DOCTORAL THESIS - oa.upm.esoa.upm.es/373/1/ELENA_LOPEZ_SUAREZ.pdf · sus siempre acertados...

UNIVERSIDAD POLITÉCNICA DE MADRID ESCUELA TÉCNICA SUPERIOR

DE INGENIEROS DE CAMINOS, CANALES Y PUERTOS

ASSESSMENT OF TRANSPORT INFRASTRUCTURE PLANS: A STRATEGIC APPROACH INTEGRATING EFFICIENCY, COHESION AND ENVIRONMENTAL

ASPECTS

DOCTORAL THESIS

Elena López Suárez Ingeniero de Caminos, Canales y Puertos

Madrid, 2007

DEPARTAMENTO DE INGENIERÍA CIVIL: TRANSPORTES

Escuela Técnica Superior de Ingenieros de Caminos, Canales y Puertos

ASSESSMENT OF TRANSPORT INFRASTRUCTURE PLANS: A STRATEGIC APPROACH INTEGRATING EFFICIENCY,

COHESION AND ENVIRONMENTAL ASPECTS

DOCTORAL THESIS

Elena López Suárez Ingeniero de Caminos, Canales y Puertos

Director:

Andrés Monzón de Cáceres Dr. Ingeniero de Caminos, Canales y Puertos

Madrid, 2007

Tribunal nombrado por el Mgfco. y Excmo. Sr. Rector de la Universidad Politécnica de Madrid, el día ___ de _______________ de 2007.

Presidente: _____________________________________________ Vocal: _____________________________________________ Vocal: _____________________________________________ Vocal: _____________________________________________ Secretario: _____________________________________________ Realizado el acto de defensa y lectura de la Tesis el día ___ de _______________ de 2007 en la E.T.S. de Ingenieros de Caminos, Canales y Puertos de la U.P.M. Calificación: ______________________________ EL PRESIDENTE LOS VOCALES

EL SECRETARIO

A mis padres, Micaela y Sebastián,

mis raíces, mis maestros

‘El hombre, en su centro, es siempre potencialmente

un hombre docto, un sabio y un maestro’

KALFRIED DÜRCKHEIM

ABSTRACT

During the last few decades there has been a shift in transport planning objectives

from economic efficiency towards strategic policy goals, such as cohesion or

environmental issues, intimately linked with the ‘sustainable transport’ paradigm.

However, the treatment of these strategic aspects is uneven and scarce among

assessment methodologies. The development of harmonized methodologies for the

strategic assessment of large scale transport infrastructure investments, such as

transport infrastructure Plans, is therefore a current challenge for the research

community.

This doctoral thesis addresses this challenge by presenting a methodology for the

assessment of transport infrastructure Plans. The proposed methodology

constitutes a strategic approach, based on the utilisation of spatial impact analysis

tools supported by a Geographical Information System (GIS). The assessment

criteria, based on the ‘sustainable transport’ paradigm, are structured into

efficiency, cohesion and environmental criteria. The procedure selected for the

integration of the assessment criteria results follows a multicriteria analysis

approach.

The suggested methodology defines a comprehensive technical procedure for the

assessment of strategic effects of transport infrastructure Plans, which is believed

to constitute a useful, transparent and flexible planning tool both for planners and

decision-makers.

The validity of the methodology is tested with its application to a case study: the

Spanish Strategic Transport and Infrastructure Plan 2005-2020 (PEIT).

RESUMEN

En las últimas décadas se viene produciendo un cambio en los objetivos que dirigen

las labores de planificación de infraestructuras de transporte, desde la eficiencia

económica hacia objetivos de carácter más estratégico, como la cohesión o los

aspectos medioambientales. Sin embargo, no existe un consenso sobre la forma en

que se deben incluir estos aspectos estratégicos en las metodologías de evaluación

oficiales, sobre todo en las que se refieren a inversiones a gran escala, como es el

caso de los Planes de infraestructura de transporte.

Esta tesis doctoral avanza en esta línea de investigación mediante la propuesta de

una metodología para la evaluación de Planes de infraestructura de transporte. La

metodología sigue un enfoque estratégico, basado en la utilización de herramientas

de análisis territorial aplicadas sobre un soporte SIG (Sistema de Información

Geográfica). Los objetivos de evaluación, basados en el paradigma del ‘transporte

sostenible’, se han estructurado en torno a criterios de eficiencia, cohesión y

medioambientales. Para su integración se ha seleccionado un método de evaluación

multicriterio.

La metodología propuesta define un procedimiento de evaluación que constituye

una herramienta útil en las labores de planificación de infraestructuras, permitiendo

la interacción entre planificadores como para decisores, así como un instrumento de

apoyo para la comunicación de resultados a la opinión pública, gracias a la cuidada

representación gráfica de resultados.

La validez de la metodología ha sido comprobada mediante su aplicación a un caso

de estudio: el Plan Estratégico de Infraestructuras y Transporte 2005-2020 (PEIT)

español.

ACKNOWLEDGMENTS

I would like to start by thanking Andrés Monzón, my thesis supervisor, for the

valuable and constant support he has given me these past four years. His

confidence in my work during difficult times has been very important help for me to

finish the thesis and my studies.

From the Transport Department and from TRANSyT-UPM I would like to thank the

teaching staff, especially Rafael Izquierdo, Aniceto Zaragoza, Oscar Martínez, José

Manuel Vassallo and the Transport Department Director, Miguel Ángel del Val. They

have all encouraged me and shared their experience with me from the first day at

the University. I also want to thank Javier Gutiérrez Puebla, from UCM, for his wise

comments and suggestions, which have served me as an invaluable guide during

the development of the research work. I also would like to thank Lawrence Baron

for his meticulous work in editing my thesis without loosing his enthusiasm and

smile.

My colleagues at TRANSyT-UPM have been there when I needed them, day after

day. Firstly, I want to thank Emilio Ortega and Belén Martín for their help in the

preparation of the maps and Santiago Mancebo for his wise comments. I also want

to specially thank Paula Vieira, Rocío Cascajo, Esther Madrigal, Mª Eugenia López,

Ana María Pardeiro, Paul Pffafenbichler, Daniel de la Hoz and Carmen Pérez. Thank

you for all the help you have given me.

Many other people have given me their support during my weak moments; I am

very lucky to have been able to depend on them during all this time. Thanks are

due to Manuel, Concha. Fernando, Cristina, Pepe, Pilar, Jose, Miren, Marta, Marieta,

Sara, Patricia, and many others: thank you for the right words and the good

gestures.

Finally, a big GRACIAS to my family. To my grandparents, Rosa and Eugenio, who

have given me serenity when I needed it most. To my brother Chano, thanks for

your advice, mi niño! And of course, to my parents, Micaela and Sebastián, for

teaching me how to get the best of myself. Thank you for showing me so much

love. For being there. Always.

AGRADECIMIENTOS

En primer lugar, quiero agradecer a mi Director de tesis, Andrés Monzón, el

respaldo decidido y constante que me ha ofrecido durante estos años. Su apuesta

por mi trabajo en los momentos difíciles ha sido muy importante para que haya

podido terminar esta tesis.

Quiero expresar también mi agradecimiento al Departamento de Transportes y a

TRANSyT-UPM, en particular a Rafael Izquierdo, Aniceto Zaragoza, Oscar Martínez y

José Manuel Vassallo, y al Director del Departamento, Miguel Ángel del Val. Todos

ellos me han infundo ánimos y me han aconsejado desde el primer día, desde la

serenidad de su experiencia. Quiero agradecer también a Javier Gutiérrez Puebla

sus siempre acertados comentarios y sugerencias, que me han servido de

inestimable guía durante el desarrollo de la investigación. Debo agradecer también

a Lawrence Baron el haberse encargado de la minuciosa tarea de edición del inglés

del texto, sin perder nunca el entusiasmo ni la sonrisa.

Mis compañeros de TRANSyT-UPM son los que me han acompañado en el día a día.

En primer lugar quiero agradecer a Emilio Ortega y a Belén Martín su gran ayuda en

la elaboración de los mapas y a Santiago Mancebo sus certeros comentarios. Quiero

dar las gracias de forma especial a Paula Vieira, Rocío Cascajo, Esther Madrigal, Mª

Eugenia López, Ana María Pardeiro, Paul Pffafenbichler, Daniel de la Hoz y Carmen

Pérez. Compañeros, gracias a todos por el enorme cariño que me han demostrado

en este tiempo.

He tenido la suerte de contar con gente que me ha dado aliento cuando me fallaban

las fuerzas. Gracias a Manuel y a Concha, maestros en el camino. Fernando,

Cristina, Pepe, Pilar, Jose, Miren, Marta, Marieta, Sara, Patricia, y tantos otros:

gracias por ayudarme con la palabra y el gesto apropiados en cada momento.

Por último, un GRACIAS a mi familia. A mis abuelos Rosa y Eugenio, que me han

dado serenidad cuando más la he necesitado. A mi hermano Chano: gracias por tus

consejos, mi niño!. Y por supuesto, a mis padres, Micaela y Sebastián, por

enseñarme a dar lo mejor de mí misma. Gracias por demostrarme tanto amor. Por

estar ahí. Siempre.

TABLE OF CONTENTS

-vii-

TABLE OF CONTENTS

DEDICATION……………………………………………………………………………………….i

ABSTRACT………………………………………………………………………………………...iii

ACKNOWLEDGMENTS………………………………………………………………………….v

CONTENTS……………………………………………………………………………………….vii

LIST OF TABLES..……………………………………………………………………………….x

LIST OF FIGURES………………………………………………………………………………xi

LIST OF ABBREVIATIONS…………………………………………………………………..xv

CONTENTS

1. INTRODUCTION ........................................................................ 1

1.1 Problem statement ................................................................... 1

1.2 Objectives................................................................................. 3

1.3 Research methodology.............................................................. 3

1.4 Structure of the thesis .............................................................. 5

2. A CHANGING PLANNING FRAMEWORK...................................... 7

2.1 Introduction.............................................................................. 7

2.2 Structuring the planning process .............................................. 9

2.2.1 Sources of conflicts in objective setting........................................... 9

2.2.2 A guiding principle: the sustainable development approach ............... 9

2.2.3 EU policy objectives ....................................................................13

2.3 The evaluation approach......................................................... 16

2.3.1 Introduction ...............................................................................16

2.3.2 Outline of an evaluation process ...................................................17

2.3.3 Current state of the practice in Europe ..........................................22

2.4 The role of evaluation in decision-making............................... 30

2.5 Conclusions............................................................................. 33

3. SPATIAL IMPACT ANALYSIS TOOLS ........................................ 35

3.1 Spatial impacts at the Plan level ............................................. 35

3.1.1 Theoretical foundations of spatial impact analysis ...........................35

TABLE OF CONTENTS

-viii-

3.1.2 Impact analysis at the Plan level...................................................36

3.1.3 The treatment of wider policy impacts at the Plan level....................38

3.2 The potential of accessibility analysis..................................... 43

3.2.1 The concept of accessibility ..........................................................43

3.2.2 The measurement of accessibility .................................................45

3.2.3 Applications in transport planning .................................................53

3.3 Spatial impact and GIS ........................................................... 61

3.3.1 GIS background..........................................................................61

3.3.2 Applications of GIS in transport planning .......................................63

3.4 Conclusions ............................................................................ 66

4. METHODOLOGY FOR THE ASSESSMENT OF TRANSPORT

INFRASTRUCTURE PLANS ....................................................... 69

4.1 Structure of the methodology ................................................. 69

4.2 Definition of the assessment framework ................................ 71

4.2.1 Assessment time horizon .............................................................71

4.2.2 Delimitation of the study area ......................................................72

4.3 Definition of assessment criteria ............................................ 72

4.3.1 Efficiency ...................................................................................73

4.3.2 Cohesion ...................................................................................73

4.3.3 Environmental sustainability.........................................................74

4.4 Definition of performance indicators ...................................... 75

4.4.1 Efficiency ...................................................................................76

4.4.2 Cohesion ...................................................................................78

4.4.3 Environmental sustainability.........................................................81

4.5 Integration ............................................................................. 84

4.5.1 Outline of the proposed approach .................................................84

4.5.2 Weight estimation.......................................................................85

4.5.3 Utility functions ..........................................................................86

4.6 Sensitivity analysis ................................................................. 86

4.6.1 Weight sensitivity .......................................................................87

4.6.2 Attribute value sensitivity ............................................................87

5. CASE STUDY DESCRIPTION..................................................... 87

5.1 Introduction ........................................................................... 87

5.2 Case study characterization.................................................... 88

5.2.1 The surface transport infrastructure networks ................................89

5.2.2 The socio-economic system..........................................................90

5.2.3 Current challenges of the Spanish transport system ........................95

TABLE OF CONTENTS

-ix-

5.2.4 The Strategic Infrastructure and Transport Plan 2005-2020

(PEIT) .....................................................................................100

5.3 The assessment framework .................................................. 101

5.3.1 Assessment time horizon and delimitation of the study area...........101

5.3.2 Definition of alternatives............................................................101

5.3.3 Generation of the GIS database..................................................104

6. ASSESSMENT RESULTS.......................................................... 111

6.1 Efficiency .............................................................................. 111

6.1.1 Network efficiency (NE) .............................................................111

6.1.2 Cross-border integration (CB).....................................................121

6.2 Cohesion ............................................................................... 129

6.2.1 Regional cohesion (RC)..............................................................129

6.2.2 Social cohesion (SC) .................................................................140

6.3 Environmental sustainability ................................................ 149

6.3.1 Global warming (GW) ................................................................149

6.3.2 Habitat fragmentation (HF) ........................................................153

6.4 Discussion on performance indicator results ........................ 156

6.4.1 Road mode ..............................................................................156

6.4.2 Rail mode ................................................................................158

6.5 Integration of results............................................................ 159

6.5.1 Description of the simplified integration procedure ........................159

6.5.2 Road mode ..............................................................................160

6.5.3 Rail mode ................................................................................162

6.5.4 Sensitivity analysis....................................................................164

7. CONCLUSIONS, CONTRIBUTIONS AND FUTURE RESEARCH ... 169

7.1 Conclusions........................................................................... 169

7.1.1 Literature review ......................................................................169

7.1.2 Methodological approach............................................................170

7.1.3 Case study application...............................................................171

7.1.4 Recommendations from a transport planning perspective...............173

7.2 Contributions ........................................................................ 175

7.3 Recommendations for future research .................................. 176

8. REFERENCES ......................................................................... 179

APPENDICES:

APPENDIX A: DEFINITION OF CRITERIA WEIGHTS…………………………..............……205

APPENDIX B: CASE STUDY APPLICATION OF THE ACCESSIBILITY MODEL…………209

TABLE OF CONTENTS

-x-

LIST OF TABLES

Table 2.1: Consideration of TEN-T territorial goals suggested in the UTS study ....26

Table 2.2: Accessibility categories (left) and evaluation matrix for distribution and

development objectives (right) of the German procedure ............................29

Table 4.1: Assessment criteria .......................................................................73

Table 4.2: Assessment criteria and performance indicators ................................76

Table 4.3: Weighting factor matrix for the cohesion criterion .............................80

Table 4.4: Structural backwardness categories.................................................80

Table 4.5: Accessibility analysis categories ......................................................80

Table 4.6: Example of the computation of PARA values .....................................83

Table 4.7: Matrix for scenario building ............................................................88

Table 5.1: Spanish administrative divisions and their NUTS correspondence ........91

Table 6.1 Network efficiency in Spanish NUTS-3 capitals. Road mode................115

Table 6.2 Network efficiency in Spanish NUTS-3 capitals. Rail mode .................120

Table 6.3: Network efficiency in Portuguese district capitals. Road mode ...........123

Table 6.4: Network efficiency in French department capitals. Road mode ..........125

Table 6.5 Network efficiency in Portuguese district capitals. Rail mode ..............127

Table 6.6: Network efficiency in French department capitals. Rail mode ............128

Table 6.7: Regional inequality indices. Road accessibility.................................133

Table 6.8: Regional cohesion performance indicator (RC). Road accessibility......134

Table 6.9: Regional inequality indices. Rail accessibility...................................138

Table 6.10: Regional cohesion performance indicator (RC). Rail accessibility ......139

Table 6.11 Travel time savings and estimated induced traffic ...........................151

Table 6.12: Forecasted induced traffic and corresponding increases in GHG

emissions. Do-nothing vs. PEIT alternative. Road and rail modes ...............152

Table 6.13 Summary of performance indicator values. Road mode ...................157

Table 6.14 Summary of performance indicator values. Rail mode .....................158

Table 6.15: Definition of value functions. Road mode ......................................160

Table 6.16: Integration of results. A0 vs. APEIT. Road mode ..............................162

Table 6.17: Definition of value functions. Rail mode........................................163

Table 6.18: Integration of results. A0 vs. APEIT. Rail mode ................................163

TABLE OF CONTENTS

-xi-

LIST OF FIGURES

Figure 2.1: The planning process ..................................................................... 7

Figure 2.2: Trade-off approach to sustainable transport ....................................13

Figure 2.3: Outline structure of the German spatial impact assessment module....28

Figure 2.4: Considerations affecting the decision-making process .......................31

Figure 3.1: Simple representation of a spatial impact system .............................36

Figure 3.2: Suggested twin approach to transport appraisal ...............................39

Figure 3.3: Activity and impedance functions ...................................................46

Figure 3.4: Example of a travel cost indicator. Road accessibility 1992 ................48

Figure 3.5: Network efficiency. Road accessibility 2005 (left) and 2020 (right) .....49

Figure 3.6: Daily accessibility indicator. Daily accessibility by rail (1993) .............51

Figure 3.7: Outline of the SASI model .............................................................57

Figure 3.8: Changes in GDP per capita as a result of the planned TEN priority

projects.................................................................................................58

Figure 3.9: Superposition of data layers in GIS for a transport study...................62

Figure 3.10: An integrated GIS approach to accessibility analysis. ......................65

Figure 4.1: Structure of the methodology ........................................................70

Figure 4.2: Comparison of alternatives ............................................................71

Figure 4.3: Performance indicators’ inputs .......................................................75

Figure 4.4. Scheme of the calculation of the PARA index....................................83

Figure 4.5: The integration procedure .............................................................85

Figure 5.1. Spanish road network (2005).........................................................89

Figure 5.2. Spanish rail network (2005)...........................................................90

Figure 5.3: Spanish NUTS divisions.................................................................91

Figure 5.4: Population density........................................................................92

Figure 5.5: Study area system of cities ...........................................................93

Figure 5.6: Growth in GDP per head in Spain, Spanish NUTS-2 regions and EU15 in

terms of EU25 average (PPS) 1995-2003...................................................94

Figure 5.7: Trends in GDP per head in Spanish NUTS-2 regions, EU15 and EU25 in

terms of Spain’ average, 1995-2003 .........................................................95

Figure 5.8: Accessibility by road (2005) ..........................................................96

Figure 5.9: Accessibility by rail (2005) ............................................................97

Figure 5.10: Trends in mobility, GDP and emissions in Spain, 1990-2003 ............99

Figure 5.11: Delimitation of the study area ....................................................102

Figure 5.12: Road network of the PEIT alternative (APEIT).................................103

TABLE OF CONTENTS

-xii-

Figure 5.13: Rail network of the PEIT alternative (APEIT) ..................................103

Figure 5.14: Sites of Community importance (SCIs) .......................................107

Figure 5.15: Special Protection Areas (SPAs) .................................................108

Figure 5.16: Spanish habitats map ...............................................................109

Figure 6.1: Network efficiency. Alternative A0. Road mode...............................112

Figure 6.2: Network efficiency. Alternative APEIT. Road mode............................114

Figure 6.3: Network efficiency. Relative differences Alternative A0 vs. APEIT. Road

mode ..................................................................................................114

Figure 6.4: Network accessibility. Alternative A0. Rail mode .............................117

Figure 6.5: Network accessibility. Alternative APEIT. Rail mode ..........................119

Figure 6.6: Network accessibility. Relative differences Alternative A0 vs. APEIT. Rail

mode ..................................................................................................119

Figure 6.7: Network efficiency in Portugal. Relative differences Alternative A0 vs.

APEIT. Road mode ..................................................................................122

Figure 6.8: Network efficiency in Southern France. Relative differences Alternative

A0 vs. APEIT. Road mode.........................................................................124

Figure 6.9: Network efficiency in Portugal. Relative differences Alternative A0 vs.

APEIT. Rail mode....................................................................................126

Figure 6.10: Network efficiency in Southern France. Relative differences Alternative

A0 vs. APEIT. Rail mode...........................................................................128

Figure 6.11: Potential accessibility. Alternative A0. Road mode .........................131

Figure 6.12: Box-plot of potential accessibility values in the do-nothing alternative.

NUTS-2 aggregation. Road mode............................................................131

Figure 6.13: Potential accessibility. Alternative APEIT. Road mode ......................132

Figure 6.14: Changes in potential accessibility. Alternative APEIT vs. A0. Road mode

..........................................................................................................133

Figure 6.15: Relative change in road accessibility inequality indices ..................134

Figure 6.16: Potential accessibility. Alternative A0. Rail mode...........................136

Figure 6.17: Box-plot of potential accessibility values in the do-nothing alternative.

NUTS-2 aggregation. Rail mode .............................................................136

Figure 6.18: Potential accessibility. Alternative APEIT. Rail mode........................137


..........................................................................................................138

Figure 6.20: Regional cohesion indices. Rail mode ..........................................139

Figure 6.21: NUTS-5 unemployment rates .....................................................140

Figure 6.22: Standardized absolute change of NUTS-5 regions in the potential

accessibility indicator. Road mode ..........................................................142

TABLE OF CONTENTS

-xiii-

Figure 6.23: Standardized relative change of NUTS-5 regions in the potential

accessibility indicator. Road mode ..........................................................142

Figure 6.24: Accessibility categories. Road mode............................................143

Figure 6.25: Structural backwardness categories ............................................144

Figure 6.26: Regional weighting factor. Road mode.........................................144

Figure 6.27: Standardized absolute change of NUTS-5 regions in the potential

accessibility indicator. Rail mode ............................................................146

Figure 6.28: Standardized relative change of NUTS-5 regions in the potential

accessibility indicator. Rail mode ............................................................147

Figure 6.29: Accessibility deficiency categories. Rail mode...............................148

Figure 6.30: Regional weighting factor. Rail mode ..........................................148

Figure 6.31: % change in the PARA index in SCIs. Road mode .........................154

Figure 6.32: % change in the PARA index in SPAs. Road mode.........................154

Figure 6.33: % change in the PARA index in SCIs. Rail mode ...........................155

Figure 6.34: % change in the PARA index in SPAs. Rail mode...........................156

Figure 6.35: Value function for the network efficiency criterion. Road mode.......161

Figure 6.36: Criterion weight sensitivity: efficiency criterion.............................164

Figure 6.37: Criterion weight sensitivity: cohesion criterion .............................165

Figure 6.38: Criterion weight sensitivity: environmental criterion......................165

Figure 6.39: Attribute value sensitivity. Road mode ........................................167

TABLE OF CONTENTS

-xiv-

LIST OF ABBREVIATIONS

-xv-

LIST OF ABBREVIATIONS

AST Appraisal Summary Table

CBA Cost-benefit analysis

CTP Common Transport Policy

DM Decision maker

DSS Decision support system

EC European Commission

ECMT European Conference of Ministers of Transport

ERDF European Regional Development Fund

ESD Environmentally Sustainable Development

ESDP European Spatial Development Perspective

ESPON European Spatial Observatory Network

EU European Union

FP Framework Programme

GDP Gross Domestic Product

GHG Greenhouse Gas

GIS Geographical Information System

HCR High Capacity Road

HSR High Speed Rail

LUTI Land use and transport interaction

MCA Multicriteria analysis

MMSS Member States

NATA New Approach to Appraisal

OJEU Official Journal of the European Union

PEIT Plan Estratégico de Infraestructuras y Transporte

PPS Purchase Power Standard

RTD Research and Technological Development

SACTRA Standing Advisory Committee on Trunk Road Assessment

SCI Site of Community Importance

SPA Special Protection Area

TEN-T Trans- European Transport Networks

TERM Transport and Environment Reporting Mechanisms

TABLE OF CONTENTS

-xvi-

Chapter 1 – INTRODUCTION

- 1 -

1. INTRODUCTION

1.1 Problem statement

The planning process of a transport infrastructure Plan entails a high degree of

complexity. Although during the past few decades there were important advances

in the development of assessment methodologies at the Plan level, today there are

still many issues for which a consensus has not been reached in the transport

research community. There are a number of reasons why the development of

assessment methodologies at the Plan level is still an area where research efforts

are needed.

First, the inclusion of the sustainable development approach (Serageldin,

1996) in transport planning processes caused a shift in transport planning

objectives towards strategic policy goals, such as network efficiency, cohesion or

environmental issues. This structure of strategic objectives is intimately linked with

the increased inclusion of transport sustainability issues (Greene and Wegener,

1997) into the planning framework. This objective shift has been translated into

policy documents by a wide variety of institutions (OECD, 1998; ECMT, 2004; EC,

2004; EC, 1999). Furthermore, it is necessary to broaden the assessment

objectives to include the above strategic impacts at the Plan level, given that the

scope of the projects might result in impacts elsewhere, either in another

transportation field, or in other sectors such as land use, energy or the

environment. Thus, national governments are increasingly demanding the inclusion

of strategic aspects in assessment methodologies (Bristow and Nellthorp, 2000).

However, both the definitions and the subsequent assessment of these strategic

impacts are uneven and scarce among official methodologies (Grant-Muller et al.,

2001).

Second, the increased importance given to consensus building, transparency

and communicative issues of the planning approach (Voogd and Woltjer, 1999)

calls for an adaptation of ‘black-box’ methodologies resulting in a single score for

each alternative, into ‘easy-to-interpret’ ones, providing relevant information on

different strategic policy aspects. It is increasingly acknowledged that the

Assessment of Transport Infrastructure Plans: a strategic approach

- 2 -

objectives of transportation policy cannot be transformed into one or two

performance criteria, but rather that there are different and competing objectives.

Indeed, decision-makers (DMs) are increasingly requiring the assessment

methodology to include relevant information which they can easily interpret, with

an enhanced graphical presentation of results, so they can make consistent

decisions on their part.

Finally, the high relevance of the political component inherent in the

assessment of transport Plans, means that the roles to be played by the technical

and the political assessment are not clear. This issue is reinforced by the frequent

presence of objective setting conflicts between the different administrative levels

(local, regional, national and European) involved in the planning process at the Plan

level (May et al., 2003; Beinat, 1998). Decision-making today is no longer seen as

an intellectual process, but as a socio-political and organizational process, where

the interest has shifted from the quality of the decision towards the quality of

decision-making (Voogd, 1997). In this context, the technical assessment enables

ranking the alternatives in terms of a set of criteria and priorities, thus making the

political decision-making stage feasible, but in no case replacing the DMs

responsibility.

The above reasons have created a need to develop a suitable methodological

basis that explicitly relates transportation policies to strategic impacts by taking

into account a wide variety of strategic aspects in a flexible and transparent way.

Increased computer capacity and the recent development of assessment tools, such

as Geographic Information Systems (GIS) has enabled the upsurge of important

methodological advances in this direction (Fotheringham and Wegener, 2000).

Hence, assessment methodologies are seen as a form of decision support to

DMs, keeping in mind that the technical assessment is important, but finally it is a

political decision ultimately derived from the consideration of a wider set of factors

than the criteria of efficiency of the transport system or the consideration of

environmental impacts (ME&P et al., 2001).

In this context, further research efforts to develop consistent methodologies

capable of assessing the strategic impacts mentioned above in a flexible and

transparent manner are needed. This thesis, ‘Assessment of transport

infrastructure Plans: a strategic approach integrating efficiency, cohesion and

environmental aspects’ is a step forward in this research line, with a proposed

assessment methodology and its subsequent validation in a case study.


- 3 -

1.2 Objectives

The overall objective of this thesis is ‘to develop a methodology capable of

complementing traditional assessment methodologies of transport infrastructure

Plans, from a strategic approach, integrating efficiency, cohesion and environmental

aspects’.

The achievement of this overall objective can be split into the following main

objectives:

� To define the set of strategic criteria, namely efficiency, cohesion and

environmental sustainability, that should be evaluated in the assessment of

transport infrastructure Plans,

� to develop a methodology, based on the use of spatial impact analysis tools,

capable of measuring the achievement of each of the criteria above,

� to integrate the results obtained in each assessment criterion in order to

provide an overall vision of the global performance of each alternative,

� to investigate the influence of the different variables present in the

methodology on the final assessment results,

� to provide DMs with policy recommendations on the basis of the contribution

of each alternative to the achievement of the assessment criteria,

� to develop a useful, transparent and flexible transport planning tool, whose

results can be easy to explain to the public.

1.3 Research methodology

In order to achieve the above objectives, the research work has been structured

into the following stages:

� Investigation of recent changes and the current situation of the transport

planning framework at the Plan level, in order to determine which are the

main strategic policy goals that any assessment methodology should handle.

� Review the current state-of-the-art assessment methodologies at strategic

levels, in order to detect possible incoherencies and methodological gaps.

� Analysis of the potential of spatial impact analysis tools, in particular GIS,

for the development of assessment methodologies and as a support system

in the planning process.

� Justification of the usefulness of accessibility indicators as a planning tool

capable of assessing strategic aspects, such as network efficiency or

cohesion impacts.


- 4 -

� Definition of a set of strategic criteria and subcriteria that should be included

in the assessment of transport infrastructure Plans, and the corresponding

procedure to assess each one of them.

� Development of a procedure, based on multicriteria analysis (MCA) capable

of ranking a set of alternatives on the basis of their performance on the set

of defined criteria.

� Test of the validity and consistency of the proposed approach through its

application in a case study. The case study corresponds to the Spanish

Strategic Transport and Infrastructure Plan 2005-2020 (PEIT), recently

launched in Spain (Ministerio de Fomento, 2005).

� Drawing of conclusions on the validity of the methodology and identification

of areas for future research.

An important part of the research work developed in this thesis is based on the

findings of different research projects which were developed during the period the

research was carried out (2002-2007). In these projects different strategic impacts

of large scale transport infrastructure investments were assessed. These are listed

below:

� Assessment of territorial impacts of transport infrastructure investments.

Application: analysis of the Spanish transport network. Research Project

funded by the 2002 Ministry of Public Works Research Programme.

� Assessment of the effects of transport infrastructure Plans on mobility, the

territory and the socio-economic system, in the context of the enlargement

of the European Union. Research project funded by the 2004/2007 Ministry

of Science and Technology Research, Development and Innovation

Programme.

� Indicators of impacts of transport infrastructure on social and territorial

equity. Supported by the 2004 Transport research aids of the Ministry of

Public Works Research Programme.


- 5 -

1.4 Structure of the thesis

In order to achieve the objectives defined in section 1.2., the thesis has been

structured into eight Chapters and two Appendices:

� Chapter 1 is this Introduction. It describes the research problem that the

thesis is aimed at solving and the main objectives of the research.

� Chapter 2 analyses recent changes in the planning framework and reviews

current research efforts and challenges in transport planning processes, with

a focus on the Plan level.

� Chapter 3 includes a review of the main spatial impacts present at the Plan

level, along with a description of recent methodological advances in spatial

impact models and tools.

� On the basis of the findings of Chapters 2 and 3, Chapter 4 describes the

proposed assessment methodology: a strategic approach integrating

efficiency, cohesion and environmental aspects.

� Chapter 5 describes the assessment framework of the case study in which

the methodology is tested.

� Chapter 6, includes the assessment results obtained from the application of

the methodology to the case study.

� Chapter 7 summarizes the main conclusions and contributions to the

literature of the thesis and identifies future research directions.

� Chapter 8 includes the Reference list.

Finally, two appendices are included. Appendix A contains the questionnaire

distributed to relevant stakeholders in order to define criteria weights to be used for

the integration stage of the MCA procedure and the resulting weights. Appendix B

includes a description of the case study application of the accessibility model and a

list with disaggregated accessibility values.


- 6 -

Chapter 2 – A CHANGING PLANNING FRAMEWORK

- 7 -

2. A CHANGING PLANNING FRAMEWORK

2.1 Introduction

The transport system can be considered as a socio-cultural complex adaptive

system (Buckley, 1967). In other words, a system in which the interchanges

between their elements may result in significant changes in the nature of the

elements themselves with important consequences for the system as a whole

(Rehfeld, 1998). Besides this ‘internal’ complexity, the transport system is also

influenced by contextual elements (Banister et al., 2000a), also referred to as

development variables (Rehfeld, 1998), which are part of other interrelated

systems, such as the environment or the economy.

Consequently, transport planning processes are unavoidably complex.

Although many approaches exist in the literature (for reviews on the topic see

Meyer and Straszheim, 1971; Button, 1993; EC, 1996c; Nijkamp et al., 1990),

there is no single best method to conduct a transport planning process. Figure 2.1

shows the approach suggested by Mackie and Nellthorp (2003), which was selected

because of its flexibility to include a wide variety of approaches. It considers the

transport planning process as a three-stage process:

� Structuring the planning problem and objective setting,

� Evaluation of the effects of each alternative course of action,

� Decision-making on the basis of the evaluation results.

Figure 2.1: The planning process

STRUCTURING EVALUATION DECISION-MAKINGSTRUCTURING EVALUATION DECISION-MAKING

Source:Mackie and Nellthorp (2003)

The process will normally entail iterative procedures (Bristow and Nellthorp,

2000; Meyer and Straszheim, 1971): the more complex the planning problem is,

the more feedback loops the evaluation process will have (Nijkamp et al., 1990).

Besides, the boundaries between the three stages are not always clear (Beuthe,

2002; Voogd, 1997).


- 8 -

In general, a hierarchy of four different planning levels can be defined (EC,

1996c) in descending order of scope and complexity: policy, programme, plan and

project levels. At the top of the hierarchy, the wide-ranging of the planning

problems, along with their long-term effects necessitate the employment of

sophisticated methods of project appraisal. Besides, they require the development

of comprehensive techniques for decision-making (Button, 1993), which are

increasingly demanding a more comprehensive consideration of uncertainty issues

(Tsamboulas et al., 1998). At the top of the hierarchy, ideally a systems planning

approach –capable of considering the independence of projects and the feedback of

the transport system on other interrelated systems- appears to be the

recommended planning procedure (Meyer and Straszheim, 1971).

Furthermore, the transport planning framework is constantly evolving. A

growing interest in the structuring stage has been developing in recent years

(Voogd, 1997), although evaluation is still a central part of the planning process.

Finally, although the decision-making stage is aimed at providing relevant

information to decision-makers, it is not a substitute for the political process, i.e. it

does not take the decision. This is especially true at planning levels situated at the

top of the hierarchy, where the political assessment is dominant and technical

appraisal very limited (EC, 1996c).

Besides, the planning process is increasingly required to be flexible and

adaptive to a highly dynamic environment, in which the political relevance of

issues, alternatives or impacts may exhibit sudden changes (Voogd, 1997).

Consensus building, transparency and communicative issues are increasingly

considered as added values (ICCR, 2002b), in a so called ‘communicative planning

approach’ (Voogd and Woltjer, 1999). This has forced all stages of the planning

process to be accessible and comprehensive in arenas such as public inquiries

(Grant-Muller et al., 2001). This is now a quality requirement which has forced the

‘technical assessment’ to be combined with educational and consensus building

tools, allowing a project to be subject to debate, consultation and participation, in a

spirit of a more open public involvement in decision-making (Small, 1999).

In this context, this Chapter reviews current research efforts and challenges

in transport planning processes, with a focus on the Plan level. For clarity reasons,

the Chapter has been structured following Figure 2.1 planning stages: structuring,

evaluation and decision-making.


- 9 -

2.2 Structuring the planning process

2.2.1 Sources of conflicts in objective setting

The definition of objectives may raise conflicts both at a ‘vertical’ level, i.e. between

the different stakeholders involved, and at a ‘horizontal’ level, i.e., between the

different systems interrelated with the transport system (Bröcker et al., 2004).

First, at a ‘vertical level’, the increased promotion of the public consultation

stage has allowed for the involvement of individuals (experts, political

entrepreneurs) or specific organizations (ad hoc structures, citizen organizations),

which have different priorities. This demands a more transparent and open

procedure for the definition of planning objectives (Voogd and Woltjer, 1999).

Furthermore, there is a risk of disagreement, lack of congruence and different

preference strength between DMs of the different territorial levels of competencies

involved, which may achieve the degree of political concerns (Tsamboulas et al.,

1998; Beinat, 1998; Ollivier-Trigalo, 2001; ICCR, 2002b). Furthermore, any

transport policy involves significant spillovers (Pereira and Roca-Sagales, 2003) and

creates further risks of overlapping benefits and double counting at different stages

of the appraisal process (Grant-Muller et al., 2001), which require a certain degree

of ‘multi-level’ coordination (Bröcker et al., 2004). In this sense, the transport

planning process of the trans-European transport networks (TEN-T) (EC, 2004c)

constitutes a successful example of integrating conflicting European, national,

regional and even local objectives (Turró, 1999; Button, 1993; Chatelus and Ulied,

1996).

Second, the existing interactions between transport and other interrelated

policies, such as spatial development, economic or energy policies, which will be

analyzed in Section 2.2.3.2, also calls for a ‘horizontal’ integration of possibly

conflicting objectives.

An integrated framework combining all these conflicting objectives is

therefore needed. In this context, the emergence of the sustainable development

concept and its subsequent application in transport planning processes has

provided a reference framework to join and integrate interests from different

approaches, as Section 2.2.2 will detail.

2.2.2 A guiding principle: the sustainable development approach

2.2.2.1 Sustainable development and transport planning

The concept of sustainable development (Brundtland Commission, 1987) emerged

in the 1980s in the environmental field, and was originally named as


- 10 -

“Environmentally Sustainable Development” (ESD). It was approached by a

triangular framework (Serageldin, 1996), representing three dimensions: economic,

social, and environmental. It was in the 1990s when the concept of sustainable

development was introduced as an overall goal for the transport sector. Since then,

the terms used to refer to the three general sustainability objectives were adapted

to suit the specific characteristics of the transport problem under consideration.

Nowadays the term ‘sustainable transport’ is a generally accepted principle

in transport planning processes (see e.g. Greene and Wegener, 1997; Nijkamp,

1994; Button and Verhoef, 1998; Feitelson, 2002; Lauridsen, 2003; OECD, 1998).

However, finding targets for these three general objectives is a complex task, as it

requires finding widely accepted statements and terms of reference from both

scientific and official policy documents which might offer a basis for target definition

(Banister et al., 2000b; Button and Verhoef, 1998). A discussion on how this issue

is dealt with in each of the three sustainability objectives is included in subsections

2.2.2.2 to 2.2.2.4.

2.2.2.2 The economic objective

The economic dimension is an area where descriptions of objectives differ

markedly: the economic objective may be also named with other terms, mainly

‘efficiency’ (Turró, 1999; Bröcker et al., 2004; Button and Verhoef, 1998),

‘competitiveness’ (Chatelus and Ulied, 1996; EC, 2004a), or ‘growth’ (Serageldin,

1996; Feitelson, 2002).

The assumption is that infrastructure network weaknesses limit the

realization of the economic growth development potential (Frybourg and Nijkamp,

1998). Under this assumption, the target is to maximise transport efficiency, a

general term which includes objectives such as an improved performance and

development of each mode and their integration into a coherent transport system,

socio-economic feasibility, or improved comfort and level of service (Giorgi and

Pohoryles, 1999).

Therefore, this objective refers to the contribution of a transport initiative to

increase the overall productivity of economic activities, in terms of increasing

opportunities for new relations and bridging existing bottlenecks (Chatelus and

Ulied, 1996). Therefore, this objective is intimately linked with the impact of

transportation costs in economic performance (SACTRA, 1999).

The economic objective lies behind the assumption that ‘missing links’ and

lack of infrastructure provision may mean a significant reduction in the potential

productivity of a region or nation. Following this rationale, investments in transport

infrastructure in backward regions help to ensure relatively equal competitive


- 11 -

advantages for all regions (Rietveld and Nijkamp, 1993; Capello and Rietveld,

1998) and therefore they have been included in national and supranational plans in

Europe. This is a controversial issue, good transport facilities, -although important-,

are not sufficient to ensure economic growth by themselves1.

2.2.2.3 The social objective

In recent years there has been an evolution of concerns and objectives of

transportation policy from efficiency to social objectives (Tsamboulas et al., 1998).

However, this is an underdeveloped field both in policy and scientific analysis

(Grant-Muller et al., 2001), where it is frequent to find many approaches included

under this heading (EC, 2004a), mostly dependant on the assessment level.

At the project level, the social dimension generally refers to objectives such

as accident reduction, noise abatement, or local emissions reduction (Bristow and

Nellthorp, 2000; Mackie and Nellthorp, 2003). This approach is rather limited at the

level of transport policies/plans, where the treatment of the social dimension

increasingly requires an analysis under the ‘cohesion’ objective (EC, 2004a; EC,

1998).

In broad terms, improved cohesion means a reduction of economic

disparities (Bröcker et al., 2004) or differences of economic and social welfare (Hey

et al., 2002) between regions or groups. In spatial policy terms, the objective is to

avoid territorial imbalances (EC, 1999), by making both sector policies which have

a spatial impact and regional policy more coherent. The concern is also to improve

‘territorial integration’ and encourage cooperation between regions or countries

(Banister et al., 1999). However, not even in official European Community

documentation is there a precise description of what is behind cohesion (INRETS,

2005; Bröcker et al., 2004; INRETS, 2005; EC, 1998). Even the term

‘convergence’, which aims at the gradual reduction of regional differences, gives

little help (EC, 2004a). This vagueness in the definition of the term frequently gives

rise to methodological problems in the evaluation stage.

2.2.2.4 The environmental objective

In the past few decades there has been an increased concern for assessing the

environmental effects of transport and developing mechanisms to report their

evolution, such as the periodic ‘Transport and Environment Reporting Mechanisms’

1 The existence and measurement of the contribution of transport to economic growth is a controversial

issue widely discussed in the literature (see e.g. Button, 1993; Oosterhaven and Knaap, 2003; Banister

and Berechman, 2003; Banister and Berechman, 2001; Vickerman et al., 1999). This issue is further

dealt with in Section 3.1.3.2.


