Content uploaded by Panayotis Christidis
Author content
All content in this area was uploaded by Panayotis Christidis on Mar 17, 2015
Content may be subject to copyright.
Measuring road congestion
P
anayo
ti
s
CHRISTIDIS
,
J
.
Ni
co
lá
s
IBAÑEZ
RIVAS
JRC - 2012
The mission of the JRC-IPTS is to provide customer-driven support to the EU policy-
making process by developing science-based responses to policy challenges that
have both a socio-economic as well as a scientific/technological dimension.
European Commission
Joint Research Centre
Institute for Prospective Technological Studies
Contact information
Address: Edificio Expo. c/ Inca Garcilaso, 3. E-41092 Seville (Spain)
E-mail: jrc-ipts-secretariat@ec.europa.eu
Tel.: +34 954488318
Fax: +34 954488300
http://ipts.jrc.ec.europa.eu/
http://www.jrc.ec.europa.eu/
Legal Notice
Neither the European Commission nor any person acting on behalf of the Commission
is responsible for the use which might be made of this publication.
A great deal of additional information on the European Union is available on the
Internet.
It can be accessed through the Europa server http://europa.eu/
JRC [PUBSY request]
Technical Note
Luxembourg: Publications Office of the European Union
© European Communities, 2012
Reproduction is authorised provided the source is acknowledged
Executive summary
Congestion is an important problem for road transport and a main challenge for transport
policy at all levels. The cost of road congestion in Europe is estimated to be equivalent to 1%
of GDP and its mitigation is the main priority of most infrastructure, traffic management and
road charging measures. Measuring congestion is however complicated due to the non-
uniform temporal and spatial distribution of traffic.
Efficiency in transport networks is a main priority for transport policy at EU level as
expressed through the European Commission's White Paper "Roadmap to a Single European
Transport Area – Towards a competitive and resource efficient transport system". The
existing links on the European road transport network vary significantly in terms of transport
volumes served and, consequently, reach different levels of capacity utilisation and
congestion. Statistics such as road counts are not widely available and are normally given on
an annual average daily traffic (AADT) basis. Such data do not allow a quantitative
estimation of congestion levels, which tend to concentrate on a limited numbers of nodes and
links during short time periods.
The methodology presented here allows to measure and monitor road congestion across
Europe using data from TomTom in-vehicle navigation systems. The approach is based on the
analysis of a large number of real vehicle speeds that have been measured on each road link
and the application of algorithms that allow the estimation of congestion indicators for
specific types of roads during selected time periods. The results include the detailed mapping
of recurrent congestion both geographically and temporally, as well as the comparison of the
quality of service of road networks between different zones.
The data used represent real speed measurements from in-vehicle navigation systems for
different time periods and days of the week. The large number of "probes" (over 1 trillion of
measurements) gives a highly accurate and representative picture of the actual driving
conditions across the European road network. The data collected are clustered in groups of
speed profiles which represent change in average speed behaviour along a road link in five-
minute time intervals over a 24-hour period. Each road link has a specific speed profile
assigned per day of the week. The average speed on a specific link during a certain time
period can be compared to the benchmark speed estimated for the link under free flow
conditions or against selected threshold values. As a result, indicators of congestion for
different time periods can be measured and compared across links, regions and countries. The
methodology can be useful for policy making in transport through various applications:
• Mapping congestion and monitoring its evolution over time, by comparing the level
and distribution of congestion in two different points in time
• Application of the congestion indicators in European transport policy, by comparing
average congestion between the peak hour and wider time periods and identifying
measures to improve the temporal distribution
• Combination of congestion indicators with traffic counts in order to improve speed
flow curves used in transport network models
The key conclusion from the analysis is that congestion mainly affects urban areas and a few
key bottlenecks in Europe. The results also suggest that the reason for congestion in many
cases is not a lack in capacity of road infrastructure, but rather an issue of demand
management.
