase in the coecient of variation,
that is, an increase in the di€erences in the accessibility
values of the total group of Spanish cities. This is not
surprising given that the new line connects a number of
cities to each other (as Madrid, Zaragoza and Barcelona) that at the national level are already highly accessible in the situation without the new line, and that
these cities are those which most bene®t from the new
line, which logically results in an increase in disparities
between cities.
Finally, if we once more change scale and consider
exclusively the centres situated along the corridor Madrid±Barcelona±French border, the results are very
di€erent: the value of the coecient of variation clearly
drops in all the accessibility indicators (Table 9). This is
because the three indicators selected re¯ect the e€ects of
a new infrastructure asymmetrically, so that the smallest
cities on the corridor (less accessible than the large ones
in the without situation) are those which obtain greatest
improvements in accessibility from the new line, above
all in the indicators which express their results in units of
activity (indicators of economic potential and daily accessibility).
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
Table 9
Changes on the coecient of variation for selected accessibility
indicators: cities along the corridor
Indicators
Location
Economic potential
Daily accessibility
Scenario, 2005
Without the
new line
With the
new line
Di€erence
21.60
46.81
29.00
16.15
28.34
12.16
)5.45
)18.47
)16.84
6. Conclusion
The e€ects of the new line on accessibility will be
relevant not only in the northeast part of Spain, but also
in other areas of the Iberian Peninsula and in the south±
southeast of France, albeit unequally, according to the
location of the cities with regard to the new line. On the
other hand, the accessibility impact of the new line will
be markedly asymmetrical, since Iberian urban agglomerations have little weight in the total group of
cities in the study area.
The three indicators used respond to di€erent conceptualizations and o€er complementary information
about the accessibility issue. The location indicator
emphasizes relationships over long distances and the
daily accessibility one emphasizes relationships over
short distances. Logically, the results are quite di€erent:
very concentrated e€ects in the daily accessibility indicator, less concentrated in the economic potential indicator and more dispersal in the location indicator (see
accessibility maps). In fact, most of the study area is
included within the 2% contour in the location indicator,
but only the Iberian Peninsula and the south±southeast
of France in the economic potential indicator and some
discontinuous areas in Spain and the south±southeast of
France in the daily accessibility indicator. Since there is
a travel time limit in the daily accessibility indicator, this
measure captures the accessibility e€ects of the new line
in a very discontinuous way: large changes near the
railway stations along or near the new line and little or
no changes in the spaces between them.
As expected, the highest average improvement in
accessibility for the total group of the main urban agglomerations in the study area is recorded by the location indicator (5%). It is surprising that a similar average
improvement (1.5%) is recorded in both economic potential and daily accessibility indicators. But in the average improvement calculation more dispersal e€ects in
the economic potential indicator are compensated by
dramatic changes near the new line in the daily accessibility indicator, so that similar average values are obtained by both indicators.
Finally, according to the changes in the coecient of
variation of the three indicators used, it seems clear that
241
at the national level the new line will lead to an increase
in inequality in the distribution of accessibility, for the
cities which have greatest increases in accessibility are
already highly accessible in the without the new line
scenario. Nevertheless, both at the corridor and at the
European level the line will reduce disparities in terms of
accessibility measures used: within the corridor because
the small and medium-sized cities will obtain greater
increases in accessibility than the large ones, which
suggests that spreading processes for economic growth
will be induced; and at the European level because better
communication of Iberian cities with each other and
with the central regions will result in a greater increase in
accessibility of Iberian cities and therefore, in a reduction of core-periphery imbalances.
In summary, caution is required when considering the
accessibility e€ects of a new infrastructure. Certain
statements could be true or false according to the geographical scale and the accessibility indicator selected.
The new infrastructure will reduce accessibility inequalities among cities at the European scale, but will
increase inequalities at the national scale. On the other
side, it will have signi®cant accessibility e€ects at the
European scale if European relationships (trips over
long distances) are emphasized (location indicator), but
the e€ects will be only minor if relationships over short
distances (potential and daily accessibility indicators)
are highlighted.
Acknowledgements
The author would like to thank J.C. Huertas, L. Jaro,
R.Knowles and two anonymous referees for their comments and G. G
omez, M. Alonso and P. L
opez for assisting him in computing and mapping accessibility
values.
References
Barber, G., 1986. Aggregate characteristics of urban transportation.