- 12 -

(TERM) Reports (EEA, 2006b). At the EU level, the transport sector is the primary

driver of the growth in total energy consumption, which is likewise directly linked

with total emissions (EEA, 2006a). Despite the important efforts devoted to

environmental abatement policies, the high rate of increase in transport demand is

outstripping the rate of improvement in environmental technology for transport

(Stead, 2001).

The result has been a significant increase in Green House Gas (GHG)

emissions from transport, which threatens European progress towards its

international commitments, such as the Kyoto targets (UNFCCC, 1997) and the

proposals by the EU Council for further emission reductions for developed countries

beyond the Kyoto Protocol period (2008–2012) (EC, 2005b).

Air pollution reduction is also on the EU agenda, although energy-related

emissions from the transport sector have decreased steadily since 1990 (EEA,

2006b), largely due to the result of increasingly strict emission standards for the

different transport modes and fuel switching. Nevertheless, further emission

reductions are also required, as recognised in the proposed Thematic Strategy on

Air Pollution (2005) (EC, 2005a), mainly because air quality in mega cities does not

yet meet the limit values set by European regulation and still has a major negative

impact on human health (EEA, 2006b). Finally, habitat fragmentation and the loss

of biodiversity associated with new transport infrastructure are also concerns for

transport policy at the strategic level (EEA, 2006b).

2.2.2.5 Trade-offs between objectives

Given this definition of the three SD objectives, it is inevitable that trade-offs

appear between them (for discussions on this issue see Feitelson, 2002; Button and

Verhoef, 1998).

These trade-offs are represented in Figure 2.1. Of particular interest in the

transport field is the conflict of efficiency (economic) vs. equity (social) objectives.

If the only objective was the maximization of economic growth, the ‘most efficient’

policy would attempt to concentrate the economic activity in several strong regional

centres and interconnect them with a high quality transport network (Gutiérrez,

2004). However, this policy would have a negative impact on equity, as it would

lead to more polarized spatial development patterns (EC, 1999): richer regions

would gain more and lagging regions would result in a comparative worse situation.

As stated by Bröcker et al. (2004): ‘In practice considerable trade-offs may be

necessary between, say, devising a transport policy to stimulate national growth

and one designed to assist the development of specified backward regions (…)’. The

design of transport strategies may need to be modified to ensure that both an


- 13 -

acceptable degree of equity is retained among the different regions, while economic

growth is maximised (Button, 1993).

Furthermore, as transport infrastructure improvements are aimed at

reducing travel costs, they may to a certain extent promote mobility and have a

negative impact on environmental objectives. This raises conflicts between

economic and environmental objectives (Bröcker et al., 2004), given the historic

link between economic growth and traffic growth. Decoupling transport from

economic growth -defined as maintaining levels of economic growth, but with lower

levels of transport intensity– is therefore a key objective in transport strategy

design (Banister et al., 2000a).

Figure 2.2: Trade-off approach to sustainable transport

SUSTAINABLE TRANSPORT

ECONOMIC

•Competitiveness

•Investment costs

SOCIAL

•Equity

•Territorial integration

ENVIRONMENTAL

•Global/Local emmissions

•Habitat preservation

Equityvs. efficiency E

fficiencyvs. Environment

Equity vs. Environment

SUSTAINABLE TRANSPORT

ECONOMIC

•Competitiveness

•Investment costs

SOCIAL

•Equity

•Territorial integration

ENVIRONMENTAL

•Global/Local emmissions

•Habitat preservation

Equityvs. efficiency E

fficiencyvs. Environment

Equity vs. Environment

Source: Adapted from Feitelson (2002)

2.2.3 EU policy objectives

2.2.3.1 Transport policy

Major transport and sector-related policy documents at an EU level respond to the

general SD framework described in Section 2.2.2. In fact, the three main basic

goals of the Common Transport Policy (CTP) are: competitiveness, cohesion and


- 14 -

environment2. However, the structural changes that are taking place at present at

the EU scale means that the current EU Transport Policy is in a ‘state of flux’

(Frybourg and Nijkamp, 1998).

The 2006 mid term review of the 2001 White Paper (EC, 2001a) summarizes

the main priorities of EU transport policy, namely ‘to help provide Europeans: with

efficient, effective transportation systems that (EC, 2006a):

� offer a high level of mobility to people and businesses throughout the EU (…),

� protect the environment, ensure energy security, promote minimum labour

standards for the sector and protect the passenger and the citizen (…),

� innovate to support the first two aims of mobility and protection by increasing

the efficiency and sustainability of the growing transport sector (…), and

� connect internationally, projecting the Union’s policies to reinforce sustainable

mobility, protection and innovation, by participating in the international

organisations’.

In terms of transport infrastructure investments, the key transport policy

instruments are the TEN-T. The implementation of the TEN-T contributes to

important objectives of the EU such as ‘the good functioning of the internal market

and the strengthening of the economic and social cohesion’ (…) or ‘to ensure a

sustainable mobility for people and goods, in the best social, environment and

safety conditions, and to integrate all transport modes’ (EC, 1996a)3. Furthermore,

the TEN-T are recognized as a key factor for the European integration process

(Turró, 1999), which relies upon the development of an efficiently operating

network connecting all nodes of the ‘European network economy’ (Frybourg and

Nijkamp, 1998).

2.2.3.2 Non-transport Policy documents

Transport policy may result in synergies or conflicts between the policy goals of

interrelated policy areas. This ‘horizontal’ policy interaction (Bröcker et al., 2004;

EC, 1999) should be taken into account in the structuring stage of strategic

transport planning problems. The following sections identify these policies.

2.2.3.2.1 Regional policy

Structural policy provides support for transport in the MMSS through the European

Regional Development Fund (ERDF) and the Cohesion Fund (OJEU, 2006). The

2 The background context for the development of the EU CTP can be found in Banister et al. (2000a), pp

58-60.

3 Amended by EC (2004b).


- 15 -

rationale behind this support is the assumption that certain transport infrastructure

investments, mainly in lagging regions, are believed to contribute in a decisive way

to the achievement of the goal of territorial and social cohesion. The performance of

the EU in terms of cohesion over a period of three years is reported in the periodic

EC Cohesion Reports (the last one (EC, 2004a) was published in 2004).

2.2.3.2.2 Spatial development policy

Spatial development policy is also of increasing concern in EU regional policy,

because of its intimate and complex relationship with transport infrastructure. In

this respect, the European Spatial Development Perspective (ESDP) (EC, 1999)

constitutes the major attempt so far to provide a Community strategy on the

spatial development of the EU, but it is in no sense a European Masterplan, which

would give rise to competency issues (Faludi, 2002). The ESDP includes among its

objectives to ‘strengthen a polycentric and more balanced system of metropolitan

regions, city clusters and city networks through closer co-operation between

structural policy and the policy on the TEN-T and improvement of the links between

international/national and regional/local transport networks’ (EC, 1999).

The ESDP proposes a movement from transport investments improving

transport links between the periphery and the core –the tendency of structural

policy- towards a new perspective for the peripheral areas through the creation of

‘several dynamic zones of global economic integration, well distributed throughout

the EU territory and comprising a network of internationally accessible metropolitan

regions and their linked hinterlands’ (EC, 1999).

2.2.3.2.3 Energy policy

These concerns from the EU on energy and transport issues have been translated

into energy policy documents. This is the case with the last EC’s Green Paper ‘A

European Strategy for Sustainable, Competitive and Secure Energy’4 (EC, 2006b),

and its predecessor (EC, 2000b). In summary, the main objective is that energy

and transport contribute to sustainable development: ‘making Europe both a

homogenous area of economic development and an area where the environment in

the broadest sense of the term is conserved’ (EC, 2004d).

2.2.3.2.4 Environmental policy

Transport as a sector is the largest single contributor to a number of environmental

problems, therefore a strong set of policy linkages occur between transport and

4 COM (2006) 105 final.


- 16 -

environmental policy (EEA, 2006b; OECD, 1998). The majority of them have

already been mentioned in the preceding section on energy policy, as both sectoral

policies are also strongly linked.

The most important environment policy document at the EU level is the

Sixth Environment Action Programme (Decision 1600/2002/EC, 22 July 2002),

which provides a strategic framework for the Commission's environmental policy up

to 2012. The programme identifies four environmental areas for priority actions:

Climate Change; Nature and Biodiversity; Environment, Health, Quality of Life; and

Natural Resources and Waste.

In the context of transport planning at the EU level, the main policy

document is the Strategic Environmental Assessment Directive (OJEU, 2001). It is

aimed at ensuring that environmental consequences of certain plans and

programmes are identified and assessed during their preparation and before their

adoption.

2.3 The evaluation approach

2.3.1 Introduction

It is beyond the scope of this thesis to conduct a review of the large number of

evaluation techniques which have been developed since their emergence in the

1960s. Therefore, this Section 2.3 includes only an outline of the major approaches,

along with an extended list of selected references containing detailed

methodological issues. Furthermore, a review of recent research developments in

the assessment field, along with an outline of the current state-of-the practice in

official evaluation approaches for transport plans in selected MMSS is included.

2.3.1.1 Evaluation, appraisal, assessment: synonyms?

It is frequent to find the terms evaluation, appraisal and assessment used

indistinctly in transport planning literature. However, although they refer to

intimately linked concepts, they are not synonymous.

In general terms, evaluation can be defined as ‘a process which seeks to

determine as systematically and objectively as possible the relevance, efficiency

and effect of an activity in terms of its objectives’ (Giorgi and Tandon, 2000b).

Appraisal can be described as ‘a process of investigation and reasoning

designed to assist DMs reach an informed and rational choice’ (Sudgen and

Williams, 1978), or as the process whereby it is determined whether a project

meets a set objectives and whether these objectives are met efficiently (Adler,

1987). Appraisal should comply with a set of requirements (EC, 1996c; Nijkamp et


- 17 -

al., 1990) and in all cases it should be viewed as an aid rather than a replacement

for the decision-making stage, which is often a political process.

Furthermore, the terms appraisal and evaluation are often used to refer to

two different forms of assessment, depending whether it is carried out before or

after a project has been implemented (May et al., 2003). If before (ex ante), the

assessment is an aid to decision-making, and then the term appraisal is more

frequently used. The term evaluation is usually reserved for an ex post assessment

after the project has been implemented, which is rather less frequent. However,

this classification is not always followed in the transport planning literature, where

it is common to find the term ‘evaluation’ referring to ex ante assessments (EC,

1996c).

2.3.2 Outline of an evaluation process

The evaluation process of a transport Plan consists of a series of logically related

modules. They are briefly described in sections 2.3.2.1 to 2.3.2.3.

2.3.2.1 Setting up the evaluation framework

Any evaluation necessarily starts with a set of preliminary tasks, including the

selection of the limits of the study area and its zonification, the definition of the

‘reference’ and assessment alternatives, and the definition of the assessment time

horizon. In strategic transport planning, this time horizon is usually long-term,

giving rise to uncertainty issues which require the evaluation stage to analyse

different possibilities of change of trends in economic, technological, environmental

and social development (EC, 1996b), i.e. to define evaluation scenarios.

Scenarios5 are ‘a kind of structures brainstorming technique, which may

widen the perceptions of researchers as well as policy-makers regarding possible

future opportunities (…) they are important tools for strategic policy analysis,

especially in situations where policy makers have too much biased and unstructured

information’ (Banister et al., 2000a).

There are two main different scenario traditions, namely (Banister et al.,

2000a):

� explorative external scenarios, i.e. external scenarios in that they describe

factors beyond the control of the transport sector, although they have a direct

effect on the sector,

5 A comprehensive review on scenario building techniques can be found in Rehfeld (1998), Banister et al.

(2000a) and Nijkamp et al. (1998).


- 18 -

� backcasting scenarios, where the scenarios are designed as ‘images of the

future’ that show desirable solutions to a major social problem (e.g. sustainable

mobility). Then one tries to find a possible path between today and the images.

2.3.2.2 Selecting the appraisal framework

2.3.2.2.1 Classification of appraisal methodologies

Despite the large number of approaches currently available, there is still

surprisingly little information regarding the specific features of the methods

available and the precise conditions under which a method is chosen in practice

(Nijkamp et al., 1990). Experience suggests that there is no single ‘best method’,

but that the choice will depend on a set of factors, of which the evaluation level if of

special importance (EC, 1996c).

Nowadays one may distinguish at least four types of evaluation styles in the

planning literature (Vreeker et al., 2002):

� A monetary decision approach, based e.g. on cost-benefit or cost-effectiveness

principles,

� A utility theory approach, based on prior ranking of the decision-maker’s

preferences,

� A learning approach, based on a sequential (interactive or cyclical) articulation

of the DM’s views,

� A collective decision approach, based on multi-person bargaining, negotiation or

voting procedures.

Depending on the style chosen, current public sector investment appraisal

can be reviewed under three broad frameworks (Bristow and Nellthorp, 2000):

Cost-benefit analysis (CBA), Multi-criteria analysis (MCA), and Descriptive

frameworks. These three frameworks will only be outlined in the following sections.

A detailed description of the theory underlying these methods can be consulted in

the many existing textbooks on the subject, several of which are referenced below.

2.3.2.2.2 Cost-benefit analysis (CBA)

The major upsurge in the development of appraisal techniques for transport

projects came in the late 1960s and early 1970s6, and they were mainly based on

CBA approaches.

6 The European Conference of Ministers of Transport (ECMT) Round Tables of this period encouraged the

discussion and development of ideas related to appraisal. A series of Round Table Reports from those

decades chart the practical development of CBA at that time (ECMT, 2005; ECMT, 2004).


- 19 -

In a CBA approach, both the potential costs and benefits of a particular

project are estimated across a set of impacts and converted into monetary units by

multiplying impact units by prices per unit. The final outcome of the appraisal is a

single value, such as a discounted net present value or a cost-benefit ratio (for

extensive reviews on CBA see Sudgen and Williams, 1978; Layard and Glaister,

1996; Boardman et al., 2001; Adler, 1987; de Rus et al., 2003).

Although CBA may, in principle, be a sound evaluation method for decisions

in the public sector, some authors claim that CBA has several limitations, which are

believed to reduce the confidence felt in the strength of CBA calculations (Beuthe et

al., 2000; Vreeker et al., 2002; Button, 1993). The most argued upon limitations of

the CBA approaches are:

� Their difficulty to arrive at monetary values for intangible effects such as

ecological risks or the fulfilment of regional planning objectives (BMVBW, 2002).

� Their impossibility to take into account explicit interest conflicts and political

priorities (Nijkamp et al., 1990; Voogd, 1997; Vreeker et al., 2002).

� Their inability to address distributive issues, given that the aggregation of all

costs and benefits implicit in CBA raises the sensitive question of the distribution

of outcomes across individuals (Beuthe et al., 2000; Nijkamp et al., 1990;

Small, 1999).

However, there is controversy in this subject. There is a substantial school of

thought that subscribes to the view that direct transport benefits measured by

means of CBA do indeed capture all the benefits of schemes and to include anything

else is to introduce double-counting (EC, 1996b).

2.3.2.2.3 Multi-criteria analysis (MCA)

In the 1970s MCA methods emerged as a result of the mentioned limitations of

CBA. The emergence of environmental problems with many qualitative dimensions

also gave MCA a particular stimulus. Its perceived ‘power of conviction’, easiness of

interpretation and transparency, compared to CBA, contributed to increase the

popularity of MCA (Voogd, 1997).

MCA aims at taking into account the heterogeneous and conflicting

dimensions of complex policy evaluations, offering an operational framework for a

multidisciplinary approach to wide-ranging (physical) planning problems (Nijkamp

et al., 1990). The method typically involves determining the extent to which

alternative proposals achieve a pre-determined set of goals or objectives. Detailed

descriptions of the theoretical foundations of the MCA method can be found in


- 20 -

Nijkamp et al. (1990), Malczewski (1999), Dodgson et al. (2001), Saaty (1990),

and Keeny and Raiffa (1976).

Many innovations of MCA, such as the treatment of qualitative weighting and

mixed data techniques, are criticised mainly due to subjective assessment and DM

judgment likely to be involved (Voogd, 1997), while CBA is considered to have

generally more objective and explicitly defined criteria (EC, 1996c). However, some

authors argue that a path needs to be steered between a ‘fixed weights’ and a

complete lack of guidance on criterion weights. This path would be to set limits for

the relative magnitudes of criterion-weights, rather than identify precise values

(Sayers et al., 2003). A possibility, as suggested by Beuthe et al. (2000), would be

to transform MCA preferences over the various criteria into equivalent money

values. This would make it possible to keep the basic and commonly accepted rules

of the CBA unchanged.

Despite this debate, MCA and CBA approaches are increasingly seen as

complementary rather than competitive (Nijkamp et al., 1990; Bristow and

Nellthorp, 2000). However, this complementarity may bring some technical

difficulties: given that some impacts are dealt with by the CBA, and others within

the MCA, there is a need for clarity within the framework as a whole (Grant-Muller

et al., 2001).

2.3.2.2.4 Descriptive frameworks

These are methods of analysis which may be objective-led and may focus on a wide

range of impacts, but within which the results are neither weighted nor valued

(Bristow and Nellthorp, 2000).

These frameworks are more flexible and respond to the difficulty stressed by

some authors in identifying the DMs preference system. Conventional appraisal

methodologies such as CBA and MCA are thus increasingly tending to become

Decision Support Systems (DSS)7 (Beuthe, 2002; Bana e Costa et al., 1999).

DSS are computer-integrated tools, usually embedded in a Geographical

Information System (GIS) platform, which help to organize the evaluation process

into a rational logical path, guaranteeing repeatability and transparency. Recent

DSS approaches include user-friendly software tools constituted by seamlessly

integrated modelling packages (see e.g. Arampatzis et al. (2004), Colorni et al.,

(1999) and Jha (2003) for existing applications).

7 A DSS is ‘a computer-based information system used to support decision-making activities in situations

where it is not possible or desirable to have an automated system perform the entire decision process’

(Dyer et al., 1992).


- 21 -

2.3.2.3 Selection of impact assessment models

The integration of different impacts from different arenas required at the Plan level,

such as environmental, energy, technology or financial, has forced the dominance

of transportation engineers to evolve towards a truly integrated assessment team

including economists, geographers, sociologists and engineers, where the value of

each component is recognized (ECMT, 2004). As stressed by Voogd (1997): ‘the

times when a couple of traffic engineers could do the job in infrastructure planning

is certainly now history’.

Furthermore, as all the above disciplines are inextricably linked, ideally

impact assessment models from different arenas should be considered as a single

interacting system (EC, 1996b; Nijkamp et al., 1990). In this sense, frameworks

developed for assessing spatial socio-economic impacts must ‘acknowledge the

different spatial and temporal dimensions of impacts, together with the interactions

which occur between transport, land use and economic development, strategic

public intervention and the environment’ (EC, 1996b).

At the Plan level, many of the policy objectives seen as of growing

importance, such as cohesion, are not readily measurable. Indeed, many are

proving elusive even as far as definition is concerned (Grant-Muller et al., 2001).

Furthermore, impact analysis is fraught with many difficulties, mainly because

some effects may be indirect, and therefore they make a detour via intermediate

variables, or because reliable data are missing (Nijkamp et al., 1990).

In this context, it is essential that the output indicators of the impact models

used refer to the achievement of wider policy objectives (Sheate, 1992; Banister et

al., 2000b). Impact models should provide acceptable ways of quantifying an

impact, so that the achievement of the objective should be in the form of a clearly

defined change in the value of the indicator chosen. The challenge is therefore to

define performance indicators, linking them to appropriate ways of categorization or

measurement, and establishing money values, weights or other ways of facilitating

their aggregation into overall indicators (Grant-Muller et al., 2001).

In response to these requirements, a wide variety of methods and models

from different disciplines has emerged in recent decades (Nijkamp et al., 1990;

Malczewski, 1999; Nellthorp et al., 1998), ranging from simple ad hoc and

correlation techniques to sophisticated models8. The selection of the most

appropriate model/s often depends on the appraisal method chosen and the data

availability.

8 An overview of spatial impact models can be found in Nijkamp et al. (1990).


- 22 -

In this context, spatial impact analysis tools (Nijkamp et al., 1990) have

proven their efficacy in transport planning processes. Spatial impact analysis can be

described as a specific type of system analysis (Buckley, 1967; Meyer and

Straszheim, 1971), as it refers to a multiplicity of sectors and it deals with an open

system, so that interactions, spatial spillovers and multi-level effects are included.

Furthermore, the development of GIS has facilitated the implementation of spatial

impact analysis models and their integration into transport planning processes

(Malczewski, 1999).

2.3.3 Current state of the practice in Europe

2.3.3.1 Towards integrated assessment methodologies

National official appraisal methodologies in EU countries vary substantially, in terms

of the impacts considered, the measurement methods and the appraisal and

evaluation techniques used (EC, 1996b; Grant-Muller et al., 2001; Bristow and

Nellthorp, 2000; ECMT, 2004; Bickel et al., 2005; ECMT, 2005; Steer Davies

Gleave, 2004). This is mainly because of their different institutional frameworks,

which lead to different requirements in political decision processes (ECMT, 2005).

Furthermore, as national transport policy has a strong influence in the resulting

appraisal methodology, it is frequent to find that developments in the

methodologies have often been brought about by changes in national, and even

supranational, transport policy objectives (Bristow and Nellthorp, 2000).

Regarding direct impacts, there is some general agreement on which of

them should be considered in the CBA, although some issues still need to be

harmonised (Nellthorp et al., 1998). In terms of environmental impacts, although

progress has been made towards their measurement (Bristow and Nellthorp, 2000),

there is less of a consensus. Finally, the assessment of the wider impacts remains

underdeveloped (Steer Davies Gleave, 2004; Grant-Muller et al., 2001).

However, there is today an intensified demand for assessment frameworks

that consider (long-term) wider policy issues rather than just their direct outputs

(EC, 1996b). Contemporary challenges, such as the debate on ‘sustainability’ both

with respect to the environment and with respect to distributional considerations or

accessibility, cannot be addressed solely by a CBA approach (ICCR, 2002b).

Moreover, separate economic efficiency or environmental assessments undertaken

in isolation are considered less efficient than integrated assessments (Nijkamp et

al., 1990) covering the whole range of impacts (ECMT, 2004).


- 23 -

These integrated assessments require (Nijkamp et al., 1990):

� Appropriate and reliable assessment of relevant impacts of policy measures or

exogenous changes.

� Complete representation of the policy areas concerned (including its feasible

decision space).

� Multidimensional representation of the diverse components or modules of the

system at hand.

� Flexible adjustment of the policy analysis to new information or new

circumstances.

� Comprehensive presentation of the results to responsible decision-makers or

actors.

� Consideration of equity aspects and spillover effects.

� Treatment of trade-offs and conflicts inherent in the choice problem at hand.

� Use of learning strategies and decision aid tools in a communication between all

participants involved in the policy problem at hand.

� Integrated approach with much attention paid to compromise procedures and

institutional dilemmas.

Current research efforts are targeted towards the development of these

integrated assessment methodologies. They are reviewed in Section 2.3.3.2.

Furthermore, Section 2.3.3.3 outlines recent advances in selected national official

assessment methodologies.

2.3.3.2 Recent research developments

Most of the research efforts so far have been aimed at developing integrated or

harmonised evaluation frameworks to apply to projects, programmes or policies

that are of common interest or added value at an EU level. This is a complex task,

in part due to the lack of harmonisation with regards to transport data, forecasts,

models or scenarios. Another reason is largely political, and is related to the

demand for flexibility by MMSS in view of the subsidiarity principle (ICCR, 2002b).

Most of these research efforts to develop this integrated assessment

framework in the EU come from institutional and academic research. Among these,

the series of research programmes and other initiatives commissioned by the EC in

the last decades are of special interest. These are briefly summarized below.

2.3.3.2.1 RTD Framework Programmes

The first attempt to integrate the results of these research activities was made by

the Transport Investment Evaluation group in 1995. They reviewed the results of


- 24 -

several projects of the 3rd Framework Programme, in particular those from EURET9

(EC, 1996c). These projects examined appraisal practice in the roads sector in the

(then) twelve members of the EU. They were projects under the APAS10/STRATEGIC

programme that examined railways, inland waterways and nodal centres for

passengers and for goods, and the APAS/ROAD study that explored socioeconomic

evaluation methods.

This research line was followed by research projects such as CODE-TEN11

(Giorgi and Tandon, 2000a), and EUNET12 (ME&P et al., 2001), which constitute

examples of interesting attempts to derive methodologies applied to the strategic

assessment of corridor developments and large-scale projects. CODE-TEN

developed a comprehensive policy assessment methodology and accompanying

decision tools, paying particular attention to the spatial distribution of

environmental and socio-economic impacts. EUNET developed a DSS combining

CBA and MCA approaches, providing a wealth of information but no simple

recommendation as in a single mono-criterion approach. Both methods advocate

the combined use of evaluation techniques, and recommend the use of scenarios to

establish and analyse the spatial distribution of impacts, indirect effects and

network effects in the long-term.

The IASON13 project (Tavasszy et al., 2004) is a continuation of the research

carried out in this field. Its wider objective is to develop rules for the social CBA of

transport projects and policies, with a focus on indirect effects, such as the spatial

distribution of benefits and the impact on cohesion. The main output was the

development of an overarching assessment framework, which was applied to the

analysis of the TEN-T implementation. Besides, a useful review of methodological

advances in project assessment methodologies was also provided (Schade et al.,

2004).

The HEATCO14 project is aimed at developing harmonised guidelines for

project assessment and transport costing at EU level. The framework is based on

welfare economics and CBA, and it is tested by applying it to selected TEN-T

9 European Research on Transport.

10 Actions de Préparation, d’accompagnement et du suivi.

11 Strategic Assessment of Corridor Developments, TEN Improvements and Extensions to the CEEC/CIS.

12 Socio-economic and spatial impacts of transport infrastructure investments and transport system

improvements.

13 Integrated appraisal of Spatial Economic and Network Effects of transport investments and policies.

Funded by the 5th Framework RTD Programme. Website: http://www.wt.tno.nl/iason/.

14 Developing Harmonised European Approaches for Transport Costing and Project Assessment. Funded

by the Sixth Framework RTD Programme. Website: http://heatco.ier.uni-stuttgart.de.


- 25 -

projects and comparing the results with those of existing CBAs. In order to ensure

that impacts without a monetary estimate are not overlooked, it is suggested that

they are evaluated separately from the CBA, in a MCA (Bickel et al., 2005)

The TRANS-TALK15 Thematic Network (ICCR, 2002b) was set up with the

objective to provide a networking platform for those involved in the field of

transport evaluation; explore the conceptual and empirical problems in

contemporary transport evaluation; and develop common guidelines (ICCR, 2002a)

that help improve transport evaluation.

TENASSESS16 (Giorgi and Pohoryles, 1999) developed a Policy Assessment

Model (PAM), capable of determining the extent to which a project achieves a

predefined set of ten policy objectives.

In summary, several projects have sought to integrate the use of standard

evaluation techniques, like CBA and MCA, through a decision framework analytical

approach. In conclusion we could state that at the European level there is a gradual

movement towards the better strategic incorporation of policy concerns in

evaluation and that this is reinforcing attempts to better co-ordinate evaluation and

policy-making. This, in turn, necessitates a better understanding of evaluation

techniques and of possible ways to integrate their results (ICCR, 2002b).

2.3.3.2.2 The ESPON

The European Spatial Planning Observation Network (ESPON)17 was set up to

increase the general body of knowledge about territorial structures, trends and

policy impacts in an enlarged EU. ESPON finances and monitors applied research

projects. Among them, ESPON Project 2.1.1.: ‘Territorial impact of EU transport

and TEN policies’ (Bröcker et al., 2004) is of particular interest.

This project developed methods to assess to what extent the TEN-T supports

territorial cohesion and a polycentric and better balanced EU territory, according to

the ESDP (EC, 1999). These methods were applied for the evaluation of different

scenarios of EU transport policy, investigating, in particular, their effects on

‘regional development potential’ and polycentricism. The results revealed three

fundamental policy goals between which trade-offs may appear: economic

efficiency, spatial equity and environmental sustainability. Furthermore, the project

15 Thematic Network Project and Policy Evaluation Methodologies in Transport. Funded under the 5th

Framework Programme. Website: http://www.iccr-international.org/trans-talk.

16 Policy assessment of TEN and Common Transport Policy. Website: http://www.iccr-

international.org/research/projects/tenassess.html.

17 Part-financed by the European Union within the Interreg III Programme. Website

http://www.espon.eu.


- 26 -

also identified four issues which require further research: socio-economic impacts,

cohesion, polycentricism and governance.

2.3.3.2.3 The UTS study

Another relevant contribution is the UTS18 study (Chatelus and Ulied, 1996),

followed by the work of Turró (1999), which is a global strategic territorial

assessment of TEN-T based on a set of territorial goals (see Table 2.1). It aims to

present relevant territorial information to DMs, helping them to optimise the

process of placing the TEN-T planned infrastructures on the territory and their

potential spatial development impacts.

Table 2.1: Consideration of TEN-T territorial goals suggested in the UTS study

COMPETITIVENESS

� Support of already existing development trends.

� Bridging capacity bottlenecks to the existing demand.

COHESION19:

� Encouragement of new development opportunities.

� Pulling demand growth in/between peripheral areas.

SUSTAINABILITY:

� Induction of environmentally friendly development patterns.

� Pushing mobility growth expectations towards more environmentally

friendly transportation modes.

Source: Chatelus and Ulied (1996)

The study is focused on a long-term and trans-European view. Therefore,

short-term impacts, partial mobility aspects or local considerations are taken into

account only when having global implications. The UTS methodology considers

quantitative models as tools to indicate significant issues and to test alternative

scenarios, rather than to predict the future evolution of the whole European

territorial system. It advocates the separate consideration of competitiveness,

sustainability and cohesion criteria in the assessment procedure, as Table 2.1

details.

18 Union´s Territorial Strategies Study Linked to Trans-European Transport Networks study.

Commissioned by DG-VII. Website: http://www.mcrit.com/UTS.

19 The cohesion concept can be extended to "territorial integration" by considering the induction of trans-

national relations.


- 27 -

2.3.3.3 National official methodologies

Some commonalities exist between national official methodologies, although most

of them have been designed for infrastructure assessment at the project level. In

general, CBA is used most widely for measuring direct transport impacts, for

prioritisation and ordering of projects, and for selecting projects from amongst a

given set of alternatives (Nellthorp et al., 1998; Bristow and Nellthorp, 2000; Steer

Davies Gleave, 2004; ICCR, 2002b).

In most countries the overall appraisal embraces not only the CBA result but

also some form of qualitative appraisal of the social, economic and environmental

effects (ICCR, 2002b; Rothengatter, 2005; Annema et al., 1999), in order to derive

more integrated assessment frameworks (Bickel et al., 2005).

This Section outlines the basic features of selected national official

methodologies that follow this tendency. A review of current national assessment

methodologies can be found in Bristow and Nellthorp (2000), Bickel et al. (2005),

and ICCR (2002b).

2.3.3.3.1 Germany

The German procedure (BMVBW, 2002) contains, in addition to a CBA, two

independent modules treated by means of a MCA. First, the environmental risk

assessment module (ERA) includes a qualitative appraisal of spatially related

environmental risks and possible conflicts with European nature conservation that

have not already been taken into account within the scope of the CBA. Second, the

spatial impact assessment (SIA) module, represented in Figure 2.3, appears as a

means to remove the regional planning component from the CBA system and to

evolve it as an independent component with comprehensive objectives and criteria,

namely “distribution and development objectives”20 and “relief and modal shift

objectives”21. Besides, a multiplier to international projects is also applied22. This

factor aims at taking into account contributions to the promotion of European

integration. For example, to evaluate the contributions to the distribution and

development objectives, the German procedure awards 1 to 5 ‘regional planning

20 To provide population with technical infrastructure throughout the country and for balanced

accessibility conditions in the regions and across modes, and the creation of locational conditions for

economic development.

21 To improve the conditions for a modal shift to environmentally friendly modes of transport in areas

and corridors of high traffic density, and in local built-up areas.

22 By awarding a separate bonus of a maximum of 10% to time and cost savings allocated to cross-

border traffic.


- 28 -

points’ depending on the combination of the accessibility deficiency and structural

backwardness features, as outlined in Table 2.2.

The German method constitutes an example of an integrated assessment

procedure, including spatial employment, equity and international effects, which

goes beyond the pure efficiency-oriented measurement of generalised costs. This

type of integrated assessment methodologies, although with the risk of some

double-counting of effects, are expected to complement CBA in the future

(Rothengatter, 2005), in order to provide better support for DMs.

Figure 2.3: Outline structure of the German spatial impact assessment module

Source: BMVBW (2002)


- 29 -

Table 2.2: Accessibility categories (left) and evaluation matrix for distribution and development objectives (right) of the German procedure

Source: BMVBW (2002)

2.3.3.3.2 United Kingdom

In the United Kingdom, the Government’s Ten Year Plan for Transport (DfT, 2000),

published in July 2000, provided a strategy for investment in infrastructure and

other policies for the period 2000-2010. The current procedure has been called the

New Approach to Appraisal (NATA) (DETR, 1998), which has at its core an Appraisal

Summary Table (AST), introducing previously excluded elements from the former

Cost Benefit Analysis procedure in a more formal manner, but retains it as one,

perhaps the key, element (Vickerman, 2000). The AST has five main criteria,

environmental impact, safety, economy, accessibility and integration, each of which

has a number of subcriteria.

The NATA represents a move to endow non-monetary criteria with an

importance and formality similar to those criteria traditionally included in the

standard CBA method, but it lacks guidance to DMs as to how the multicriteria

information about alternative projects should be used to identify the preferred

option. This could lead to a lack of clarity, consistency and accountability in a

crucial part of the decision-taking process, despite the care taken to assess all the

various impacts of the alternatives (Sayers et al., 2003).

2.3.3.3.3 France

In France, the process of project assessment combines a quantitative, strict

application of CBA approach, with a qualitative, rather loose use of an MCA

approach (Quinet, 2000; Sayers et al., 2003). Current research is focused in long-

term considerations; the redistributive effects between territories and individuals;

the conciliation of spatial equity and economic profitability; the impacts of


- 30 -

improvements in accessibility on the more distant localisation of housing from

activities; risk, irreversibility and cumulative impacts (Seligmann, 2005).

2.3.3.3.4 The Netherlands

In The Netherlands all major national infrastructure projects have to be given the

OEI23 approach (Annema et al., 1999). The OEI is a new, relatively sophisticated

and integrated version of the CBA. It is more advanced and comprehensive because

it also covers wider safety, environmental and other impacts, and emphatically

avoids providing highly suggestive and arbitrary final CBA ratios, but instead aims

to give overviews of relevant social effects (Stoelinga and Luikens, 2005).

2.3.3.3.5 Scandinavian countries

The current planning evaluation procedure of the Scandinavian countries (Norway,

Sweden, Denmark) has a ‘strategic nature’ (Lauridsen, 2003). Scandinavia

currently applies a new ‘third generation national transport planning systems’. This

approach, considered the most relevant for strategic transport planning, is

objective-oriented and cross-sector. This implies that planning is seen as a problem

solving process that will respond to a set of goals, objectives and criteria which, to

some extent, may be contradictory. Nevertheless, it aims to achieve these in the

best possible way.

2.4 The role of evaluation in decision-making

At the Plan level, decision-making is the result of an interaction between many

actors influenced by a complex environment, in which many other considerations

than the evaluation/appraisal results affect the final decision taken, as shown in

Figure 2.4.

Conventional decision theory has adapted in order to be able to deal with

conflicting behaviour of the increasing number of stakeholders involved (Pearman

et al., 2003). A communication and learning process between planners and DMs

may therefore be necessary (Nijkamp et al., 1990), in order to develop and employ

flexible decision-support systems, as Section 2.3.2.2.4 stressed. In this context, it

is seen as a healthy trend to find DSS focusing on finding ‘good’ instead of

23 Onderzoeksprogramma Effecten Infrastructuur (Research Programme on the Impacts of Investments

in Infrastructure), introduced by the Ministry of Transport, Public Works and Water Management and the

Ministry of Economic Affairs.


- 31 -

‘optimal’ solutions, and on supporting the entire decision-making process from

problem structuring through solution implementation (Dyer et al., 1992).

Figure 2.4: Considerations affecting the decision-making process

Source: Guitouni and Martel (1998)

Decision-making today is no longer seen as an intellectual process, but as a

socio-political and organizational process, where the interest has shifted from the

quality of the decision towards the quality of decision-making (Voogd, 1997).

Furthermore, it is also argued that behavioural convergence is far more important

than mathematical convergence (Dyer et al., 1992).

Generally speaking, decision-making could be grouped into three broad

behavioural categories, depending on the objective of the evaluation procedure, i.e.

‘optimizing’, ‘satisficing’ and ‘justificing’ categories (Nijkamp et al., 1990). The

many conflicting aspects to be handled means that today, in general, the decision is

no longer an optimal one but a satisfactory one (Guitouni and Martel, 1998).

Furthermore, in European (or national) policy practice evaluation, although a

rigorous approach for technical appraisal is a vital input to the decision-making

process, the decision is ultimately a political one (EC, 1996c). Appraisal is at this

level often used as a means of justifying decisions, even if the actual decisions are

not in agreement with optimizing or satisficing principles (Tsamboulas et al., 1998).

Therefore, the exact role of the technical appraisal in the process of decision

making is a controversial issue. It depends partly on the quality of the appraisal

process, and partly on the roles assigned to the planner and the politician in the


- 32 -

decision process (Grant-Muller et al., 2001). Evidence from recent applications of

MCA of large scale infrastructure projects have shown that appraisal results played

only a minor role in the political discussions (Voogd, 1997).

If appraisal is seen as a tool to assess only value for money, the decision

could be directly derived from the result of a CBA. This approach could be valid only

if two conditions are met: that all relevant effects can be measured as monetary

equivalents and DMs are fully agreed on those measurements (Small, 1999).

However, at the Plan level, there is evidence that wider strategic issues are

increasingly more important in decision-making than the CBA results from

appraisals (Steer Davies Gleave, 2004). Thus it is clear that any CBA must be

extended through consideration of MCA, but more importantly complemented by a

brainstorming/discussion on key issues (ICCR, 2002b). Furthermore, it is argued

(Beuthe, 2002) that the role of MCA is reduced to ranking projects, leaving the task

of choosing among them to the DMs.

Finally, the presence of risk or uncertainty adds complexity in the decision-

making stage. Uncertainty has always been, and remains, a key concern in

appraisal. The always arbitrary, weighting of judgement criteria (Beuthe, 2002;

Tsamboulas et al., 1998; Keeny and Raiffa, 1976; Beuthe and Scanella, 1998), as

well as the uncertainty caused by the technical assumptions of the evaluation

method used (Tsamboulas et al., 1998; Voogd, 1997) are two of the most

important sources of uncertainty present at the decision-making stage.

Uncertainty is usually only dealt with performing a sensitivity analysis in

order to explore their dependency on assumptions taken during the definition of the

decision environment (Tsamboulas et al., 1998). However, a sensitivity analysis

does not, in principle, completely solve the DM’s problem (Beuthe et al., 2000).

Ideally, a preliminary research of the possible future scenarios and of the possible

range of variation of the different impacts and of their associated probabilities,

along with a thorough analysis of all the factors which affect the outcome of a

project should be conducted (Beuthe, 2002).

In summary, recent approaches tend to consider that appraisal is not a

substitute for political decisions (ECMT, 2004; Lauridsen, 2003; Small, 1999).

Rather they constitute a useful tool for DMs, presenting them with the information

they need to make an adequately well informed decision so that they can make

their implications more transparent, in a context in which uncertainty issues should

be recognized. Appraisal results are therefore seen as the starting point for

negotiation and deliberation; providing a tool for reflection and discussion by


- 33 -

planners and the numerous political DMs (Beuthe, 2002; Tsamboulas et al., 1998;

ICCR, 2002a).

2.5 Conclusions

This Chapter is a review of the recent evolution of the planning framework focusing

on transport Infrastructure Plans. This evolution has prompted significant changes

in the planning approach, which in turn has resulted in a number of challenges for

the appraisal community (Mackie and Nellthorp, 2003; Pearman et al., 2003;

Grant-Muller et al., 2001; ICCR, 2002b; Voogd, 1997; Voogd and Woltjer, 1999).

These changes and challenges are briefly summarized below:

� The planning framework has witnessed a growing importance of

‘communicative’ and ‘consensus building’ issues. This is mainly due to the

increase in the number of stakeholders and government structures involved,

with potentially conflicting interests, the growing importance of public opinion

and an observed greater social awareness on the impact of large transport

infrastructure investments. As well as finding ways of improving the quality of

technical appraisal, it is also needed to find ways of communicating its meaning

effectively. All this calls for a more ‘transparent and easy-to-explain’ planning

process.

� Regarding the aim of the evaluation procedure, appraisal results are

increasingly required to act as a starting point for negotiation and deliberation

between planners and DMs, rather than the end of the planning process. It is

not that important to find a ‘single best' solution, but to provide the DMs with

the information they need in order to take a decision.

� The paradigm for appraisal used to be that of a single project for a single mode

has now moved towards strategic transport policies covering all modes. This

move has prompted the need for integrated assessment methodologies related

more securely to the overall objectives of transport policy, such as

environmental, economic development and equity issues. The research

challenge is now to derive appropriate and harmonized indicators or procedures

to measurements of the achievement of these wider policy objectives. Besides,

many of these impacts readily double-count with direct impacts, which provides

a further challenge.

� There is a growing importance in the definition of alternatives and judgment

criteria. This is to the detriment of the mathematical structure of the appraisal

methods, which can be considered as a sign of the 'technical maturity' of the

latter (Voogd, 1997). Regarding the appraisal framework, there is a tendency to

see CBA and MCA as complementary rather than competing approaches, but


- 34 -

there is no harmonized procedure to integrate their corresponding evaluation

outputs.