3
Table of Contents
Executive summary............................................................................................................................................... 3
List of figures......................................................................................................................................................... 6
List of tables........................................................................................................................................................... 6
1. Introduction................................................................................................................................................ 7
2. Methodology ............................................................................................................................................... 7
3. Application for the European road network.......................................................................................... 12
4. Application for Spain............................................................................................................................... 20
5. Ways to fight congestion in Spain........................................................................................................... 26
6. Conclusions............................................................................................................................................... 30
7. References.................................................................................................................................... 31
5
6
List of figures
Figure 1: Typical speed profiles.......................................................................................8
Figure 2: Map of road congestion in Europe, indicator c(3h) for roads with free flow speed over 80km/h 10
Figure 3: Congestion indicator during different calculation periods (length of moving average), EU weighted
average................................................................................................................14
Figure 4: Inter-urban congestion in the EU, 1 hour peak period, average delays higher than 10 seconds per km,
free flow speed higher than 100 km/hour................................................................15
Figure 5: Local/urban congestion in the EU, 1 hour peak period, average delays higher than 45 seconds per km,
free flow speed below 50 km/hour.........................................................................16
Figure 6: Inter-urban congestion in Spain: Average delay per km (in seconds) during the 1-hour peak, for roads
with a free flow speed higher than 100 km/h...........................................................21
Figure 7: Urban congestion in Spain: Average delay per km (in seconds) during the 1-hour peak, for roads with a
free flow speed lower than 50 km/h.......................................................................22
Figure 8: Most congested roads in Spain: Average delay per km (in minutes), all types of roads 23
Figure 9: Most congested roads in Seville: Average delay per km (in minutes), all types of roads 24
Figure 10: Most congested roads in Madrid: Average delay per km (in minutes), all types of roads 25
List of tables
Table 1: Main congestion indicators per member state, for roads with free flow speed <50km/h, >80km/h and
>100km/h, 1h and 3h moving averages ..................................................................13
Table 2: Annual cost of congestion per EU member state, in absolute terms and as share of GDP 19
1. Introduction
Congestion is an important problem for road transport and a main challenge for transport
policy at all levels. The cost of road congestion in Europe is estimated to be over €110 billion
a year and its mitigation is the main priority of most infrastructure, traffic management and
road charging measures. Measuring congestion is however complicated due to the non-
uniform temporal and spatial distribution of traffic.
Efficiency in transport networks is a main priority for transport policy at EU level as
expressed through the European Commission's "Roadmap to a Single European Transport
Area – Towards a competitive and resource efficient transport system" [1]. The existing links
on the European road transport network vary significantly in terms of transport volumes
served and, consequently, reach different levels of capacity utilisation and congestion.
Statistics such as road counts are not widely available and are normally given on an annual
average daily traffic (AADT) basis. Such data do not allow a quantitative estimation of
congestion levels, which tend to concentrate on a limited numbers of nodes and links during
short time periods.
The methodology presented here allows to measure and monitor road congestion across
Europe using data from TomTom in-vehicle navigation systems. The approach is based on the
analysis of a large number of real vehicle speeds that have been measured on each road link
and the application of algorithms that allow the estimation of congestion indicators for
specific types of roads during selected time periods. The results include the detailed mapping
of recurrent congestion both geographically and temporally, as well as the comparison of the
quality of service of road networks between different zones.
2. Methodology
The data used represent real speed measurements from in-vehicle navigation systems for
different time periods and days of the week [2]. The large number of "probes" (over 1 trillion
of measurements) collected during 2008 and 2009 gives a highly accurate and representative
picture of the actual driving conditions across the European road network. The data collected
are clustered in groups of speed profiles which represent change in average speed behaviour
along a road link in five-minute time intervals over a 24-hour period [3, 4]. Each road link has
a specific speed profile assigned per day of the week. The average speed on a specific link
during a certain time period can be compared to the benchmark speed estimated for the link
under free flow conditions or against selected threshold values. As a result, indicators of
congestion for different time periods can be measured and compared across links, regions and
countries.
Average speed for a road link can be calculated at any time of the day by combining the free
flow speed of the link with the appropriate profile for the current time period. Figure 1 shows
a sample of typical speed profiles. The value in the y-axis of the profile is a percentage value
of the free flow speed for a road link, while the value on the x-axis represents the change in
time over a 24-hour period.
7
Congestion indicators
The average speed on a road link can be calculated as the moving average of probe speeds
over a period p:
,1ip
ip j
ji
v
p
+
=
=∑v [Eq. 1]
Where
p is the duration f the time period p
i is the start of the period
vj is the average probe speed in time slot j
04812 16 20 24
30%
40%
50%
60%
70%
80%
90%
100%
HOUR OF THE DAY
RELATIVE SPEED (% of free flow speed)
Figure 1: Typical speed profiles
Free flow speed is considered as the maximum measured average speed over the period:
0
[,]
max
f
ree j
jtt
v
∈
=v [Eq. 2]
while the period with the lowest average speed would be:
8
0,
[,]
min
p
itt
v
∈
=ip
v [Eq.3]
There are two main indicators of congestion that can be derived from the above. The ratio of
the average speed during the most congested period p with respect to the maximum speed:
p
p
f
ree
v
cv
= [Eq. 4]
and the average delay during period p expressed in minutes per km (or mile):
60
11
p
pfree
d
vv
=
−
[Eq. 5]
These indicators allow a direct mapping of recurrent road congestion in Europe, offering a
flexible framework of analysis. Modifying the period p can limit the estimation on specific
peak periods of e.g. 1 hour, or extend it to wider periods (e.g. 3, 6 hours or even a whole 24
hour period). Filtering in terms of the free flow speed can help focus on specific types of
roads such as highways/freeways or urban roads. Figure 2 presents an example of a map of
congestion in Europe, for the most congested 3-hour period in each link and for roads with
free flow speeds over 80 km/h.