In: Hanson, S. (Ed.), The Geography of Urban Transportation.
New York, The Guildford Press, pp. 73±90.
Bonnafous, A., 1987. The regional impact of the TGV. Transportation
14, 127±137.
Bruinsma, F.R., Rietveld, P., 1993. Urban agglomerations in European infrastructure networks. Urban Studies 30, 919±934.
Bruinsma, F.R., Rietveld, P., 1998. The accessibility of European
cities: theoretical framework and comparison approaches. Environment and Planning A 30, 449±521.
Campenon, G., 1995. 2nd report of the working group of the
European Commission High-Speed Europe. In: Paper presented
at the PTRC, Proceedings of Seminar A, Pan-European Transport
Issues, Warwick, pp. 53±60.
Commission of the European Communities, 1990. The European
High-Speed Train Network. Brussels/Luxembourg.
Cook, T., 1996. European Timetable. London.
242
J. Gutierrez / Journal of Transport Geography 9 (2001) 229±242
Dundon-Smith, D.M., Gibb, R.A., 1994. The Channel Tunnel and
regional economic development. Journal of Transport Geography
2 (3), 178±189.
Forslund, U.M., Johansson, B., 1995. Assessing road investments:
accessibility changes cost bene®t and production e€ects. The
Annals of Regional Science 29, 155±174.
Frost, M.E., Spence, N.A., 1995. The rediscovery of accessibility and
economic potential: the critical issue of self-potential. Environment
and Planning A 27, 1833±1848.
Geertman, S.C.M., Ritsema van Eck, J.R., 1995. GIS and models of
accessibility potential: an application in planning. International
Journal of Geographical Information Systems 9 (1), 67±80.
Gutierrez, J., Jaro, L., 1999. Impacto de la nueva lõnea de alta
velocidad Madrid±Barcelona±frontera francesa en la accesibilidad
del sistema de ciudades espa~
nol. Estudios de Construcci
on,
Transportes y Comunicaciones 85, 51±81.
Gutierrez, J., Urbano, P., 1996. Accessibility in the European Union:
the impact of the Transeuropean road network. Journal of
Transport Geography 4 (1), 15±25.
Gutierrez, J., Gonz
alez, R., G
omez, G., 1996. The European highspeed train network: predicted e€ects on accessibility patterns.
Journal of Transport Geography 4 (4), 227±238.
Harris, C.D., 1954. The market as a factor in the localisation of
industry in the United States. Annals of the Association of
American Geographers 44, 315±348.
Jones, S.R., 1981. Accessibility measures: a literature review. Crowthorne, Transport and Road Research Laboratory, TRRL Laboratory Report, 967.
Keeble, D., O€ord, J., Walker, S. 1988. Peripheral Regions in a
Community of Twelve. Oce for Ocial Publications of the
European Communities, Brussels/Luxembourg.
Linneker, B., Spence, N.A., 1992. Accesibility measures compared in
an analysis of the impact of the M25 London Orbital Motorway on
Britain. Environment and Planning A 24, 1137±1154.
Lutter, H., P
utz, T., Spangenberg, M., 1992. Accessibility and
Peripherality of Community Regions: The Role of Road Long
Distance Railway and Airport Networks. Commission of the
European Communities, Brussels.
Morris, J.M., Dumble, P.L., Wigan, M.R., 1978. Accessibility indicators for transport planning. Transportation Research 13A, 91±109.
Pirie, G.H., 1979. Measuring accessibility: a review and proposal.
Environment and Planning A 11, 299±312.
Reggiani, A., 1998. Accessibility, trade and locational behaviour: an
introduction. In: Reggiani, A. (Ed.), Accessibility Trade and
Locational Behaviour. Aldershot, Ashgate, pp. 1±16.
Smith, D.M., Gibb, R.A., 1993. The regional impact of the Channel Tunnel. A return to potential analysis. Geoforum 24 (2),
183±192.
Spence, N., Linneker, B., 1994. Evolution of the motorway network
and changing levels of accessibility in Great Britain. Journal of
Transport Geography 2 (4), 247±264.
Vickerman, R.W., 1974. Accessibility, attraction and potencial: a
review of some concepts and their use in determining mobility.
Environment and Planning 6, 675±691.
Vickerman, R.W., 1995. The regional impacts of trans-European
networks. The Annals of Regional Science 29, 237±254.