� Furthermore, at the Plan level, the existence of hierarchical and sometimes

overlapping Government structures (EU, national, regional, local) requires the

inclusion of spillover effects and distributive issues; mainly under the ‘cohesion’

objective, as part of an integral assessment methodology. However, both of

them are usually lacking in national official methodologies, due to the fact that

there is limited knowledge on how handle them at the appraisal stage, even at

the level of indicators or qualitative assessment.

� In this context, spatial impact analysis tools are specially suited for impact

analysis at the Plan level. Increased computing power, facilitating increased

capability for information processing and presentation of results, such as GIS

have supported technical improvements.

In summary, modern transport planning has been forced to adapt to this new

planning framework and has become significantly more flexible: today, one is more

likely to talk about a ‘framework of analysis for decision-making rather than an

appraisal methodology in the usual sense’ (Beuthe, 2002). Spatial impact models

and GIS have a major role to play towards this adaptation of assessment

methodologies. They are both described in the next Chapter.

Chapter 3 – SPATIAL IMPACT ANALYSIS TOOLS

- 35 -

3. SPATIAL IMPACT ANALYSIS TOOLS

Chapter 2 concluded that spatial impact analysis techniques are especially suited to

support all the stages of the planning process of Transport Infrastructure Plans.

This Chapter goes one step forward and includes a review of the main spatial

impacts present at the Plan level, along with a description of spatial impact models

and tools. The objective is to draw the line between the theoretical foundations of

spatial impact analysis in strategic planning and recent methodological advances

(section 3.1), in order to give scientific evidence of the potential of accessibility

indicators as a tool to measure spatial impacts of transport infrastructure Plans

(section 3.2). These issues are complemented with a description of Geographic

Information Systems (GIS) (section 3.3), the computation software used in most

spatial impact applications. Finally, conclusions are included in section 3.4.

3.1 Spatial impacts at the Plan level

3.1.1 Theoretical foundations of spatial impact analysis

Spatial impact analysis can be considered as a specific type of system analysis

(Buckley, 1967; Meyer and Straszheim, 1971), which has two distinctive features

that make it specially suited as a transport planning tool, i.e.: it refers to a

multiplicity of policy sectors, and it deals with an open system, so that spatial

spillover effects and multi-level effects can be considered. The compound

representation of a spatial impact system suggested by Nijkamp et al. (1990) is

included in Figure 3.1, in which A represents the set of attributes of each of the n

profiles characterizing the successive parts of a spatial system, and B represents

the policy measures of each of the j policy fields which constitute part of the

environment of the spatial system. The existing relationships and interactions are

represented by S –if between all elements within the spatial system- and R –if

between elements within and outside the spatial system, i.e. the responses of

spatial systems to external policy measures.


- 36 -

Figure 3.1: Simple representation of a spatial impact system

Source: Nijkamp et al. (1990)

Spatial planning models attempt at representing the system in Figure 3.1. in a

simplified way. There is a long tradition of theoretical approaches from different

disciplines which attempt at modelling the complex interactions of the above spatial

system1. These specialised models come from different fields such as economics,

geography, transport engineering or environmental science.

Since the 1960s these models were integrated by ‘synthetic’ disciplines such

as regional science or planning (Wegener, 2001; Miller, 1999b). This is particularly

true as transportation systems are becoming increasingly integrated, global-scale

and multi-modal. Furthermore, the inclusion of the sustainability approach is

increasingly demanding the use of integrated spatial models (Wegener, 2001), in

which two or more specialised models are combined.

These past few decades have witnessed a major resurgence and

enhancement of the traditional subfield of spatial models, partly due to the

development of Geographical Information Systems (GIS), as will be discussed in

Section 3.3.2.

3.1.2 Impact analysis at the Plan level

There is a wide variety of sets of headings of possible impacts of transport

Infrastructure Plans ( for reviews on this topic see EC, 1996b; Nijkamp et al., 1990;

Nellthorp et al., 1998; Bristow and Nellthorp, 2000; ICCR, 2002; Schade et al.,

2004). In general, they may have a range of dimensions in time (from short to long

1 For definitions and existing classifications of spatial models see Wegener, 2001;Nijkamp et al., 1990.


- 37 -

term), space (from local to global), and by sector. The majority of literature reviews

agree in their classification into direct transport, strategic environmental and wider

policy impacts. Each of these impact categories is described in the following

subsections.

3.1.2.1 Direct transport impacts

Direct transport impacts usually include components such as construction and

maintenance costs, operating cost changes and travel time savings; (for reviews on

direct impact categories see Nellthorp et al., 1998; SACTRA, 1999; Bickel et al.,

2005).

On this impact category much technical work has been done so far.

However, although the principle of valuing the direct transport impacts in monetary

terms is generally accepted, the conventions and values used differ between

countries (Grant-Muller et al., 2001; Bristow and Nellthorp, 2000; EC, 1996c).

3.1.2.2 Strategic environmental impacts

This category includes a wide spectrum of global effects, such as habitat

fragmentation or global warming. There is a lot of work still to be done in defining

the indicators to measure these impacts, mainly due to the difficulty of setting

targets for the environmental dimension (Hey et al., 2002; Banister et al., 2000).

However, their assessment at the Plan level is increasingly required by DMs

because of their significant ‘sustainability’ implications.

Methodological issues in valuing environmental impacts constitute a

sensitive area in current research, mainly aimed at finding their equivalent money

values (Grant-Muller et al., 2001), which still show a very wide range of variation

among MMSS (Nellthorp et al., 1998; ICCR, 2002; Bickel et al., 2005) or where

there is even a disagreement over the legitimacy of monetizing such impacts

(Bristow and Nellthorp, 2000). Some authors even argue that quantifying certain

environmental effects in monetary terms may add considerable uncertainty to the

resulting evaluation (Small, 1999; Beuthe, 2002).

3.1.2.3 Wider policy impacts

The measurement of wider policy impacts is one of the priorities of current research

agenda, especially in the field of spatial economic impacts (for existing reviews on

the topic see Banister and Berechman, 2003; Rietveld and Nijkamp, 1993; SACTRA,

1999; Oosterhaven and Knaap, 2003; Evers et al., 1987).

This is an impact category where there is a wide variety of assessment

approaches. Wider policy impacts are also frequently referred to as secondary


- 38 -

benefits (Adler, 1987), indirect socio-economic effects (Bristow and Nellthorp,

2000) or wider economic effects (SACTRA, 1999). They include ‘intangible benefits’

(Adler, 1987), such as a more equal distribution of income or a more effective

international integration, spatial socio-economic and economic development effects,

or the contribution to economic and social cohesion.

Despite this lack of sound and commonly agreed methods to assess these

wider impacts, they are increasingly required to be included in assessment

methodologies at the Plan level (Grant-Muller et al., 2001; Bristow and Nellthorp,

2000). Subsection 3.1.3 justifies the need for a complementary analysis to handle

this requirement.

3.1.3 The treatment of wider policy impacts at the Plan level

The assessment of wider policy impacts is frequently carried out with the support of

spatial impact models and subsequently included as a complementary analysis to a

‘conventional’ appraisal method, such as CBA (see e.g. Salling et al., in press;

Tsamboulas et al., 1998; BMVBW, 2002; INRETS, 2005). This complementary

analysis enables a wider view to be taken of the investment proposal and therefore

it is claimed that it should become an integral part of all evaluations at strategic

levels (Banister & Berechman, 2003; Banister & Berechman, 2001; Beuthe, 2002).

Furthermore, it is argued that this more complex type of analysis seems to be

increasingly important where there is already a high quality transport network, as

the ‘conventional benefits’ may be providing an ever decreasing proportion of the

total returns (Rietveld and Nijkamp, 1993). According to a proposal by Banister and

Berechman, (2003), and as represented in Figure 3.2., this complementary analysis

would include the assessment of three impact categories:

� Network effects: measurement of the contribution of the concerned

infrastructure improvement to the transport network as a whole, evaluating

issues such as ‘network integration’ or ‘network efficiency’.

� Value added: mainly economic development effects, including long-term indirect

changes in income, factor productivity and employment.

� Distributional impacts: analysis of the distribution of impacts among regions

and/or social groups.

The detailed description of the aforementioned impacts, along with existing

scientific approaches for their measurement are described in subsections 3.1.3.1 to

3.1.3.3.


- 39 -

Figure 3.2: Suggested twin approach to transport appraisal

TRADITIONAL CBA

•User transport benefits

•Cost of investment

COMPLEMENTARY ANALYSIS

• Network effects-Accessibility analysis

• Value added from project

•Changes in employment

•Changes in factor productivity

• Distributive impacts

PROJECT

TRADITIONAL CBA

•User transport benefits

•Cost of investment

COMPLEMENTARY ANALYSIS

• Network effects-Accessibility analysis

• Value added from project

•Changes in employment

•Changes in factor productivity

• Distributive impacts

PROJECT

Source: Adapted from Banister and Berechman (2003)

3.1.3.1 Network effects

The addition or improvement of a single link of the transport network can

significantly affect demand on competitive and complementary links, therefore

changing interconnections and the resulting patterns of network usage and

performance. These effects on the performance of the transport network as a whole

are termed ‘network effects’ (Chatelus and Ulied, 1996). Network effects are

therefore related to issues such as ‘network efficiency’, (Gutiérrez and Monzón,

1998), ‘network synergy’ (Capello and Rietveld, 1998; Capineri and Kamann, 1998)

or ‘network integration’ (Banister et al., 1999; Turró, 1999; Peters, 2003).

The importance of assessing network effects is obviously higher at strategic

planning levels. For example, at the EU level, the existence of network effects is the

basis of transport initiatives such as the TEN-T and certain transborder projects

(EC, 2004c; High Level Group of the Trans-European Transport Network, 2003).

The concept of network effects is also intimately linked with that of spillover

effects (Pereira and Roca-Sagales, 2003), i.e. effects that occur in any assessment

methodology outside its corresponding study area. At the EU level, initiatives such

as the TEN-T involve significant spillovers between MMSS (Bröcker et al., 2004),

which have been measured through concepts such as the ‘European added value’

(van Exel et al., 2002) or the ‘community component’ (Roy, 2003) of certain


- 40 -

projects. Clear examples are the national sections of transborder projects2, where

the principle of subsidiarity forces the assessment to be carried out separately

within each country. The exclusion of these effects is argued to make the

profitability of transborder projects, and therefore the public financing they require,

to be systematically under-estimated relatively to purely national projects (Roy,

2003). Moreover, their exclusion also underestimates the benefits related to the

opportunities that transborder projects provide for the promotion of the parallel

development of cross-border regions, which is of crucial importance both for the

EU3 and for particular MMSS (Ollivier-Trigalo, 2001; EC, 1999; EC, 2004b; Turró,

1999).

Hence, it is not surprising to find that so far the analysis of network effects

has been of interest mainly at this EU level (for studies on the subject see van Exel

et al., 2002; Turró, 1999; Pearman et al., 2003; Laird et al., 2005; Bröcker et al.,

2002; Bröcker et al., 2004; Frybourg and Nijkamp, 1998). However, at the national

level, these effects are not usually included in official assessment methodologies

(Bristow and Nellthorp, 2000; Grant-Muller et al., 2001), although they may be of

crucial importance in the assessment of national transport Plans (Condeço and

Gutiérrez, 2006; López et al., 2006a).

3.1.3.2 Value added

This impact category refers mainly to economic development effects, including

those long-term indirect changes in income, output, productivity, and employment,

which are induced by the new opportunities offered by an improvement of the

transport network. The relationship between transport infrastructure improvements

and regional economic development has been and still is a controversial issue for

the research community and the subject of much theoretical and political debates4.

Most authors argue that, under certain market conditions, transport accessibility

improvements can potentially trigger several major positive externalities, which, in

turn, can boost productivity, reduce production costs and promote more efficient

use of resources and, collectively, bring about additional economic development

2 An example of a transborder project is the PBKAL (High-Speed Rail Project Paris-Brussels-Cologne-

Amsterdam-London). The solution for this case was to produce for the EU an European evaluation to

complement the national evaluations of the participating MMSS with a ‘community component’ (Roy,

2003), which provided a rational basis for determining the appropriate level of EU subsidy.

3 Initiatives such as the INTERREG include specifically the objective of cross-border cooperation (EC,

2004b).

4 See Oosterhaven & Knaap, 2003; Banister & Berechman, 2003; Vickerman et al., 1999; Button 1993;

EC 1996a; Banister & Berechman, 2001; Rietveld & Nijkamp, 1993; SACTRA, 1999.


- 41 -

benefits. These benefits must be in addition to the primary accessibility

improvement benefits and not merely their market capitalisation. This is the

approach followed in the analytical framework developed by Banister and

Berechman (2003), which highlights the idea that the main output from a transport

investment is network accessibility improvement. Subsequently, two additional

effects may arise:

� Activity spatial redistribution, which, if it ensues, may improve spatial patterns

and economic efficiency,

� Economic development, predicated on the presence of certain market conditions

or ‘allocative externalities’5.

There is consensus in that only in the presence of these positive externalities a

transport project can potentially promote regional economic development (Beuthe,

2002; Banister and Berechman, 2003; SACTRA, 1999). If these externalities are

not present, adding accessibility benefits and potential development benefits would

amount to substantial double-counting of benefits.

Furthermore, there are many other factors influencing the final economic

development effect. One of them is the quality of the transport network, as the

magnitude of the effect seems to depend strongly on the already existing level of

accessibility (Bröcker et al., 2004). It is in regions with low infrastructure qualities

that one expects the highest impacts of infrastructure investments on regional

development, (Button, 1993; Evers et al., 1987; Rietveld and Nijkamp, 1993). In

countries with an already highly developed transport infrastructure further

transport network improvements bring only marginal benefits (Vickerman et al.,

1999). Moreover, the alleged linkage between accessibility improvements and

economic development is being lessened by current trends, as stressed by Banister

and Berechman (2003), namely:

� The declining role of accessibility improvements in the contemporary economies

of cities and regions (Copus et al., 2002),

� the marked change in the relative importance of work related trips and the

increased complexity of commuting patterns6,

5These include a favourable market environment, availability of funds, and supporting legal,

organisational and institutional policies and processes. Under these conditions, the non-compensatory

action of one economic entity can affect the utility level of another, which in turn, can affect the efficient

allocation of resources in the economy (Banister and Berechman, 2003). Traffic congestion is an

example of negative allocative externalities, whereas agglomeration of firms represents positive ones.

6Apparently, the market operates through the relocation of firms and households to achieve the balance

of keeping commuting times within tolerable limits (Banister and Berechman, 2003).


- 42 -

� the growing importance of information and communications technology in

improved production and distribution processes, compared with transport,

� the concern of reducing travel as a means to achieve the objectives of

sustainable development, which links environmental, economic development

and equity arguments.

Finally, the spatial level of analysis is a key variable influencing the results. If the

analysis is carried out at the highest level of spatial aggregation it may appear that

there is no significant impact on economic growth, while it is possible that spatial

effects do exist but that these relate to distribution effects within regions (Bruinsma

et al., 1997). This issue is discussed in Section 3.1.3.3.

In an attempt to move forward in this research field, there has been in

recent decades an upsurge of different, and sometimes contradictory,

methodological attempts to describe parts of the transport-economy system7. Key

selected models include production functions, (Biehl, 1986; Blum, 1982; ME&P et

al., 2001); accessibility models (Keeble and Owens, 1982; Schürmann et al., 1997;

Bröcker et al., 2002); land-use transport interaction (LUTI) models (van Wee,

2002); and spatial computable general equilibrium (SCGE) models (Salling.K.A. et

al., in press; Laird et al., 2005).

3.1.3.3 Distributive impacts

Improvement of transport infrastructure leads to a reduction of transport costs

which may give rise to substantial redistribution effects among economic groups

and also among regions. (Rietveld and Nijkamp, 1993). This issue is linked with the

trade-off between ‘generative vs. distributive growth’ (Rietveld and Nijkamp,

1993), ‘efficiency vs. equity’ (Bröcker et al., 2004; Feng and Wu, 2003), or

‘competitiveness vs. cohesion’ (EC, 2004a; Gutiérrez, 2004) effects of transport

infrastructure. The three terms distributive, equity and cohesion impacts are used

as almost synonyms in the literature.

Distributive impacts do not refer to the global effect of transport investment,

such as global improvement in accessibility levels, but to its distribution among

regions, frequently addressed as regional equity or cohesion effects, or groups of

individuals, in this case under the social perspective on equity or cohesion

(Schürmann et al., 1997).

7 It is not the purpose of this subsection to describe the theoretical aspects of existing methodological

attempts to describe the complex network of transport-economy linkages (for studies reviewing relevant

methodologies, along with evidence of their fundamental weaknesses and omissions see Nijkamp et al.,

1990; Oosterhaven and Knaap, 2003; Bröcker et al., 2002; Rietveld and Nijkamp, 1993; Aschauer,

1989).


- 43 -

Equity motivations have provided the main justification for financing

infrastructure investments in peripheral and/or landlocked regions at the EU level,

as stated in different EU policy documents (EC, 2004a; EC, 2004c; EC, 1999), but

also at national planning levels (see e.g. the case of Germany (BMVBW, 2002), or

Spain (Ministerio de Fomento, 2005). However, their inclusion in appraisal

methodologies is uneven and scarce (Bristow and Nellthorp, 2000; Grant-Muller et

al., 2001), as most CBA studies concentrate on efficiency considerations. However,

it has been suggested that some allowance for distributional impacts should be

incorporated in CBA studies (Button, 1993), or in a MCA framework complementing

the CBA (Beuthe, 2002; ME&P et al., 2001; SACTRA, 1999; Banister and

Berechman, 2003).

The treatment of equity effects is a current challenge for spatial planning

models at strategic levels. Recent research approaches suggest analyzing

distributive impacts in terms of spatial equity impacts8, e.g. via changes in the

spatial distribution of accessibility among regions (Schürmann et al., 1997; Martín

et al., 2004; Bröcker et al., 2004; López, 2005; López et al., in press; INRETS,

2005). Results obtained from these studies show that certain investments may lead

to increasing rather than reducing regional disparities in accessibility, i.e. to a more

polarized distribution of accessibility.

3.2 The potential of accessibility analysis

The use of the accessibility concept has repeatedly been referred to in previous

sections as a useful planning tool which is increasingly included in spatial impact

models. In order to give the reader a more complete understanding of these

references, Section 3.2.1 includes a brief review of the concept of accessibility and

its measurement, while evidence of their potential for spatial impact analysis is

included in Section 3.2.3.

3.2.1 The concept of accessibility

There are many definitions of accessibility and many ways for measuring it. The

concept of accessibility came to the fore in the early 1950s and has a long tradition

in regional science and transport economics, where practical concepts of

accessibility have been widely used for the assessment of the impact of transport

policies on regional economic performance.

8 Distributive issues can also be analysed in terms of social impacts, which requires detailed classification

of the affected population (commonly in terms of low income, disadvantaged groups, etc.). These are

not easy to identify (Banister and Berechman, 2003), especially at strategic planning levels.


- 44 -

The first basic concept of accessibility related with the ease to reach goods,

services or activities from any given location. One of the first steps forward

appeared when accessibility was also considered in the sense of opportunity or

possibility that people in a certain location have to participate in specified activities.

More recent approaches incorporate social and economic aspects, as they identify

accessibility as the net benefit people of a specific location obtain from the use of

the existing transport and land use system. Still today the concept of accessibility is

still evolving with new approaches that continuously enrich the concept with new

connotations9. This subsection is not aimed at providing an exhaustive list of all

scientific contributions to the concept of accessibility, but a brief review of a

selective list.

Different authors (Gutiérrez, 2004; Martellato et al., 1998; Ney, 2001) have

suggested using the approach used in the accessibility analysis as a tool to review

existing definitions of accessibility. A selection of these approaches is described

bellow.

First, from an infrastructural approach. In this case, accessibility is

exclusively aimed at measuring the performance of the transport system in a

specified area; with accessibility measures such as network density or average

network speed.

Second, from a locational/geographical approach, accessibility is referred to

the degree of separation between locations. This is the approach followed by Morris

et al., (1979), who define accessibility as ‘some measure of spatial separation of

human activities, which denotes the ease with which activities may be reached

using a particular transportation system’.

Third, from a potential of opportunities approach, accessibility is related to

the volume of economic activity that can be reached from any given location,

following the Hansen, (1959) approach to accessibility as ‘the potential of

opportunities for interaction’, or ‘the possibilities of using the opportunities that the

economic, social, cultural and political facilities and institutions provide’ (Domanski,

1979).

Fourth, the utility approach (Koenig, 1980; Ben-Akiva and Lerman, 1979), is

founded in microeconomic welfare theory and it is related to the outcome

individuals obtain from the utilization of the transport system. The latter approach

9 For a review on the theoretical foundations of the concept of accessibility, see Morris et al., 1979; Ney,

2001; Bruinsma and Rietveld, 1998; Reggiani, 1998; Geurs and van Wee, 2004; Geurs and Ritsema van

Eck, 2001.


- 45 -

follows a different perspective, as accessibility is not defined as a characteristic of a

location, but that of individuals at a specific location.

The approach used to define the concept of accessibility has motivated aun

upsurge in different formulations for its measurement; i.e. different accessibility

indicators/measures. These are summarized in subsection 3.2.2.

3.2.2 The measurement of accessibility

3.2.2.1 Theoretical background

There is a wide spectrum of existing formulations which attempt to measure the

concept of accessibility. Extensive reviews and existing classifications of

accessibility indicators/measures can be found in Baradaran and Ramjerdi, 2001;

Bruinsma and Rietveld, 1998; Gutiérrez, 2001; García-Palomares, 2000; Handy and

Niemeier, 1997; Reggiani, 1998; Izquierdo and Monzón, 1992; Wegener et al.,

2000; Schürmann et al., 1997; Martellato et al., 1998; David Simmonds

Consultancy et al., 1998; Geurs and Ritsema van Eck, 2001. It is not the purpose

of this subsection to describe the theoretical aspects of accessibility measurement,

but to give an overview of most frequently used indicators. For this purpose, the

approach followed by Schürmann et al., (1997) has been selected from the above

list because of its flexibility to include a wide variety of formulations.

In Schürmann et al. (1997) accessibility indicators are classified according to

their complexity into two broad groups. Simplest accessibility indicators are those

‘infrastructure-based’ (Geurs and Ritsema van Eck, 2001) and only consider the

characteristics of the transport network of the area under consideration. The main

disadvantage of these indicators is that they fail to recognize that many

destinations of interest may lie far away from this area (Wegener et al., 2000;

Geurs and Ritsema van Eck, 2001). There are other types of accessibility indicators

which study the characteristics of the transport network as a whole, but these only

consider the topological properties of network, such as its connectivity. These are

called topological indicators (Mackiewicz and Ratajczak, 1996).

Transport literature is increasingly claiming a ‘paradigm shift’ from the more

traditional infrastructure-based measures towards more complex accessibility

indicators, also called ‘activity-based’ measures (Geurs and Ritsema van Eck,

2001). The common feature of these more complex indicators is that they take into

account not only the characteristics of the transport network but also those of the

land use system which the network is intended to connect. Despite their perceived

complexity, their added value is that they provide complementary information for


- 46 -

more comprehensive analyses, as they allow testing the efficiency of both land-use

patterns and transport network configurations and their interdependencies.

Schürmann et al., (1997) further classify the most frequently used

formulations of activity-based indicators. Their classification encompasses a great

variety of possible indicators in three generic types, as Equation ( 3.1 ) shows:

( ) ( )ij

j

iji cfWgA ⋅=∑ ( 3.1 )

Following this approach, the accessibility (A) of a given location i is a

construct of two functions, the activity function g, representing the activities or

opportunities (W) to be reached at given locations j, and the travel impedance

function f, representing the ‘effort, time, distance or cost’ (c) needed to reach

them. This is a general form of a gravity model, where the attractors are the

activities or opportunities in areas j, and the distance term is the spatial impedance

cij. The most frequent forms of g and f are represented in Figure 3.310. The different

combinations of both functions result in the corresponding different types of

accessibility indicators.

Figure 3.3: Activity and impedance functions

Source: Schürmann et al. (1997)

Each indicator has specific advantages and drawbacks, although a general

observation is that there seems to be a trade-off between ‘soundness’, i.e.

theoretical and empirical insights, and ‘plainness’, i.e. ease of understanding of

existing formulations (Bertolini et al., 2005; Wegener et al., 2000). Furthermore,

potential improvements in the theoretical foundations of most popular indicators

10 Examples of early formulations of f using are the utilization of a negative potential function (Hansen,

1959) or the negative exponential function conceived by Wilson (1971). Both formulations used distance

as the impedance variable.


- 47 -

may imply a loss of their interpretability and therefore they are not being

implemented in most practical accessibility studies (Geurs and Ritsema van Eck,

2001).

Further key improvements of activity-based measures mainly relate to a

dissagregation of individuals according to their socio-economic characteristics

(Bertolini et al., 2005) or the consideration of competition effects11 (Geurs and

Ritsema van Eck, 2001; van Wee et al., 2001). Another weakness comes from the

fact that an increase in destination masses yields a proportional increase in

accessibility, which does not take into account capacity constrains and congestion

risks (Martellato et al., 1998).

Another interesting perspective on accessibility is given by utility-based

measures, which offer an ‘economic’ background for the potential approach. Utility

based measures relate accessibility to the notion of consumer surplus in

microeconomic theory (see Koenig, 1980; Ben-Akiva and Lerman, 1979). This

approach requires accessibility to be measured at the individual level and to model

travel behaviour and the (net) benefits of the users of a transport system. Despite

its ‘methodological significance’ and the many theoretical studies on the subject

(see e.g. Martínez, 1995), these measures are rarely used in empirical applications

(Martellato et al., 1998; Geurs and Ritsema van Eck, 2001).

Finally, another approach is the utilization of space-time prisms (Miller,

1999a). These measures come from space-time geography and take into account

the availability of activities at different times of the day and the times in which

individuals participate in specific activities, given their time-budgets and

restrictions. The applications of these indicators fall beyond the scope of this thesis,

as their large data requirements forces practical applications to be restricted to

relatively small regions and small subsets of the population.

A description of the most frequently used indicators in long-range transport

planning studies, along with a summary of relevant applications are included in

subsections 3.2.2.2 to 3.2.2.4.

3.2.2.2 Travel cost indicators

There is a wide spectrum of formulations under this heading12. Most frequent

formulations measure total or average travel cost to a predefined set of

destinations, as in Lutter et al. (1992) and Schürmann et al. (1997). Other

11 Using inverse balancing factor of singly or double constrained spatial interaction models. A review of

existing approaches can be found in Geurs and Ritsema van Eck, (2001), or Martellato et al.,(1998).

12 Sometimes they are referred to as contour measures with fixed opportunities (Geurs and Ritsema van

Eck, 2001)


- 48 -

approaches weight destinations according to their size. An example is the location

indicator (Gutiérrez, 2001; Gutiérrez and Urbano, 1996), in which the population of

each destination is used as a weighting factor.

The results of these indicators are inevitable and heavily influenced by the

geographical position, and usually result in core-periphery patterns. This point can

be verified in Figure 3.4, which shows an example of the utilization of a travel cost

indicator for the calculation of European road accessibility13 in 1992.

Figure 3.4: Example of a travel cost indicator. Road accessibility 1992

Source: Gutiérrez and Urbano (1996)

Travel cost indicators are popular because they are expressed in familiar units and

they are easy to interpret, although they lack a behavioural foundation because

they ignore the fact that more distant destinations are visited less frequently

(Schürmann et al., 1997). Furthermore, their values depend heavily on the selected

set of destinations (Wegener et al., 2000).

Other approaches substitute the customary notion of travel cost with that of

network efficiency (Gutiérrez and Monzón, 1998; Monzón et al., 2005). This

substitution highlights the infrastructure effect from that of having a peripheral

geographic location. Figure 3.5 shows an example of an application of a network

13 The indicator is computed as the average road travel time to destinations over 300,000 inhabitants,

expressed in minutes.


- 49 -

efficiency indicator in Spain14, for the road mode, before (left) and after (right) the

implementation of the Spanish Transport and Infrastructure Plan (PEIT) (Ministerio

de Fomento, 2005). The indicator is computed as the ratio between real and ‘as the

crow flies’ travel times15 to main population centres. Resulting accessibility contours

show how high-accessibility corridors are concentrated along high capacity

networks, whereas extremely high accessibility values appear in the nodes where

this network converges. Least accessible areas are located outside these corridors,

particularly in mountainous regions. Accessibility contours do not follow core-

periphery patterns: central regions may have low accessibility, whereas peripherally

located regions with efficient connections with the road network may exhibit high

accessibility values.

Figure 3.5: Network efficiency. Road accessibility 2005 (left) and 2020 (right)

Source: López et al. (2006b)

Other examples of travel cost indicators are isochrones, which respond to

the simplest case of a travel cost indicator, where the impedance term is travel

time and only one destination is considered. Other types of indicators frequently

used in planning studies are situational– also called distance (Geurs and Ritsema

van Eck, 2001)- indicators, where the number of destinations are also limited to

14 The map shows predicted accessibility values for the road mode in Spain due to the adoption of the

Infrastructure Master Plan 2005-2020 (PEIT) (Ministerio de Fomento, 2005). This indicator was

previously used in the 2000-2007 Plan (PIT) and the 1993-2007 Master Plan (PDI) (Ministerio de Obras

Publicas y Transportes, 1993).

15 For a detailed description on how these travel times are computed see Gutiérrez and Monzón (1998)

and López et al. (2006).


- 50 -

one16, but a different destination is selected for each origin area, such as each

inhabitant must be able to reach a bus stop in 5 minutes or should have a bus stop

within 500 metres from their home.

3.2.2.3 Cummulative opportunities indicators

Cummulative opportunities, also called contour (Geurs and Ritsema van Eck, 2001)

indicators, measure the number of opportunities reachable within a given travel

time or distance.

One of the most frequent applications of contour measures at the interurban

scale is the daily17 accessibility indicator. Several accessibility studies have used

this indicator (see Schürmann et al., 1997; Wegener et al., 2000; Gutiérrez, 2001;

Martín et al., 2004; Vickerman et al., 1999). This indicator measures total activity,

usually in terms of population, jobs or GDP, reachable in a given travel time

threshold which is frequently set up between 3 and 5 hours.

Other applications at an international levels use higher time thresholds, such

as 8 hours (Dupuy and Stransky, 1996) or three days (Chatelus and Ulied, 1996).

At the urban scale these values are obviously different, and it is frequent to find

travel time thresholds of 30 minutes (Bertolini et al., 2005).

16 According to the early work of Ingram (1971), these indicators fall into the ‘relative accessibility’

measures, defined as those related to the degree to which two places or points are connected, in

contraposition to ‘integral accessibility’ measures, which take more than one destination into

consideration.

17 The concept of daily accessibility was developed by Törnqvist (1970), from the case of a business

traveler who wishes to travel to a certain city, conduct business there and return home in the evening.


- 51 -

Figure 3.6: Daily accessibility indicator. Daily accessibility by rail (1993)

Source: Spiekermann and Wegener (1994)

As an example, Figure 3.6 shows the three-dimensional rail accessibility

surfaces results for a daily accessibility indicator18, computed using raster-based

GIS technology (Schürmann et al., 1997). The map shows how the location of high

accessibility nodes is heavily dependant on the spatial distribution of both

population and railway stations. However, accessibility sharply declines with

distance from these nodes, giving rise to interstitial zones of low accessibility even

in highly accessible central regions, such as the Benelux.

These indicators share with the travel cost indicators their ease of

interpretability, e.g. ‘population reachable in three hours’, and the methodological

drawback of their heavy dependence on the ‘arbitrarily’ selected set of destinations

to be taken into consideration in the analysis. Furthermore, another disadvantage is

that improvements of travel time which do not reduce travel time bellow the

specified threshold do not lead to an improvement in accessibility (Geurs and

Ritsema van Eck, 2001). This drawback is solved if the accessibility measure could

allow a decreasing influence of each destination as travel time increases. This is the

case of potential accessibility measures, which are described in the next section.

18 Results correspond to the rail mode, with population as the mass activity variable and a travel time

threshold of five hours (Wegener et al., 2000).


- 52 -

3.2.2.4 Potential indicators

Potential indicators are the most popular type of accessibility indicators found in the

literature. The first attempt to use the potential concept to describe accessibility

was developed by Hansen (1959); see Equation ( 3.2 ):

∑=j ij

j

ic

WA ( 3.2 )

Hansen used population and distance as activity (W) and impedance (c)

variables, and introduced a parameter reflecting distance deterrence (beta) equal

to -1. In recent years there has been a widespread application of different

adaptations of this basic Hansen’s formulation, mainly in accessibility studies from

an economic perspective, generally under the assumption that accessibility

deficiencies of certain locations may have influenced their structurally lagging

situation (see e.g. Hansen, 1959; Keeble and Owens, 1982; Bruinsma and Rietveld,

1993; Bruinsma and Rietveld, 1997; Copus, 1999).

In general, potential indicators take into account both the size of

destinations and the travel cost to reach them, under the gravitational based

assumption that the attraction of a destination increases with its size and declines

with travel cost. Most frequent applications follow an exponential negative

impedance function of the form showed in Equation ( 3.3 ):

∑⋅−

⋅=j

c

ji

ijeWAβα

( 3.3 )

where all the terms are already known except for the agglomeration parameter α,

which is usually set up equal to one (Geurs and Ritsema van Eck, 2001; Schürmann

et al., 1997; Gutiérrez, 2001; Martín et al., 2004). The above formula is an

example of the general formula included in Equation ( 3.1 ), with linear and

exponential activity and impedance functions, respectively.

Although potential indicators are widely used in empirical applications, they

have certain limitations (Martellato et al., 1998; Bröcker, 1989; Bruinsma and

Rietveld, 1993), such as their aggregated approach, which implies taking into

account that all individuals in the same zone have the same level of accessibility;

the significant influence that the form of the distance decay function, which should

be tested in an empirical setting, has in the result; the treatment of the ‘self-

potential,’ i.e. the internal component of accessibility, measuring the contribution of

each area to its own accessibility19, or the fact that they are not measured in

familiar units, as it is the case of travel cost measures.

19 The treatment of the self-potential may have considerable influence in the resulting accessibility

values, especially in large agglomerations, as it has important implications for the study of


- 53 -

3.2.3 Applications in transport planning

3.2.3.1 Background

It is widely claimed that the potential of accessibility analysis for transport planning

purposes is not fully exploited (Halden, 2003; Geurs and Ritsema van Eck, 2003;

David Simmonds Consultancy et al., 1998; David Simmonds Consultancy et al.,

1998). Indeed, although the concept of accessibility has been widely reported in

geographical studies, it has rarely been used for policy evaluation. It had very little

practical impact on policies (Handy and Niemeier, 1997; Halden, 2002; Ney, 2001),

despite recent interesting attempts to draw formulations related to wider policy

objectives; (see e.g. the work by Bertolini et al., 2005 on the achievement of

‘sustainable accessibility’). Moreover, accessibility analysis has major presentational

advantages by describing the impacts of transport investment in terms that people

can easily understand (Halden, 2000), which is an added value given the increasing

influence of public opinion on these issues.

The lack of practical applications of accessibility analysis in transport

appraisal methodologies is mainly due to concerns about double counting of effects

(Beuthe, 2002; David Simmonds Consultancy et al., 1998; Geurs and Ritsema van

Eck, 2001), the perceived complexity of their formulations and their resulting

difficulty of interpretation (Geurs and Ritsema van Eck, 2003). Each of the

formulations of accessibility described in Section 3.2.2 is particularly suited to

address a certain transport planning problem. However, the selection of the

appropriate indicator for a particular case is a complex task. Moreover, there is

evidence that the formulation chosen, mainly the choice of the distance decay

function, has a strong influence in the results obtained (Baradaran and Ramjerdi,

2001). In general there is no single best ‘ideal’ indicator, but it is argued that the

analysis is enriched if a set of indicators is computed and their results analyzed in a

complementary way (see e.g. Gutiérrez, 2001; Martín et al., 2004; Schürmann et

al., 1997).

An analysis of the use of accessibility indicators for transport planning

purposes is given in the following sections. First, the use of accessibility

improvements as a policy goal by itself is analyzed in section 3.2.3.2.

Subsequently, and following the SD approach, the following subsections include

agglomeration economies (Bruinsma and Rietveld, 1993; Bruinsma and Rietveld, 1997). It requires the

computation of an ‘internal travel time.’ Although there is a wide variety of approaches that handle the

self-potential problem, it is still a controversial issue (for studies dealing with this issue see, for example

Bruinsma and Rietveld, 1997; Bruinsma and Rietveld, 1993; Frost and Spence, 1995; Bröcker, 1989).


- 54 -

evidence of the utility of accessibility indicators for the assessment of wider policy

impacts from the economic (subsection 3.2.3.3) and social (subsection 3.2.3.4)

sustainability dimensions. The use of accessibility analysis from an environmental

perspective is less frequent and therefore has not been addressed here (for one of

the few existing approaches see e.g. Sánchez and Zamorano, 2006).

3.2.3.2 Accessibility improvement as a policy goal

The need for plans to evaluate their accessibility impacts is emphasized in many

national planning policy guidance (see e.g. DETR, 2000; Scottish Executive, 2000;

Ministerio de Fomento, 2005). ‘Accessibility improvement’ as a policy goal by itself

has been widely included among the key objectives of transport infrastructure plans

and programmes. For example, in Spain the impact of transport infrastructure

improvements on accessibility was included in the evaluation of the Plan Director de

Infraestructuras 1993-2007 (PDI) (Gutiérrez and Monzón, 1998) and the recently

launched Plan Estratégico de Infraestructuras y Transporte (PEIT) (López et al.,

2006b). Another example is the UK case, in which ‘accessibility’ is one of the five

assessment criteria of the ‘New approach to appraisal’ (NATA) (DETR, 1998). At the

European level, accessibility impacts of TEN-T have been the subject of many

studies (see e.g. Gutiérrez and Urbano, 1996; Schürmann et al., 2004; Chatelus

and Ulied, 1996; Lutter et al., 1992).

3.2.3.3 Economic perspective

In most economic studies public infrastructure is measured according to the

amount of public capital stock. This implies some limitations, mainly that the quality

of public infrastructure in each region is only measured in terms of its infrastructure

stock and that spatial spillover effects are usually ignored. These problems may be

remedied using some measure of the economic accessibility of each region –the

economic potential model-, instead of the stock of infrastructure (Rietveld and

Nijkamp, 1993; Bruinsma and Rietveld, 1997; Oosterhaven and Knaap, 2003; de

Orellana-Pizarro, 1994). The economic potential measure gives an aggregate

measure of the market area of each region, resulting in a deceptively reduction in

potential accessibility as we move away from the centre (Vickerman et al., 1999).

Obviously, the potential accessibility formulation gives a more realistic

approximation of the increase in economically useful opportunities available to a

certain region that will result from a transport network improvement. Moreover,

using accessibility instead of transport stock makes it possible to show that not only

the region where the actual investment takes place will profit from improved

accessibility, but also to take into account network and spillover effects, since they


- 55 -

are capable of assessing the effects of a single link on a transport network

(Gutiérrez and Monzón, 1998; Halden, 2003; Geurs and Ritsema van Eck, 2001).

Accessibility analysis can be used to assess network effects of improved links in a

certain country on neighbouring countries (López et al., 2006a). Furthermore, it

can be applied to any transboundary project, i.e. referring to any administrative

boundaries, where competency issues are also of concerns for DMs (for recent

applications at the national level see e.g. Condeço and Gutiérrez, 2006; López et

al., 2006a).

Moreover, it is even argued that the wide establishment of comprehensive

composite accessibility methods may provide a more accurate assessment of the

economic value of transport improvements than the one derived from travel

demand models, which are frequently less than fully comprehensive at national

planning levels (Halden, 2003; David Simmonds Consultancy et al., 1998). Most

scientific contributions have attempted to provide evidence of the relationship

between accessibility and economic development (for a review of relevant

contributions see Rietveld and Nijkamp, 1993). However, the few satisfactory

empirical investigations of the role of accessibility as a means to promote regional

economic activities provide uncertain (and controversial) results (Beuthe, 2002).

Some examples of quite modest effects are impact of the motorway system on the

regional distribution of employment in 28 regions in the UK (Botham, 1983), the

analysis of the effect of the M62 in the UK (Dodgson, 1974), or the London M25

orbital (Linneker and Spence, 1996), the extension of the USA highway system

(Kau, 1976), a new railway line in the Amsterdam-Hamburg corridor (Evers et al.,

1987), the fixed link across the Great Belt in Denmark (Illeris and Jakobsen, 1991),

the relationship between accessibility and the attractiveness of Dutch cities for

situating economic activities (Bruinsma and Rietveld, 1997), or the investigation

carried out by Ozbay et al. (2003) on the impacts of accessibility changes on the

level of economic development of 18 counties in the New Jersey/New York region.

Moreover, accessibility indicators can be used as a tool for the economic

evaluation of transport planning initiatives, as reviewed by Geurs and Ritsema van

Eck (2001). This research work includes applications of accessibility indicators in

three approaches, following Bruinsma et al. (1997): a CBA approach, a production

function approach, and an employment approach. They are briefly described below.

3.2.3.3.1 Accessibility and CBA

In a social cost benefit analysis (CBA) several authors have shown that utility-based

accessibility measures can be used to estimate consumer surplus (Neuburger,

1971; Williams, 1976; Handy and Niemeier, 1997). Essentially, the idea is that a


- 56 -

measure of consumers surplus (expressed in utility) can be derived by taking the

natural logarithm of a potential accessibility measure with a negative exponential

distance decay function.

The use of utility-based measures in the measurement of consumer surplus

has advantages compared to standard CBA, as reviewed in Geurs and Ritsema van

Eck (2001). One of them is that it enables the measurement of both the benefits of

an improved transport system and changes in the land-use system. Another

strength is its potential to analyse equity aspects (i.e. it can be used to analyse

which individuals or groups of individuals living in certain locations benefit from

changes in accessibility), since the benefits of changes in accessibility can be

located to regions.