9
Figure 2: Map of road congestion in Europe, indicator c(3h) for roads with free flow speed over
80km/h
10
Apart from the map itself, the data also allow the estimation of aggregate congestion levels
for each geographic zone, from country level to a local area.
The congestion indicator for a zone based on the ration of average speed to free flow speed
would be the weighted average:
,
[] ,
[]
()
rp
r
r zone rfree
vzone r
rzone
v
Lv
cL
∈
∈
=
∑
∑ [Eq. 6]
The congestion indicator for a zone using the average delay would also be a weighted average
of the form:
,
[]
[]
rrp
rzone
tzone r
r zone
Ld
cL
∈
∈
=
∑
∑ [Eq. 7]
where Lr is the length of each individual link
The measured speeds for each link can also be used as an indicator. The share of the length of
the network with an average speed within a certain range can be used to evaluate the overall
performance of the network:
min max
[]
,, min max
[]
,
rp
r zone
pv v p
r
r zone
Lv
Sforv
L
∈
∈
=<
∑
∑vv< [Eq. 8]
In a similar fashion, the share of the network for which average delays during a period are
within a certain range can give an indication of the distribution of congestion:
min max
[]
,, min ma
[]
,
pp
rp
r zone
pd d p p p
r
r zone
Ld
Wford
L
∈
∈
=<
∑
∑x
dd< [Eq. 9]
11
3. Application for the European road network
The selection of the suitable indicator to measure congestion depends on the specific
application and on the policy variable of interest. The formulation of the congestion indicators
depends on:
• The expression of delay used, ie ratio of measured to free speed or delay per km
• The time period for which congestion is measured, ie. the moving average used
• The type of road (urban, inter-urban, etc.) or the characteristics the road has (eg free
flow speed)
• The threshold values used to define congestion
• The geographic area for which the analysis is made
For transport policy at European level, an indicator at both member state and EU level is
probably the most relevant level of analysis. After testing various moving averages, the 1h
and 3h moving averages are the ones that give a reasonable representation of the duration of
peak periods. Although not accurately reflecting whether a road is urban or not, the free flow
speed can be a good proxy as regards its type and characteristics. Roads can be therefore
divided in three groups, one below 50km/h and one over 100km/h to represent local and inter-
urban links. A third group, roads with a free flow speed over 80km/h is also useful in order to
analyse inter-urban traffic in areas with few highways.
What is more relative, though, is the threshold value for which congestion is important. The
ratio of average to free flow speed is an easy to understand metric, especially when comparing
zones or types of roads between them or over time, but does not necessarily give a clear
indication on how important congestion is.
Average delay per km is an indicator with a simpler physical interpretation, since it allows a
direct estimate of total delay for a given trip. Measuring congestion in such a way also
highlights the fact that the impact on delays from a similar drop in average speed is much
more profound at lower speeds. Reducing the speed from 50 km/h to 40 km/h generates a
delay of 0.3 minutes per km, while a proportional 20% reduction from 100 km/h to 80 km/h
leads to a delay of only 0.15 km/h.
Table 1 summarizes the results for the main indicators at EU level. It is evident that
congestion is more pronounced at local level than when considering inter-urban links (defined
as the ones with a free flow speed of more than 100km/h). Comparing the real delays though
shows that in practice the delays in roads with higher speeds are far less important than in
local roads. It is also obvious that the moving average of 1h gives significantly higher
congestion levels than the one of 3 hours (figure 3 gives the whole range of moving averages
from 0.5 to 24 hours for the speed based congestion indicator, EU wide weighted average).
12
13
Table 1: Main congestion indicators per member state, for roads with free flow speed <50km/h,
>80km/h and >100km/h, 1h and 3h moving averages
Ratio of average to free flow speed (%) Average delay per km (seconds)
Free flow
speed
<50km/h
Free flow
speed
>80km/h
Free flow
speed >100
km/h
Free flow
speed
<50km/h
Free flow
speed
>80km/h
Free flow
speed >100
km/h
Country Mov
avg 1h Mov
avg 3h
Mo
v
avg
1h
Mov
avg
3h
Mov
avg
1h
Mov
avg
3h
Mov
avg
1h
Mov
avg
3h
Mov
avg
1h
Mov
avg
3h
Mov
avg
1h
Mov
avg
3h
Austria 86.7 88.2 92.8 94.1 93.9 95.1 15.7 13.7 3.0 2.4 2.0 1.6
Belgium 84.0 85.9 89.8 91.3 88.9 90.3 19.9 17.1 4.4 3.6 4.3 3.6
Czech
Republic 88.0 89.5 92.5 94.2 92.5 94.3 13.3 11.5 3.1 2.4 2.5 1.9
Germany 85.0 86.6
89.9 91.2 89.9 90.9 17.3 15.3 4.2 3.6 3.5 3.1
Denmark 85.9 87.7 92.5 93.9 92.4 94.2 17.1 14.5 3.2 2.6 2.8 2.0
Spain 89.5 91.2 92.5 94.1 92.0 93.7 12.7 10.4 2.9 2.2 2.8 2.2
Estonia 87.3 88.9 95.6 96.8 93.5 95.2 14.6 12.5 1.9 1.3 2.5 1.8
Finland 87.4 88.9 94.3 95.5 94.4 95.6 14.5 12.6 2.4 1.9 2.0 1.6
France 87.6 89.1 93.0 94.3 92.7 94.3 15.2 13.1 2.9 2.3 2.5 1.9
United
Kingdom 80.9 82.7
90.4 92.2 90.2 92.4 24.3 21.5 4.3 3.4 3.9 2.9
Hungary 82.2 84.0
91.3 92.9 91.0 92.8 22.1 19.3 3.5 2.8 3.1 2.4
Ireland 80.7 82.1
93.1 94.3 94.7 96.0 26.7 24.4 3.1 2.5 1.9 1.4
Italy 87.7 89.3 93.2 94.5 93.8 95.0 15.0 12.8 2.8 2.3 2.2 1.7
Lithuania 84.0 85.7 93.8 95.5 93.4 95.3 19.6 17.2 2.5 1.8 2.3 1.6
Luxem-
bourg 85.8 87.7
87.3 89.6 84.4 87.0 16.8 14.1 5.7 4.4 6.6 5.1
Nether-
lands 85.6 87.2
88.2 90.0 86.6 88.7 17.8 15.6 5.4 4.3 5.6 4.5
Poland 82.2 83.6
92.9 94.4 93.2 94.7 23.1 20.9 3.1 2.4 2.4 1.8
Portugal 88.7 90.3 93.3 94.8 93.3 95.0 13.3 11.2 2.6 2.0 2.3 1.7
Slovakia 84.9 86.5 91.6 93.2 91.3 93.1 17.5 15.3 3.6 2.8 3.0 2.3
Sweden 86.5 87.9 94.3 95.4 94.3 95.4 15.8 13.9 2.4 1.9 2.1 1.6
EU
weighted
average 86.3 87.9
92.2 93.6 91.8 93.4 16.6 14.4 3.3 2.6 3.0 2.3
The congestion indicators are a relative measure and the use of thresholds is necessary in
order for them to be used for the actual evaluation of the situation as regards congestion.
Table 1 allows the comparison between member states:
• Seven member states have a 1h speed based congestion indicator below 85% for local
roads, four of which remaining below that threshold even for the 3 hour most
congested period. All four have average delays of more than 20 seconds per km during
the 1 hour peak, while three of them still do so during the 3 hour peak. By most
measures Ireland, United Kingdom, Poland and Hungary are identified as the member
states where local/urban congestion is highest.
• No member state presents similar congestion levels for the inter-urban links/highways
(roads with free flow speed of more than 100km). Four member states have a speed
based congestion indicator below 90% for the 1 hour peak period, while two of them
remain below 90% during the 3 hour peak period as well. The average delays per km
are rather low however, the maximum being 6.6 seconds per km. Luxembourg and the
Netherlands demonstrate the highest inter-urban congestion levels. Data for Belgium,
Germany and United Kingdom suggest that inter-urban congestion may be an issue.
84
86
88
90
92
94
96
98
0 6 12 18 24
length of moving average period (hours)
measured speed as share of free flow speed (%)
free flow speed
>=100km/h
free flow speed
<=
50km/h
Figure 3: Congestion indicator during different calculation periods (length of moving average),
EU weighted average
Mapping the average delays on the road network across Europe gives a similar message. The
interurban network (Figure 4) presents a limited number of links with a delay of over 10
seconds per km1, mainly located in the Netherlands, parts of Belgium, Luxembourg, inter-
urban connection in the United Kingdom, parts of Germany, around main population centres
(Paris, Madrid, Rome, Milano) and on the links between Copenhagen and the rest of Denmark
and with Sweden.