However, the theoretically correct use of utility-based measures for

measuring consumer surplus requires a transport model which properly forecasts

the combined land-use transport equilibrium, including land-use and transport

feedbacks (van Wee, 2002; Geurs and Ritsema van Eck, 2003).

3.2.3.3.2 Accessibility and the production function approach

In literature, most empirical studies using the production function approach have

estimated the effects of total public capital on economic growth. Much less the

impacts of transport infrastructure development. Very few of the impacts of

accessibility changes on productivity and economic growth (see the review on this

topic carried out by Geurs and Ritsema van Eck, 2001).

One of the few examples of a study on the relationship between

infrastructure, accessibility and economic development was carried out in the

context of the SASI project (Wegener et al., 2000) and its follow-up project EUNET

(ME&P et al., 2001) and IASON (Tavasszy et al., 2004), where an ‘extended’ SASI

model was used. The SASI recursive simulation model estimates the socio-

economic impacts of transport infrastructure investments and transport system

improvements, based on empirically derived production functions with activity-

based accessibility indicators as a variable, as outlined in Figure 3.7.


- 57 -

Figure 3.7: Outline of the SASI model

Source: Schürmann et al. (1997)

As an example of an application of the SASI model, Figure 3.8. shows changes in

GDP per capita as a result of the (at that time) planned TEN-T priority projects for

the year 2020.


- 58 -

Figure 3.8: Changes in GDP per capita as a result of the planned TEN priority projects

Source: Tavasszy et al. (2004)

Of particular interest are the Spanish studies carried out in the 1980s

(Izquierdo et al.,1980; SENDA 3, 1986) on the economic implications of

accessibility improvements, and the study by de Orellana-Pizarro (1994), who

developed a production function approach to assess regional economic development

effects of the Spanish Master Plan 1993-2007 based on changes in accessibility

values.

3.2.3.3.3 Accessibility and employment

Although employment effects are often considered important in the analysis of

infrastructure projects, especially in regions with chronic underemployment,

employment effects of a typical transport project are generally considered to be

temporary and distributive (Geurs and Ritsema van Eck, 2001).

There are few studies on employment effects based on the use of

accessibility indicators. An example is the study by Bruinsma et al. (1997) who

investigated the relationship between highway construction, potential accessibility

to employment and regional employment growth in the Netherlands, found no

simple mono-casual relationships. Another approach is the one followed by Ozbay

et al. (2006), based on the approach suggested by Berechman and Paaswell


- 59 -

(2001). They used an accessibility function, transportation data’ weighted travel

time, as an input to develop an employment function in terms of several socio-

economic variables. These functions were tested in the New York/New Jersey

metropolitan area and the Bronx County in New York, respectively. The main

results showed that accessibility parameters are statistically significant for the

employment function, but that changes in accessibility have rather modest effects

on net employment growth.

3.2.3.4 Social perspective

Despite the debate on the effects of accessibility on economic development,

detailed in Section 3.1.3.2, regional development studies have traditionally been

based on the assumption that the uneven spread of development is a function of

spatial inequalities in accessibility (EC, 1996b). Accessibility is therefore seen as an

added value of a location and an important factor of quality of life (Schürmann et

al., 1997), and in a sense a proxy for measuring welfare, if we accept that the

welfare of individuals is related with the ease which they can access essential

services (Hay, 1995).

Early examples of the use of accessibility to assess cohesion impacts date

back to late 1970s, such as the study by Domanski (1979), who relates the

increase of accessibility to spatial concentration. This author uses accessibility as a

measure to represent spatial equity, essentially by applying the potential formula to

a hypothetical spatial system. Under this general approach, accessibility is often

considered in regional planning as a means to economic activity and cohesion,

rather than a desirable good by itself (Vickerman et al., 1999).

Therefore, changes in the distribution of accessibility values among regions

may be used as a proxy for measuring regional cohesion; whereas differences in

accessibility among individuals or groups of individuals may play the same role in

assessing of social cohesion impacts. The regional cohesion approach is more

generally used at wider geographical levels. The social perspective is more widely

used at a local level, given the large amount of information that the disaggregation

of the population requires according to their socio-economic characteristics.

3.2.3.4.1 Accessibility and regional cohesion

Spatial distribution of accessibility is one of the variables included in the ‘check-list’

of the periodical Cohesion Reports of the European Union (EC, 2004a), among

which are included macroeconomic indicators such as GDP per capita, employment

levels or R&D investments. The rationale behind the inclusion of accessibility in this

list is that the ‘equality of access to services of general economic interest’ is


- 60 -

considered a key condition for territorial cohesion (Peters, 2003). Special interest is

placed in regions with geographical handicaps characterized by problems of

accessibility and integration with the rest of the EU. Infrastructure investment is

thus considered a key factor in order to provide a fair distribution of accessibility to

all its regions and to reduce existing disparities in accessibility between them

(Schürmann et al., 1997).

However, transport investment and increased regional cohesion do not

follow a causal relationship. Some scientists even argue, that better transport links

between strong and competitive centres and economically weak peripheries may

increase polarisation instead of cohesion (Hey et al., 2002). It can be concluded

that at the regional level redistribution will take place, often as additional

advantage to the already accessible parts of the country (Banister and Berechman,

2003).

Accessibility differences can therefore be used to assess equity impacts. It is

claimed that horizontal and vertical equity should be considered to minimize the

intraregional difference in accessibility for the main cities in the same region and to

minimize that interregional difference among regions, respectively (Feng and Wu,

2003). Existing attempts to use accessibility for the computation and visualization

of regional cohesion impacts can be found mainly at the European level. An

example is the study by Bruinsma and Rietveld (1993), who used a population

potential indicator to analyse if disparities of accessibility values in 42 major

European agglomerations increased with the implementation of planned

infrastructure investments. They found that inequalities in accessibility are least

pronounced in the road network, than in the rail network. A similar study using four

different accessibility indicators was carried out by Martín et al. (2004), in this case

to assess cohesion effects of a HSR line. Changes in the spatial distribution of

accessibility were also used to measure regional cohesion impacts at the EU level in

the SASI (Schürmann et al., 1997), ESPON (Bröcker et al., 2004), and IASON

(Tavasszy et al., 2004) projects. This type of analysis becomes less frequent as the

geographical scale is narrowed (for existing approaches at the national and regional

levels see e.g. López, 2005; Condeço and Gutiérrez, 2006).

The above studies show that many factors influence the final impact on

regional cohesion, such as the geographical scale of the analysis (Martín et al.,

2004), given that there may be positive equity impacts at the regional level but not

at a national level, or the existing development level of the network considered

(Martín et al., 2004; López, 2005), as in under-developed regions, efficiency is


- 61 -

usually the main concern and it is difficult for DMs to justify investments on the

basis of equity considerations.

The assessment of regional cohesion effects following this accessibility

approach can be included in strategic assessment methodologies, although existing

attempts are mostly found only in research applications (see ME&P et al., 2001;

INRETS, 2005; Bröcker et al., 2004; López and Monzón, 2006).

3.2.3.4.2 Accessibility and social cohesion

Accessibility indicators are increasingly used in social cohesion studies, mainly at

the local level with a focus on public transport. These studies are related to the

assessment of social exclusion issues of transport policies (Lucas, 2006). For this

purpose, accessibility is usually dissagregated by population groups, what allows

focusing in the situation of disadvantaged sectors of the population, such as aging

population or low income workers. Recent examples are: Lucas, 2006; Preston and

Rajé, in press; Lau and Chiu, 2003 and Alsnih and Hensher, 2003. Other

applications focus in the situation of population of rural areas and low density zones

(Nutley, 2003), equity mapping (Talen and Anselin, 1996; Talen, 1998),

distribution of job accessibility (Cervero et al., 1995; Geurs and Ritsema van Eck,

2003), or social changes (Halden, 2002; Bröcker et al., 2004).

3.3 Spatial impact and GIS

3.3.1 GIS background

Geographical Information Systems (GIS) can be defined20 as ‘a suite of methods for

capturing, storing, analysing and communicating geo-referenced information’

(Miller, 1999b). The history of GIS dates back to the 1950s21, when early

applications of GIS emerged for land use management and environmental impact

assessment activities supported by government agencies in Canada and the United

Sates. The emergence of computer-assisted cartography in the 1970s constituted a

crucial step in the development of GIS. Then, during the 1980s and 1990s there

were several important technical and organizational developments that greatly

assisted the current wide application and appreciation of GIS.

Today, although GIS technology is still evolving, it has already reached a

certain maturity, with existing applications in a wide spectrum of fields such as

20 See Burrough and McDonnell (1998) for a review of existing definitions of GIS.

21 See Foresman, 1998; Burrough and McDonnell, 1998 for a detailed chronology of GIS evolution.


- 62 -

agriculture, tourism, navigation, or archaeology. Recent applications show a

tendency towards the integration of GIS and Global Positioning Systems (GPS) (see

e.g. Taylor et al., 2000).

GIS database management capabilities and the usefulness of advanced

operations and functions allow performing integrated analysis of spatial and

attribute data22, especially useful in spatial impact analysis (Nijkamp et al., 1990).

Figure 3.9: Superposition of data layers in GIS for a transport study.

Source: Taylor et al. (2000)

As an example, Figure 3.9 shows a diagram illustrating the use of GIS as a

database integrator for a transport study area. The GIS is able to integrate a

mixture of spatial, numerical, and perhaps textual datasets, and display them by

superposition of separate map layers, such as topographical and land use, transport

network infrastructure, socioeconomic and demographic, traffic flow, and pollution

22 A comprehensive review of GIS operations can be found in Burrough and McDonnell (1998) and

DeMers (1997).


- 63 -

and environmental impacts. The Figure also indicates the areas of modelling and

analysis involving the use of different subsets layers, such as travel demand

modelling or impact analysis. There is a wide list of literature on the integration of

GIS and spatial models (Wegener, 2001; Nijkamp et al., 1990; Fotheringham and

Wegener, 2000). Spatial analysis is currently benefiting from the new geo-

computational tools that are emerging from geographic information science (GISci),

a new interdisciplinary field that focuses on the theory and techniques behind GIS

and related technologies (Goodchild, 1992). However, the tools offered by current

GIS do not usually include the analytical and modelling capabilities needed for

spatial modelling (Wegener, 2001). Indeed, the potential synergies between GIS

and spatial models are not fully exploited (Miller, 1999b).

Furthermore, recent development of computer programs have enabled GIS

platforms to provide user friendly interfaces so that analysts and planners can

execute model runs and analyze model results without the need for GIS

technicians, which is considered one of current’s primary benefits (Miller and Storm,

1996). All the above capabilities make GIS specially suited at providing support in

situations where decision making is required (Nijkamp et al., 1990; Malczewski,

1999; Arampatzis et al., 2004; Miller, 1999b). This is the case with transport

planning processes.

3.3.2 Applications of GIS in transport planning

3.3.2.1 GIS and transport planning processes

GIS capabilities make them specially suited for each of the three stages of transport

planning process described in Chapter 2. First, GIS can have a relevant role to play

in the structuring stage. The capabilities of GIS are useful for data organisation,

ease of data entry, data processing and visualisation, and for the detection of

deficiencies/problems and subsequently set the corresponding planning objectives.

This may constitute a critical issue. Indeed, it is argued that the way the planning

problem is represented has a major importance for DMs, as seemingly minor

variations in the way the problem is represented can lead to different

recommendations (Guitouni and Martel, 1998). The above advantages are making

both private and public agencies increasingly rely on GIS as indispensable tools for

planning and decision making.

Second, the use of GIS in the evaluation stage is also widespread (Nijkamp

et al., 1990; Malczewski, 1999; Gómez and Bosque, 2004; Arampatzis et al.,

2004). The capability of GIS to provide, at any time, appropriate information


- 64 -

regarding trade-offs and efficiency of proposed solutions has created an increase in

their popularity in recent years (Miller, 1999b). Key approaches use GIS to support

accessibility calculations (Liu and Zhu, 2004; Zhu and Liu, 2004; Gutiérrez and

Monzón, 1998; Miller and Wu, 2000), environmental assessments (Antunes et al.,

2001; Colorni et al., 1999; Li et al., 1999; Brown and Affum, 2002), network

demand models (Miller, 1999b; Miller and Storm, 1996) , traffic congestion studies

(Taylor et al., 2000), alignment optimization models (Jha and Schonfeld, 2004;

Sadek et al., 1999) or regional (Bröcker et al., 2002) and urban spatial planning

models (Wegener, 2001).

Third, in the decision-making stage a GIS ‘serves to contribute at solving,

organizing and rationalizing complex choice and decision problems’ (Nijkamp et al.,

1990). The development of GIS has exerted a deep on-going impact on modern

decision analysis, enabling the design of interactive user-oriented multiple criteria

decision models (MCDM) (Malczewski, 1999; Sadek et al., 1999; Klungboonkrong

and Taylor, 1999) or Decision Support Systems (DSS) (Arampatzis et al., 2004;

Colorni et al., 1999; Jha, 2003) especially useful in transport decision-making

processes, where the stakeholders can be a large and diverse group (Miller and

Storm, 1996). In this sense, the ‘communicative’ abilities of GIS due to their

capabilities to create high standard visual images could be more fully exploited in

order to stimulate discussion and evaluation of different project designs in new and

more informative ways (Grant-Muller et al., 2001).

However, despite the above mentioned advances, a more integrated work of

spatial analysis, transportation and GIS research communities could have a

substantial positive impact on the theory and practice of transportation analysis and

planning (Miller, 1999b; Wegener, 2001). This constitutes a current challenge for

the research community.

3.3.2.2 GIS and accessibility analysis

Finally, this subsection describes the use of GIS to compute most of the

accessibility indicators described in section 3.2. Given the spatial nature of

accessibility, GIS have become a useful tool for accessibility analysis, which

provides capabilities for data collection, data management and manipulation,

spatial analysis, network analysis, and cartographical presentation of accessibility

measures.

A description of the steps required for the use of GIS to compute

accessibility indicators can be found in Zhu and Liu (2004), who developed an

integrated GIS approach to accessibility analysis, which is illustrated in Figure 3.10.

This approach consists of six major processes or elements: problem definition and


- 65 -

data collection, query and data retrieval, measure selection and specification, travel

impedance measurement, calculation of accessibility measures and visualization of


Figure 3.10: An integrated GIS approach to accessibility analysis.

Source: Zhu and Liu (2004)


- 66 -

3.4 Conclusions

At the Plan level, additional assessments are necessary to address wider policy

impacts such as network effects or distributive impacts, usually not covered by

traditional appraisal methodologies such as CBA (SACTRA, 1999; Beuthe, 2002).

Most of these impacts have a spatial component and therefore modern spatial

impact models are especially suited for these tasks.

Accessibility indicators could potentially be used as a criterion for project and

policy appraisal (David Simmonds Consultancy et al., 1998; Halden, 2003; Ney,

2001) which could complement existing methodologies. Further research is needed

to take advantage of this unexploited potential and develop formulations combining

a theoretically sound foundation and a relative ease of interpretation for DMs and

the public opinion.

Furthermore, special care is needed for the inclusion of accessibility analysis

in appraisal frameworks, as the addition of accessibility impacts to the CBA results

would generally have a double-counting of effects. However, accessibility indicators

can be introduced in MCA frameworks, although it is essential to define with

precision what the accessibility analysis is meant to represent or to which policy

objective it corresponds in order to attribute a sensible weight in the MCA. For

example, if weights varying inversely with the level of economic development are

used, accessibility results may be useful to compute the projects impact on social

cohesion under the assumption that improved accessibility induces economic

development (ME&P et al., 2001; BMVBW, 2002). A possible alternative is to

compute inequality indices of the spatial distribution of accessibility among regions,

in order to measure regional cohesion impacts (Schürmann et al., 1997; López and

Monzón, 2006). Another possibility is to use accessibility results in a MCA to

measure political issues, such as international network integration, but in this case

it should be made explicit that the objective is political and not economic (Beuthe,

2002).

Finally, a current challenge for the research community is to take full

advantage of recent developments of GIS to develop a more integrated work of

spatial analysts, transportation planners and GIS capabilities. This integration could

result in significant synergies and have a substantial positive impact on the theory

and practice of transportation analysis and planning (Miller, 1999b; Wegener,

2001).

Chapter 4 – METHODOLOGY FOR THE ASSESSMENT OF TRANSPORT INFRASTRUCTURE PLANS

- 67 -

4. METHODOLOGY FOR THE ASSESSMENT OF

TRANSPORT INFRASTRUCTURE PLANS

Chapter 2 concluded that national transport Infrastructure Plans assessment

methodologies need to adapt to a new planning framework. Chapter 3 reviewed

recent developments in assessment tools and techniques. In the light of these

findings, this Chapter describes a methodology for the assessment of Transport

Infrastructure Plans. The structure of this methodology is outlined in Section 4.1.

The structure of the proposed procedure is built on the definition of the assessment

framework, described in subsection 4.2, and assessment criteria, defined in

subsection 4.3. Performance indicators linking model outputs and assessment

criteria are defined in subsection 4.4. Finally, subsection 4.5 outlines the

foundations of a MCA framework for the integration of assessment results.

4.1 Structure of the methodology

The proposed methodology is suggested as a tool to complement traditional

assessment methodologies at the Plan level. The proposed approach suggests a

procedure to assess those strategic aspects which are usually not included in official

methodologies, despite the fact that they are very important for the achievement of

transport policy goals.

The methodology is flexible in the sense that the final integration of their

results with that of the traditional methodologies, such as a CBA, has intentionally

been left for the consideration of DMs. This responds to the fact that, at the Plan

level, as discussed in Section 2.4., the technical appraisal plays an important role

but the final decision is usually influenced by a set of external constraints, mainly of

a political nature.

Figure 4.1 shows the structure of the proposed methodology, which is fully

implemented in a GIS. The process is built from two starting points, following a

twin approach (Brown et al., 2001):

� A ‘top-down’ approach: The identification of strategic policy objectives

constitutes the main guidelines to the definition of assessment criteria,


- 68 -

� A ‘bottom-up’ approach: The definition of the different alternatives to be

assessed is necessary for the models to forecast the impacts of the transport

projects under consideration.

Figure 4.1: Structure of the methodology

DEFINITION OF ALTERNATIVES

POLICY OBJECTIVES

GIS

PERFORMANCE INDICATORS

ASSESSMENT CRITERIA

ANALYSIS OF RESULTS

ACCESSIBILITY/TRANSPORT MODELS

INTEGRATION (MADM)

INPUT DATABASE

SENSITIVITY ANALYSIS

DEFINITION OF ALTERNATIVESDEFINITION OF ALTERNATIVES

POLICY OBJECTIVES

POLICY OBJECTIVES

GIS


ASSESSMENT CRITERIA

ANALYSIS OF RESULTS

ANALYSIS OF RESULTS



INTEGRATION (MADM)

INPUT DATABASE

INPUT DATABASE

SENSITIVITY ANALYSIS

The accessibility and the transport models provide the main inputs for

impact assessment. Subsequently, in order to measure the performance of each

alternative in each assessment criteria, a set of performance indicators1, linking

model outputs with each assessment criteria is defined.

The added value of the methodology is mainly included in the design of the

assessment criteria and the corresponding set of performance indicators. The main

singularity of the proposed approach is that the accessibility model is the driving

engine of the assessment methodology. Indeed, most performance indicators

include accessibility indicators in their formulation. The rationale behind their

inclusion is that changes in the values and the spatial distribution of accessibility

indicators can be used as proxy variables for the assessment of strategic impacts,

taking advantage of their unexploited potential in transport assessment

methodologies (see Section 3.2.3.).

The results obtained in each performance indicator need to be subsequently

integrated in order to obtain a single score for each alternative. This is dealt with in

the application of a multiatribute decision making (MADM) method. The process is

subsequently complemented with a sensitivity/robustness analysis of the results to

1 Also referred to as measures of performance (Brown et al., 2001).


- 69 -

key parameters and external factors of the evaluation process. Finally, conclusions

and recommendations on the best performing alternative are provided to the DMs2.

4.2 Definition of the assessment framework

4.2.1 Assessment time horizon

The assessment of alternatives is carried out at the planning time horizon (tp),

previously defined in the Plan. The performance of each alternative is compared

against the ‘do-nothing’ alternative (A0), defined as that with the same network as

the one existing in the base year (t0). This comparison is outlined in Figure 4.2.

External inputs for the accessibility and transport models, such as population

growth, are defined identical between alternatives, in order to isolate the effects

stemming from the infrastructure changes from those derived from the

development of external variables.

Figure 4.2: Comparison of alternatives

NETWORK BASE YEAR A0

NETWORK ALTERNATIVE A1

NETWORK ALTERNATIVE AN

NETWORK BASE YEAR A0

BASE YEAR t0PLANNING TIME

HORIZON tp

NETWORK BASE YEAR A0NETWORK BASE YEAR A0

NETWORK ALTERNATIVE A1NETWORK ALTERNATIVE A1

NETWORK ALTERNATIVE ANNETWORK ALTERNATIVE AN

NETWORK BASE YEAR A0NETWORK BASE YEAR A0

BASE YEAR t0PLANNING TIME

HORIZON tp

This scheme requires making a prognosis concerning the future development

of these external variables. The assessment procedure therefore relies on the

accuracy of this prognosis which, given the long-term planning horizons of most

Plans, introduces high levels of uncertainty in the results. The methodology handles

2 On the basis of these conclusions, the possibility for the inclusion of feedback loops (see e.g. Nijkamp

et al., 1990) in the procedure is represented by the dotted lines in Figure 4.1.


- 70 -

this issue in the sensitivity analysis step, with the definition of assessment

scenarios.

4.2.2 Delimitation of the study area

The analysis done in Section 3.1.3.1 shows that the effects of transport Plans

extend beyond the administrative boundaries of the region or nation concerned,

generating spillover effects. In order to assess these effects, the limits of the study

area should be widened over these administrative boundaries. However, this

extension of the study area faces some political and technical drawbacks, mostly if

the additional territories belong to other countries. On the one hand, DMs are

usually exclusively concerned with the effects produced inside the frontiers of the

territories in their competency, even if the Plan is supported with ‘external’ funds,

as it is the case of national Plans supported by European funds. On the other hand,

the inclusion of information from different countries makes the assessment

procedure more data demanding and complex, as these data usually need to be

harmonised with national data sources.

This issue is dealt with in the methodology with the adoption of a compromise

solution for the delimitation of the study area. The proposed approach consists in

widening the study area to include those regions in which spillover effects may

appear. As will be detailed in this Chapter, the specific treatment of these ‘external’

regions in the assessment procedure is different than that of the territory for which

the Plan has been originally designed.

4.3 Definition of assessment criteria

Assessment criteria have been selected based on the review of recently developed

research studies, national official assessment methodologies and most relevant EU

policy documents carried out in Section 2.2.3. These studies mostly agree that at

the Plan level current traditional appraisal procedures, mainly based on CBA

approaches, should be complemented with the assessment of certain wider

strategic objectives/criteria and be consistent with the triangular SD approach.

Hence, this methodology suggests complementing traditional assessment

methodologies with the assessment of strategic assessment criteria. The structure

of these criteria and corresponding subcriteria is included in Table 4.1. This

structure responds to the SD approach: i.e efficiency is related to the economic,

cohesion to the social and environmental sustainability to the environmental SD

dimensions. The policy relevance of each subcriterion and their detailed description

is included in subsequent sections.


- 71 -

Table 4.1: Assessment criteria

CRITERIA SUB-CRITERIA

Network efficiency Efficiency

Cross-border integration

Regional cohesion Cohesion

Social cohesion

Global warming Environmental sustainability

Habitat fragmentation

4.3.1 Efficiency

The term efficiency embraces different concepts, such as competitiveness, network

efficiency, regional development, economic development or growth. The efficiency

of transport links between major economic activity centres is considered as one of

the factors determining competitiveness (EC, 2004a). These activity centres may be

located inside or outside the national boundaries; therefore the improvement of

cross-border links is frequently included as a policy goal for improved

competitiveness, particularly in peripheral countries, such as Spain (Ministerio de

Fomento, 2005). Moreover, cross-border cooperation is intended to develop cross-

border economic and social centres through joint strategies for sustainable

territorial development (EC, 2004b).

The methodology takes into account this issue by splitting the efficiency

criterion into two subcriteria, namely (national) network efficiency and cross-border

integration. This is done in order to split efficiency benefits accruing to the study

area under consideration from those that occur in neighbouring cross-border

regions, as they respond to different policy goals and therefore the DM will probably

attach a different preference strength to each one of them.

4.3.2 Cohesion

Among the wide variety of existing approaches included under the cohesion heading

(see Section 3.1.3.3.), the methodology focuses on two issues which are of

increasing interest among transport planners at strategic levels. These issues are

regional and social cohesion.

Regional cohesion effects are assessed with the analysis of whether the Plan

increases or reduces existing disparities in the spatial distribution of accessibility

among regions. This increase or reduction can be interpreted as a negative or

positive regional cohesion effect, respectively. The concern derived from the


- 72 -

tendency of certain transport infrastructure extensions to induce polarising effects

are therefore also addressed.

The contribution of the Plan to the social cohesion objective is handled in the

methodology with the exploration of a possible contribution of accessibility

improvements to regional development in lagging regions. This contribution would

result in a more balanced distribution of socio-economic conditions among social

groups: those living in lagging regions and those that do not. This balancing effect

is usually termed as an enhanced social cohesion.

The social cohesion criterion is aimed at measuring to what extent

accessibility benefits are accruing to structurally backward regions, whilst the

regional cohesion criterion analyses effects stemming from changes in the spatial

distribution of accessibility; i.e. the objective is to investigate whether the

accessibility improvements derived from the Plan increase or reduce existing

disparities in accessibility among regions. The focus of the regional cohesion

criterion is therefore exclusively limited to the changes in the regions’ relative

positions in terms of accessibility, independently from their economic development

level.

4.3.3 Environmental sustainability

Climate change, loss of biodiversity due to habitat fragmentation, effects on human

health (e.g. local emissions) and well-being due to accidents, air quality and noise

are the most important environmental concerns related to transport activity3 (EEA,

2003). Only the first two -climate change and habitat fragmentation- have been

selected from this list because of their strategic nature; the rest are frequently

addressed at the project level (see Section 3.1.2.2.).

First, the climate change phenomenon is directly linked to energy

consumption and directly related to green house gas (GHG) emissions. These are

strategic environmental aspects of great interest due to both the need to comply

with international environmental commitments and the urgency to reduce energy

consumption, which has a greater economic component.

Second, the assessment of habitat fragmentation due to transport

infrastructure has been included since it is recognised globally as one of the biggest

threats to the conservation of ecological biodiversity, which should be ideally

addressed at strategic levels. The assessment of habitat fragmentation i.e. the

‘dynamic process of splitting of habitats into smaller and more isolated areas, called

3 Transport and Environment Reporting Mechanism. Periodical reports are published by the European

Environment Agency (EEA) and are available at http://reports.eea.eu.int/.


- 73 -

patches’ (Burel and Baudry, 2002) remains a contentious issue, where there is a

lack of commonly accepted procedures for its assessment4.

The selection of fragmentation indices5 is an important aspect of assessment

procedures (Riitters et al., 2004). The perimeter and the area of each patch are

probably its most important and useful variables from an ecological point of view

and therefore are usually included in the formulation of fragmentation indices. On

the one hand, the shape of the patches is directly linked with the ‘border effect’:

the higher the perimeter, the higher the threat from external factors. On the other

hand, the presence and wealth of many species is intimately linked with the area of

the patch (Robbins et al., 1989).

4.4 Definition of performance indicators

Figure 4.3 describes the interdependence between the outputs of the models and

performance indicators.

Figure 4.3: Performance indicators’ inputs

INFRASTRUCTURE NETWORKSOCIO-ECONOMIC VARIABLES

ACCESSIBILITY CALCULATIONS TRAFFIC FLOWS

TRAVEL IMPEDANCES

LANDSCAPE QUALITYINPUT

GW

Global warming

HF

Habitatfragmentation

NE

Networkefficiency

CB

Cross-borderintegration

RC

Regional cohesion

SC

Social cohesion

MODELLING


INFRASTRUCTURE NETWORKSOCIO-ECONOMIC VARIABLES

ACCESSIBILITY CALCULATIONS TRAFFIC FLOWS

TRAVEL IMPEDANCES

LANDSCAPE QUALITYINPUT

GW

Global warming

HF


NE

Networkefficiency

CB


RC

Regional cohesion

SC

Social cohesion

MODELLING


4 For a review on impacts of transport infrastructure on habitat fragmentation, see Riitters et al., (2004)

and Burel and Baudry (2002).

5 Usually based on raster land-cover information derived from satellite imagery. The value of the index

obtained is particularly sensitive to the spatial and theme distribution of the input map. National

ecological assessment procedures typically handle this issue computing a small number of indices and

analysing their results in a complementary way.


- 74 -

Depending on their input data requirements, performance indicators can be

classified as follows:

� Based on accessibility results: network efficiency, cross-border integration,

regional cohesion and social cohesion.

� Independent from modeling results: habitat fragmentation.

� Based on transport flows in the whole system: global warming.

Table 4.2 gives an overview of the assessment criteria and their corresponding

performance indicators. Their detailed formulation is explained below.

Table 4.2: Assessment criteria and performance indicators

CRITERIA PERFORMANCE INDICATOR

EFFICIENCY

Network efficiency Change of network efficiency in national

territories

Cross-border integration Change of network efficiency in cross-border

regions

COHESION

Regional cohesion Change in synthetic inequality index of

accessibility indicators

Social cohesion Standardised increase in potential

accessibility of isolated and/or lagging

regions

ENVIRONMENTAL SUSTAINABILITY

Global warming Change in GHG emissions

Habitat fragmentation Change in habitat fragmentation index

4.4.1 Efficiency

4.4.1.1 Network efficiency

The main input variable for this performance indicator is the network efficiency

accessibility indicator (E) (Gutiérrez and Monzón, 1998; Gutiérrez, 2001), using

Equation ( 4.1):

∑∑

⋅

=j

j

j

j

ij

ij

iP

PII

I

E ( 4.1 )


- 75 -

This indicator is used to calculate, for each i-j pair, a weighted mean of

ratios between travel time using the network (Iij) and an ‘ideal’6 travel time (IIij).

The population of each destination (Pj) has been selected as the weighting factor.

The set of origins i is restricted to those belonging to the national territory, whereas

the set of potential destinations j includes national as well as cross-border economic

activity centres.

For each alternative s, the performance indicator – NEs - calculates the

percentage change, compared with the do-nothing alternative A0, of a population-

weighted aggregated value of the network efficiency accessibility indicator,

according to the formulation detailed in Equation ( 4.2 ):

100

0

0 ⋅

−

=

∑∑

∑∑

∑∑

r

i

i

i

i

sr

i

i

i

i

i

i

i

i

i

s

P

PE

P

PE

P

PE

NE ( 4.2 )

4.4.1.2 Cross-border integration

This performance indicator is also measured using the network efficiency

accessibility indicator as the main input variable. The main difference lies in the

definition of the spatial coverage, as the set of origins only includes those nodes (r’)

located in cross-border regions of neighbouring countries, whilst the set of

destinations is the same as that used for the assessment of national network

efficiency impacts. Hence, benefits accruing outside the national boundaries

(spillover effects) are accounted for.

Equation ( 4.3 ) includes the formulation of the performance indicator – CBs -

, where all the terms have already been defined:

100

0

0

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

⋅

−

=

∑∑

∑∑

∑∑

rr

r

rr

srr

r

rr

rr

r

rr

s

P

PE

P

PE

P

PE

CB ( 4.3 )

6 Measured as travel time as the crow flies, with 120 km/h speed for road mode and 220 km/h for rail

mode.


- 76 -

4.4.2 Cohesion

4.4.2.1 Regional cohesion

This performance indicator is based on the analysis of disparities in the spatial

distribution of accessibility. The rationale behind this analysis is that accessibility

can be considered as an ‘added value’ of locations, in a way related to their level of

welfare. Hence, disparities in accessibility among regions can be used as a proxy

variable for the measurement of territorial cohesion effects.

The choice of the accessibility indicator depends on the purpose of the study,

as analyzed in Section 3.2.2. In this context, where cohesion is analysed under a

welfare perspective, the most suited accessibility indicator is the potential indicator.

Among existing formulations (see Section 3.2.2.4.), the one selected has proved its

validity in similar studies (Martín et al., 2004). It is described in Equation ( 4.4 ):

∑=j ij

j

rI

PPP ( 4.4 )

There is a wide variety of statistical indices capable of characterising the

level of dispersion of any given variable; they are referred to as inequality

measures/indices7. The variable under analysis here is potential accessibility, hence

each inequality index is computed combining the accessibility values of individual

regions (PPr) into one single measure of their spatial concentration. The population

of each individual region (Pr) has been selected as the weighting variable.

The choice of the inequality index may have a strong influence on the results

(Bröcker et al., 2004; Schürmann et al., 1997). Therefore, four of the most

commonly used inequality indices are computed and their results analysed in a

complementary way.

The four inequality indices are compared with their corresponding values in

the ‘do-nothing’ alternative. Then, the regional cohesion performance indicator

(RCs) is computed as the mean value of the resulting relative change in the four

inequality indices, as expressed in Equation ( 4.5 ). In every case, a lower value of

the index represents a more equal distribution, and vice versa. A brief description

of each of the indices is included below8.

−+

−+

−+

−⋅=

0

0

0

0

0

0

0

0

4

1

Th

ThTh

At

AtAt

Gi

GiGi

CoV

CoVCoVRC ssss

s ( 4.5 )

7 Inequality measures are commonly used in the economic literature for the analysis of income

distribution. For a comprehensive review on inequality measurement, see Cowell (1995).

8 For an extensive description of these and other inequality indices, see Cowell (1995).


- 77 -

The first one is the coefficient of variation (CoV): it is computed as the ratio

between the standard deviation and the mean (µ).

µ

deviation standard=CoV ( 4.6 )

The second one is the GINI coefficient (Gi). It is a summary statistic of the

Lorenz curve, a cumulative frequency curve that compares the distribution of a

specific variable with the uniform distribution that represents equality. The Gini

coefficient ranges from a minimum value of zero, when all individuals are equal, to

a theoretical maximum of one in an infinite population in which every individual

except one has a size of zero. It corresponds to twice the area between the Lorenz

curve and the diagonal (perfect equality).

Third, the Atkinson index (At) is computed using Equation ( 4.7):

εε

µε

−−

⋅−= ∑

1

11

1)(i

i

i

xpAt if 1≠ε

⋅−= ∑

i

i

i

xpAt

µε logexp1)( if 1=ε

( 4.7 )

Where pi is the percentage of population and xi is the proportion of total

accessibility enjoyed by the ith group, respectively, and ε is the so-called inequality

aversion parameter. The parameter ε reflects the strength of society's preference

for equality, and can take values ranging from zero to infinity. When ε > 0, there is

a social preference for equality (or an aversion to inequality). As ε rises, society

attaches more weight to income transfers at the lower end of the distribution and

less weight to transfers at the top. Typically used values of ε include 0.5 and 2.

Finally, the Theil index (Th) is part of a larger family of measures referred to

as the General Entropy class. Its formulation is as follows:

∑

⋅=

i i

ix

pThµ

log ( 4.8 )

4.4.2.2 Social cohesion

This performance indicator calculates a weighted sum of regional accessibility

changes. Each region’s weighting factor (Φr) depends on its level of structural

backwardness and of that of accessibility deficiencies in the do-nothing situation, as

Table 4.3 shows. Weighting factors vary from zero-point minimum to a four-point


- 78 -

maximum in case of coincidence of accessibility deficits with high levels of economic

backwardness, according to the following matrix:

Table 4.3: Weighting factor matrix for the cohesion criterion

Accessibility deficiencies

Structural backwardness

category

None Not very

significant

Significant Very

significant

Non-lagging regions 0 1 1 2

Potentially lagging regions 1 1 2 3

Lagging regions 1 2 3 4

Source: Adapted from BMVBW (2002) and Bröcker et al. (2004)

The typology of lagging regions is based on a single or a combination of

socio-economic indicators9, usually unemployment rates and GDP per capita levels,

depending on data availability. All regions are ranked and classified according to the

standards defined in Table 4.4.

Table 4.4: Structural backwardness categories

CATEGORY Structural backwardness Cells per type

Substandard 0 Non-lagging regions Best 50%

Substandard 1 Potentially lagging 50% to 30%

Substandard 2 Lagging Worst 30%

On the other hand, the classification in accessibility levels is split into four

categories. The definition of the substandard of accessibility categories is included

in Table 4.5.

Table 4.5: Accessibility analysis categories

CATEGORY Accessibility deficiencies Cells per type

Substandard 0 None Best 50%

Substandard 1 Not very significant 50% to 25%

Substandard 2 Significant 25 to 10%

Substandard 3 Very significant Worst 10%

For the selection of accessibility indicators to be used in the model, three

possibly conflicting objectives are considered to be relevant (Bröcker et al., 2004):

first, the accessibility indicators should contribute as much as possible to explaining

9 For example, a possible combination is the one followed by Bröcker et al. (2004): the unemployment

rate weighted by 60 percent, GDP per head weighted by 40 percent.


- 79 -

regional economic development. Second, the accessibility indicators should be

meaningful in themselves as indicators of regional quality of life. Third, the

accessibility indicators should be consistent with theories and empirical knowledge

about human spatial perception and behaviour. In the light of these objectives the

formulation of the population potential accessibility indicator (PPr), included in

Equation ( 4.4 ), was adopted.

The social cohesion performance indicator –SCs- is calculated as the ratio

between the weighted (by the corresponding population and the social cohesion

weighting factor) accessibility increase and the population-weighted accessibility

value of the do-nothing alternative, expressed in percentage terms.

100

)(

0

0

⋅⋅

−⋅⋅

=

∑

∑

∑

∑

r

r

r

r

r

r

r

r

s

rr

r

r

s

P

PPP

P

PPPPP

SC

φ

( 4.9 )

Hence, above average accessibility improvements in lagging regions are

given higher weights in the final value, therefore measuring the relative

contribution of each alternative to social cohesion. The interpretation of this

indicator is as follows: the higher values, the more the concentration of higher

accessibility improvements for those individuals of structurally lagging and/or

inaccessible regions, i.e. the higher contribution to the social cohesion objective.

4.4.3 Environmental sustainability

4.4.3.1 Gobal warming

The selected performance indicator for the global warming criterion is total

greenhouse gas emissions (GHG), measured in equivalent tons of CO2. CO2

emissions are included as indicators for climate change issues in different

environmental indicator lists (see Banister et al., 2000b; EEA, 2003; OECD, 1998).

Annual CO2 emissions are computed and summed up to calculate total tons of CO2

emitted in each alternative, using the vehicle-kilometres-travelled generated by the

transport model per mode and the information on the national vehicle fleet (drives,

car categories and emission standards).


- 80 -

4.4.3.2 Habitat fragmentation

The availability and level of disaggregation of environmental information on

landscape quality is uneven among MMSS. Therefore, for the definition of the

patches the methodology proposes to use existing EU-level databases containing

harmonised land-cover information and to complete them with additional country-

specific information, in case this is available10. In particular, the definition of

patches is suggested to be based on existing landscape information on habitats

included in the Habitats Directive 92/43/EEC11. All the available land cover

information is subsequently aggregated to obtain a final GIS vectorial map with the

spatial classification of habitat types.

The selected fragmentation index is the Perimeter/Area RAtio (PARA). It is a

widely used ‘spatial configuration index’ 12 to calculate (see Equation ( 4.6 )) the

ratio between the perimeter (Pe) and the area (Ar) of each patch (i).

i

i

iAr

PePARA = ( 4.10 )

The GIS capabilities allow analysing how the perimeters and areas of all

patches comprising each habitat are affected when they are crossed by the road

and rail infrastructure network extensions included in the Plan. Transport

infrastructure is considered as a total barrier, i.e. a new infrastructure is considered

to divide every patch it crosses into smaller patches.

Figure 4.4. represents an example of the fragmentation produced in a

specific patch (drawn in the left Figure), when it is crossed by a highway and hence

fragmented into 5 smaller patches (drawn in the right Figure).

10 In this sense, CORINE (Land cover dataset firstly developed in the 1990s as part of the European

Commission programme to COoRdinate INformation on the Environment ) datasets can, to a certain

extent, be interpreted with regard to its potential land use and hence recognise areas being mainly

natural or semi-natural, receiving low human impact. However, some anomalies may appear when

assessing landscape fragmentation based only on CORINE information (EC, 2000).

11 Council Directive 92/43/ECC of 21 May 1992 on the conservation of natural habitats and of wild fauna

and flora, O. J. L 206, 22.07.92.

12 It is beyond the scope of this thesis to include a review on the concept and measurement of habitat

fragmentation. The reader is referred to Jaeger (2000), Riitters et al. (2004) and McGarigal and Marks

(1995) for selected literature reviews on the topic.


- 81 -

Figure 4.4. Scheme of the calculation of the PARA index

The corresponding values of the PARA index in the original situation (Figure 4.4

left) and the new situation (Figure 4.4 right) have been included in Table 4.6.

Table 4.6: Example of the computation of PARA values

PATCH Area (m2) Perimeter (m) PARA % increase in PARAOriginal patch 1,697,521.84 9,877.70 0.0058 -Patch # 1 920,980.58 6,169.39 0.0067 15.12Patch # 2 406,845.16 2,980.34 0.0073 125.89Patch # 3 66,975.51 1,083.20 0.0162 277.94Patch # 4 58,430.32 1,097.32 0.0188 322.74Patch # 5 63,480.74 1,258.97 0.0198 340.83

This index is measured at the patch level and subsequently aggregated

using the area of each patch as the weighting variable.