The map for local road congestion (Figure 5) suggests that it is spread more uniformly across
the EU and is highly dependent on urban densities and the quality and capacity of
infrastructure. A large number of links with an average delay of 45 seconds per km is visible2,
corresponding to urban road in most of the medium and big cities across the EU.
1 Setting 10 seconds per km as a threshold allows a comprehensive coverage of congested links at inter-urban level. Lowering
the thresholds results in blurring the map with too many links appearing, while raising the thresholds results in a limited
number of links appearing, not sufficient to draw conclusions.
2 The threshold here is 45 seconds per km for practical reasons, in order to obtain a readable map. Obviously the resulting
map cannot be directly comparable to the one on inter-urban links, it only serves to identify where urban congestion is most
likely to occur.
14
The key conclusion from the analysis is that congestion mainly affects urban areas and a few
key bottlenecks in Europe. The results also suggest that the reason for congestion in many
cases is not a lack in capacity of road infrastructure, but rather an issue of demand
management. Average delays are important during peak periods, but since demand is not
spread uniformly during the day, a large part of road capacity is underutilised outside the one
or two peaks of 3 to 6 hours a day.
Figure 4: Inter-urban congestion in the EU, 1 hour peak period, average delays higher than 10
seconds per km, free flow speed higher than 100 km/hour
15
Figure 5: Local/urban congestion in the EU, 1 hour peak period, average delays higher than 45
seconds per km, free flow speed below 50 km/hour
16
Table 2: Share of road network for each average speed level
Total length of network for
which data is available
(kms)
Roads with
average speed
below 50km/h
Roads with
average between
80-and 100 km/h
Roads with
average speed
above 100km/h
Austria 27907 16.9% 19.9% 12.9%
Belgium 29984 22.5% 11.0% 12.8%
Czech
Republic 25158 13.7% 21.1% 9.0%
Germany 225605 19.4% 21.0% 11.6%
Denmark 18762 9.9% 35.5% 11.7%
Spain 112945 16.5% 22.4% 28.6%
Estonia 5178 5.5% 65.9% 2.9%
Finland 37002 5.7% 52.7% 6.1%
France 274426 18.0% 24.4% 10.7%
United
Kingdom 94155 17.7% 18.2% 13.9%
Hungary 9208 7.1% 39.3% 22.9%
Ireland 13567 9.6% 23.1% 10.0%
Italy 129103 24.7% 10.6% 12.9%
Lithuania 5793 5.6% 57.9% 12.8%
Luxembourg 2456 20.7% 11.3% 10.3%
Netherlands 36840 25.8% 11.9% 13.0%
Poland 49358 8.3% 38.8% 6.2%
Portugal 28559 19.1% 16.1% 19.5%
Slovakia 6063 6.1% 33.4% 15.2%
Sweden 43098 7.8% 45.5% 13.0%
Total EU
(available
countries) 1175167 17.5% 23.3% 13.3%
Table 2 summarizes the results for the indicator of Eq.8, the share of the road network in each
EU member state that has an average measured speed of below 50 km/h, between 80 and 100
km/h and over 100 km/h. The differences among member states reflect the local conditions as
regards the quality of the road network infrastructure (e.g. the large share of highways in
Spain results in more than 28% of the network having an average speed of more than 100
km/h) or speed limits (e.g. in Finland where only 6% is higher than 100 km/h, but more than
half the network has an average speed between 80 and 100 km/h).
Table 3 shows the results for the indicator of Eq.9, the share of the network with average
delays within a certain range, ranked according to the share of roads with an average delay of
more than 10 seconds per km. This indicator gives a clearer picture of overall congestion,
since it is based on the difference between the average and the free flow speeds.