∑∑∈

∈

⋅=

hi

hi

i

ii

hAr

ArPARAPARA

( 4.11 )

For each alternative, the value of the PARA index is computed as the

weighted mean of PARA indices among habitats, using each habitat area as the

weighting variable:

∑∑

⋅=

h

h

h

hh

Ar

ArPARAPARA

( 4.12 )


- 82 -

Finally, for each alternative, the habitat fragmentation performance indicator

is computed as the percentage change in the PARA index, if compared to that of the

do-nothing alternative, as Equation ( 4.13 ) shows:

1000

0⋅

−=

PARA

PARAPARAHF

s

s ( 4.13 )

4.5 Integration

4.5.1 Outline of the proposed approach

The objective is to obtain a ranking of the n alternatives on the basis of the results

of the performance indicators (X) in each of the m (6 in this case) criteria. The

aggregation of performance indicators to obtain an overall score for each

alternative constitutes a multiattribute decision making (MADM) problem.

The value/utility function approach has been selected from the vast list of

available MADM methods13. The standard assumptions underlying this method

involve preferential independence (i.e. the relative preferences of attributes are not

altered by changes in other attributes) and utility independence (i.e. the utility

function over a single attribute does not depend on the other attributes).

The utility function method is based on multiattribute utility theory (Keeny

and Raiffa, 1976). The term utility function is restricted to a probabilistic criterion

or decision under uncertainty, when the DMs’ attitudes toward risk are an important

determinant of the final results. This approach involves the estimation of the value

(utility) function f and the scaling constant (weight) wj for each attribute14. As

Equation ( 4.14) shows, by multiplying the utilities by the weights, the trade-offs

among the attributes utilities are taken into account in the multiattribute utility

function:

∑ ⋅=j

sjjs uwU ( 4.14 )

Where Us is the overall utility of the s-th alternative, wj is a normalized

weight or scaling constant for attribute j, and usj is the utility of the sth alternative

with respect to the jth attribute, measured by means of the utility function. The

procedures to determine criteria’s weights and utility functions are explained in

Sections 4.5.2 and 4.5.3, respectively. The procedure is outlined in Figure 4.5.

13 A comprehensive review of MADM methods can be found in Malczewski, 1999; Keeny and Raiffa,

1976; Goodwin and Wright, 1991; Nijkamp et al., 1990.

14 Criterion is a generic term including the concepts of attribute and objective, whereas an attribute is

used to measure performance in relation to an objective (Malczewski, 1999).


- 83 -

Figure 4.5: The integration procedure

ALTERNATIVE 1 ... j ... m

1

...

s Xs1 Xsj XSm

...

n

CRITERIA


1

...

s us1 usj usn

...

n

CRITERIA

U1

...

Us

...

Un

PERFORMANCE MATRIX

PARTIAL UTILITY MATRIX

GLOBAL UTILITY VECTOR

)( sjsj Xfu =

∑ ⋅=j

sjjs uwU


1

...

s Xs1 Xsj XSm

...

n

CRITERIA


1

...

s us1 usj usn

...

n

CRITERIA

U1

...

Us

...

Un

PERFORMANCE MATRIX

PARTIAL UTILITY MATRIX

GLOBAL UTILITY VECTOR

)( sjsj Xfu =

∑ ⋅=j

sjjs uwU

4.5.2 Weight estimation

The suggested method to derive the base weight profile is the REMBRANDT

procedure (Lootsma, 1992), based on the Analytical Hierarchy Process (AHP),

originally developed by Saaty (1990). REMBRANDT is one of the best known

attempts to retain the strengths of AHP while avoiding some of its objections

(Tsamboulas et al., 1998).

The procedure requires conducting a questionnaire in which the respondent

is asked to express his/her strength of preferences between pairs of criteria on a

qualitative scale, which corresponds to numerical values in a -8/+8 interval. These

values form the elements of the pairwise comparison matrix of jxj elements, which

are subsequently transformed15 into the estimated criteria weights.

15 The procedure uses a direct rating system which is based on a logarithmic scale to replace the 1 - 9

scale of AHP and exchanges the eigenvector-based synthesis approach for one which is based on use of

the geometric mean.


- 84 -

The required information was obtained from a survey conducted in different

transportation-related events16. The questionnaire and the average resulting weight

profile obtained are included in Appendix A.

The procedure described above is aimed at computing a base weight profile.

These weights will be subsequently used as a starting point to define a set of

weight profiles. The sensitivity of the results to changes in criteria weights will be

assessed in the sensitivity analysis step of the methodology, as it is detailed in

Section 4.6.

4.5.3 Utility functions

The selected procedure to derive each criterion’s utility function is the indifference

technique (Keeny and Raiffa, 1976). It requires the DM to assess an outcome that

will make him/her indifferent between this outcome and a gambling of two other

values that already have a utility value.

A pointwise approximation of the utility function can be obtained by asking

the DM a series of questions such as: for attribute j, what certain outcome xj would

be equally desirable as realizing the highest outcome with a probability p and the

lowest outcome with a probability of 1-p? This can be expressed in utility terms

using the extreme xj+ and xj- as:

( ) ( )( ) ( )−+ ⋅−+⋅== jjjjjj xupxupxu )1(? ( 4.15 )

where uj(xj) is the utility function associated with the jth attribute, and uj(xj+)=1 and

uj(xj-)=0 are the utilities of the best and the worst outcomes for the jth attribute,

respectively. The extreme values of the outcomes, xj+ and xj- are therefore

necessary to construct the functions. To construct the utility curve, a set of

questions need to be asked to the DM until enough discrete points have been

assessed to give an accurate picture. Usually three points is enough (Martínez-

Falero and González-Alonso, 1995).

4.6 Sensitivity analysis

Sensitivity analysis is a collection of methods used for evaluating how sensitive the

outputs are to small changes in the input values (Malczewski, 1999). The most

important elements to consider in sensitivity analysis are criterion weights and

criterion (attribute) values.

16 VI Transportation Engineering Congress (CIT 2004), Doctorate Transport Policy Course (academic year

2003/2004) of the Civil Engineering School of the Polytechnic University of Madrid. The sample contains

a total of 38 questionnaires.


- 85 -

4.6.1 Weight sensitivity

Sensitivity to attribute weights is perhaps more important than to attribute values,

because they are the essence of value judgments (Malczewski, 1999). The

suggested approach consists in systematically carrying out sensitivity tests varying

the weights attached to each criterion, in order to detect the values that provoke a

shift in the ranking of alternatives. The comparison of the ranking of alternatives

obtained using each weight profile makes it possible to analyse the robustness of

the results to changes in the preference strength attached to each criterion.

4.6.2 Attribute value sensitivity

The other sensitivity concern is sensitivity due to errors in estimating the attribute

values. The values obtained depend on many factors beyond the control of the DM.

These uncontrollable factors are referred to as states of nature or states of

environment (Malczewski, 1999). The states of nature, such as the state of the

economy (e.g. recession, inflation) reflect the degree of uncertainty about decision

outcomes. For each alternative there is a set of possible outcomes, depending on

the corresponding state of nature.

One popular procedure to deal with different states of nature is the design of

scenarios. A scenario represents a plausible assumption on the future development

of external factors: i.e. it represents a particular state of nature. Therefore,

scenario construction techniques (see Rehfeld, 1998; Banister et al., 2000a; Hey et

al., 2002) allow isolating the impacts caused by transport infrastructure

improvements from the ones stemming from the, unknown, future development of

external variables to the assessment procedure.

The assessment procedure defines nxm scenarios (S) as a combination of n

network alternatives (A) and m assumptions on the development of these external

variables (E), as Table 4.7. shows.

The reference scenario (S00) combines a business-as-usual (BAU) prognosis

of external variables, and a ‘do-nothing’ network alternative, i.e. that with the

network of the base year. The comparison of scenarios of the same line, when

compared to the ones of the reference scenario, allow for the isolation of the effects

of a new infrastructure. In contrast, the comparison of scenarios of the same row

gives insight into the effects of changes in the development of external variables.


- 86 -

Table 4.7: Matrix for scenario building

NETWORK ALTERNATIVES

A0

(Do-nothing) A1 … … An

E0

(BAU)

S00

(Reference

scenario)

S01 S0n

E1 S10

…

…

EXTERNAL

VARIABLES

Em Sm0

Finally, the resulting ranking of alternatives under different scenarios allows

for the assessment of the robustness of the results to external factors of the

methodology.

Chapter 5 – CASE STUDY DESCRIPTION

- 87 -

5. CASE STUDY DESCRIPTION

The case study will deal with the assessment of the extension of the Spanish High

Capacity Road (HCR) and High-Speed Rail (HSR) networks as included in the

Spanish Transport and Infrastructure Strategic Plan 2005-2020 (PEIT) (Ministerio

de Fomento, 2005). The PEIT objectives have a strategic nature and they include

the enhancement of the transport system’s efficiency and its general sustainability,

the contribution to social and territorial cohesion and the promotion of economic

development and competitiveness. This strategic nature makes the PEIT especially

appropriate for the application of the proposed methodology. This Chapter first

includes an Introduction, and Section 5.1, with some specific aspects of the

application of the methodology. The case study is subsequently briefly

characterized in Section 5.2. Finally, the assessment framework is described in

Section 5.3.

5.1 Introduction

The methodology described in Chapter 4 is aimed at providing an integrated

assessment framework for a hypothetical situation in which the planner is

confronted with the problem of selecting one from a set of alternatives for the

extension of a national transport infrastructure network.

The case study described in this Chapter provides a first approximation to

the full application of the methodology, which allows testing its validity. However,

the full application of the methodology is beyond the scope of this thesis, as it

would require a great amount of data processing work as well as the existence of

already developed modeling tools, which is not the case for Spain. One of these

modeling tools is a national transport demand model (see Figure 4.3.).

Unfortunately, Spain has not yet developed its own transport demand model,

although its development is currently included in the Government’s research

agenda. This problem has been solved with the utilization of average travel time

elasticities in order to roughly estimate traffic growth derived from the capacity

extension. The calculation of GHG emissions faces a similar problem, as it requires

the existence of an inventory of vehicle fleet composition and related emission


- 88 -

factors. This issue has been solved using TREMOVE (Transport & Mobility Leuven

and K.U.Leuven, 2006), a pre-existing model at a EU scale.

Another issue is the selection of one out of a set of network alternatives. The

case study does not handle this issue, as it assesses the effects of the already

approved alternative: the one included in the PEIT. However, the methodology has

included a MCA framework in order to allow for the possibility to select one among

a set of alternatives. The design of a fictitious set of alternatives would have

constituted an interesting theoretical research exercise but it would have implied

again a great amount of modeling work.

In summary, although a full application of the methodology is hindered by the

above limitations, the validity of the fundamental scientific added value of the

suggested approach is considered to be sufficiently tested with its application to the

case study, which has made it possible to test the robustness of the method and its

explanatory potential of strategic effects.

5.2 Case study characterization

Spain constitutes an interesting case study in an EU context. Due to its status of

cohesion country it has received substantial support from European Funds for its

infrastructure development in the last two decades. This was particularly the case

for transport, in which Spain received a third of the total investment in improving

the transport network in Objective 1 regions over the periods 1994-99 (CEC, 2001)

and 2000-2006 (EC, 2004), contributing in an average of some 20%-30% of the

Ministry of Public Works and Transport infrastructure expenditure (Ministerio de

Fomento, 2005). These investments were mainly dedicated to the extension of the

Spanish HCR and HSR networks.

The result has been that Spain has reduced its disparities in network

endowment with the rest of the EU significantly. This fact, along with the

progressive convergence of Spanish GDP per capita values has meant that this

financial support will be substantially reduced in the near future.

This subsection provides the reader with contextual information on the case

study so that the assessment results included in Chapter 6 can be correctly

interpreted and analyzed. For this purpose, it includes a review of the current

situation of the transport infrastructure networks (subsection 5.2.1) and the socio-

economic system (subsection 5.2.2). Subsequently, the main challenges of the

Spanish transport system are described in subsection 5.2.3.


- 89 -

5.2.1 The surface transport infrastructure networks

5.2.1.1 Road network

Figure 5.1 represents the Spanish1 structural road network, which comprises a total

of 24,797 km divided into national, regional and local roads. The High Capacity

Road (HCR) network has a marked radial nature and comprises 10,200 kilometers

of highways, dual carriageways and toll motorways.

Figure 5.1. Spanish road network (2005)

Source: Ministerio de Fomento (2005)

The improvements of the Spanish road network were historically done

through the development of a hierarchical radial road structure, which gave rise to

a higher polarization of the spatial system. This has motivated the corresponding

economies of scale and contributed towards economic development and European

integration objectives (Ministerio de Fomento, 2005). However, although this radial

structure has a number of advantages in terms of network efficiency, its spatial

equity effects are not desirable (EC, 1999). The main challenges for the road

network are its transformation from its radial structure into a grid mesh and a

reduction of the surface of the areas with accessibility deficiencies.

1 Canary Islands, Ceuta, Melilla and the Balearic Islands are not included in this description of the

transport system. In what follows, the term “Spanish” will be referred to as the national territory

included in the Iberian Peninsula.


- 90 -

5.2.1.2 Rail network

The Spanish rail network is basically constituted by nearly 14,000 km of

conventional rail network (Iberian track gauge) and nearly 1,100 km of high-speed

rail network (UIC track gauge). The HSR network is therefore in an underdeveloped

stage compared to that of the HCR. A huge investment effort is currently been

made to transfrom the conventional rail network into a HSR network, in order to

improve rail accessibility and to harmonize it with the rest of the European network.

Similarly to the HCR network, the rail network also presents a marked radial

nature, as Figure 5.2 shows.

Figure 5.2. Spanish rail network (2005)


5.2.2 The socio-economic system

5.2.2.1 Administrative divisions

The classification of the Spanish territory into administrative boundaries, its number

and its correspondence with the EU’s NUTS2 nomenclature have been summarized

in Table 5.1 and has been represented in Figure 5.3.

2 Nomenclature of Territorial Units for Statistics


- 91 -

Table 5.1: Spanish administrative divisions and their NUTS correspondence

Administrative unit NUTS level # of divisions

Group of Autonomous Regions NUTS-13 7

Autonomous region NUTS-2 17

Province NUTS-3 50

Municipality NUTS-5 8,109

Figure 5.3: Spanish NUTS divisions

5.2.2.2 Spatial distribution of population

Spain suffers from a rather polarized spatial distribution of population. The two

main urban agglomerations are Madrid and Barcelona, which together account for

nearly 26% of total population4. These two cities and coastal areas concentrate the

higher population densities (see Figure 5.4), whereas less populated areas are

located in inner and/or rural areas, which furthermore suffer from progressive

population falls.

3 The areas belonging to NUTS-1 level are not governed or controlled by a specific national entity. This

division was only made with statistical aims. 4 Data obtained from the INE (National Statistical Institute) database.


- 92 -

Figure 5.4: Population density

Source: Ministerio de Vivienda (2005)

The Spanish system of cities has a hierachical structure, constituted of

(Ministerio de Fomento, 2004a):

� International metropolitan areas: Madrid and Barcelona, which represent the

two major attraction poles of the system.

� National metropolitan areas, defined as those with population between 500,000

and 1,500,000 inhabitants. Their scarcity and highly polarized spatial

distribution hinders a balanced development and favors the appearance of weak

areas throughout the country.

� Sub-regional capitals: this level includes province capitals and urban

agglomerations over 50,000 inhabitants. In this level there is a higher spatial

balance. In general, coastal areas are sufficiently structured around these sub-

regional capitals, whereas inner areas suffer from higher deficiencies with large

areas with unstructured urban systems.

� The rest of capitals and urban areas: they include the rest of province capitals

and cities or urban agglomerations under 50,000 inhabitants.

Figure 5.5 represents the system of cities of the study area.


- 93 -

Figure 5.5: Study area system of cities

Source: National Statistics Institute (INE)

During the last few decades there has been a trend towards the

consolidation of this polarizing process of the spatial distribution of the population.

In this context, the current main challenges for the Spanish spatial system are to

reverse the processes of concentration, agglomeration and conurbation of

settlements; densification of coastal areas; preservation of North-South

imbalances, and the upsurge of large underpopulated or low density areas

(Ministerio de Vivienda, 2005).

5.2.2.3 Spatial distribution of economic activity

The Spanish economy has grown at higher rates than EU average in the last two

decades, which has facilitated its convergence with average GDP per capita UE

levels, from 87% of EU25 in 1996 to 97% in 2003. Northeast and Madrid NUTS-2

regions are those concentrating higher GDP per capita levels, above EU25 average,

whereas South, Centre and Northwestern regions suffer from below average GDP

levels, as Figure 5.6 shows.


- 94 -

Figure 5.6: Growth in GDP per head in Spain, Spanish NUTS-2 regions and EU15 in terms of EU25 average (PPS5) 1995-2003

60

70

80

90

100

110

120

130

140

150

1995 1996 1997 1998 1999 2000 2001 2002 2003

year

GDP per head in PPS (Spain=100)

EU25 EU15 Spain Noroeste Noreste

C. Madrid Centro Este Sur

Source: EUROSTAT

These marked disparities in regional income per capita levels within Spanish

regions do not show a balancing trend. This can be observed in Figure 5.7 : Spanish

higher and lower income regions have kept their relative position in terms of the

national average for the 1995-2003 period, whereas EU25 and EU15 values have

converged with that of Spain. From a cohesion perspective, this unequal regional

development is unacceptable6.

5 Purchasing Power Standard. 6 Donaghy (2003) analyzes differences in Spanish regional economies and how these differences might

be taken into account in designing policies to reduce regional inequality.


- 95 -

Figure 5.7: Trends in GDP per head in Spanish NUTS-2 regions, EU15 and EU25 in terms of Spain’ average, 1995-2003

60

70

80

90

100

110

120

130

140

150

1995 1996 1997 1998 1999 2000 2001 2002 2003

year

GDP per head in PPS (Spain=100)

EU25 EU15 Spain Noroeste Noreste

C. Madrid Centro Este Sur

Source: EUROSTAT

5.2.3 Current challenges of the Spanish transport system

The main challenges that the Spanish transport system is now confronted with can

be summarized under three main categories, namely: accessibility spatial

imbalances, the high increase in mobility and the related transport environmental

impacts, and the improvement of cross-border connections, due to the Spanish

peripheral location in the EU context. They are briefly described in subsections

5.2.3.1 to 5.2.3.3.

5.2.3.1 Accessibility imbalances

The presence of marked spatial imbalances in accessibility is illustrated below with

maps extracted from the ‘Diagnosis of the transport system’ carried out in the PEIT

(Ministerio de Fomento, 2005), both for road and rail modes.


- 96 -

Figure 5.8 shows accessibility contours7 for the road mode. As can be

observed in this Figure, outstanding high-accessibility corridors (dark brown,

maximum accessibility) concentrate along the radial corridors of the HCR network.

Within these corridors, extremely high accessibility values can be found in the

nodes where motorways converge. On the other hand, regions located outside

these high accessibility corridors, particularly in mountainous areas (yellow, low

accessibility) have persistent deficient access.

Figure 5.8: Accessibility by road (2005)


For the rail mode these marked imbalances caused by the spatial distribution

of high and low performance infrastructures are more significant. Figure 5.9 shows

that the high level of accessibility attained along the corridors of the HSR lines

Madrid-Seville and Madrid-Lleida are outstanding, with the nodes with maximum

7 An adaptation of the network efficiency accessibility indicator (Gutiérrez and Monzón, 1998) including a

gravitational component has been used for this purpose. This indicator has been used in previous studies

at a national scale (see e.g Monzón et al., 2005). Its theoretical foundations can be found in Subsection

3.2.2. and its formulation in Equation (4.1.).


- 97 -

accessibility concentrated in the stations along these HSR lines. Besides, the lack of

cross-border permeability becomes particularly stressed.

Figure 5.9: Accessibility by rail (2005)


Both for road and rail modes, these high-performance infrastructures are

less permeable for the territory as a whole and define a dual territory where

efficient access is restricted to a few nodes. This is particularly stressed for the rail

mode, where spatial separation between stations is necessarily higher than that of

motorway junctions. Moreover, following the guidelines of the ESDP (EC, 1999), the

PEIT stresses that the role of infrastructures with lower levels of performance is

underestimated, as they may provide capillary access in certain regions and

contribute to address local development objectives, although they are wrongly seen

as incompatible with the region’s development expectations.

5.2.3.2 High mobility increase and related environmental impacts

Transport demand has significantly increased in recent decades in Spain. This

increase has been estimated to be 88.7% for passenger and 50% for freight


- 98 -

interurban transport8 in the 1990-2003 period. For passenger transport, air has

been the fastest growing mode (164.8% increase), followed by road (91.1%) and

rail (26.6%). Traffic in the rail mode is slightly increasing mainly due to the

completion of HSR lines, and therefore rail is becoming progressively more

specialized in certain connections offering high-quality service, in which it can

compete with road and/or air transport. In the same period, goods transport

volume has almost doubled, with road mode experiencing a 123.8% increase,

followed by pipeline (142.5%), maritime (22.1%) and rail (6.9%).

Modal share in Spain has experienced a significant change in the last fifty

years. While in the 50’s rail was the most used mode of transport, both for

passengers (60%) and freight (36%), this position has been occupied by the road

mode in both cases. For passenger transport, road transport represents a 91% of

total passenger-km, followed by rail (5%) and air (4.3%). Freight transport is also

dominated by the road mode (85%), followed by the maritime mode (10%).

This high increase in motorised mobility has given rise to a number of

environmental concerns. At the national level, the increase in GHG emissions

threatens the compliance with international commitments. At the urban level, the

impact of transport on health (air quality, noise, healthy mobility habits, etc.) is

also on the Spanish political agenda. The evolution of the above mentioned trends

in conjunction with that of mobility and GDP have recently been analyzed by Pérez-

Martínez and Monzón, (2006). Their findings are summarized in Figure 5.109.

As can be seen in the Figure, both passenger and freight mobility have

increased at higher rates than GDP, which is contrary to the Spanish objectives is,

in line with EU guidelines to decouple this mobility increase from economic growth

(Aparicio et al., 2004). The Figure also highlights the bad performance of transport-

related GHG emissions, which have increased in a 47% in the 1990-2003 period.

Finally, another environmental concern at the national level is related to the

progressive occupation of land and the subsequent habitat fragmentation, with its

extremely negative effects on biodiversity. This is a crucial issue in Spain, given the

wealth of its natural heritage and the vast amount of environmentally vulnerable

areas.

8 Inter-urban mobility represents 81 % and 83 % of this total passenger and freight mobility,

respectively. 9 An extensive review of major environmental indicators related to the transport sector can be found in

the TRAMA Report, recently published by the Spanish Ministry for Environment (Pérez-Martínez and

Monzón, 2005).


- 99 -

Figure 5.10: Trends in mobility, GDP and emissions in Spain, 1990-2003

1990 1992 1994 1996 1998 2000 2002 2004

Index

(1990=

100)

60

80

100

120

140

160

180

200Green house gases Acidifying substancesOzone precursorsParticlesPassengersFreightGDP

Source: Pérez-Martínez and Monzón (2006)

5.2.3.3 Cross-border relations

The peripheral situation of Spain in the EU has been aggravated after the

Enlargement, which has moved the EU centre of gravity eastwards (Spiekermann

and Neubauer, 2002). Furthermore, the growing integration of European economies

has caused international transit traffic in Spain to rise significantly in recent years

(Sánchez and Aparicio, 2004). Moreover, there is also potential for expansion of the

flows between the Maghreb and Europe10. Finally, the promotion of the parallel

development of cross-border regions is in the EU agenda11 and it is considered an

‘European added value’ (van Exel et al., 2002). Given all the above reasons, it is

not surprising to find that the PEIT considers the reinforcement of cross-border

links as a crucial prerequisite for the promotion of economic development and

competitiveness (Ministerio de Fomento, 2005). However, environmental quality of

cross-border areas and the related impacts of new connections cannot be taken out

of the analysis, especially in the case of the Pyrenees area (Sánchez and

Zamorano, 2006).

10 Here the PEIT deals with and encourages the promotion of the technical studies and work begun by

Spain and Morocco in connection with the Fixed-Link project across the Straits of Gibraltar. This is, in

any event, a long-term project, which may exceed the PEIT horizon. 11 Initiatives such as the INTERREG stress the concern from the EU that national borders should not be a

barrier to the balanced development and integration of the European territory. Section B of INTERREG

III concentrates on cross-national co-operation, contributing to an integrated and harmonious territory

across the EU.


- 100 -

5.2.4 The Strategic Infrastructure and Transport Plan 2005-2020 (PEIT)

The definitive version of the PEIT was approved after subjecting of the PEIT

proposal document (Ministerio de Fomento, 2004b) to a public consultation

procedure. The PEIT objectives have been structured in four fields, which have

been summarized below (Ministerio de Fomento, 2005):

� To enhance the system’s efficiency in terms of the quality of the services

actually provided, and to deal with the needs for the mobility needs of people

and flows of goods (…)’ .

� To enhance social and territorial cohesion by ensuring equitable conditions of

accessibility throughout the country (…) and identifying the potential

beneficiaries of infrastructure and transport policy, avoiding regressive transfers

of income.

� To contribute to the system’ general sustainability by compliance with the

international commitments in the European environmental provisions, in

particular in relation to GHG emissions.

� To promote economic development and competitiveness, by enhancing the role

of Spanish urban and metropolitan areas, reinforcing cross-border links and

promoting R&D&I programmes and technological advances (…).

The PEIT establishes a set of guidelines to achieve these objectives. In relation

to territorial policy objectives, the focus is set on the achievement of a ‘territorial

balance and enhanced accessibility’. For this purpose, the development of land

transport networks should aim at correcting ‘the radial systems of the past,

establishing connections with the other networks, limiting territorial concentration

of high-capacity infrastructures and adjusting services to the intensity of flow’

(Ministerio de Fomento, 2005, p 58). Furthermore, it also demands the

development of ‘cross-border links between Autonomous Communities with land

borders and the regions of Portugal and the South of France (…) to channel their

economic and cultural relations’. However, it stresses that this development should

follow ‘specific criteria which avoid their de facto transformation into alternative

corridors for large transport flows’ (Ministerio de Fomento, 2005, p 58).

Finally, it is also important to include the economic estimate of PEIT projects.

Total PEIT planned actions amount for over €248 billion, of which high-performance

rail network investments amount to €83.450 billion (33.5% of total budget), whilst

high-capacity road networks is €32.105 billion (12.9 % of total budget). The

financial framework of the Ministry of Public Works and Development will

presumably have to deal with a possible cut in European funds. If the investment

levels of the last few decades are to be maintained, this may ultimately demand an


- 101 -

increase in off-budget financing sources12, in order to comply with the requisite that

the existing budgetary stability commitment is fulfilled (Ministerio de Fomento,

2005).

5.3 The assessment framework

In this section are outlined the basic characteristics of the assessment framework.

These include the definition of the assessment time horizon and the delimitation of

the study area; the definition of alternatives and the description of the procedure

for the generation of the GIS database.

5.3.1 Assessment time horizon and delimitation of the study area

The assessment base year is 2005 and the assessment time horizon is set to 2020,

as established in the PEIT (Ministerio de Fomento, 2005).

The study area basically comprises the Spanish mainland. This basic study

area has been extended to include cross-border regions in neighbouring countries,

which include Portugal and the three southern French NUTS-2 regions, although

with a higher spatial aggregation level. The study area and the corresponding lower

aggregation level used (municipalities in Spain, districts in Portugal and

departments in France) is represented in Figure 5.11.

Widening the potential destinations to include these ‘external’ zones reduces

non desired border effects in Spanish cross-border regions. Moreover, some

improvements in Spanish links have to be assessed under a cross-border

perspective to take into account spillover effects, as justified in Section 4.3.1.

5.3.2 Definition of alternatives

The evaluation is carried out for the assessment time horizon 2020, on the basis of

the comparison of the ‘do-nothing alternative’ (A0) with the ‘PEIT alternative’

(APEIT). Both alternatives share their corresponding socio-economic data, which has

been obtained from estimates from existing time series data, as detailed in section

5.3.3.3.

The only difference between alternatives corresponds to the Spanish

transport infrastructure networks. Land transport infrastructure networks in

Portugal and France are also identical between both alternatives, and they

12 The PEIT financing strategy sets a 40.5% as the amount of the aforementioned off-budget financing

source.


- 102 -

correspond to the estimates of the European Commission for 202013. This way the

effects from the Spanish network extension can be isolated from those derived from

the development of socio-economic variables and the infrastructure extension in

neighbouring countries in the period 2005-2020. Hence, in the ‘do-nothing

alternative’ the land infrastructure networks in Spain have been modeled as those

existing in the base year (2005), whilst in the ‘PEIT alternative’ road and rail

infrastructure networks correspond to those defined in the PEIT for 2020.

The planned investments of the PEIT in terms of high-performance land

transport infrastructure network extension include the development of the HCR

network from 10,200 km to nearly 15,000 km in 2020 (Figure 5.12) and from

1,100 km to 7,200 km of the HSR network (Figure 5.13).

Figure 5.11: Delimitation of the study area

13 Maps obtained from http://europa.eu.int/comm/


- 103 -

Figure 5.12: Road network of the PEIT alternative (APEIT)

Figure 5.13: Rail network of the PEIT alternative (APEIT)


- 104 -

5.3.3 Generation of the GIS database

5.3.3.1 Transport infrastructure networks

In order to calculate accessibility values, a dense intermodal (road and rail)

network was modeled with the support of a GIS; in this case the ArcGis software

was used. Accessibility values are obtained for each node of the network, which

coincide with the nodes of the road network (nearly 12,000). Using interpolation

techniques, aggregated NUTS-5 values in Spain, and NUTS-3 values in Portugal and

France, are derived from node values.

The first step was to model the road network of the do-nothing alternative. A

vectorial GIS was used, in which the network is modeled as a graph comprised of a

set of nodes and arcs. For each arc on the road network, the length, estimated

speed according to type of road 14 (120 km/h for motorways, 110 for expressways,

90 for interregional roads, 80 for other roads and 50 for urban roads) and resulting

travel time were also recorded. For the rail mode, each arc is given a commercial

speed according to both infrastructure and quality of service characteristics. Rail

network modeling tasks are significantly more complex than those of the road

mode, as it is necessary to include track gauge (Iberian/UIC) data, the location of

the stations and frequency of service information in order to calculate travel times,

as detailed in the following section.

5.3.3.2 Travel time calculations

An O-D matrix with travel times is necessary for the accessibility calculations. The

set of origins coincides with that of the destinations, and is made up of the nearly

8,000 municipalities of Spain, the 19 French departments and the 18 Portuguese

districts. This results in approximately 8,000*8,000= 64,000,000 travel time

calculations for each alternative.

For the road mode, this matrix is directly calculated from arc speeds, using

minimum path algorithms of the GIS software used. For the rail mode, calculations

are more complex. The spatial separation between stations makes the modeled rail

network unavoidably multimodal, as in the majority of cases, neither the origin i

nor the destination j have a train station. Hence, rail travel times between each i-j

14 The use of free flow speeds is common in long-scale studies (see Martín et al., 2004; Gutiérrez and

Monzón, 1998; Schürmann et al., 1997). The introduction of congestion effects is in a sense included in

this case via the low speed attached to urban roads (50 km/h). This speed reduction is rather accurate

in the surroundings of large agglomerations (given the high density of the modeled road network), which

are precisely those with congestion problems in Spain.


- 105 -

pair - IRAIL (i, j)- is made up of five terms, following the approach used in previous

similar studies (see e.g. López et al., 2006a), as Equation ( 5.1 ) shows:

� Road travel time form the origin i to the nearest train station (Si): IROAD (i, Si),

� a penalty for the intermodal change road-rail of 20 min (IC),

� rail travel time between Si and the nearest train station to the destination j (Sj):

IRAIL (Si, Sj),

� a second penalty for the intermodal change road-rail of 20 min (IC), and

� road travel time from Sj to the final destination j: IROAD (Sj, j).

),(),(),(),( jSIROADIChSSIRAILIChSiIROADjiIRAIL jjii ++++= ( 5.1 )

Furthermore, IRAIL (Si, Sj) is computed following Equation ( 5.2 ):

IFrIBoIGaITrSSIRAILSSIRAIL jiNji ++++= ),(),( ( 5.2 )

where:

� IRAILN(Si, Sj) is computed as the total travel time between stations Si and Sj,

computed from stored average speeds on the rail arcs linking both stations,

� ITr is a penalty due to train transfers. It has been estimated as 15 minutes for

each hour travel time, in those trips exceeding 4 hours travel time 15,

� IGa is a time penalty of 20 minutes in order to simulate the change from Iberian

broad gauge to European standard gauge,

� IBo is a time penalty due to the crossing of national boundaries, established in

30 minutes, and

� IFr is a time penalty due to frequency of rail service. It is more complex to

calculate and is explained in the next paragraph.

The estimation of a train frequency table for the year 2020 would require

carrying out a prognosis on a large amount of data of train frequencies between

each origin-destination pair. Instead, a simplified procedure has been applied,

following the approach developed by (Megía, 2002). It estimates the frequency of

train services (Ns) between each pair of stations (Si, Sj) using a gravitational model,

as Equation ( 5.3 ) shows:

15 The assumption behind this decision is that trips exceeding this 4 hour length have a high possibility

to require a train transfer. This approach has been used in previous similar studies (López et al., 2006b;

Gutiérrez et al., 2006).


- 106 -

( )( )

β

α

⋅

⋅=2

2

),(

),(),(

ji

ji

SS

jisSSIRAIL

SSD

PP

SSN

ji

( 5.3 )

Where A and B are parameters to be calibrated, P represents the population

(expressed in thousands of inhabitants) and D the Euclidean distance (in km).

The time penalty IFr is considered inversely proportional to this estimated

frequency of service Ns, with a maximum value of one hour if there is only one train

service between the stations. The model has been calibrated based on existing train

frequencies in Spain (2005 data)16. The resulting values of the parameters were

3.011 for A and 0.366 for B, with a correlation coefficient R2 of 0.69.

5.3.3.3 Socio-economic data

Population is the selected variable to measure each destination’s attractiveness in

the accessibility model. The population for Spain and cross-border regions for 2020

has been estimated on the basis of prognosis of available historical data series. The

information sources used were the corresponding national Statistical Institues,

namely the INE 17 in Spain, the INSEE 18 in France, and the INE19 in Portugal. In the

three countries, population data correspond to prognosis based on past trends of

these variables for 2020, based on linear regression models20.

In Spain, the selected destination centres correspond to the centroids of the

approximately 8,000 municipalities of the Spanish mainland. Centroids in Portugal

and the three southern French regions have been included as destination centres at

a more aggregated level, namely the 18 districts (distritos in Portuguese) in

mainland Portugal and the 19 departments (departments in French) in the three

southern French regions.

16 The sample contained train frequencies between Madrid and Barcelona to the rest of capitals of

Spanish provinces. 17 Instituto Nacional de Estadística (www.ine.es) 18 Institut National de la Statistique et des Études Économiques (www.insee.fr) 19 Instituto Nacional de Estatística (www.ine.pt) 20 A simple linear regression model to estimate future population development was used in order to

derive 2020 population data.


- 107 -

In the accessibility calculations with origins in Spain, populations in France

and Portugal have been reduced by a factor of 0.25, to take into account that

destinations in neighbouring countries are visited less than national ones21.

5.3.3.4 Environmental data

The first step in order to introduce the environmental data in the GIS consisted in

joining all the available information on Spanish habitats included in the Annex I of

the Habitats Directive 92/43/EEC22. In Spain, this information is provided by the

Ministry for the Environment23. According to the classification of the Natura 2000

network defined in the Directive, Spain counts on 1,068 Sites of Community

Importance (SCIs), (see Figure 5.14) and 425 Special Protection Areas (SPAs) (see

Figure 5.15), which represent a 22.04% and a 17.87% of the Spanish territory.

Figure 5.14: Sites of Community importance (SCIs)

Source: Ministry for the Environment

21 The same approach, with similar reduction factors is widely applied in international accessibility

studies (Bruinsma and Rietveld, 1993; Gutiérrez and Urbano, 1996). 22 Council Directive 92/43/ECC of 21 May 1992 on the conservation of natural habitats and of wild fauna

and flora, O. J. L 206, 22.07.92. 23 Information available on the Ministry for the Environment’ webpage (www.mma.es) split by NUTS-3

divisions.


- 108 -

Figure 5.15: Special Protection Areas (SPAs)

Source: Ministry for Environment

All the above data have been joined in a single map of Spanish Habitats

(represented in blue colour in Figure 5.16). The information source for the

delimitation of habitats is the Habitat Map developed by the Biodiversity General

Directorate of the Ministry for the Environment (Ministerio de Medio Ambiente,

2005). These habitats were divided into 120 classes which correspond to the

aforementioned typology established in the Habitats Directive. The total area

classified as Habitats represents a 28.43% of Spanish mainland area. This

information made it possible to define the patches of the do-nothing alternative.

For fragmentation calculations, road and rail networks are considered as

equal, independently from their typology. They have been modeled as 100 m width

corridors. The area of each habitat crossed is therefore reduced in the amount of

the area of the corresponding corridor that crosses it.


- 109 -

Figure 5.16: Spanish habitats map

Source: Ministry for the Environment


- 110 -

Chapter 6 – ASSESSMENT RESULTS

- 111 -

6. ASSESSMENT RESULTS

This Chapter includes the results obtained for the PEIT road and rail alternatives, in

each performance indicator. The results were not restricted to the final computation

of performance indicators, but were also complemented with an introductory

assessment of the correspondent indicator related aspects. For this purpose, a

graphical analysis with the presentation of key maps and intermediate results, has

been included. For each performance indicator, the results were split into road and

rail PEIT alternatives.

6.1 Efficiency

The efficiency assessment starts with an overall analysis of the accessibility

patterns in the do-nothing and PEIT alternatives. The objective of this ‘initial’

assessment is to explain in more detail the reasons behind the final value of the

performance indicators. However, it is beyond the scope of this thesis to analyse

comprehensively each of the resulting accessibility values and maps. Hence, the

analysis focused on those aspects considered most relevant for the subsequent

calculation of the performance indicators.

6.1.1 Network efficiency (NE)

6.1.1.1 Road mode

The resulting network efficiency accessibility values in the do-nothing alternative

have been mapped1 in Figure 6.1. The formulation of the accessibility indicator

chosen –the network efficiency accessibility indicator (Gutiérrez and Monzón,

1998)- is the one included in Equation (4.1.).

The map clearly shows the contrasts between areas with better and worse

accessibility values. Areas with best accessibility values are concentrated along the

axes of the HCR network, therefore resulting in a radial pattern of high accessibility

1 In order to improve the graphical representation of the maps, an interpolation using the Inverse

Distance Weighted (IDW) option (Exponent=1, Number of neighbours=6) of the ArcGis software has

been used. This interpolation has been used for graphical presentation purposes only.


- 112 -

corridors. This effect is particularly visible in those nodes in which these axes

intersect, such as in Madrid, Barcelona, or Zaragoza. The existence of these

‘corridor effects’ (surface accessibility) is a characteristic feature of road

infrastructure extensions, contrary to the tunnel effect characteristic of HSR

extensions (point accessibility) (Gutiérrez et al., 1996; Gutiérrez, 2004).

Figure 6.1: Network efficiency. Alternative A0. Road mode

It can also be observed that the geographical distance to main population

centres does not affect the results. Indeed, certain locations of the geographical

periphery result in acceptable network efficiency values, such as Barcelona, whilst

others located in more central areas suffer from accessibility deficiencies, such as

Cuenca. The radial nature of the HCR network creates an upsurge of inaccessible

areas, mainly concentrated in spaces between corridors. This effect is clearly seen

in the isolated ‘islands’ that appear between the motorways that access Madrid.

Figure 6.2 represents the resulting accessibility patterns of the road PEIT

alternative. The overall spatial pattern is similar to that of the do-nothing

alternative represented in Figure 6.1. The comparison of both Figures shows that

due to the HCR extension both areas with higher accessibility values are extended

and areas with accessibility deficiencies are reduced.

The above mentioned improvements can be more easily detected with the

analysis of Figure 6.3, in which percentage change in accessibility values (compared


- 113 -

to the do-nothing situation) have been mapped. It can be observed that, on the

one hand, higher accessibility benefits mainly concentrate in those regions in which

new infrastructure is included in the PEIT, although in some cases this effect is

spread to cover adjacent regions. These regions are located mainly in the west

Portuguese frontier, spread to Badajoz, Córdoba and Ciudad Real; Teruel-Cuenca

and their surroundings; western Pyrenees-Navarra-La Rioja; and some coastal

areas in Asturias and Cantabria. On the other hand, Madrid, eastern Andalucía and

eastern Cataluña concentrate lower relative accessibility gains.

A selection of network efficiency accessibility values corresponding to NUTS-

3 capitals in the do-nothing and PEIT alternatives, as well as the percentage change

between them, is included in Table 6.1. The capitals with better accessibility levels

are concentrated in the corridors of the axes of the HCR network, in particular in

those were they intersect. This is the case of cities such as A Coruña, Barcelona,

Burgos, Girona, Madrid, Murcia, Tarragona or Valencia, with results below 1.30. On

the other extreme of the list appear capitals which are out of the main

infrastructure corridors and therefore lack any efficient connections with the main

population centres. This is the case of e.g. Cuenca and Teruel, with values over

1.40.

Higher relative accessibility gains concentrate, as the maps already show, in

those capitals nearer the Portuguese frontier: Zamora, Salamanca, Badajoz,

Cáceres and Huelva; as well as in inner capitals such as Cuenca, Teruel or Ciudad

Real. They all obtain over 5% accessibility improvements.


- 114 -

Figure 6.2: Network efficiency. Alternative APEIT. Road mode

Figure 6.3: Network efficiency. Relative differences Alternative A0 vs. APEIT. Road mode


- 115 -

Table 6.1 Network efficiency in Spanish NUTS-3 capitals. Road mode

Name E0 EPEIT % change

A Coruña 1.273 1.246 2.10 Albacete 1.333 1.285 3.57 Alicante 1.303 1.276 2.10 Almería 1.354 1.326 2.05 Ávila 1.344 1.312 2.33 Badajoz 1.337 1.264 5.45 Barcelona 1.292 1.273 1.48 Bilbao 1.318 1.290 2.12 Burgos 1.258 1.225 2.62 Cáceres 1.366 1.289 5.64 Cádiz 1.376 1.330 3.35 Castelló de la Plana 1.333 1.289 3.30 Ciudad Real 1.393 1.298 6.85 Córdoba 1.390 1.312 5.62 Cuenca 1.438 1.364 5.15 Girona 1.279 1.259 1.51 Granada 1.309 1.288 1.56 Guadalajara 1.335 1.291 3.29 Huelva 1.342 1.257 6.31 Huesca 1.360 1.286 5.40 Jaén 1.346 1.313 2.46 León 1.296 1.254 3.20 Lleida 1.329 1.297 2.36 Logroño 1.381 1.310 5.16 Lugo 1.300 1.263 2.81 Madrid 1.279 1.264 1.18 Málaga 1.340 1.319 1.55 Murcia 1.286 1.265 1.62 Ourense 1.349 1.298 3.77 Oviedo 1.310 1.262 3.65 Palencia 1.320 1.283 2.85 Pamplona 1.371 1.297 5.42 Pontevedra 1.355 1.316 2.85 Salamanca 1.343 1.261 6.12 San Sebastián 1.346 1.301 3.32 Santander 1.319 1.275 3.36 Segovia 1.382 1.344 2.78 Sevilla 1.338 1.288 3.72 Soria 1.383 1.312 5.12 Tarragona 1.279 1.251 2.16 Teruel 1.474 1.385 6.04 Toledo 1.336 1.281 4.16 Valencia 1.300 1.265 2.70 Valladolid 1.308 1.262 3.48 Vitoria 1.322 1.290 2.47 Zamora 1.323 1.256 5.07 Zaragoza 1.311 1.256 4.18


- 116 -

After this initial assessment, the network efficiency performance indicator

(NE) was computed, following Equation (4.2.), as the percentage improvement in

the network efficiency accessibility indicator values.

Average accessibility values were computed by aggregating NUTS-5 values,

using the population as the weighting variable. This results in a 1.334 value for the

do-nothing alternative, whereas in the PEIT alternative this value is reduced to

1.299. The resulting value of the network efficiency performance indicator,

following Equation (4.2.), is NE= 2.637; which represents a 2.637% improvement

of network accessibility.

637.2100334.1

299.1334.1=⋅

−=−RoadPEITNE

This value is consistent with the results obtained in previous studies

assessing network efficiency improvements of the Spanish road network (López and

Monzón, 2004; López et al., 2006b). This percentage improvement is low if

compared with e.g. that of the percentage increase in the length of the HCR

network between both alternatives, which is nearly 50% (from 10,200 km to nearly

15,000 km). This is mainly due to the relatively good starting situation in terms of

network efficiency, which leaves reduced room for high percentage improvements.

As the network becomes denser, the marginal increases in its efficiency are

reduced. In the 1980s and early 1990s, in the first stages of the development of

the Spanish HCR network, percentage improvements were significantly higher with

similar increases in network length (Gutiérrez and Monzón, 1998). This fact will be

confirmed with the comparison of road and rail improvements in network efficiency,

carried out in section 6.1.1.2, where the network’s starting situation is

comparatively worse.

6.1.1.2 Rail mode

For the rail mode, the network efficiency indicator highlights more the differences in

transport infrastructure quality than in the road mode, due to the larger difference

in their speeds, as Figure 6.4 shows.

First, it can be observed in Figure 6.4 that the values of the accessibility

indicator have indeed, a much larger range of variation than for the road mode.

HSR Madrid-Sevilla and Madrid-Lleida corridors (along with those areas indirectly

served by them) appear highlighted as zones with significantly higher accessibility

levels than the rest of the territory. Furthermore, the good results obtained in the

Mediterranean and Madrid-Valencia corridors, which enjoy efficient train services,

are also highlighted. As happened with the road mode, given that the network


- 117 -

efficiency accessibility indicator eliminates the influence of the geographic location,

more accessible regions do not necessarily coincide with centraly located ones. This

is the case, for example, of western Andalucía or eastern Cataluña, which despite

being located in the geographic periphery of the Iberian Peninsula, enjoy good

accessibility levels.

Figure 6.4: Network accessibility. Alternative A0. Rail mode

Second, the location of the stations has a strong influence in the resulting

spatial patterns of rail accessibility. On the one hand, the spatial distribution of

train stations, mainly those of the HSR network, determine the presence of ‘islands’

and ‘corridors’ with better accessibility than their surroundings. On the other hand,

interstitial areas within corridors are exposed to the ‘tunnel effect’ (point

accessibility) (Plassard, 1991; Plassard, 1992), characteristic of HSR lines, as

higher accessibility points (stations) alternate with lower accessibility areas in the

stretches between stations. This tunnel effect can be observed in Figure 6.4 in the

HSR corridor Madrid-Sevilla.

This situation experiences a significant improvement with the extension of

the HSR network included in the PEIT alternative, as Figure 6.5 shows. The location

of the stations has a strong influence in the final results, therefore resulting in a

highly marked tunnel effect (see e.g. the Zaragoza-French border stretch).


- 118 -

In general, best accessibility results appear concentrated in the surroundings

of HSR stations, except in those of very large urban agglomerations, such as

Madrid2. Some inner areas in Extremadura, Castilla y León and Aragón still appear

as landlocked areas. They correspond mainly to areas which are not crossed by an

HSR infrastructure.

Relative percentage differences between PEIT and do-nothing alternatives

are mapped in Figure 6.6. It can be seen how higher relative benefits concentrate

in the northwest and southeast quadrants, along with some minor areas in País

Vasco, Cuenca and Teruel. In this case, the range of benefits extends up to nearly

60% in the best cases, due to the ambitious extension of the HSR network

proposed in the PEIT.

The results obtained by NUTS-3 capitals are included in Table 6.2. Their

analysis shows that lower values in the do-nothing alternative concentrate in cities

with a HSR station, such as Madrid, Córdoba, Ciudad Real or Lleida, with values of

the network efficiency indicator below 4.000, whereas worst values concentrate, as

Figure 6.4 suggested, in capitals such as Almería, Santander or Soria, with values

above 5.000. In the PEIT alternative, there is a huge improvement in accessibility.

Furthermore, as in this alternative all province capitals have a HSR station, the

range of variation of accessibility values included in Table 6.2 has significantly been

narrowed. Higher relative gains -note the difference in the magnitude of these

differences when compared with those of the road mode- concentrate in Galicia,

eastern Andalucía, western Castilla y León and costal northern capitals, with

percentage benefits around 50% and even above 55% in some cases.

2 This effect appears due to the high level of dissagregation used for the origins and destinations nodes

(NUTS-5 centroids). For example, in the case of Madrid, this causes nodes located in its vicinity, which

do not have a station, to have a very inefficient connection with Madrid (which has a significant weight,

given that it accounts for nearly 1/6 of total population). This is why the surroundings of Madrid appear

with deficient accessibility values. However, although this is not graphically noticeable, the result

obtained in the Madrid node is good (see Table 6.2), as 1/6 of its destinations are reached with the

better efficiency value: 1. This effect is related to the conflicting issue of the self potential explained in

Section 3.2.2.4.


- 119 -

Figure 6.5: Network accessibility. Alternative APEIT. Rail mode

Figure 6.6: Network accessibility. Relative differences Alternative A0 vs. APEIT. Rail mode


- 120 -

Table 6.2 Network efficiency in Spanish NUTS-3 capitals. Rail mode


A Coruña 4.396 2.086 52.55 Albacete 3.956 2.515 36.43 Alicante 4.207 2.832 32.69 Almería 5.074 2.265 55.36 Ávila 4.813 2.608 45.82 Badajoz 4.108 2.632 35.93 Barcelona 3.993 2.877 27.93 Bilbao 4.765 2.561 46.25 Burgos 4.276 2.360 44.81 Cáceres 4.366 2.623 39.92 Cádiz 3.960 2.369 40.19 Castelló de la Plana 4.034 2.409 40.27 Ciudad Real 3.435 2.341 31.84 Córdoba 3.272 2.212 32.39 Cuenca 4.796 2.514 47.58 Girona 3.643 2.376 34.79 Granada 4.672 2.530 45.84 Guadalajara 4.431 3.418 22.87 Huelva 3.633 2.241 38.31 Huesca 4.079 2.446 40.03 Jaén 4.713 2.624 44.33 León 4.499 2.257 49.84 Lleida 3.422 2.302 32.72 Logroño 4.320 2.514 41.81 Lugo 4.721 2.307 51.12 Madrid 3.800 2.662 29.93 Málaga 4.074 2.188 46.30 Murcia 4.436 2.442 44.96 Ourense 4.282 2.129 50.27 Oviedo 4.900 2.273 53.61 Palencia 4.834 2.327 51.87 Pamplona 4.584 2.398 47.69 Pontevedra 4.208 2.139 49.18 Salamanca 4.221 2.443 42.13 San Sebastián 4.523 2.460 45.62 Santander 5.083 2.266 55.42 Segovia 4.686 2.552 45.54 Sevilla 3.262 2.301 29.46 Soria 5.390 2.812 47.82 Tarragona 3.538 2.350 33.59 Teruel 4.801 2.835 40.95 Toledo 4.215 2.533 39.89 Valencia 4.136 2.852 31.03 Valladolid 4.251 2.181 48.68 Vitoria 4.536 2.456 45.85 Zamora 4.485 2.283 49.09 Zaragoza 3.365 2.187 35.01


- 121 -

As with the road mode, this initial assessment of graphical and key node

values provides the foundations for the calculation of the performance indicator.

The network efficiency performance is computed following Equation (4.2.), as the

percentage improvement in the network efficiency accessibility indicator values.

533.34100395.4

877.2395.4=⋅

−=−RaadPEITNE

Average accessibility values have been computed aggregating NUTS-5

values, using the population as the weighting variable. This results in a 4.395 value

for the do-nothing alternative, whereas in the PEIT alternative this value is reduced

to 2.877. Hence, the resulting value of the performance indicator, NE= 34.533,

can be interpreted as a 34.533% improvement of network efficiency.

As mentioned in Section 6.1.1.1, this percentage increase is significantly

higher than that of the road mode. This is due to the comparatively worse network

efficiency of the rail network in the do-nothing alternative compared to that of the

road mode. Furthermore, the difference in the planned average speeds of HSR

compared to average speeds of existing conventional rail lines is significantly higher

than the one between a motorway and a conventional road, which also causes a

comparatively higher increase in rail improvement values.

6.1.2 Cross-border integration (CB)

The presentation of results follows the same procedure used for the national

network efficiency criterion: first, a graphical analysis and the calculation of values

in key nodes is carried out; subsequently, the performance indicator is computed.

In this case, only those maps showing relative percentage improvements with

respect to the do-nothing situation have been drawn, as the focus of this criterion is

the assessment of the contribution of the PEIT to cross-border integration, rather

than the analysis of the accessibility situation of neighbouring countries. This latter

analysis should be carried out by their corresponding competent authorities.

6.1.2.1 Road mode

First, the analysis starts with the assessment of accessibility improvements in

Portugal. Figure 6.7 represents relative percentage improvements in network

efficiency accessibility values in Portugal, due to the completion of the PEIT

alternative, for the road mode.


- 122 -

Figure 6.7: Network efficiency in Portugal. Relative differences Alternative A0 vs. APEIT. Road mode


- 123 -

It is beyond of the scope of this subsection to analyse comprehensively the

spatial distribution of the resulting changes. However, some general considerations

can be made. The key issue that arises when interpreting the maps is that the

spatial pattern followed by relative improvements is a result of both the planned

cross-border links included in the PEIT and, to a lesser extent, the network distance

to most important destinations3.

Hence, as the northern and southern links (via Porto and Faro, respectively)

already existed in the do-nothing alternative, these regions are those with lower

benefits. In contrast, the central Portuguese regions, such as Guarda, Castelo

Branco or Portalegre, are the ones which benefit more from the new cross-border

links.

The values obtained in each Portuguese district capital, in the do-nothing

(E0) and PEIT (EPEIT) alternatives, have been included in Table 6.3 and are coherent

with what the maps have pointed out. Percentage changes vary from the 4.10%

improvement achieved by Portalegre to the 1.15% obtained by Beja. In summary,

the population-weighted average accessibility improvement in Portugal results in a

2.032%.

Table 6.3: Network efficiency in Portuguese district capitals. Road mode


Aveiro 1.363 1.327 2.64 Beja 1.333 1.317 1.15 Braga 1.369 1.350 1.41 Bragança 1.363 1.320 3.18 Castello Branco 1.464 1.421 2.94 Coimbra 1.363 1.325 2.77 Évora 1.338 1.313 1.82 Faro 1.286 1.250 2.80 Guarda 1.420 1.374 3.30 Leiria 1.339 1.298 3.07 Lisboa 1.329 1.313 1.23 Portalegre 1.421 1.363 4.10 Porto 1.370 1.335 2.56 Santarém 1.332 1.297 2.63 Setúbal 1.308 1.291 1.26 Viana do Castelo 1.362 1.341 1.50 Vila Real 1.428 1.395 2.31 Viseu 1.428 1.387 2.87

3 The lower weight attached to international destinations in the accessibility model also influences the

results. This issue is further analyzed by López et al. (2006a).


- 124 -

Second, accessibility improvements in the three southern NUTS-2 regions of

France were also evaluated. These are mapped in Figure 6.8, which represents

percentage change in network efficiency.

Figure 6.8: Network efficiency in Southern France. Relative differences Alternative A0 vs. APEIT. Road mode

In the French case, given that the motorway connection with Perpignan

already existed in the do-nothing alternative, lower percentage increases

concentrate in the eastern part of the French territory. This means that, as we

move westwards, higher accessibility improvements are achieved. Moreover,

accessibility improvements are progressively reduced with the distance to the

frontier. Table 6.4 includes the network efficiency results obtained in NUTS-3

centroids (department capitals) of southern France, in the do-nothing (E0) and PEIT

(EPEIT) alternatives, as well as the percentage change between them. Indeed, it can

be observed that lower (below 1%) percentage changes concentrate in eastern

departments capitals, such as Ales, Mende, Nimes, Montpellier or Perpignan. Higher

percentage increases do not surpass the 2.60% value recorded in Pau, with the

lowest value (0.45%) being recorded in Ales. In summary, the population-weighted

average accessibility improvement in France results in a 1.479%.


- 125 -

Table 6.4: Network efficiency in French department capitals. Road mode


Agen 1.328 1.307 1.64 Albi 1.367 1.342 1.85 Ales 1.357 1.351 0.45 Auch 1.360 1.333 1.96 Bordeaux 1.272 1.250 1.72 Cahors 1.378 1.357 1.57 Carcassonne 1.336 1.320 1.20 Foix 1.355 1.326 2.16 Mende 1.409 1.397 0.84 Montauban 1.339 1.312 1.98 Mont-de-Marsan 1.349 1.320 2.18 Montpellier 1.262 1.256 0.52 Nimes 1.257 1.251 0.49 Pau 1.347 1.312 2.60 Périgueux 1.350 1.333 1.30 Perpignan 1.300 1.292 0.60 Rodez 1.384 1.364 1.49 Tarbes 1.376 1.346 2.20 Toulouse 1.297 1.268 2.21

After this initial assessment, and as detailed in Section 4.4.1.2., the cross-

border integration performance indicator is computed as a population-weighted

average percentage change in the network efficiency accessibility indicator. This

average has been computed for both Portuguese and French territories, resulting in

an aggregated value of the indicator of CB= 1.771.

This means that average network accessibility improvements in cross-border

regions, due to the completion of the PEIT, accounts for a 1.771% change. This is a

relatively high value, if compared with the 2.637% value obtained in the Spanish

territory. This result confirms the importance of the assessment of spillover effects,

as the literature review suggests.

6.1.2.2 Rail mode

Figure 6.9 represents relative percentage improvement in network efficiency

accessibility values in Portugal due to the completion of the PEIT, for the rail mode.


- 126 -

Figure 6.9: Network efficiency in Portugal. Relative differences Alternative A0 vs. APEIT. Rail mode


- 127 -

The interpretation of the resulting values requires a combined analysis of

each centroid’s population, its starting situation in terms of accessibility, and its

proximity to new HSR stations. Although this analysis is beyond the scope of this

thesis, general considerations can be made, similarly to the road mode. In this

case, the location of HSR stations is a key factor influencing the final results.

Indeed, it can be observed in Figure 6.9 that those centroids in which a HSR station

is not planned, such as Portalegre or Castello Branco, suffer from lower accessibility

gains. Moreover, the effect of the new links spreads through the corridors of the

already existing HSR network. The corresponding values obtained in district capitals

are included in Table 6.5. The average population-weighted accessibility

improvement in Portugal is a 17.2435%.

Table 6.5 Network efficiency in Portuguese district capitals. Rail mode


Aveiro 3.833 3.030 20.95 Beja 3.598 2.686 25.34 Braga 4.667 3.868 17.10 Bragança 6.853 5.817 15.11 Castello Branco 5.311 4.599 13.41 Coimbra 3.775 2.952 21.80 Évora 3.721 2.888 22.39 Faro 3.592 2.437 32.16 Guarda 4.215 3.202 24.03 Leiria 3.832 3.058 20.21 Lisboa 4.619 3.920 15.12 Portalegre 4.849 4.124 14.94 Porto 5.610 4.804 14.36 Santarém 4.560 3.881 14.90 Setúbal 4.369 3.650 16.46 Viana do Castelo 3.751 2.887 23.03 Vila Real 3.902 3.151 19.23 Viseu 4.473 3.567 20.27

Figure 6.10 shows percentage change of network efficiency accessibility in

the three southern regions of France. As happened in Portugal, the proximity to the

stations of the HSR network is one of the main factors determining the final

percentage improvement. This is reflected in Figure 6.10 in that higher percentage

gains, in some cases above 25%, are located in those regions with a better

connection with the stations of the three cross-border planned links (through both

frontier extremes and the one connecting with Tarbes).


- 128 -

Figure 6.10: Network efficiency in Southern France. Relative differences Alternative A0 vs. APEIT. Rail mode

Table 6.6: Network efficiency in French department capitals. Rail mode


Agen 3.459 2.667 22.91 Albi 3.980 3.238 18.64 Ales 3.211 2.591 19.31 Auch 4.694 3.809 18.84 Bordeaux 2.911 2.118 27.23 Cahors 3.872 3.149 18.69 Carcassonne 3.281 2.467 24.81 Foix 4.466 3.538 20.78 Mende 4.226 3.613 14.49 Montauban 3.502 2.725 22.19 Mont-de-Marsan 4.077 3.096 24.06 Montpellier 2.843 2.177 23.42 Nimes 2.851 2.227 21.89 Pau 4.238 3.076 27.41 Périgueux 3.608 2.909 19.37 Perpignan 3.152 2.285 27.51 Rodez 4.238 3.567 15.84 Tarbes 4.496 3.200 28.82 Toulouse 3.299 2.480 24.82

These observations are verified with the numerical results included in Table

6.6. Indeed, Bordeaux, Pau, Tarbes and Perpignan appear as those capitals with


- 129 -

higher percentage increases (in all cases above 25%), whilst capitals located far

from the planned HSR links, such as Rodez or Mende, suffer from lower accessibility

increases, with values around 15%. The resulting value of the average population-

weighted accessibility improvement in southern France is a 23.4663%.

After this initial assessment, and as detailed in Section 4.4.1.2., the cross-

border integration performance indicator is computed as a population-weighted

average percentage change in the network efficiency accessibility indicator. This

average has been computed jointly for both Portuguese and French territories,

resulting in a value of the indicator of CB= 20.179. As happened with the road

mode, the value obtained confirms the significant spillover effects in neighbouring

countries due to the extension of the Spanish HSR network.

This percentage change is significantly higher than that of the road mode (i.e.

1.771, see section 6.1.2.1). The same happened when comparing road and rail

performance indicator values in sections 6.1.1.1 and 6.1.1.2. As justified in these

sections, the main causes for this phenomenon are the differences between the

initial situation of both networks and the higher differences between HSR and

conventional rail speeds, when compared to those of motorways and conventional

roads.

6.2 Cohesion

6.2.1 Regional cohesion (RC)

As detailed in Section 4.4.2.1, regional cohesion effects are assessed analyzing

changes in the spatial distribution of the potential accessibility indicator. In this

case study, the dissagregation level used for the regional cohesion analysis is the

municipality (NUTS-5 equivalent). The sample containing the nearly 8,000 values of

the potential accessibility indicator in the do-nothing and the PEIT alternative is

characterized through the calculation of the set of four inequality indices described

in Section 4.4.2.1.: the coefficient of variation, the Atkinson, GINI and Theil indices.

The assessment results are included below, both for road and rail modes.

6.2.1.1 Road mode

Potential accessibility values are mapped in Figure 6.11. The resulting spatial

distribution pattern shows that the distribution of accessibility clearly has

imbalances between large urban agglomerations, such as Madrid, Barcelona or

Valencia, with above average accessibility levels, and regions located in less


- 130 -

densely populated areas and far from major transport infrastructure corridors. The

latter have accessibility deficiencies, as evident, for example, in Extremadura,

Galicia or the central Pyrenees areas. These results provide a clear example of the

agglomeration effect stemming from the concentration of population and high level

transport infrastructure networks, widely mentioned in the literature review carried

out in Chapter 4.

The interpretation of the results provided by the potential indicator needs to

be carried out taking into account the joint effect of distance (travel time) and

attraction masses (population of the destination) in the relations of each node with

activity centres. Hence, those nodes with better accessibility conditions (a higher

potential) will presumably be those nearer to and better linked with major densely

populated areas.

The general picture follows, therefore, certain core-periphery patterns in the

surroundings of these nodes where the agglomeration effects appear. This effect is

particularly visible in the case of Madrid and Barcelona. Distortions due to the

existence of transport infrastructure (see how in Figure 6.11 accessibility contours

appear distorted in the direction of major network axes e.g. in the surroundings of

Madrid), are combined with those derived from the effect of the attraction masses.

Hence, it appears that although geographically located areas tend to have low

accessibility values, their major cities usually enjoy a higher potential than their

surroundings. An example is the case of Sevilla, whose high potential overshadows

its neighbours.

In order to analyse this distribution in more detail, Figure 6.12 includes a

box-plot4 graph showing the distribution of potential accessibility values in each

NUTS-2 region. It can be observed how there are significant differences between

Autonomous Communities: larger values concentrate in the Madrid Region, followed

by Cataluña and Comunidad Valenciana. Extremadura and Galicia are the regions

showing lower accessibility values. Furthermore, there are significant differences in

the distribution of accessibility values within regions: e.g. the box corresponding to

Cataluña shows that values in this region present a wide range of variation,

whereas in other regions, such as in Galicia or Extremadura, differences between

more and less accessible regions are significantly lower.

4 A box-plot (also known as a box-and-whisker diagram) is a convenient way of graphically showing the

five-number summary, which consists of the smallest non-outlier observation, lower quartile, median,

upper quartile, and largest non-outlier observation. Box-plots are able to visually show different types of

populations, without any assumptions of the statistical distribution. The spacings between the different

parts of the box help indicate variance, skew and identify outliers.


- 131 -

Figure 6.11: Potential accessibility. Alternative A0. Road mode

Figure 6.12: Box-plot of potential accessibility values in the do-nothing alternative. NUTS-2 aggregation. Road mode

1 2 3 6 7 8 9 10 11 12 13 14 15 16 17

100000

200000

300000

400000

500000

600000

700000

800000

900000

Potential accessibility

NUTS-2 Region

Code Name1 Andalucía2 Aragón3 P. de Asturias6 Cantabria7 Castilla y León8 Castilla-La Mancha9 Cataluña

10 C. Valenciana11 Extremadura12 Galicia13 C. de Madrid14 Región de Murcia15 C. Foral de Navarra16 País Vasco17 La Rioja

1 2 3 6 7 8 9 10 11 12 13 14 15 16 17

100000

200000

300000

400000

500000

600000

700000

800000

900000


NUTS-2 Region




- 132 -

As Figure 6.13 shows, this imbalance in the distribution of accessibility does

not seem to change significantly with the PEIT’s HCR extension.

Figure 6.13: Potential accessibility. Alternative APEIT. Road mode

In order to facilitate the comparison between alternatives, the percentage

change of the PEIT alternative compared to the do-nothing alternative is

represented in Figure 6.14. It can be observed how higher accessibility gains

concentrate in those corridors in which a new infrastructure connecting with large

urban and/or nearby destinations is planned, such as western Castilla y León,

Extremadura, the surroundings of Teruel province, or the central Pyrenees. Given

that these areas had accessibility deficiencies in the do-nothing alternative, a

positive cohesion effect can be expected. The validity of this first suggestion on the

cohesion sign is subsequently verified below, with the computation of inequality

indices.


- 133 -


Table 6.7 includes the resulting values of the four inequality indices of the

regional distribution of the potential accessibility indicators. All of the four indices

yield a positive regional cohesion effect.

Table 6.7: Regional inequality indices. Road accessibility

Inequality index A0 APEIT

Coefficient of variation 45.041 43.922Gini index 0.241 0.235Atkinson index 0.045 0.043Theil index 0.093 0.089

A comparison of the percentage change in the values of the four indices, in

terms of the do-nothing alternative, is represented in Figure 6.15. In relative terms

this change is slightly more significant if measured with the Atkinson (At) or the

Theil (Th) indices.


- 134 -

Figure 6.15: Relative change in road accessibility inequality indices

50556065707580859095

100

CoV Gi At Th

Inequality index

Index (Do-nothing=100)

Do-nothing PEIT

The regional cohesion performance indicator (RC) is then computed using

Equation (4.5.) as the mean value of the resulting relative change in the four

indices, resulting in a positive value: a 3.430% reduction in the inequality indices.

These values are summarized in Table 6.8.

Table 6.8: Regional cohesion performance indicator (RC). Road accessibility

Inequality index Changea

Coefficient of variation 2.484Gini index 2.490Atkinson index 4.444Theil index 4.301RC 3.430

a Measured in percentage change of the corresponding do-nothing value

This performance indicator value signals a slight positive regional cohesion

effect derived from the extension of the HCR network. In other words, the new

planned links included in the PEIT, which attempt at transforming the radial

network into a grid mesh, reduce the polarization of the territory, contributing to a

slight reduction in the accessibility disparities between the most and the least

accessible regions.

6.2.1.2 Rail mode

Figure 6.15 shows potential accessibility values in the do-nothing alternative, for

the rail mode. In the case of the potential indicator, the effect of the high speed is

accentuated and the accessibility of large cities and their hinterlands is highlighted.

The tunnel effect, a characteristic of HSR, is also made visible. This effect is


- 135 -

graphically made visible e.g. in the Madrid-Sevilla HSR corridor, where it can be

seen that high accessibility levels in the surroundings of stations alternate with

lower accessibility values in the spaces between them.

In an overall picture, the Madrid-Barcelona and Madrid-Sevilla corridors

appear as axes with a privileged accessibility situation. This corridor concentrates

some large cities, such as Madrid, Barcelona, Sevilla and Zaragoza, as well as a set

of medium size cities, well connected to those and therefore resulting in high

potential values, such as Ciudad Real or Córdoba. It can also be observed that a

high potential is achieved in the Mediterranean and the Madrid-Valencia corridors,

both enjoying acceptable commercial speeds in the do-nothing alternative.

The worst values are recorded in peripheral regions with inefficient rail

infrastructure networks, such as Galicia, Asturias, Cantabria and part of

Extremadura and Andalucía. In the latter region, the effect of the HSR Madrid-

Sevilla is clearly shown: Sevilla and Córdoba appear among the cities with higher

potential, whilst Almería has one of the worst positions.

The corresponding box-plot graphs showing the distribution of potential

accessibility values in each NUTS-2 division are included in Figure 6.17. Madrid is

the region with higher accessibility values, in this case showing a wide range of

variation between lower and higher accessibility values. It is followed by Cataluña,

Aragón, Castilla La Mancha and Comunidad Valenciana, which also show significant

variations between higher and lower accessibility values.


- 136 -

Figure 6.16: Potential accessibility. Alternative A0. Rail mode

Figure 6.17: Box-plot of potential accessibility values in the do-nothing alternative. NUTS-2 aggregation. Rail mode

1 2 3 6 7 8 9 10 11 12 13 14 15 16 17

50000

100000

150000

200000

250000

300000

350000

400000


NUTS-2 Region



1 2 3 6 7 8 9 10 11 12 13 14 15 16 17

50000

100000

150000

200000

250000

300000

350000

400000


NUTS-2 Region




- 137 -

This imbalance in the distribution of accessibility values seems to change

with the PEIT’s planned extension of the HSR network. As Figure 6.18 shows, the

area with higher accessibility values is extended from Madrid in the direction of the

axes of the planned HSR network. The tunnel effect is clearly shown, in some cases

overlapped with the agglomeration effect, which causes an extension of the areas

of high accessibility values in large agglomerations with a HSR station, such as the

case of Madrid. The situation of high accessibility locations almost coincides with

that of the planned HSR stations. On the other hand, worst values are concentrated

in peripheral areas such as Galicia, the Pyrenees or regions with low density of HSR

infrastructure, such as Extremadura, although their situation has dramatically been

improved with respect to that of the do-nothing alternative.

Figure 6.18: Potential accessibility. Alternative APEIT. Rail mode

Figure 6.19 includes the percentage change of the PEIT alternative

compared to the do-nothing alternative. The first observation relates to the

magnitude of the average percentage increase, which although it does not influence

the cohesion effect, it is significantly higher than that of the road mode. Higher

accessibility gains concentrate mainly in the northwest and southeast quadrants,

which are precisely those which had a relatively worse position in the do-nothing

alternative. Thus, the graphical analysis signals the existence of a positive regional


- 138 -

cohesion effect. The validity of this positive sign needs to be verified with the

computation of inequality indices.


Resulting values of the four inequality indices of the regional distribution of

the potential accessibility indicators are included in Table 6.9. Results show that all

of the four indices yield a positive effect in regional cohesion.

Table 6.9: Regional inequality indices. Rail accessibility

Inequality index A0 APEIT

Coefficient of variation 43.170 30.609Gini index 0.227 0.159Atkinson index 0.040 0.021Theil index 0.084 0.043


- 139 -

Figure 6.20: Regional cohesion indices. Rail mode

50556065707580859095

100

CoV Gi At Th

Inequality index

Index (Do-nothing=100)

Do-nothing PEIT

The values of the percentage change in the inequality indices and the

resulting regional cohesion performance indicator (RC= 38.841) are included in

Table 6.10.

Table 6.10: Regional cohesion performance indicator (RC). Rail accessibility

Inequality index Changea

Coefficient of variation 29.097Gini index 29.956Atkinson index 47.500Theil index 48.810RC 38.841

As the graphical analysis suggested, the planned extension of the HSR will

result in a significant positive effect in regional cohesion, which in terms of the

performance indicator is measured as a 38.841% average reduction in the selected

inequality indices. Indeed, there is a dramatic change in the differences between

potential accessibility of the most and the least accessible regions. In the do-

nothing alternative, stations located in the Sevilla-Madrid-Lleida corridor benefited

from dramatically higher accessibility levels than the rest of the territory, which

resulted in a highly polarized spatial distribution pattern. The ambitious HSR

extension changes this situation and makes it possible that all province capitals

have a HSR station, therefore reducing the comparative advantage of the Sevilla-

Madrid-Lleida corridor and resulting in a more balanced distribution of accessibility

among regions, i.e. a significant positive regional cohesion effect.


- 140 -

6.2.2 Social cohesion (SC)

As described in Section 4.4.2.2., the assessment of social cohesion is carried out by

analyzing the distribution of accessibility improvements among different socio-

economic groups. In the Spanish case study, because of the data5 available, the

unemployment rate of the base year (2005)6 was selected as the socio-economic

variable to classify the population of NUTS-5 regions. Given the spatial distribution

of unemployment rates, a positive social cohesion effect will presumably take place

if above average accessibility benefits concentrate in those municipalities with

above average unemployment rates, whilst negative social cohesion effects are

expected in the opposite situation.

Figure 6.21: NUTS-5 unemployment rates

Source: Fundación La Caixa (2006)

5 At the NUTS-5 level, income or GDP related variables are not available in Spain. 6 The source of information is the Social Annual report published by Fundación La Caixa (2006), in which

unemployment rates are computed from INEM (Instituto de Empleo), and based on the unemployment

data of the Ministry of Employment and Social Affairs. This variable computes the percentage of

unemployed people registered with INEM of the corresponding total population of each municipality. This

unemployment rate related to total population is a good comparative indicator between municipalities,

although it is not related to the active population. A technical statistics explanation of the main

differences between unemployment rates using the EPA (Encuesta de Población Activa- Active Population

Survey) can be found in the methodological notes of the 2006 Social Report (Fundación La Caixa, 2006).


- 141 -

Unemployment data are only available for those municipalities over 1,000

inhabitants, which makes up approximately 96% of the total Spanish population.

These data have been standardized7 and their spatial distribution is represented in

Figure 6.21. It can be observed how above average unemployment rates (in red in

the Figure) concentrate in Western Andalucía, Extremadura, coastal and cross-

border territories in Galicia, along with inner areas of Southern Castilla-La Mancha,

whilst below average rates (represented in blue) concentrate in the north and

northeast of the Peninsula, the Mediterranean and Cantabric coastal zones.

Before the computation of the social cohesion performance indicator is

carried out, and in order to give some insight into the assessment of social

cohesion effects, a preliminary analysis is included below. The analysis is split into

road (section 6.2.2.1) and rail (section 6.2.2.2) modes.

6.2.2.1 Road mode

Standardized8 absolute road accessibility improvements in each municipality with

unemployment data availability, expressed as the difference between the PEIT and

the do-nothing alternative, is mapped in Figure 6.22. The comparison of this map

represented values with the corresponding values of unemployment rates of Figure

6.21 may allow detecting possible biases of accessibility benefits towards regions

with high or low unemployment rates.

The same graphical analysis has been carried out for standardized relative

accessibility improvements9, as the results may be different than the corresponding

to absolute improvements (as suggested by Bröcker et al., 2004; Schürmann et al.,

1997). Relative improvements are mapped in Figure 6.23. The overall spatial

pattern is similar of that of absolute improvements; i.e. no significant biases of

accessibility improvements towards lagging regions can be detected in the graphical

analysis.

On the one hand, Figure 6.22 shows that above average accessibility gains

concentrate in regions with high unemployment rates, such as Western

Extremadura, and Southern Castilla-La Mancha. However, the opposite situation

appears due to the high accessibility improvements of the north coast, northern

Aragón and Northwestern Cataluña, which enjoy below average unemployment

levels.

7 Mean unemployment rate=100. 8 Mean absolute improvement =100. 9 Expressed as the percentage change of the do-nothing alternative.


- 142 -

Figure 6.22: Standardized absolute change of NUTS-5 regions in the potential accessibility indicator. Road mode

Figure 6.23: Standardized relative change of NUTS-5 regions in the potential accessibility indicator. Road mode


- 143 -

The social cohesion performance indicator formulated in Equation (4.9.) is

now computed after the insight obtained from the graphical analysis.

The first step consists in the calculation of each municipality’s weighting

factor Φ. Using the values of the potential road accessibility indicator and the

unemployment rate levels previously recorded in the GIS, each municipality is first

classified in a substandard of accessibility deficiency (see Table 4.5.) and structural

backwardness category (see Table 4.4.). The resulting classifications are mapped in

Figure 6.24, for substandard of accessibility deficiency and Figure 6.25, for

structural backwardness category.

Based on the above categories, each municipality is assigned a

corresponding weighting factor, following the specifications in Table 4.3. The

resulting values are represented in Figure 6.26.

Figure 6.24: Accessibility categories. Road mode


- 144 -

Figure 6.25: Structural backwardness categories

Figure 6.26: Regional weighting factor. Road mode


- 145 -

Higher weighting factors concentrate in those areas combining road

accessibility deficiencies and high unemployment rates: these mainly concentrate in

Galicia, Extremadura, western Andalucía and southern Castilla-La Mancha. In

contrast, lower weighting factors appear in Madrid and the Mediterranean coast.

The next step consists in calculating each municipality change in its potential

accessibility indicator, formulated in Equation (4.4.). These calculations have

already been described in Section 6.2.1, and the corresponding accessibility

changes are mapped in Figure 6.14.

Finally, the social cohesion performance indicator is computed using

Equation (4.9.) as the weighted increase in the population potential accessibility

indicator, expressed in percentage terms of the non-weighted accessibility of the

do-nothing alternative. The corresponding weighted increase is 8,288.103 inh./min,

whereas the non-weighted potential accessibility value accounts for 396,507.93

inh/min. This results in a final value of the performance indicator of SC=2.091.

The interpretation of this indicator is as follows: the higher its value, the

more concentrated the benefits in lagging and/or inaccessible regions, i.e. the more

the corresponding alternative contributes to the social cohesion objective. In this

case, the result obtained means that weighted potential accessibility increase

represents a 2.091% of the population potential of the do-nothing alternative.

6.2.2.2 Rail mode

Figure 6.27 represents standardized absolute accessibility improvements in each

municipality for the rail mode. Higher accessibility benefits concentrate in

Cantabria, Asturias and South western Andalucía, whilst Extremadura and Cataluña

concentrate most below average improvements. As happened with the road mode

analysis, the combined analysis of these results with the unemployment rates

represented in Figure 6.21 does not allow drawing any conclusion on the cohesion

effect produced.


- 146 -

Figure 6.27: Standardized absolute change of NUTS-5 regions in the potential accessibility indicator. Rail mode

Moving now to relative accessibility improvements, represented in Figure

6.28, the distribution of accessibility changes is more polarized than that of

absolute changes, i.e. the density of both dark red and blue colors is higher.

Regions with above average accessibility gains concentrate in the north west

Iberian quadrant (Galicia, Cantabria and Asturias) and south and eastern Andalucía.

In most cases, these regions improve their accessibility over 200% of the average

relative accessibility improvement. This is partly because they started from a

deficient situation, so that the same absolute increase represents for those regions

a higher relative accessibility benefit than for those regions with high initial



- 147 -

Figure 6.28: Standardized relative change of NUTS-5 regions in the potential accessibility indicator. Rail mode

Following the same procedure as for the road mode, the first step for the

calculation of the performance indicator is the classification of each municipality in

its corresponding accessibility deficiencies category. This classification is

represented in Figure 6.29.

Each municipality weighting factor can be calculated by combining the values

in Figure 6.29 with those mapped in Figure 6.25, using the specifications in Table

4.2. The results are mapped in Figure 6.30.


- 148 -

Figure 6.29: Accessibility deficiency categories. Rail mode

Figure 6.30: Regional weighting factor. Rail mode


- 149 -

The conclusion of the analysis of Figure 6.30 is that a higher weight will be

assigned to accessibility improvements in municipalities located mainly in Galicia,

the Cantabric coast, and most of the southern half of the Iberian Peninsula.

The interpretation of this indicator is as follows: the higher its value, the more

concentrated the benefits in lagging and/or inaccessible regions, i.e. the more the

corresponding alternative contributes to the social cohesion objective.

The weighted average accessibility improvement that results is 90,712.980

inh./min, whereas the non-weighted accessibility value accounts for 186,832.831

inh./min. This results in a final value of the performance indicator of SC= 48.553.

As mentioned already in the previous section, the higher the value of this

performance indicator, the better is the performance of the social cohesion

criterion. In this case, the result obtained means that weighted potential

accessibility increase represents 48.553% of the population potential of the do-

nothing alternative.

6.3 Environmental sustainability

6.3.1 Global warming (GW)

6.3.1.1 Travel demand forecasts

As already explained in the Introduction to Chapter 5, a national transport model is

not available in Spain to date, although its development is currently on the Spanish

research agenda (ETT and EPYPSA, 2006). This fact made it difficult to calculate the

total CO2 emissions in each alternative. A simplification has been made in order to

obtain an approximate value of these emissions. A brief literature review on this

topic is therefore necessary in order to justify the selected approach.

It is well reported that induced travel is an important component of travel

demand (see e.g. Goodwin, 1996; Cervero and Hansen, 2002; Lee, 2002; Litman,

2004; Guirao, 2000). With improved transportation conditions, short run effects

(e.g., route switches, mode switches, changes of destination, and new trip

generation) and long term effects (e.g., change in household car ownership, and

spatial reallocation of activities) will be observed.

Many studies have estimated travel time elasticities, mostly related to

highway expansions, but one of the difficulties in interpreting these results is the

uncertainty of the time frame that is applicable to the data (Lee, 2002). Goodwin(

1996), Noland and Lem (2002) and Cervero and Hansen (2002) provide reviews of

many empirical studies on induced demand due to road capacity expansions. For


- 150 -

example, Goodwin (1996) found that proportional savings in travel time were

matched by proportional increases in traffic on almost a one-to-one basis. Other

works suggest an average value for the elasticity of travel volume with respect to

travel time of about -0.5 to -1.0 in the short term and up to -2.0 in the long term

(Lee, 2002).

Rail related studies are less frequent. They mostly agree in that demand for

rail services is much more sensitive to changes in cost and travel time than the

demand for automobile or airline travel. Morrison and Winston (1985) found that

rail demand is elastic with respect to time, estimating it as -1.67 for business trips

and -1.58 in vacation trips. Bel (1997) carried out a study with Spanish data and

estimated rail travel time elasticities of -2.66 (for daytime traffic trains below 400

km) and -2.37 for trips over 400 km. Other works of intercity HSR projects planned

in Japan, computing short term induced travel elasticities are presented by Yao and

Morikawa (2005). In summary, for the rail mode ‘across all relations, account being

taken of the weight of each relation, an approximate travel-time elasticity of -2.2

emerges’ (Savelberg and Vogelaar, 1987).

In order to take into account the uncertainty of travel demand prognosis,

instead of selecting a single value for travel time elasticities, a range between -0.5

and -2.0 will be used for the road mode, and between -1.7 and -2.7 for the rail

mode.

6.3.1.2 Calculation of travel time savings

The approach used to compute travel time savings is based on the calculation of

accessibility indicators. The selected formulation is that of the location accessibility

indicator, a ‘travel cost indicator’ (see Section 3.2.2.2), which computes average

travel time to the set of destinations. This indicator was previously used in similar

studies at the Spanish national level (see e.g. Monzón et al., 2005). The

formulation chosen is included in Equation ( 6.1 ). The location indicator (Li) as

computed as the average travel time (in minutes) to the set of destinations, using

the population of each destination as the weighting variable.