17
Table 3: Congestion classification: Average delay per km during 1 hour peak period (share of
total road network)
Average delay per km (seconds)
Higher
than 20 Higher
than 10
1 to 5 5 to 10 10 to 20
United
Kingdom 48.2% 25.7% 11.1% 8.8% 19.9%
Belgium 42.7% 35.1% 12.6% 6.4% 19.1%
Netherlands 46.3% 32.0% 11.6% 6.4% 18.0%
Luxembourg 44.5% 36.2% 9.6% 5.8% 15.3%
Germany 46.7% 36.8% 9.5% 4.3% 13.8%
Italy 50.7% 25.2% 7.9% 4.7% 12.6%
Hungary 65.7% 19.0% 7.3% 4.1% 11.4%
Poland 60.8% 21.7% 6.4% 4.5% 10.9%
Slovakia 57.8% 26.6% 7.6% 2.6% 10.2%
Ireland 61.8% 18.7% 5.2% 4.1% 9.3%
Czech
Republic 52.8% 28.0% 6.3% 2.5% 8.8%
Austria 55.7% 28.4% 5.8% 2.7% 8.5%
France 61.1% 19.4% 5.3% 2.7% 7.9%
Portugal 57.3% 21.0% 5.5% 2.3% 7.9%
Denmark 62.8% 20.9% 5.2% 2.3% 7.5%
Sweden 70.7% 13.6% 3.5% 1.5% 5.0%
Spain 68.2% 16.8% 3.7% 1.2% 4.9%
Lithuania 78.6% 9.4% 1.9% 1.7% 3.6%
Estonia 74.4% 8.3% 1.9% 1.2% 3.2%
Finland 74.8% 13.4% 2.1% 0.8% 2.9%
The data available allow the quantification of average delays per km for specific roads and
across the road network of wider zones. The indicators presented here measure the share of
the network that is congested during various time periods, but do not give a precise picture of
the share of traffic that is on the road during congested periods. The distribution of traffic
volume is not uniform either during a specific time period, or among the links of a road
network. There is a clear correlation between distribution of traffic volumes and congested
network links, but the relationship is not linear and does not allow a direct transformation.
The estimation of the total congestion delays, for all trips, requires an additional step that
takes into account the distribution of trips during each time period and each type of network:
,, ,,
,
tt
zone zone p w zone p w
pw
DcV
∀
=∑ [Eq. 9]
where the total delay for all trips in the zone, the weighted average of delay
per unit of length for each time period p and type of network w for the specific zone, and
the total traffic for the specific combination of period, network type and zone (in
vehicles or passenger or tonnes x length).
tzone
D
w
,,
tzone p w
c
,,zone p
V
18
Since data on the distribution of traffic volumes at this level is not available, an
approximation is made using results of the TRANSTOOLS model v 2.5 [5, 6]. For the term
the results of the model as regards the share of total traffic during the various peak
and off-peak periods for each of the three types of network (<50km/h, >80km/h and
>100km/h) is used. These shares are then applied of the total transport activity for 2009 in
each country as published by Eurostat [
,,zone p w
V
7]. The term corresponds to the figures in
,,
tzone p w
c
Table 4 for average delay per km.
The estimate of the total delay for each zone can also be used as a basis for the estimation of
the cost of congestion. Applying the time values proposed by HEATCO [6], adjusted for
inflation, produces the results that are summarised in Table 2. On average, road congestion
costs for passenger and freight transport represent 1% of GDP, with important variations
among EU member states.
Table 4: Annual cost of congestion per EU member state, in absolute terms and as share of
GDP Annual
cost of
congestion
(€ billion)
Cost of
congestion
as % of
GDP 2009
Austria 1.8 0.6%
Belgium 3.4 1.0%
Czech
Republic 0.8 0.6%
Germany 24.2 1.0%
Denmark 1.5 0.7%
Spain 5.5 0.5%
Estonia 0.1 0.8%
Finland 1.4 0.8%
France 16.5 0.9%
United
Kingdom 24.5 1.6%
Hungary 0.7 0.8%
Ireland 1.8 1.1%
Italy 14.6 1.0%
Lithuania 0.5 1.7%
Luxembourg 0.3 0.7%
Netherlands 4.7 0.8%
Poland 4.8 1.6%
Portugal 1.2 0.7%
Slovakia 0.3 0.5%
Sweden 2.6 0.9%
Total EU
(available
countries) 111.3 1.0%
19
4. Application for Spain
Applying the methodology described above for Spain, some interesting conclusions can be
drawn. The road network in Spain is less congested than in the EU on average, mainly due to
the lower population density and the better quality of the road network compared to other EU
member states.
• The average delay on Spanish urban roads is 12.7 seconds per km during peak periods,
compared to 16.6 seconds per km in the EU as a whole.
• For inter-urban roads congestion levels are much lower, on average 2.9 seconds per
km during peak periods, compared to 3.3 seconds per km in the EU as a whole.
• In terms of quality, the Spanish road network is among the best in the EU, with 28.6%
of the network allowing travel at an average speed higher than 100 km/h.
• Congestion is, however, still a main problem in Spain. About 5% of the network
suffers delays of more than 10 seconds per km, concentrated mainly in Madrid,
Barcelona and other main cities.