∑∑

⋅=

j

j

j

jij

iP

PIL ( 6.1 )

The set of destinations includes those of Portugal and the three southern

regions of France. The location indicator is therefore used as a proxy for the

evaluation of travel time savings, when its results in the PEIT alternative are

compared to those of the do-nothing alternative.


- 151 -

Hence, a single aggregated value of the location indicator for all Spain has

been computed and compared to that of the do-nothing alternative. The result, in

percentage changes, has been translated into the corresponding increases in travel

demand with the use of the range of travel time elasticities selected in Section

6.3.1.1. These values are summarized in Table 6.11.

Table 6.11 Travel time savings and estimated induced traffic

Location indicator (min) % induced traffic Transport

mode Do-nothing

alternative

PEIT

alternative

% reduction Minimum Maximum

Road 156.809 153.178 2.31 1.12 4.62

Rail 325.813 213.859 34.36 58.41 92.72

6.3.1.3 Computation of the global warming performance indicator

The next step for the calculation of the performance indicator consists in

transforming the estimated increase in travel demand into the corresponding

increase in GHG emissions. This estimation has been carried out with version 2.44

of the TREMOVE model (Transport & Mobility Leuven and K.U.Leuven, 2006).

TREMOVE is a policy assessment model designed to study the effects of different

transport and environment policies on the emissions of the transport sector10.

Model runs were carried out with the data on induced traffic included in

Table 6.11, resulting in the corresponding CO2 emissions. Results obtained for the

road and rail modes are summarized in Table 6.12. The percentage change results

obtained by the road and rail alternatives are obviously not directly comparable, as

these values are heavily influenced by the emission levels of the do-nothing

alternative. A ‘global’ relative percentage change has therefore been computed,

representing the percentage change compared to the sum of road and rail

emissions.

10 The model estimates the transport demand, the modal split, the vehicle fleets, the emissions of air

pollutants and the welfare level under different policy scenarios. All relevant transport modes are

modeled, including air transport. Maritime transport is treated in a separate model. TREMOVE models

both passenger and freight transport, and covers the period 1995-2020. TREMOVE consists of 21 parallel

country models. Each country model consists of three inter-linked ‘core’ modules: a transport demand

module, a vehicle turnover module and an emission and fuel consumption module, to which a welfare

cost module and a well-to-tank emissions module is added. This model was developed by Transport &

Mobility Leuven and the K.U.Leuven in a service contract for the European Commission, DG

Environment. More information can be found at http://www.tremove.org.


- 152 -

Table 6.12: Forecasted induced traffic and corresponding increases in GHG emissions. Do-nothing vs. PEIT alternative. Road and rail modes

Road Rail332,359.16 275.43

Minimum 336,081.58 436.28Maximum 347,714.15 530.75Mean 340,036.65 370.17

72,513,765.92 234,275.13Minimum 73,279,001.33 365,755.62Maximum 75,670,365.08 442,987.18Mean 74,474,683.47 404,371.40

Minimum 765,235.41 131,480.49Maximum 3,156,599.16 208,712.05Mean 1,960,917.55 170,096.27

Minimum 1.06 56.12Maximum 4.35 89.09Mean 2.70 72.61

Minimum 1.05 0.18Maximum 4.34 0.29Mean 2.70 0.23

Traffic (million vkm)

Do-nothing alternative

PEIT alternative

Global relative*

(%)

Increase in GHG

emissions

GHG emissions (t CO2)

Do-nothing alternative

PEIT alternative

Absolute (t CO2)

Relative+

(%)

(+) Percentage change of each mode emissions of the do-nothing alternative

(*) Percentage change of total road and rail emissions of the do-nothing alternative

Source: TREMOVE 2.44 model (Transport & Mobility Leuven and K.U.Leuven, 2006)

On the one hand, the road mode do-nothing alternative accounts for over

72.5 million tons of CO2. It can be seen how the mean increase in GHG emissions

due to the extension of the HCR network included in the PEIT accounts for near 2

million tons of CO2, which represents a 2.70% increase compared to both the do-

nothing alternative value for the road mode, and total road and rail emissions of

the do-nothing alternative. This comparison could be carried out if different road

and/or rail alternatives were assessed.

On the other hand, the rail mode do-nothing alternative accounts for only

234,000 tons. The extension of the HSR network included in the PEIT accounts for

nearly 170,000 tons of CO2, which represents over a 72% increase, in terms of the

do-nothing alternative value for the rail mode, whereas it represents only a 0.23%

increase of total road and rail emissions of the do-nothing alternative.

The comparison of the absolute increases in GHG emissions between road

and rail modes (2 million vs. 170,000 tons of CO2) gives us an idea of the

significant difference in the contribution of the above transport modes to GHG

emissions, which is obviously proportional to their corresponding traffic volumes.


- 153 -

6.3.2 Habitat fragmentation (HF)

The assessment of habitat fragmentation is carried out in a vectorial GIS support.

Two inputs are necessary for the calculation of the PARA index (see Equation 4.10):

� The spatial delimitation of the patches, directly derived form the Spanish Map of

Habitats (see Figure 5.16).

� The modelling of the transport infrastructure networks, which has already been

implemented in the GIS for the accessibility calculations.

The PARA index was computed for the do-nothing alternative and the PEIT, road

and rail, alternatives. The PARA index is computed for each patch (t) as a

percentage change compared to that of the do-nothing alternative:

0

0

)(

)()(

t

tPEITt

tPARA

PARAPARAHB

−= ( 6.2 )

These values were subsequently aggregated for each habitat type, using the

area of each patch as the weighting variable, as explained in Section 4.4.3.2.

Furthermore, in order to graphically represent the PARA values, percentage change

in PARA values in SCIs and SPAs were mapped.

The resulting PARA values for the road and rail PEIT alternatives are

included in the following subsections, where the corresponding habitat

fragmentation performance indicators are also computed.

6.3.2.1 Road mode

Percentage changes in the PARA index in SCIs and SPAs, due to the implementation

of the road PEIT alternative are represented in Figure 6.31 and Figure 6.32 . It can

be observed how, in both cases, the new infrastructure networks scarcely cross the

protected areas. Indeed, the areas of SCIs and SPAs containing at least one

fragmented patch represent only a 0.1114% and a 0.1297% of their corresponding

areas, respectively.


- 154 -

Figure 6.31: % change in the PARA index in SCIs. Road mode

Figure 6.32: % change in the PARA index in SPAs. Road mode


- 155 -

The changes in the PARA index in each habitat type were aggregated using the area

of each habitat as the weighting variable. The resulting value of the habitat

fragmentation performance indicator, following Equation (4.13), is HF=0.259

The interpretation of this indicator is as follows: the higher the value, the

more the impact of the corresponding alternative on habitat fragmentation. In this

case, this result can be interpreted as a 0.2586% mean increase in habitat

fragmentation, a relatively low value, predictable due to the already mentioned fact

that the HCR extension planned in the PEIT scarcely crosses high environmental

quality areas.

6.3.2.2 Rail mode

Figure 6.33 and Figure 6.34 map percentage changes in the PARA index in SCIs and

SPAs, due to the implementation of the rail PEIT alternative, with respect to PARA

indices in the do-nothing alternative.

As can be seen in both Figures, planned links cross only a small proportion

of protected sites. Indeed, the areas of SCIs and SPAs cointaining at least one

fragmented patch represent only a 0.2630% and a 0.2301% of their corresponding

areas, respectively.

Figure 6.33: % change in the PARA index in SCIs. Rail mode


- 156 -

Figure 6.34: % change in the PARA index in SPAs. Rail mode

The same procedure used for the road mode (see section 6.3.2.1) is used here,

i.e. the changes in the PARA index in each habitat type were also aggregated using

the area of each habitat as the weighting variable. The resulting value of the

habitat fragmentation performance indicator is HF=0.462.

The interpretation of this value is that the new links of the HSR network

included in the PEIT contribute in a 0.4623% to habitat fragmentation. Although

this value is higher than the one corresponding to the road mode (0.2586), it is still

a relatively low value. Again, this is due to the fact that only a small proportion of

the new links cross protected areas.

6.4 Discussion on performance indicator results

6.4.1 Road mode

Table 6.13 summarizes the results obtained by the PEIT alternative in each of the

six performance indicators. Subcriteria shaded with green colour means that the

higher the value of the indicator, the better the alternative performs in the

corresponding criterion, whilst the opposite holds for red-coloured subcriteria. The

positive sign means that the alternative contributes to an improvement of the


- 157 -

corresponding subcriterion, whereas the opposite holds for the negative sign. Their

interpretation is as follows:

� In network efficiency terms, the PEIT road alternative results in a 2.6%

improvement of Spanish network efficiency and a 1.8% improvement of

network efficiency in cross-border regions.

� In cohesion terms, a 3.4% improvement of regional inequality indices and a

2.1% improvement in potential accessibility of inaccessible and/or structurally

lagging regions.

� In environmental terms, a 2.7% increase in GHG emissions and a 0.26%

increase in habitat fragmentation.

Table 6.13 Summary of performance indicator values. Road mode

Criteria Subcriteria Value Interpretation

Network efficiency 2.637 + Efficiency

Cross-border integration 1.771 +

Regional cohesion 3.430 + Cohesion

Social cohesion 2.090 +

Global warming 2.705 - Environment

Habitat fragmentation 0.259 -

In terms of efficiency, the network efficiency accessibility indicator has

confirmed its validity as a planning tool capable of measuring the contribution of

the PEIT planned network extension to the efficiency of the network as a whole.

Furthermore, it has made it possible to determine the improvement of efficiency

allocated in neighbouring regions, whose value has resulted in relatively high level

(1.8% increase) if compared with the one obtained in the Spanish territory (2.6%

increase).

The starting situation of the network and the specific characteristics of road

infrastructure have demonstrated to be the main factors determining the final

efficiency value. In the Spanish case, the high investment efforts during recent

decades have meant that the base HCR network of the do-nothing alternative had a

relatively good quality. This fact reduces the ‘marginal’ efficiency increase of further

network extensions. It is precisely in already developed networks, like the Spanish

one, where the efficiency criterion needs to be complemented with that of a more

balanced distribution of accessibility. In other words, if most of the links connecting

major agglomerations have been built, the risk of polarisation of the territory

increases if the efficiency criterion continues to dominate over the cohesion one.


- 158 -

The methodology enables the explicit consideration of this ‘efficiency vs

cohesion’ trade-off and therefore provides the DM with information on the

consequences of the planned infrastructure extension on both criteria. As the

cohesion results show, the road PEIT alternative sorts out this polarisation risk with

the definition of a grid mesh network, which in consequence results in a more

balanced distribution of accessibility.

In summary, the road PEIT alternative, if compared with the do-nothing

alternative, results in an improvement of efficiency and cohesion criteria, whereas it

brings about a negative contribution to environmental criteria.

6.4.2 Rail mode

The results of the performance indicators of the PEIT rail alternative are included in

Table 6.14. As with the road mode, a summary of their values is included below:

� In network efficiency terms, the PEIT rail alternative results in a 34.5%

improvement of Spanish network efficiency and a 20.2% improvement of

network efficiency in cross-border regions.

� In cohesion terms, a 38.8% improvement of regional inequality indices and a

48.5% improvement in potential accessibility of structurally lacking regions.

� In environmental terms, a 0.23% increase in GHG emissions and a 0.46%

increase in habitat fragmentation.

Table 6.14 Summary of performance indicator values. Rail mode

Criteria Subcriteria Value Interpretation

Network efficiency 34.533 + Efficiency

Cross-border integration 20.179 +

Regional cohesion 38.841 + Cohesion

Social cohesion 48.553 +

Global warming 0.230 - Environment

Habitat fragmentation 0.462 -

As happened with the road mode, the rail PEIT alternative results in an

improvement of efficiency and cohesion criteria and a negative impact on

environmental criteria.

However, contrary to the situation of the HCR network, the HSR network

was underdeveloped in the do-nothing alternative. Thus, the PEIT planned network

extension results in drastic improvements of accessibility (34.5%) if compared to

that of the road mode (2.6%). The same holds for the increases obtained in

neighbouring regions (20.2% vs. 1.8%). Furthermore, the differences in the speed


- 159 -

of conventional rail, which served the majority of the Spanish territory in the do-

nothing alternative, and HSR lines, which are dramatically increased (from nearly

1,100 km to nearly 7,200 km) to link major agglomerations and provide them with

HSR stations.

Another important result is that the PEIT HSR network extension contributes

to a significant improvement of regional cohesion, i.e. a more balanced distribution

of accessibility. This result is especially important given the ‘polarisation proneness’

of HSR links, derived from the commented ‘point accessibility’ feature of this

transport mode.

Furthermore, as happened with the road mode, special care needs to be

taken when interpreting these results, as it is when different rail network extension

alternatives are compared against the do-nothing alternative that the values of the

performance indicators gain more interpretability.

Thus, the same remarks made when interpreting the road results on the

potential of the methodology to provide more conclusions on the basis of the

resulting performance indicator values, are applicable for the rail mode.

6.5 Integration of results

6.5.1 Description of the simplified integration procedure

As discussed in section 5.1., the full development of the MCA procedure, described

in Section 4.5., is not possible in this case study application. However, a

simplification was made in order to obtain an approximation of the integrated value

representing the performance of the PEIT alternative against the do-nothing

alternative. It consisted in the application of a linear additive model (see e.g.

Nijkamp et al., 1990; Dogson et al., 2001 for a description), which has been used

in similar MCA studies (see e.g. Fiorello et al, 2006; Monzón et al., 2003). In the

linear additive model, following the nomenclature used in Equation (4.14.), the

partial utility of each alternative s in each criterion j (usj), frequently termed value

score, is multiplied by the weight of that criterion (wj) and all those weighted scores

are added together to obtain the overall utility value for the alternative s (Us).

This method defines linear value functions (u) translating the performance

indicator values (x) into a value score, on a 0-100 scale. The specification of each

of these value functions requires the definition of two reference points, i.e. the least

preferred (llj) and the most preferred (xmj) values of the corresponding performance

indicators, which are assigned a 0 and a 100 value score, respectively. This implies

that the general formulation of utility functions included in Equation (4.15) is

transformed into a linear one, as Equation (6.1) shows:


- 160 -

( )( )

ljsj

ljmj

sj xxxx

u −⋅−

=100

(6.1)

The definition of these reference points was estimated, only for the purpose of this

simplified integration procedure, from the performance indicator values obtained in

the PEIT and do-nothing alternatives. As mentioned before, a consistent

correspondence between performance indicator values and scores should be carried

out from the results of a series of alternatives, which are not available in this case

study application.

Finally, criteria weights were obtained from a survey conducted among

researchers and relevant stakeholders, as detailed in Section 4.5. The details of the

questionnaire distributed and the corresponding weights obtained are included in

Annex A.

6.5.2 Road mode

The first step for the integration of results consists in defining the reference points

of the value functions, i.e. the least and most preferred values of the performance

indicators, which are included in Table 6.15.

Table 6.15: Definition of value functions. Road mode

Criterion xlj xmj

Network efficiency 0.000 8.110Cross-border integration 0.000 7.130Regional cohesion -50.000 50.000Social cohesion 0.000 5.900Global warming 4.340 0.000Habitat fragmentation 0.693 0.000

The rationale behind the suggested thresholds of the least and most preferred

values is described below:

� Network efficiency: a 100 score is given to the highest percentage improvement

in network efficiency accessibility achieved in the PEIT alternative of all Spanish

provinces, i.e. a 8.11 performance indicator value, whereas a 0 score is given to

a 0 value. As an example, Figure 6.35 includes a graphical representation of the

road mode network efficiency value function, in which the procedure to

translated the PEIT performance indicator value (2.64) into the corresponding

score (32.52) has been represented with the dotted arrows.


- 161 -

Figure 6.35: Value function for the network efficiency criterion. Road mode

Network efficiency

0

20

40

60

80

100

0 2 4 6 8

xj

uj

xPEIT

uPEIT

Network efficiency

0

20

40

60

80

100

0 2 4 6 8

xj

uj

xPEIT

uPEIT

� Cross-border integration: the highest percentage improvement in network

efficiency accessibility achieved in the PEIT alternative of all Portuguese districts

and French departments, i.e. a 7.13 value is given a 100 score. As with the

previous indicator, a 0 score is given to a 0 value.

� Regional cohesion: the specification of this value function has some

peculiarities. On the one hand, road and rail mode results can be directly

compared (López et al., in press). On the other hand, the sign of the effect may

be positive (i.e. increased cohesion) or negative (i.e. reduced cohesion). Hence,

the range of variation of performance indicator values has been defined from

the results obtained in both modes and includes the possibility for negative

values to appear. The most preferred value has been assigned as 50 (i.e. a 50%

reduction in inequality indices) and the least preferred value a -50 (i.e. a 50%

increase in inequality indices), as these are the range of variation

� Social cohesion: the most preferred value corresponds to that obtained if the

average increase in potential accessibility (5,875 inh./min) accrued to regions

with the highest (i.e. 4) weighting factor. This means that a 100 score is

assigned to a 5.9 value of the performance indicator, whereas as 0 score is

assigned to a 0 indicator value.

� Global warming: the maximum value of the range of forecasted increases in

GHG emissions (see Table 6.12) has been used as the threshold to define the

alternative with a 0 value: it would correspond to an alternative in which GHG

emissions would increase in a 4.34%.

� Habitat fragmentation: as happened with the regional cohesion indicator, road

and rail results are comparable. Hence, the same value function is suggested for

both modes: the 0 score corresponds to the highest fragmentation index


- 162 -

obtained among both modes incremented in a 50%, which corresponds to a

0.693 performance indicator value, whereas a 100 score corresponds to a 0

value.

With the values included in Table 6.15, and applying Equation (6.1.), it is

possible to apply the linear additive model. As a result, Table 6.16 includes the

scores obtained after the application of the model to the do-nothing and PEIT

alternatives, for the road mode.

Table 6.16: Integration of results. A0 vs. APEIT. Road mode

A0 APEIT A0 APEIT A0 APEIT

Network efficiency 0.00 2.64 0.00 32.52 0.00 8.49Cross-border integration 0.00 1.77 0.00 24.84 0.00 3.23Efficiency score 0.00 11.72Regional cohesion 0.00 3.43 50.00 53.43 10.60 11.33Social cohesion 0.00 2.09 0.00 35.43 0.00 6.84Cohesion score 10.60 18.16Global warming 0.00 2.71 62.33 0.00 6.36 0.00Habitat fragmentation 0.00 0.26 37.37 0.00 3.81 0.00Environment score 10.17 0.00

Integrated score 20.77 29.88

CriterionPerformance

indicator valuesUnweighted scores Weighted scores

The scores included in the Table are consistent with the conclusions

extracted from the analysis of results carried out in Section 6.4., i.e. the PEIT

appears with better performance than the do-nothing alternative in efficiency and

cohesion criteria, whereas the opposite holds for environmental criteria. If the

integrated scores of both alternatives are compared, it can be observed that the

PEIT alternative score is a 44% higher than the do-nothing one.

6.5.3 Rail mode

For the rail mode, the same rationale used for the road mode was followed for the

definition of the least and most preferred values defining the value functions. Table

6.17 includes the suggested values.


- 163 -

Table 6.17: Definition of value functions. Rail mode

Criterion xlj xmj

Network efficiency 0.000 52.540Cross-border integration 0.000 48.075Regional cohesion -50.000 50.000Social cohesion 0.000 166.490Global warming 0.290 0.000Habitat fragmentation 0.693 0.000

The integrated scores of the do-nothing and PEIT alternatives were obtained

after transforming performance indicator values into unweighted scores and the

subsequent use of the base weight profile. They are included in Table 6.18.

Table 6.18: Integration of results. A0 vs. APEIT. Rail mode

A0 APEIT A0 APEIT A0 APEIT

Network efficiency 0.00 34.53 0.00 65.73 0.00 17.15Cross-border integration 0.00 20.18 0.00 41.97 0.00 5.46Efficiency score 0.00 22.61Regional cohesion 0.00 38.84 50.00 88.84 10.60 18.83Social cohesion 0.00 48.55 0.00 29.16 0.00 5.63Cohesion score 10.60 24.46Global warming 0.00 0.23 79.31 0.00 8.09 0.00Habitat fragmentation 0.00 0.46 66.67 0.00 6.80 0.00Environment score 14.89 0.00

Integrated score 25.49 47.07

Criterion Performance

indicator valuesUnweighted scores Weighted scores

As happened with the road mode results, the consistency between the

individual assessment results and the final integrated score is checked: efficiency

and cohesion criteria are improved with the implementation of the PEIT rail

alternative, whereas environmental criteria are deteriorated.


- 164 -

6.5.4 Sensitivity analysis

As detailed in section 4.6, the next step of the MCA procedure consists in analyzing

the sensitivity of the results to changes in the criteria weights and the attribute

(performance indicator) values.

Given the limitations of the case study application, only two simplified

examples of how this sensitivity analysis should be conducted were carried out.

They have been grouped, following the classification made in sections 4.6.1. and

4.6.2., into weight and attribute value sensitivity analysis. The analysis was only

carried out for the road mode, as its extension to cover the rail mode does not add

any significant insight to the case study conclusions.

6.5.4.1 Weight sensitivity

As described in section 6.5.2, if the base weight profile is used, the ratio between

the scores of the PEIT and the do-nothing alternatives results in a value of 1.44. In

this section, there is an analysis of how this ratio changes when the base weight

profile is systematically modified.

For this purpose, three tests were made: each one consists in modifying the

weights of each of the three broad criteria categories: efficiency, cohesion and

environment, while leaving the proportion of the weights between the other two

invariable. The results are plotted in Figure 6.36, Figure 6.37 and Figure 6.38,

respectively.

Figure 6.36: Criterion weight sensitivity: efficiency criterion

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Efficiency criterion weight

Ratio score APEIT/A0

Weight Ratio0.0 0.870.1 0.970.2 1.080.3 1.230.4 1.440.5 1.720.6 2.140.7 2.840.8 4.240.9 8.441.0 1.78

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Efficiency criterion weight


Weight Ratio0.0 0.870.1 0.970.2 1.080.3 1.230.4 1.440.5 1.720.6 2.140.7 2.840.8 4.240.9 8.441.0 1.78


- 165 -

Figure 6.37: Criterion weight sensitivity: cohesion criterion

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

2.00

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Cohesion criterion weight


Weight Ratio0.0 1.150.1 1.240.2 1.320.3 1.390.4 1.440.5 1.520.6 1.580.7 1.640.8 1.690.9 1.73

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

2.00

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Cohesion criterion weight


Weight Ratio0.0 1.150.1 1.240.2 1.320.3 1.390.4 1.440.5 1.520.6 1.580.7 1.640.8 1.690.9 1.73

Figure 6.38: Criterion weight sensitivity: environmental criterion

0.00

0.50

1.00

1.50

2.00

2.50

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Environmental criterion weight

Ratio scoreAPEIT/A0 Weight Ratio

0.1 2.030.2 1.460.3 1.080.4 0.800.5 0.590.6 0.420.7 0.280.8 0.170.9 0.08

0.00

0.50

1.00

1.50

2.00

2.50

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Environmental criterion weight

Ratio scoreAPEIT/A0 Weight Ratio

0.1 2.030.2 1.460.3 1.080.4 0.800.5 0.590.6 0.420.7 0.280.8 0.170.9 0.08

The situation corresponding to the base weight profile is represented with a

vertical dotted line and corresponds to the above mentioned 1.44 value of the ratio

between the PEIT and do-nothing alternatives. The situation in which the ranking of

alternatives is reversed is represented with a horizontal dotted line, providing in the

three cases the value of the corresponding weight threshold.

On the one hand, the comparison of the three Figures shows the different

influence of the three criteria categories on the final scores. Hence, while changes

in the efficiency weight appear as those with a higher influence in the results, the

weight attached to the cohesion criteria has a significantly lower influence in the


- 166 -

variation of the scores’ ratio. The influence of the environmental weight holds an

intermediate position between the other two.

On the other hand, the analysis of each individual Figure makes it possible

to detect the value of the weight that would result in a reversal of the ranking of

alternatives, i.e. in that the do-nothing alternative would obtain a higher score than

the PEIT one. This is the case for both a decrease of the importance given to the

efficiency criterion (see Figure 6.36), if its weight is reduced below a 0.1 value, and

an increase in the importance of the environmental criterion (see Figure 6.38), in

case its weight exceeds a 0.3 value. In contrast, this rank reversal is not possible

for any variation in the cohesion criterion’ weight, as Figure 6.37 shows.

6.5.4.2 Attribute value sensitivity

In this subsection there is a description of an example of how the sensitivity of the

results to errors in the estimation of performance indicator values should be carried

out. This exercise constitutes an example of the attribute value sensitivity analysis

proposed in section 4.6.2.

For this purpose, the values of the road PEIT performance indicators were

systematically ‘worsened’ by 10, 20, 30, 40 and 50% and the corresponding

changes in the ratio between the scores of the PEIT and do-nothing alternatives

were computed. They are plotted in Figure 6.39, in which the dotted line represents

the original value of the ratio between the scores of the PEIT and do-nothing

alternatives (a 1.44 value).

It can be concluded from the analysis of Figure 6.39 that errors in the

estimation of the performance indicator values corresponding to the network

efficiency and regional cohesion criteria are the ones that most and least influence

the final scores, respectively. This is due to the combination of their corresponding

weights and the formulation of their value functions. In any case, the influence of

individual errors in the resulting ratio is slight, as even in the case of significant

errors in sensitive criterions (e.g. a 50% reduction in network efficiency

improvement of the PEIT alternative), the score ratio does not fall below a 1.20

value.


- 167 -

Figure 6.39: Attribute value sensitivity. Road mode

1.20

1.25

1.30

1.35

1.40

1.45

1.50

Networkefficiency


Regionalcohesion

Social cohesion

Global warming


% change in performance indicator values

Ratio score A

PEIT

/A0

0% 10% 20% 30% 40% 50%

This exercise constitutes a basic example of how the attribute value

sensitivity analysis should be conducted. This analysis could be extended if different

hypothesis of simultaneous errors in the six criteria were analyzed and a set of

alternatives were assessed.


- 168 -

Chapter 7– CONCLUSIONS, CONTRIBUTIONS AND FUTURE RESEARCH

- 169 -

7. CONCLUSIONS, CONTRIBUTIONS AND FUTURE

RESEARCH

7.1 Conclusions

The general objective of this research work was to develop a methodology capable

of addressing strategic effects of transport infrastructure Plans which are not

usually covered by traditional assessment methodologies. The proposed procedure

has been validated with its application for the assessment of the Strategic

Transport and Infrastructure Plan 2005-2020 (PEIT) (Ministerio de Fomento, 2005).

The conclusions from the research work have been grouped into four

categories. The first group relates to the conclusions drawn from the literature

review carried out in Chapters 2 and 3. The second group includes conclusions from

the development of the proposed methodology. The third group summarizes the

main findings resulting from the application of the methodology to the case study.

Finally, the last group includes some general conclusions from a global spatial

planning perspective.

7.1.1 Literature review

The following are the most relevant conclusions drawn from the literature review:

� The planning framework in which transport infrastructure Plans are assessed is

in constant evolution. This has created a lack of harmonized assessment

methodologies and the subsequent important research efforts from both

research institutions and governments in order to fill in this research gap.

� The large number of stakeholders and government structures involved, the

increasing importance of public opinion issues, and the observed greater social

awareness on the impacts of transport infrastructure Plans has meant a growing

importance of ‘communicative’ and ‘consensus building’ issues when developing

assessment methodologies. In this context, data processing and graphical

presentation capabilities of spatial impact analysis tools, such as GIS, make

them especially useful as planning supporting tools, as they are capable of

facilitating both the interaction between planners and DMs and the presentation


- 170 -

of assessment results to the public opinion. However, an enhanced integration

of the work of spatial analysts, transportation planners and GIS capabilities is

still a current challenge for the research community.

� DMs require flexible, transparent and ‘easy-to-explain’ methodologies, as the

results of the assessment are increasingly seen as a starting point for

negotiations and deliberation between planners and DMs, rather than the end of

the planning process. This is reinforced by the high relevance that the political

assessment has at the Plan level when compared to the technical assessment.

� The application of the sustainable development concept is basic for the

assessment of strategic effects of transport infrastructure Plans. However, the

translation of the three sustainability dimensions into a harmonized set of

assessment criteria, and the definition of the procedures to evaluate their

performance is still on the research agenda.

� In particular, additional research work is necessary to address wider policy

impacts such as network effects or distributive impacts, usually not covered by

traditional appraisal methodologies. Most of these strategic impacts have a

spatial component and therefore modern spatial impact tools are especially

suited for these tasks.

� There is an unexploited potential of accessibility indicators to be used to assess

the above wider policy impacts. Although important research efforts have been

made towards the development of new formulations of accessibility indicators,

further research is needed to develop formulations combining a theoretically

sound foundation and a relative easiness of interpretation for DMs and the

public opinion.

7.1.2 Methodological approach

� The sustainable development concept has constituted a useful guide in order

to structure the strategic impacts that should be evaluated in the

assessment of transport infrastructure Plans. After the literature review, it

was concluded that these criteria should include efficiency, cohesion and

environmental aspects.

� Accessibility indicators are useful tools in order to measure the performance

of each alternative in most of the criteria defined. In particular, the different

formulations of accessibility indicators selected have enabled the

measurement of impacts on efficiency and cohesion. They have also proven

their utility as a proxy to measure travel time savings, in the absence of a

calibrated transport demand model.


- 171 -

� Efficiency aspects have been assessed with the calculation of the network

efficiency accessibility indicator, which has proved its great potential in

spatial planning tasks. The formulation of this indicator has shown to be

especially useful for the assessment of the quality of transport infrastructure

links connecting activity centres, eliminating the effect of the geographical

location.

� The issue of spillover effects, which is frequently missing in similar studies,

is dealt with in the methodology with the inclusion of a ‘cross-border

integration’ subcriterion in the efficiency criterion group. The methodology

proposes a procedure for widening the study area to cover cross-border

regions, so that spillover effects in neighbouring countries can be assessed

through the calculation of network efficiency accessibility improvements in

these regions.

� The methodology has suggested a procedure to assess regional cohesion

effects based on the calculation of changes in the spatial distribution of

potential accessibility levels among regions. This approach is useful in order

to analyze the risk of polarizing effects due to the extension of transport

infrastructure. This is the situation of the new MMSS of the recently enlarged

EU, where the transport network is currently under-developed and the main

priorities are efficiency rather than cohesion goals.

� On the other hand, the potential accessibility indicator has proven its

usefulness to assess social cohesion effects, when a higher importance to

potential accessibility improvements experienced in lagging and/or

inaccessible regions is assigned.

� A MCA-based procedure has been considered the most appropriate one for

the integration of the different aspects covered by the methodology. The

methodological proposal includes the suggested approach for the definition

of criteria weights, including a questionnaire which was distributed among

researchers and planners, utility functions and a final sensitivity analysis.

7.1.3 Case study application

These conclusions have been structured according to the six subcriteria groups of

the methodology.

� In terms of network efficiency, road mode results have shown modest network

accessibility improvements due to the completion of the PEIT, i.e. a 2.6% was

achieved. The reason for this low value is the relatively good starting situation

of the network in the do-nothing alternative. The analysis of results therefore


- 172 -

shows that, as the network becomes denser, reductions in marginal accessibility

improvements appear. The opposite holds for the rail mode: the situation of the

rail network in the do-nothing alternative in terms of network efficiency was

considerably worse than for the road mode and therefore the PEIT planned

extension of the HSR network results in a significant percentage improvement

(a 34.5%) of the network efficiency.

� Regarding cross-border integration issues, the analysis has shown that, both for

the road and rail modes, network efficiency improvements experienced in

neighbouring countries due to the PEIT should not be overlooked. Indeed, in

Portugal and southern France, a 1.8% and a 20.2% road and rail network

accessibility improvements were achieved. If these values are compared with

the 2.6% and 34.5% improvements of national road and rail network efficiency,

it can be concluded that spillover effects are significant. The main factors

determining the spatial distribution and the magnitude of accessibility

improvements are the distance to the Spanish border and the quality of the

network of the neighbouring country under consideration, both for road and rail

modes.

� In terms of regional cohesion, both road and rail infrastructure extensions

included in the PEIT result in positive regional cohesion effects; i.e. accessibility

differences between regions are reduced. This cohesion effect is significantly

higher for the rail mode (38.8 %) than for the road mode (3.4%), due to the

already mentioned fact that the HSR extension results in a significant change in

accessibility levels and provides all province capitals with a HSR station,

resulting in a more balanced distribution of accessibility. The extension of the

HCR network results in much lower accessibility improvements, in which the

regions’ relative position scarcely moves.

� The procedure for the calculation of social cohesion effects based on the

weighting of regional accessibility improvements according to structural and

accessibility backwardness categories has been validated. The social cohesion

performance indicator yields a result of a differential improvement in

accessibility potential of structurally lagging regions for road mode (41.1%) and

rail mode (16.6%), compared to national average accessibility improvements.

Additional considerations on the social cohesion effect could be made if different

alternatives were compared.

� The calculation of the global warming performance indicator has required for the

definition of an ‘ad-hoc’ procedure for the case study. The simplification made,

consisting in using the results of the location indicator and an estimation of


- 173 -

travel time elasticities in order to obtain the volume of induced traffic and the

corresponding increase in GHG emissions has proved its efficiency. Following

this simplification, for the road mode, the 2.3% average travel time savings are

translated into an increase of nearly 2 million tons of CO2, which represents a

2.7% increase in GHG emissions. For the rail mode, the induced traffic

generated due to the 34.4% reduction in travel times result in approximately

170,000 tons of CO2, which represents a 72.6% increase in GHG emissions if

expressed in terms of increase in rail-related emissions, and a 0.23% if

expressed in terms of percentage increase of global road and rail emissions. It

is obvious that direct comparisons of the relative increases in mode-specific

emissions between both transport modes should therefore be avoided.

� Habitat fragmentation issues have been dealt with through the calculation of a

fragmentation indicator. The calculations show that the PEIT planned extension

yields a result of a 0.26% and a 0.46% increase in habitat fragmentation in the

road and the rail modes, respectively. These relatively low values reflect the

fact that the majority of the PEIT new links do not cross protected areas, in

accordance with environmental policy recommendations.

7.1.4 Recommendations from a transport planning perspective

Finally, some general considerations which could be used as recommendations for

transport planning processes at the strategic level have arisen:

� The specific characteristics of the transport mode of the network extension

under consideration have proven to be key factors determining its

corresponding spatial impacts. The road network shapes the space in a

relatively continuous way (surface accessibility): the longer the distance, the

more travel time needed. But HSR is creating a space that is becoming more

and more discontinuous (point accessibility): more distance may result in less

travel time, depending on the location of HSR stations. In other words, whereas

new highways shape the territory in a relatively continuous way, due to the high

density of junctions, new HSR links introduce spatial discontinuities, as the

access to the infrastructure, and the corresponding accessibility benefits,

concentrates in the surroundings of HSR stations, which inevitably need to be

separated from one another.

� On the one hand, HCR links introduce ‘corridor effects’ in the territory; in the

Spanish case study, given the radial nature of the HCR network, this corridor

effect is translated into the fact that high efficiency areas appear concentrated

in radial axes and their junctions. As this indicator reduces the effect of the


- 174 -

geographical location, it is the quality of the transport network which

determines the accessibility level obtained, so that peripheral regions do not

necessarily show low accessibility values (Gutiérrez and Monzón, 1998).

� On the other hand, HSR links result in ‘tunnel effects’, i.e. accessibility is not

constant in a given HSR corridor, but it is significantly higher in the

surroundings of HSR stations, whilst interstitial areas between stations appear

with comparatively worse accessibility (Martín et al., 2004; Gutiérrez et al.,

1996). This effect makes HSR a transport mode with a potential ‘polarisation

risk’ (Bruinsma and Rietveld, 1993; Gutiérrez, 2001; Vickerman et al., 1999).

Consequently, for the rail mode, the absolute location of the nodes (whether

they are in the core or on the periphery) is less important, whilst the relative

location takes on a greater relevance (whether they are in the HSR network or

not). In this situation, there is the danger of an increase in tendencies towards

polarization of space (Plassard, 1991; Plassard, 1992) with negative cohesion

effects. Within this new situation, there is no doubt that a decisive role is to be

played by improvements in regional transport infrastructures that link HSR

stations to the rest of the region. Thus, those spaces that are situated outside

the HSR network, but efficiently linked to it, could also benefit from the HSR

(Gutiérrez et al., 1996; EC, 1999).

� The level and quality of transport infrastructure provision is another key factor

influencing the final results. In under-developed networks, infrastructure

investments are mainly ‘efficiency-oriented’. Once a basic network is completed,

and as the network becomes denser, it is possible to move from efficiency

towards cohesion goals. Investments to extend the network are then justified in

order to achieve a more balanced spatial distribution of accessibility. This is the

case of the majority of the HCR extension included in the PEIT, which is mainly

aimed at the transformation of the existing radial network into a grid mesh, with

the construction of new cross and longitudinal links, and the completion of the

existing axis to connect peripheral regions to the HCR network.

� The results of the cross-border integration criterion confirms the importance of

a joint planning process between neighbouring countries (Ollivier-Trigalo,

2001); where co-operation and establishment of priorities are fundamental

issues in order to take advantage of the full potential of new links to improve

network efficiency. The possibility to integrate the assessed benefits into official

assessment methodologies of national Plans, in order to translate these benefits

into monetarized effects, may provide the justification for the financial support

from neighbouring countries and/or the EU.


- 175 -

� The trade-off between the three main sustainability criteria needs to be taken

into account explicitly in assessment methodologies. In the first stages of a

network development, transport policy is more ‘efficiency-oriented’, but as

infrastructure provision increases, cohesion and environmental issues come into

the fore of the transport planning process. It is in the latter cases where

strategic assessment methodologies have an important role in order design

‘optimal’ alternatives that maximize the transport system’s efficiency while

minimizing spatial polarization processes and the deterioration of environmental

conditions.

7.2 Contributions

The research work carried out includes a number of contributions to the strategic

transport planning research field. These are included below:

� Development and validation of a methodology capable of assessing strategic

effects of transport infrastructure Plans, in terms of efficiency, cohesion and

environmental effects, based on the application of spatial impact tools and a

MCA approach.

� Definition of a set of performance indicators in order to measure the

contribution of each alternative to each of the criteria. The formulation of these

performance indicators constitutes the core added value of the methodological

approach. Their definition deals with the existing trade-off between their

theoretical soundness and their easiness for interpretation.

� Definition of a useful, transparent and flexible methodology capable of

integrating the strategic aspects present in the assessment of transport

infrastructure Plans. The application of the methodology provides the technical

basis that DMs are currently demanding, providing them with transport planning

recommendations and a global view of the consequences of the Plan.

� The methodology provides an enhanced vision of strategic impacts and effects

stemming from the implementation of a transport infrastructure Plan, which can

complement the economic-oriented approach, providing a framework in which

territorial and environmental aspects are integrated.

� The use of GIS results in an enhanced graphical presentation of results,

improving their interpretability and in this way facilitating the communication

between transport planners, DMs, stakeholders and public opinion and their

participation in the planning process.


- 176 -

7.3 Recommendations for future research

The research work has identified a number of issues which could be the object of

future research. These are summarized below:

� Establishment of a feedback channel between the outputs of the transport

demand model and the formulation of the accessibility indicators, in the

framework of a doubly-constrained spatial interaction model (Wilson, 1971). In

particular, the possibility to use the results of the distribution stage as the basis

to define each destination’ weight could be explored.

� A research field that is intimately linked with this issue is the use of accessibility

as a concept with a strict economic entity, related with the concepts of

consumer surplus and welfare. This way, the measurement of accessibility

would represent transport users’ benefits, as suggested by Martínez (1995).

� Development of a model capable of forecasting regional economic impacts of

transport infrastructure Plans, using the computed accessibility improvements

as input variables, as discussed in Section 3.2.3.3.

� Substitution of travel time with generalized cost of travel as the impedance term

of formulation of the accessibility indicators, as suggested by Bröcker et al.,

(2004). This would make it possible to assess the effects on accessibility of

other transport policy instruments, such as pricing.

� Analysis of the strengths and weaknesses of each inequality index and

exploration of the possibility to use a different mathematical procedure to derive

a synthetic inequality index.

� Further development of the frequency of service penalty estimation procedure in

the rail accessibility calculations. This would require the collection of additional

service data, in order to obtain a more accurate estimation of the time penalty.

� Definition of a multimodal accessibility indicator. This would allow defining a

single classification of accessibility deficiency substandards, and a combined

analysis of the effects of the extension of different infrastructure networks,

including air and sea modes. Possible options are the utilization of the modal

shares of a multimodal transport demand model, or the calibration of an

indicator which best correlates with regional GDP values, as suggested by

Schürmann et al. (1997).

� Inclusion of capacity constrains issues in the determination of link speeds, in

order to address the possibility of congestion effects in the surroundings of large

agglomerations. Again, a transport demand model is required for this

improvement of the model to take place.


- 177 -

� Integration of recent EU-wide transport demand models, such as TREMOVE

(Transport & Mobility Leuven and K.U.Leuven, 2006), with the corresponding

national models. This would enable a more realistic estimation of the

attractiveness of regions in neighbouring countries.

� Broadening the scope of the case study to assess a set of alternatives. This

would enable the full application of the MCA-based integration procedure

described in Section 4.5.

� Application of the methodology to a set of scenarios on the development of

external variables, such as population growth, composition of the vehicle fleet

and its emission rates. The influence of the development of these external

variables in the ranking of alternatives and therefore the treatment of

uncertainty issues would be improved with this scenario approach.

� Assessment of alternatives which include the network extension of different

transport modes. This way, the performance of the transport system as a whole

could be assessed. The development of a multimodal transport model, in order

to estimate potential shifts between competing modes, is again a prerequisite

for this possibility to be explored.