• Drivers and passengers in Spain spend more than 420 million hours a year in
congestion, with an estimated annual cost of more than € 5.5 billion.
Figure 6 presents the most congested points of the urban road network, while the main points
where inter-urban road congestion is concentrated can be seen in
Figure 7. Figure 8 gives an overview of congestion across the whole road network.
Focusing on specific cities, the maps for Sevilla (Figure 9) and Madrid (Figure 10) identify
the main bottlenecks of the respective urban road networks. In the case of Sevilla, congestion
is concentrated on the main points of entrance to the city centre, with highest delay levels in
the Tablada area (where construction works caused significant delays during 2009). In
Madrid, congestion is spread throughout the city centre, including the main ring roads (M-30,
M-35 and M-40) and the links with the inter-urban network.
20
Figure 6: Urban congestion in Spain: Average delay per km (in seconds) during the 1-hour
peak, for roads with a free flow speed lower than 50 km/h
21
Figure 7: Inter-urban congestion in Spain: Average delay per km (in seconds) during the 1-hour
peak, for roads with a free flow speed higher than 100 km/h
22
Figure 8: Most congested roads in Spain: Average delay per km (in minutes), all types of roads
23
Figure 9: Most congested roads in Seville: Average delay per km (in minutes), all types of
roads
24
Figure 10: Most congested roads in Madrid: Average delay per km (in minutes), all types of
roads
25
5. Ways to fight congestion in Spain
The results for Spain suggest that congestion is a serious issue for urban areas. Reducing
congestion requires a combination of policy measures that can modify the patterns of demand
and/ or increase the efficiency of the use of road infrastructure. Increasing or improving road
infrastructure itself is probably a counter-productive approach: apart from the lack of space
and the high costs it would entail, the existing situation in Spain in terms of both quantity and
quality of roads is very good, compared to other EU member states.
The main options to combat congestion seem to be on the demand management side, by either
motivating users to change the time or route of their trip, or by re-directing them to alternative
transport modes, especially public transport (bus, tram, metro and train) or slow modes
(bicycle, walking).
A particularly successful way of changing urban travel patterns is congestion pricing, a
measure successfully applied in London, Stockholm and other European cities. The concept
behind congestion pricing is for the users to pay a toll if they use a congested part of the
network during a certain (peak) period. Ideally, the total amount of charges for congestion
should be subtracted from the total taxation for road transport (vehicle or fuel taxes). This
way, on aggregate level the measure would be budget neutral but at individual level users
avoiding congestion would face a lower tax burden. Such measures normally result in a re-
distribution of traffic in time and space that leads to a "peak shaving" of demand and
significantly lower levels of congestion. Figure 11 gives an example of how a congestion
charge can lead to a more uniform distribution of traffic levels during a day.
0
1000
2000
3000
4000
4 7 10 13 16 19 22
time
traffic volume (vehicles per hour
)
before
after
Figure 11: Example of peak shaving as a result of congestion charging
26
The other main policy measure is related to making public transport and other alternative
modes more attractive. Particularly in the case of Spain, there seems to be significant margin
for improvement. The responses to the Eurobarometer survey on the future of transport [8]
highlight several differences among EU member states, but also help in identifying possible
measures that can decrease road congestion and improve transport in general.
The majority of respondents in Spain (53%) and the EU as a whole (50%) express a
favourable opinions as regards a replacement of existing car charges such as registration and
circulation taxes with charging schemes that take into account the actual use of the car such as
the kilometres driven, or the use of it in peak hours.
The survey also identifies possible improvements in public transport (Figure 12). The lack of
connections and the low frequency of service are the main reasons for not using public
transport. Respondents in Spain highlighted these reasons more frequently than in the rest of
the EU, something that may imply the need for an increase in the supply of public transport
services in Spanish cities. The lack of information on schedules is a more important factor in
Spain, with more than half of the respondents identifying it as a main obstacle for more
frequent use of public transport. Cost seems to worry users in Spain at comparable levels as in
the rest of the EU, but security concerns are significantly higher, probably as a result of the
relatively recent terrorist attacks in Madrid. Spanish public transport seems to compare
positively as regards reliability, which was identified as the less important reason for not
using public transport in Spain, but was still important in the other member states.
0 20406080
Lack of reliability
Security concerns
Too expensive
Lack of information on
schedules
Low frequency of
service
Lack of connections
Spain
EU-27
Figure 12: Euro-barometer survey: Reasons for not using public transport (Very important +
rather important)
27
Other options to increase the use of public transport are related to technological and
organisational solutions. Using a single ticket for all public transport modes in a journey is
considered as a good measure by most respondents (Figure 13).
Better (online) information on schedules and the possibility to by tickets online are
technological options that can certainly encourage the use of transport modes other than car
(Figure 14). Respondents in Spain identified the need for such solutions more frequently than
in other member states, possibly due to the more limited availability of such options. The
need for more attractive terminals and better transfers between modes were also identified as
important more frequently than in the rest of the EU.
0 20406080
DK/DA
No
Yes, maybe
Yes, definitely
Spain
EU-27
Figure 13: Euro-barometer survey: Would you consider using public transport more frequently
if it were possible to buy a single ticket covering all possible transport modes (such as bus,
train or tram) for your journey?
28
0 20406080
Possibility to buy
tickets online
Easy transfer from one
transport mode to
another
Attractive terminals
Better (online)
information on
schedules
Spain
EU-27
Figure 14: Euro-barometer survey: options that encourage the use of different modes of
transport instead of car
29
6. Conclusions
The methodology presented here allows the measurement and mapping of congestion levels
across the network and the comparison between different zones and countries. The results and
maps that were presented give a picture of the existing situation and allow the quantification
of congestion levels and of the costs for the users. The approach can be also applied to
monitor the evolution of congestion over time by comparing indicators for the same road link
or area over time. The results can be useful for policy makers in several ways:
• Mapping congestion and monitoring its evolution over time, by comparing the level
and distribution of congestion in two different points in time. This would allow the
identification of bottlenecks in the road network, the exploration of measures to avoid
congestion and the monitoring of the progress.
• Application of the congestion indicators in European transport policy, by comparing
average congestion between the peak hour and wider time periods and identifying
measures to improve the temporal distribution. Such analysis can help in identifying
whether and how demand management measures, as well as in estimating congestion
charging levels and periods of application.
• Combination of congestion indicators with traffic counts in order to improve speed
flow curves used in transport network models. Apart from allowing the monitoring of
traffic flows, this combination can also be useful for modelling the future evolution of
congestion and help anticipate potential new bottlenecks.
30
31
7. References
1. European Commission,
White Paper: Roadmap to a Single European Transport Area
– Towards a competitive and resource efficient transport system, COM(2011) 144
final. 2011: Brussels.
2. TomTom,
White Paper: How TomTom’s HD Traffic™ and IQ Routes™ data provides
the very best routing: Travel Time Measurements using GSM and GPS Probe Data.
2009.
3. TeleAtlas (2008)
White Paper: Developing Mapping Applications with Historical
Speed Profile Data.
4. Schafer, R.P.,
IQ Routes and HD Traffic - Technology Insights about TomTom's Time-
Dynamic Navigation Concept. 7th Joint Meeting of the European Software
Engineering Conference and the Acm Sigsoft Symposium on the Foundations of
Software Engineering, 2009: p. 171-172.
5. Burgess, A. and O. Nielsen. European Transtools Transport Model. in Transportation
Research Board 87th Annual Meeting 08-0262. 2008. Washington D.C.:
Transportation Research Board.
6. HEATCO project,
Developing Harmonised European Approaches for Transport
Costing and Project Assessment, Del 5 Proposal for Harmonised Guidelines. Contract
No. FP6-2002-SSP-1/502481 for European Commission, DG TREN. 2005.
7. European Commission,
EU transport in figures, Statistical Pocketbook 2011. 2011.
8. European Commission,
Flash Eurobarometer 312, Future of Transport, Analytical
Report. 2011.
European Commission
JRC Pubsy Nr. – Joint Research Centre – Institute for Prospective Technological Studies
Title: Measuring road congestion
Authors: Panayotis Christidis, J. Nicolás Ibañez Rivas
Luxembourg: Publications Office of the European Union
2012
Technical Note
Abstract
The methodology presented here allows to measure and monitor road congestion across Europe using data from TomTom
in-vehicle navigation systems. The approach is based on the analysis of a large number of real vehicle speeds that have
been measured on each road link and the application of algorithms that allow the estimation of congestion indicators for
specific types of roads during selected time periods. The results include the detailed mapping of recurrent congestion both
geographically and temporally, as well as the comparison of the quality of service of road networks between different
zones.
The mission of the Joint Research Centre is to provide customer-driven scientific an
d
technical support for the conception, development, implementation and monitoring o
f
European Union policies. As a service of the European Commission, the Joint Research
Centre functions as a reference centre of science and technology for the Union. Close to the
policy-making process, it serves the common interest of the Member States, while being
independent of special interests, whether private or national.