� Further refinements of the cross-border integration performance indicator, in

order to define a procedure to monetize spillover effects, in line with the

approach by Condeço and Gutiérrez (2006), and this way provide a justification

for financial aids in cross-border projects.

� Development of a procedure to integrate the results from the application of the

proposed methodology with those of a CBA-based approach, as suggested by

e.g. BMVBW (2002) and ME&P et al. (2001). This would require defining a

procedure for the transformation of the results of the performance indicators

into their monetary equivalent, or the inclusion of the result of the CBA in the

MCA procedure. Special care should be taken in this integration in order to

avoid double-counting effects.


- 178 -

Chapter 8– REFERENCES

- 179 -

8. REFERENCES

Adler, H. A. (1987). Economic appraisal of transport projects, Economic

Development Institute of the World Bank, Washington, D.C.

Alsnih, R. and Hensher, D. A. (2003). The mobility and accessibility expectations of

seniors in an aging population. Transportation Research A, 37, 903-916.

Annema, J. A., van Wee, B., van Hoek, T., and van der Waard, J. (1999).

Evaluation of Dutch public investment plans. Environmental Impact Assessment

Review, 19, 305-317.

Antunes, P., Santos, R., and Jordáo, L. (2001). The application of Geographical

Information Systems to determine environmental impact significance.

Environmental Impact Assessment Review, 21, 511-535.

Aparicio, A., Menéndez, F., and Sánchez, A. (2004). Disociar Crecimiento

Económico y Demanda de Transporte. Caso de Estudio de España. Proyecto OCDE.

Ministerio de Medio Ambiente.

Arampatzis, G., Kiranoudis, C. T., Scaloubacas, P., and Assimacopoulos, D. (2004).

A GIS-based decision support system for planning urban transportation policies.

European Journal of Operational Research, 152, 465-475.

Aschauer, D. A. (1989). Is public expenditure productive? Journal of Monetary

Economy, 23, 177-200.

Bana e Costa, C., Ensslin, L., Corréa, E. C., and Vansnick, J. C. (1999). Decision

Support Systems in action: Integrated application in a multicriteria decision aid

process. European Journal of Operational Research, 113, 315-335.

Banister, D. and Berechman, Y. (2003). The economic development effects of

transport investments. In Transport Projects, Programmes and Policies: Evaluation

needs and capabilities, A. Pearman, P. Mackie, and J. Nellthorp, (Eds.), Ashgate,

Aldershot.


- 180 -

Banister, D., Maggi, R., Nijkamp, P., and Vickerman, R. (1999). Actors and factors

in the Integration of the Strategic Infrastructure Network in Europe. In New

Contributions to Transportation Analysis in Europe, M. Beuthe and P. Nijkamp,

(Eds.), Avebury, Aldershot.

Banister, D., Stead, D., Steen, P., Akerman, J., Dreborg, K., Nijkamp, P., and

Scheicher-Tappeser, R. (2000a). Scenario building and methodological framework.

In European Transport Policy and Sustainable Mobility, Spon Press, London.

Banister, D., Stead, D., Steen, P., Akerman, J., Dreborg, K., Nijkamp, P., and

Scheicher-Tappeser, R. (2000b). Targets for sustainable mobility. In European

Transport Policy and Sustainable Mobility, Spon Press, London.

Banister, D. and Berechman, Y. (2001). Transport investment and the promotion of

economic growth. Journal of Transport Geography, 9(3), 209-218.

Baradaran, S. and Ramjerdi, F. (2001). Performance of accessibility measures in

Europe. Journal of Transportation and Statistics, 4(2/3), 31-48.

Beinat, E. (1998). A methodology for policy analysis and spatial conflicts in

transport policies. Final Report of the project DTCS (Spatial Decision Support for

negotiation and conflict resolution of environmental and economic effects of

transport policies), Institute for Environmental Studies, Vrije Universiteit,

Amsterdam.

Bel, G. (1997). Changes in travel time across modes and its impact on the demand

for inter-urban rail travel. Transportation Research E, 33(1), 43-52.

Ben-Akiva, M. and Lerman, S. R. (1979). Disaggregate travel and mobility choice

models and measures of accessibility. In Behavioral Travel Modelling, D. A. Hensher

and P. R. Spher, (Eds.), Croom Helm, Andover, Hants.

Berechman, J. and Paaswell, R. (2001). Accessibility Improvements and Local

Employment: An Empirical Analysis. Journal of Transportation Statistics, 4(2/3),

49-66.

Bertolini, L., le Clercq, F., and Kapoen, L. (2005). Sustainable accessibility: a

conceptual framework to integrate transport and land use plan-making. Two test-

applications in the Netherlands and a reflection on the way forward. Transport

Policy, 12(3), 207-220.


- 181 -

Beuthe, M. (2002). Transport Evaluation Methods: from Cost-benefit Analysis to

Multicriteria Analysis and the Decision Framework. In Project and Policy Evaluation

in Transport, L. Giorgi and A. Pearman, (Eds.), Ashgate, Aldershot.

Beuthe, M. and Scanella, G. (1998). Review of some multicriteria methods in

uncertainty. Deliverable D10, Volume 2, Attachment III-1, of the EUNET (Socio-

Economic and Spatial Impacts of Transport) project of the 5th Framework

Programme of the European Commission. Les Facultés Universitaires Catholiques de

Mons -FUCAM, Mons.

Beuthe, M., Eeckhoudt, L., and Scannella, G. (2000). A practical multicriteria

methodology for assessing risky public investments. Socio-Economic Planning

Sciences, 34(2), 121-139.

Bickel, P., Burgess, A., Hunt, A., Laird, J., Lieb, C., Lindberg, G., and Odgaard, T.

(2005). State-of-the art in project assessment. Deliverable D2 of the HEATCO

(Developing Harmonised European Approaches for Transport Costing and Project

Assessment) project of the 6th RTD Framework Programme of the European

Commission.

Biehl, D. (1986). The Contribution of Infrastructure to Regional Development. Final

Report of the Infrastructure Studies Group to the Commission of the European

Communities. Office for Official Publications of the European Communities,

Luxembourg.

Blum, U. (1982). Effects of transportation investments on regional growth: a

theoretical and empirical investigation. Papers of the Regional Science Association,

49, 169-184.

BMVBW (2002). Federal Transport Infrastructure Plan 2003: Basic Features of the

Macroeconomic Evaluation Methodology. Federal Ministry of Transport, Building and

Housing, Berlin.

Boardman, A. E., Greenberg, D. H., Vining, A. R., and Weimer, D. L. (2001). Cost-

benefit analysis. Concepts and practice., Prentice Hall, New Jersey.

Botham, R. (1983). The road programme and regional development: the problem of

the counterfactual. In Transport, location and spatial policy, K. J. Button and D.

Gillingwater, (Eds.), Avebury, Aldershot.


- 182 -

Bristow, A. L. and Nellthorp, J. (2000). Transport project appraisal in the European

Union. Transport Policy, 7(1), 51-60.

Brown, A. L. and Affum, J. K. (2002). A GIS-based environmental modelling system

for transportation planners. Computers, Environment and Urban Systems, 26, 577-

590.

Brown, M., Milner, S., and Bulman, E. (2001). Assessing Transport Investment

Projects: A Policy Assessment Model. In Transport Policy and Research: what

future?, L. Giorgi and R. J. Pohoryles, (Eds.), Ashgate, Aldershot.

Bröcker, J. (1989). How to eliminate certain defects of the potential formula.

Environment and Planning A, 21, 817-830.

Bröcker, J., Capello, R., Lundqvist, L., Meyer, J., Rouwendal, J., Schneekloth, N.,

Spairani, A., Spangenberg, M., Spiekermann, K., van Vuuren, D., Vickerman, R.,

and Wegener, M. (2004). Final Report of Action 2.1.1. of the European Spatial

Planning Observatory Network (ESPON) 2000-2006. Christian-Albreschts-

Universität Kiel, Kiel.

Bröcker, J., Kancs, A., Schürmann, C., and Wegener, M. (2002). Methodology for

the assessment of spatial economic impacts of transport projects and

policies.Deliverable 2 of the IASON (Integrated Appraisal of Spatial Economic and

Network Effects of Transport Investments and Policies) project of the 5th RTD

Framework Programme of the European Commission. Berichte aus dem Institut für

Raumplanung 54, Kiel/Dortmund.

Bruinsma, F. R., Rienstra, S., and Rietveld, P. (1997). Economic Impacts of the

Construction of a Transport Corridor: A Multi-level and Multi-approach Case Study

for the Construction of the A1 Highway in The Netherlands. Regional Studies, 31(4),

391-402.

Bruinsma, F. R. and Rietveld, P. (1993). Urban agglomerations in European

infrastructure networks. Urban Studies, 30(6), 919-934.

Bruinsma, F. R. and Rietveld, P. (1997). The impact of accessibility on the valuation

of cities as location for firms. Vrije Universiteit Amsterdam, Amsterdam.

Bruinsma, F. R. and Rietveld, P. (1998). The accessibility of European cities:

theoretical framework and comparison of approaches. Environment and Planning A,

30, 499-521.


- 183 -

Brundtland Commission (1987). Our Common Future. World Commission on

Environment and Development, New York.

Buckley, W. (1967). Sociology and Modern Systems Theory, Prentince Hall.

Burel, F. and Baudry, J. (2002). Ecología del paisaje. Conceptos, métodos y

aplicaciones, Ediciones Mundiprensa, Madrid.

Burrough, P. A. and McDonnell, R. A. (1998). Principles of Geographical Information

Systems, Oxford University Press, New York.

Button, K. J. (1993). Transport Economics, University Press, Cambridge.

Button, K. J. and Verhoef, E. (1998). Transport at the Edge of Mobility and

Sustainability. In Transport networks in Europe: concepts, analysis and policies, K.

J. Button, P. Nijkamp, and H. Priemus, (Eds.), Edward Elgar Publishing,

Cheltenham/Massachusetts.

Capello, R. and Rietveld, P. (1998). The Concept of Network Synergies in Economic

Theory: Policy Implications. In Transport networks in Europe: concepts, analysis

and policies, K. J. Button, P. Nijkamp, and H. Priemus, (Eds.), Edward Elgar

Publishing, Cheltenham/Massachusetts.

Capineri, C. and Kamann, D.-J. F. (1998). Synergy in Networks: Concepts. In

Transport networks in Europe: concepts, analysis and policies, K. J. Button, P.

Nijkamp, and H. Priemus, (Eds.), Edward Elgar Publishing,

Cheltenham/Massachusetts.

Cervero, R. and Hansen, M. (2002). Induced travel demand and induced road

investment. A simultaneous equation analysis. Journal of Transport Economics and

Policy, 36(3), 469-490.

Cervero, R., Rood, T., and Appleyard, B. (1995). Job accessibility as a performance

indicator: An analysis of trends and their social policy implications in the San

Francisco Bay area. University of California Transportation Center Working Paper

366, University of California, Berkeley, California.

Chatelus, G. and Ulied, A. (1996). Union Territorial Strategies linked to the Trans-

European Transportation Networks (UTS). CEC, DG VII.


- 184 -

Colorni, A., Laniado, E., and Muratori, S. (1999). Decision support systems for

environmental impact assessment of transport infrastructures. Transportation

Research D, 4, 1-11.

Condeço, A. M. and Gutiérrez, J. (2006). Medición de efectos de desbordamiento de

las infraestructuras de transporte a partir de indicadores de accesibilidad .XIV

Congreso Panamericano de Ingeniería de Tránsito y Transporte. Las Palmas de

Gran Canaria.

Copus, A. K. (1999). A New Peripherality Index for the NUTS III Regions of the

European Union. Report for General Directorate XVI.A.4. (Regional Policy and

Cohesion), Rural Policy Group, Management Division, Scottish Agricultural College,

Aberdeen.

Copus, A. K., Spiekermann, K., and Wegener, M. (2002). Aspatial peripherality?

Journal of Nordregio, 3(2), 13-19.

Cowell, F. A. (1995). Measuring inequality, Prentice-Hall, London.

David Simmonds Consultancy, ITS University of Leeds, MVA, and Oxford Brookes

University (1998). Accessibility as a criterion for project and policy appraisal. Final

Report to the Department of the Environment, Transport and Regions (DETR),

Cambridge.

de Orellana-Pizarro, H. (1994). Evaluación de las infraestructuras de transporte y

sus efectos sobre el desarrollo regional mediante la aplicación de indicadores de

accesibilidad., Doctoral Dissertation. Escuela Técnica Superior de Ingenieros de

Caminos, Canales y Puertos. Universidad Politécnica de Madrid.

de Rus, G., Campos, J., and Nombela, G. (2003). Economía del Transporte, Antoni

Bosch, editor,S.A., Barcelona.

DeMers, M. N. (1997). Fundamentals of Geographic Information Systems, Wiley,

New York.

DETR (1998). A New Deal for Trunk Roads in England: Understanding the New

Approach to Appraisal (NATA). Department of the Environment, Transport and the

Regions, London.

DETR (2000). Transport and Planning. Planning Policy Guidance. Department of the

Environment, Transport and the Regions, London.


- 185 -

DfT (2000). Transport Ten Year Plan 2000-2010. Department for Transport,

London.

Dodgson, J. S. (1974). Motorway investment, industrial transport costs and sub-

regional growth, a case study of the M62. Regional Studies, 8, 75-91.

Dodgson, J., Spackman, M., Pearman, A., and Philips, L. (2001). Multi-criteria

analysis: a manual. Department of the Environment, Transport and the Regions,

London.

Domanski, R. (1979). Accessibility, efficiency and spatial organization. Environment

and Planning, 11, 1189-1206.

Donaghy, K. P. D. S. (2003). Structural and Spatial Aspects of Regional Inequality

in Spain: Growth Rates, Spatial Gradients, and Regional Policies. University of

Illinois Working Paper, Vol. 2, No. 2.

Dupuy, G. and Stransky, V. (1996). Cities and highway networks in Europe. Journal

of Transport Geography, 4(2), 107-121.

Dyer, J. S., Fishburn, P. C., Steuer, R. E., Wallenius, J., and Zionts, S. (1992).

Multiple Criteria Decision Making, Multiattribute utility Theory: The next Ten Years.

Management Science, 38(5), 645-654.

EC (1996a). Decision No 1692/96/EC on 23rd July 1996 on Community guidelines

for the development of the trans-European transport network.

EC (1996b). Transport Research- APAS- Methodologies for transport impact

assessment. Office for Official Publications of the European Communities,

Luxembourg.

EC (1996c). Transport Research-EURET Concerted Action 1.1. Cost-benefit analysis

and multi-criteria analysis for new road construction, Office for Official Publications

of the European Communities, Luxembourg.

European Commission (1996d). Methodologies for transport impact assessment.

Transport Research APAS - Strategic transport. Office for Official Publications of the

European Communities, Luxembourg.

EC (1998). Communication from the Commission: Cohesion and Transport COM

(1998) 806 final.


- 186 -

EC (1999). European Spatial Development Perspective (ESDP): Towards balanced

and sustainable development of the territory of the European Union. Committee of

Spatial Development, Luxembourg.

EC (2000a). From Land Cover to Landscape diversity in the European Union.

European Commission, DG Agriculture.

EC (2000b). Green Paper :Towards a European strategy for the security of energy

supply. COM (2000) 769 final.

EC (2001a). Commission of the European Communities. White Paper: "European

Transport Policy for 2010: time to decide". COM (2001) 370 final.

EC (2001b). Unity, solidarity, diversity for Europe, its people and its territory.

Second report on economic and social cohesion. Luxembourg, Office for Official

Publications of the European Communities.

EC (2004a). A new partnership for cohesion: convergence, competitiveness,

cooperation: Third Report on Economic and Social Cohesion. Office for Official

Publications of the European Communities, Luxembourg.

EC (2004b). Communication of the Commission to the Member States of 2

September 2004 laying down guidelines for a Community Initiative concerning

trans-European cooperation intended to encourage harmonious and balanced

development of the European territory (INTERREG III). Official Journal C 226/2 of

10.09.2004.

EC (2004c). Decision No 884/2004/EC of the European Parliament and of the

Council amending Decision No 1692/96/EC on 21 April 2004 on Community

guidelines for the development of the trans-European transport network.

EC (2004d). Energy and Transport: Report 2000-2004. Office for Official

Publications of the European Communities, Luxembourg.

EC (2004e). Policy Support Tools for Transport Issues. European Commission, Joint

Research Centre.

EC (2005a). Thematic Strategy on Air Pollution. COM (2005) 446 final.

EC (2005b). Winning the Battle Against Global Climate Change. COM (2005) 35

final.


- 187 -

EC (2006a). Communication from the Commission to the Council and the Eurpean

Parliament- Keep Europe moving -Sustainable mobility for our continent. Mid-term

review of the European Commission's 2001 Transport White Paper. COM (2006)

314 final.

EC (2006b). Green Paper: A European Strategy for Sustainable, Competitive and

Secure Energy. COM (2006) 105 final. Commission of the European Communities,

Brussels.

ECMT (2004). Assessment & decision making for sustainable transport. OECD

Publications Service, Paris.

ECMT (2005). National systems of transport infrastructure planning. Round Table

128, European Conference of Ministers of Transport, Paris.

EEA (2003). Europe's environment: the third assessment. Environmental

assessment report No 10. European Environment Agency, Copenhagen.

EEA (2006a). Energy and environment in the European Union. Tracking progress

towards integration. EEA Report N0 8/2006. European Environment Agency,

Copenhagen.

EEA (2006b). Transport and Environment: Facing a Dilemma. TERM 2005:

indicators tracking transport and environment in the European Union. EEA Report

N0 3/2006. European Environment Agency, Copenhagen.

ETT and EPYPSA (2006). Desarrollo de un modelo de transporte interurbano de

pasajeros y mercancías en España. Financed by the CEDEX (Centro de Estudios y

Experimentación de Obras Públicas). NEC: 306053. BOE 28/07/2006.

Evers, G. H. M., Meer, P. H., Oosterhaven, J., and Polak, J. (1987). Regional

impacts of new transport infrastructure: a multisectoral potentials approach.

Transportation, 14, 113-126.

Faludi, A. (2002). Positioning European Spatial Planning. European Planning

Studies, 10(7), 897-909 .

Feitelson, E. (2002). Introducing environmental equity considerations into the

sustainable transport discourse: issues and pitfalls. Transportation Research D, 7,

99-118.


- 188 -

Feng, C.-M. and Wu, J. Y.-J. (2003). Highway Investment Planning Model for Equity

Issues. Journal of Urban Planning & Development, 129(3), 161-176.

Fiorello, D., D.Huismans, G., López, E., Marques, C., Steenberghen, T., Wegener,

M., and Zografos, K. (2006). Transport strategies under the scarcity of energy

supply, Buck Consultants International, The Hague.

Foresman, T. W. (1998). GIS early years and the threads of evolution. In The

History of Geographic Information Systems: Perspectives from the Pioneers, T. W.

Foresman, (Eds.), Prentice Hall, NJ.

Fotheringham, A. S. and Wegener, M. (2000). Spatial Models and GIS: New

Potential and New Models, Taylor and Francis, London.

Frost, M. E. and Spence, N. A. (1995). The rediscovery of accessibility and

economic potential: the critical issue of self-potential. Environment and Planning A,

27, 1833-1948.

Frybourg, M. and Nijkamp, P. (1998). Assessing changes in Integrated European

Transport Network Operations. In Transport networks in Europe: concepts, analysis

and policies, K. J. Button, P. Nijkamp, and H. Priemus, (Eds.), Edward Elgar

Publishing, Cheltenham/Massachusetts.

Fundación La Caixa (2006). Anuario Social de España. Fundación La Caixa.

García-Palomares, J. C. (2000). La medida de la accesibilidad. Estudios de

Construcción y Transportes, 88, 95-110.

Geurs, K. and Ritsema van Eck, J. R. (2001). Accessibility measures: review and

applications. Evaluation of accessibility impacts of land-use transportation

scenarios, and related social and economic impacts. RIVM Rapport 408505006,

RIVM, Bilthoven.

Geurs, K. and Ritsema van Eck, J. R. (2003). Evaluation of accessibility impacts of

land-use scenarios: the implications of job competition, land-use, and infrastructure

developments for the Netherlands. Environment and Planning B: Planning and

Design, 30, 69-87.

Geurs, K. T. and van Wee, B. (2004). Accessibility evaluation of land-use and

transport strategies: review and research directions. Journal of Transport

Geography, 12(2), 127-140 .


- 189 -

Giorgi, L. and Pohoryles, R. J. (1999). The TENASSESS Final Report. Final Report of

the TENASSESS project (Policy assessment of TEN and Common Transport Policy)

of the 4th RTD Framework Programme of the European Commission, ICCR -The

Interdisciplinary Centre for Comparative Research in the Social Sciences, Vienna.

Giorgi, L. and Tandon, A. (2000a). The DECODE Method: Theory and Applications.

Final Report of the CODE-TEN project (Strategic Assessment of Corridor

Developments, TEN Improvements and Extensions to the CEEC/CIS) of the 4th RTD

Framework Programme of the European Commission, ICCR -The Interdisciplinary

Centre for Comparative Research in the Social Sciences, Vienna.

Giorgi, L. and Tandon, A. (2000b). The Theory and Practice of Evaluation.

Conclusions from the First TRANSTALK Workshop. Deliverable 2 of the TRANS-TALK

project (Thematic Network Project and Policy Evaluation Methodologies in

Transport) of the 5th Framework Programme of the European Commission, ICCR -

The Interdisciplinary Centre for Comparative Research in the Social Sciences,

Vienna.

Goodchild, M. F. (1992). Geographical Information Science. International Journal of

Geographical Information Systems, 6, 31-45.

Goodwin, P. and Wright, G. (1991). Decision Analysis for Management Judgment,

Wiley, Chichester.

Goodwin, P. B. (1996). Empirical Evidence on Induced Traffic: A Review and

Synthesis. Transportation, 23, 35-54.

Gómez, M. and Bosque, J. (2004). Sensitivity Analysis in Multicriteria Spatial

Decision-Making: A Review. Human and Ecological Risk Assessment, 10, 1173-

1187.

Grant-Muller, S., Mackie, P., Nellthorp, J., and Pearman, A. (2001). Economic

appraisal of European transport projects: the state-of-the-art revisited. Transport

Reviews, 21(2), 237-261.

Greene, D. L. and Wegener, M. (1997). Sustainable transport. Journal of Transport

Geography, 5(3), 177-190.

Guirao, B. (2000). El cálculo del tráfico inducido como herramienta en la

planificación de infraestructuras de transporte. Aplicación a la puesta en servicio de

las nuevas líneas de alta velocidad en España, Doctoral Dissertation. Escuela


- 190 -

Técnica Superior de Ingenieros de Caminos, Canales y Puertos. Universidad

Politécnica de Madrid.

Guitouni, A. and Martel, J.-M. (1998). Tentative guidelines to help choosing an

appropriate MCDA method. European Journal of Operational Research, 109, 501-

521.

Gutiérrez, J. (2001). Location, economic potential and daily accessibility: an

analysis of the accessibility impact of the high-speed line Madrid-Barcelona-French

border. Journal of Transport Geography, 9(4), 229-242.

Gutiérrez, J. (2004). Carreteras y equidad territorial: el papel de los indicadores de

accesibilidad. XXV Semana de la Carretera, Palma de Mallorca.

Gutiérrez, J., González, R., and Gómez, G. (1996). The European high-speed train

network : Predicted effects on accessibility patterns. Journal of Transport

Geography, 4(4), 227-238.

Gutiérrez, J., Gómez, G., García-Palomares, J. C., and López, E. (2006). Análisis de

los efectos de las infraestructuras de transporte sobre la cohesión regional. VII

Congreso de Ingeniería de los Transportes (CIT), Ciudad Real.

Gutiérrez, J. and Monzón, A. (1998). Accessibility, network efficiency, and transport

infrastructure planning. Environment and Planning A, 30, 1337-1350.

Gutiérrez, J. and Urbano, P. (1996). Accessibility in the European Union: the impact

of the trans-European road network. Journal of Transport Geography, 4(1), 15-25.

Halden, D. (2003). Accessibility analysis: concepts and their application to transport

policy, programme and project evaluation. In Transport Projects, Programmes and

Policies: Evaluation needs and capabilities, A. Pearman, P. Mackie, and J. Nellthorp,

(Eds.), Ashgate, Aldershot.

Halden, D. C. (2000). Accessibility: Review of Measuring Techniques and their

Application. Scottish Executive, Edinburgh.

Halden, D. (2002). Using accessibility measures to integrate land use and transport

policy in Edinburgh and the Lothians. Transport Policy, 9(4), 313-324.

Handy, S. L. and Niemeier, D. A. (1997). Measuring accessibility: an exploration of

issues and alternatives. Environment and Planning A, 29, 1175-1194.


- 191 -

Hansen, W. G. (1959). How accessibility shapes land use. Journal of the American

Institute of Planners, 25, 73-76.

Hay, A. (1995). Concepts of equity, fairness and justice in geographical studies.

Transactions of the Institute of British Geographers, New Series, 20(4), 500-508.

Hey, C., Nijkamp, P., Rienstra, S., and Rothenberger, D. (2002). Assessing

Scenarios on European Transport Policies by Means of Multicriteria Analysis. In

Project and Policy Evaluation in Transport, L. Giorgi and A. Pearman, (Eds.),

Ashgate, Aldershot.

High Level Group of the Trans-European Transport Network (2003). Priority

projects for the Trans-European Transport Network up to 2020. High-level group

Report. EC DG-TREN, Brussels.

ICCR (2002a). Improving Evaluation Practices in Transport. Guidelines prepared in

the framework of the TRANS-TALK (Thematic Network Project and Policy

Evaluation Methodologies in Transport) Thematic Network of the 5th Framework

Programme of the European Commission, The Interdisciplinary Centre for

Comparative Research in the Social Sciences (ICCR), Vienna.

ICCR (2002b). Thematic Network: Policy and Project Evaluation Methodologies.

Final Report of TRANS-TALK (Thematic Network Project and Policy Evaluation

Methodologies in Transport) of the 5th Framework Programme of the European

Commission, The Interdisciplinary Centre for Comparative Research in the Social

Sciences (ICCR), Vienna.

Illeris, S. and Jakobsen, L. (1991). The effects of the Fixed Link across the Great

Belt. In Infrastructure and Regional Development, R. Vickerman, (Eds.), Pion,

London.

Ingram, D. R. (1971). The concept of accessibility: a search for an operational

form. Regional Studies, 5, 101-107.

INRETS (2005). Assessment of Relevant Indicators: Indicators of Regional

Accessibility and Social Cohesion. Deliverable 3 of EUNET (Socio-economic and

spatial impacts of transport infrastructure investments and transport system

improvements) of the 4th Framework Programme of the European Commission.

Izquierdo, R. and Monzón, A. (1992). La accesibilidad a las redes de transporte

como instrumento de evaluación de la cohesión económica y social. TTC, 56, 33-56.


- 192 -

Izquierdo, R., Pesquera, M. A., and Ibeas, A. (1980). Accesibilidad demográfica y

renta en municipios de Cantabria. XII Semana de la Carretera, Santander.

Jaeger, J. A. G. (2000). Landscape division, splitting index and effective mesh size:

new measures of landscape fragmentation. Landscape Ecology, 15(2), 115-130.

Jha, M. K. (2003). Criteria-based Decision Support System for selecting highway

alignments. Journal of Transportation Engineering, 129(1), 33-41.

Jha, M. K. and Schonfeld, P. (2004). A highway alignment optimization model using

geographical information systems. Transportation Research A, 38, 455-481.

Kau, J. B. (1976). The interaction of transportation and land use. In Forecasting

Transportation Impacts and Land Use, P. F. Wendt, (Eds.), Martinus Nijhoff, Leiden.

Keeble, D. and Owens, P. L. (1982). Regional accessibility and economic potential

in the European Community. Regional Studies, 16, 419-432.

Keeny, R. L. and Raiffa, H. (1976). Decisions with multiple objectives: preferences

and value trade-offs, Wiley, New York.

Klungboonkrong, P. and Taylor, M. A. P. (1999). An Integrated Planning Tool for

Evaluating Road Environmental Impacts. Computer-Aided Civil and Infrastructure

Engineering, 14, 335-345.

Koenig, J. G. (1980). Indicators of urban accessibility: theory and applications.

Transportation, 9, 145-172.

Laird, J., Nellthorp, J., and Mackie, P. (2005). Network effects and total economic

impact in transport appraisal. Transport Policy, 12, 537-544.

Lau, J. C. Y. and Chiu, C. C. H. (2003). Accessibility of low-income workers in Hong

Kong. Cities, 20(3), 197-204.

Lauridsen, H. (2003). Strategic Transport Planning Evaluation-The Scandinavian

Experience. In Transport Projects, Programmes and Policies: Evaluation needs and

capabilities, A. Pearman, P. Mackie, and J. Nellthorp, (Eds.), Ashgate, Aldershot.

Layard, R. and Glaister, R. (1996). Cost-benefit analysis, Cambridge University

Press, Cambridge.


- 193 -

Lee, D. B. (2002). Appendix B: Demand Elasticities for Highway Travel. In HERS-ST

v20: Highway Economic Requirements System - State Version Technical Report.

FHWA-IF-02-060. Federal Highway Administration, Office of Asset Management.

Washington DC.

Li, X., Wang, W., Li, F., and Deng, X. (1999). GIS based map overlay method for

comprehensive assessment of road environmental impact. Transportation Research

D, 147-158.

Linneker, B. and Spence, N. (1996). Road transport infrastructure and regional

economic development. The regional development effects of the M25 London orbital

motorway. Journal of Transport Geography, 4(2), 77-92.

Litman, T. (2004). Generated Traffic and Induced Travel: Implications for Transport

Planning. Victoria Transport Policy Institute. Available at:

http://www.vtpi.org/gentraf.pdf..

Liu, S. and Zhu, X. (2004). An Integrated GIS Approach to Accessibility Analysis.

Transactions in GIS, 8(1), 45-62.

Lootsma, F. A. (1992). The REMBRANDT system for multicriteria decision analysis

via pairwise comparisons or direct rating. Rep. No. 92-05, Faculty of Technical

Mathematics and Informatics, Delft University of Technology, Delft, The

Netherlands.

López, E. (2005). Measuring regional cohesion effects of large-scale transport

infrastructure investments: an accessibility approach. 45th Congress of the

European Regional Science Association (ERSA), Amsterdam.

López, E., Gutiérrez, J. y Gómez, G. (in press): Measuring regional cohesion effects

of large-scale transport infrastructure investments: an accessibility approach,

European Planning Studies.

López, E., Mancebo, S., Ortega, E., and Monzón, A. (2006a). El cálculo del valor

europeo añadido mediante la utilización de indicadores de accesibilidad: aplicación

a la evaluación de Planes de Infraestructura de Transporte .XIV Congreso

Panamericano de Ingeniería Tránsito y Transporte. Las Palmas de Gran Canaria.

López, E. and Monzón, A. (2004). Impactos territoriales del Plan de Infraestructuras

2000-2007.Ponencia del VI Congreso de Ingeniería del Transporte -CIT 2004.

Zaragoza.


- 194 -

López, E. and Monzón, A. (2006). The STEPs assessment approach. The

assessment methodology. In Transport strategies under the scarcity of energy

supply, A. Monzón and A. Nuijten, (Eds.), Buck Consultants International, The

Hague.

López, E., Monzón, A., Mancebo, S., Ortega, E., Gutiérrez, J., and Gómez, G.

(2006b). Impactos territoriales del PEIT: Plan Estratégico de Infraestructuras y

Transporte 2005-2020. VII Congreso de Ingeniería del Transporte (CIT), Ciudad

Real.

Lucas, K. (2006). Providing transport for social inclusion within a framework for

environmental justice in the UK. Transportation Research A, 40, 801-809.

Lutter, H., Pütz, T., and Spangenberg, T. (1992). Accessibility and Peripherality of

Community Regions: the Role of Road, Long-Distance Railways and Airport

Networks. Report to the European Commission DG XVII, Bundesforschungsanstalt

für Landeskunde und Raumordnung, Bonn.

Mackie, P. and Nellthorp, J. (2003). Transport Appraisal in a Policy Context. In

Transport Projects, Programmes and Policies. Evalaution Needs and Capabilities, A.

Pearman, P. Mackie, and J. Nellthorp, (Eds.), Ashgate.

Mackiewicz, A. and Ratajczak, W. (1996). Towards a new definition of topological

accessibility. Transportation Research B, 30(1), 47-79.

Malczewski, J. (1999). GIS and multicriteria decision analysis, Wiley, New York.

Martellato, D., Nijkamp, P., and Reggiani, A. (1998). Measurement and Measures of

Network Accessibility: Economic Perspectives. In Transport networks in Europe:

concepts, analysis and policies, K. J. Button, P. Nijkamp, and H. Priemus, (Eds.),

Edward Elgar Publishing, Cheltenham/Massachusetts.

Martín, J. C., Gutiérrez, J., and Román, C. (2004). Data Envelopment Analysis

(DEA) Index to Measure the Accessibility Impacts of New Infrastructure

Investments: the Case of the High-Speed Train Corridor Madrid-Barcelona-French

Border. Regional Studies, 38(6), 697-712.

Martínez, F. J. (1995). Access: the transport-land use economic link. Transportation

Research B, 29(6), 457-470.


- 195 -

Martínez-Falero, E. and González-Alonso, S. (1995). Quantitative techniques in

landscape planning, CRC Press, Florida.

May, A. D., Karlstrom, A., Marler, N., Metthews, B., Minken, H., Monzón, A., Page,

M., Pfaffenbichler, P., and Shepherd, S. (2003). PROSPECTS Decision Maker's

Guidebook. Deliverable 15 of PROSPECTS (Procedures for Recommending Optimal

Sustainable Planning of European City Transport Systems) of the 5th Framework

Programme of the European Commission.

McGarigal, K. and Marks, B. J. (1995). FRAGSTRATS: Spatial pattern analysis

program for quantifying landscape structure. Gen. Tech. Report PNW-GTR-351,

ASDA Forest Service, Pacific Northwest Research Station, Portland, OR.

ME&P, NTUA, ITS, and IRPUD (2001). The EUNET/SASI Final Report. Deliverable 19

of the EUNET (Socio-Economic and Spatial Impacts of Transport) project of the 5th

Framework Programme of the European Commission, ME&P, Cambridge.

Megía, M. J. (2002). Oferta de transporte de viajeros por ferrocarril entre ciudades

importantes de la Unión Europea. Estudios de Construcción y Transportes, 95, 71-

92.

Meyer, J. and Straszheim, M. R. (1971). Techniques of Transport Planning, The

Brooking Institution. Transport Research Program, Washington, D.C.

Miller, H. J. (1999a). Modelling accessibility using space-time prism concepts within

Geographical information Systems. International Journal of Geographical

Information Systems, 5(3), 287-301.

Miller, H. J. (1999b). Potential contributions of spatial analysis to geographic

information systems for transportation (GIS-T). Geographical Analysis, 31, 373-

399.

Miller, H. J. and Storm, J. D. (1996). Geographic information system design for

network equilibrium-based travel demand models. Transportation Research C, 4(6),

373-389.

Miller, H. J. and Wu, Y. H. (2000). GIS Software for Measuring Space-Time

Accessibility in Transportation Planning and Analysis. GeoInformatica, 4(2), 141-

159.


- 196 -

Ministerio de Fomento (2004a). Plan Estratégico de Infraestructuras y Transporte.

Documento Diagnóstico. Ministerio de Fomento. Secretaría de Estado de

Infraestructuras y Planificación. Dirección General de Planificación y Coordinación

Territorial, Madrid.

Ministerio de Fomento (2004b). Plan Estratégico de Infraestructuras y Transporte.

Documento Propuesta. Ministerio de Fomento. Secretaría de Estado de Planificación

y Concertación Territorial, Madrid.

Ministerio de Fomento (2005). PEIT: Plan Estratégico de Infraestructuras y

Transporte 2005-2020. Secretaría General Técnica. Ministerio de Fomento, Madrid.

Ministerio de Medio Ambiente (2005). Mapa de Habitats (versión octubre 2005).

Dirección General de la Biodiversidad. Ministerio de Medio Ambiente, Madrid.

Ministerio de Obras Públicas y Transportes (1993). Plan Director de Infraestructuras

1993-2007. Secretaría General de Planificación y Concertación Territorial, Madrid.

Ministerio de Vivienda (2005). Atlas Estadístico de las Áreas Urbanas en España

2004, Ministerio de Vivienda. Secretaría General de Vivienda. Dirección General de

Urbanismo y Política de Suelo, Madrid.

Monzón, A., Cascajo, R., Sammer, G., Kementschitz, R., and Roider, O. (2003).

Multi-criteria evaluation results and sensitivity analysis. Deliverable 5 of the

TRANSECON (Urban transport and socio-economic development) project of the 5th

Framework Programme of the European Commission, TRANSyT-UPM, Madrid.

Monzón, A., Gutiérrez, J., López, E., Madrigal, E., and Gómez, G. (2005).

Infraestructuras de transporte terrestre y su influencia en los niveles de

accesibilidad de la España peninsular. Estudios de Construcción y Transportes, 103,

97-112.

Morris, J. M., Dumble, P. L., and Wigan, M. R. (1979). Accessibility indicators for

transport planning. Transportation Research A, 13A, 91-109.

Morrison, S. A. and Winston, C. (1985). An Econometric Analysis of the Demand for

Intercity Transportation. Research in Transportation Economics, 2, 213-237.

Nellthorp, J., Mackie, P., and Bristow, A. (1998). Measurement and Valuation of the

Impacts of Transport Initiatives. Deliverable D9 of the EUNET (Socio-Economic and


- 197 -

Spatial Impacts of Transport) project of the 5th Framework Programme of the

European Commission, Institute for Transport Studies, University of Leeds, Leeds.

Neuburger, H. (1971). User benefits in the evaluation of transport and land use

plans. Journal of Transport Economics and Policy, 5(1), 52-75.

Ney, S. (2001). Understanding accessibility. In Transport Policy and Research: what

future?, L. Giorgi and R. J. Pohoryles, (Eds.), Ashgate, Aldershot.

Nijkamp, P. (1994). Roads toward environmentally sustainable transport.

Transportation Research A, 28A(4), 261-271.

Nijkamp, P., Rienstra, S., and Vleugel, J. M. (1998). Design and Assessment of

Long Term Sustainable Transport System Scenarios. In Transport networks in

Europe: concepts, analysis and policies, K. J. Button, P. Nijkamp, and H. Priemus,

(Eds.), Edward Elgar Publishing, Cheltenham/Massachusetts.

Nijkamp, P., Rietveld, P., and Voogd, H. (1990). Multicriteria Evaluation in Physical

Planning, North Holland, Amsterdam.

Noland, R. B. and Lem, L. L. (2002). A Review of the Evidence for Induced Travel

and Changes in Transportation and Environmental Policy in the US and the UK.

Transportation Research D, 7, 1-26.

Nutley, S. (2003). Indicators of transport and accessibility problems in rural

Australia . Journal of Transport Geography, 11, 55-71.

OECD (1998). Towards Sustainable Development-Environmental Indicators.

Organisation for Economic Co-operation and Development, Paris, France.

OJEU (2001). Directive 2001/42/EC of the European Parliament and of the Council

of 27 June 2001 on the assessment of the effects of certain plans and programmes

on the environment . Rep. No. Official Journal L 197 , 21/07/2001 P. 0030 - 0037.

OJEU (2006). Council Regulation (EC) No 1083/2006 of 11 July 2006 laying down

general provisions on the European Regional Development Fund, the European

Social Fund and the Cohesion Fund and repealing Regulation (EC) No 1260/1999.

OEJU 31.7.2006.


- 198 -

Ollivier-Trigalo, M. (2001). The implementation of major infrastructure projects:

conflicts and co-ordination. In Project and Policy Evaluation in Transport, L. Giorgi

and A. Pearman, (Eds.), Ashgate, Aldershot.

Oosterhaven, J. and Knaap, T. (2003). Spatial Economic Impacts of Transport

Infrastructure Investments. In Transport Projects, Programmes and Policies:

Evaluation needs and capabilities, A. Pearman, P. Mackie, and J. Nellthorp, (Eds.),

Ashgate, Hampshire.

Ozbay, K., Ozmen, D., and Berechman, J. (2003). Empirical analysis of the

relationship between accessibility and economic development. Journal of Urban

Planning & Development, 129(2), 97-117.

Ozbay, K., Ozmen, D., and Berechman, J. (2006). Modeling and analysis of the link

between accessibility and employment growth. Journal of Transportation

Engineering, 132(5), 385-393.

Pearman, A., Mackie, P., and Nellthorp, J. (2003). Foreword. In Transport Projects,

Programmes and Policies: Evaluation needs and capabilities, A. Pearman, P. Mackie,

and J. Nellthorp, (Eds.), Ashgate, Aldershot.

Pereira, A. M. and Roca-Sagales, O. (2003). Spillover effects of public capital

formation: evidence from the Spanish regions. Journal of Urban Economics, 53(2),

238-256.

Peters, D. (2003). Cohesion, policentricity, missing links and bottlenecks:

conflicting spatial storylines for Pan-European transport investments. European

Planning Studies, 11(3), 317-339.

Pérez-Martínez, P. and Monzón, A. (2005). Informe sobre transporte y medio

ambiente (TRAMA 2005). Centro de Publicaciones Secretaría General Técnica

Ministerio de Medio Ambiente, Madrid.

Pérez-Martínez, P. and Monzón, A. (2006). Transporte y medio ambiente en

España: un análisis que exige mayor control de la demanda. XIV Congreso

Panamericano de Ingeniería de Tránsito y Transporte (PANAM), Las Palmas de Gran

Canaria.

Plassard, F. (1991). Le train à grande vitesse et le réseau des villes. Transports,

345, 14-22.


- 199 -

Plassard, F. (1992). L'impact territorial des transports a grande vitesse. In Espace

et dinamiques territoriales, P. H. Derycke, (Eds.), Economica, Paris.

Preston, J. and Rajé, F. (in press). Accessibility, mobility and transport-related

social exclusion . Journal of Transport Geography.

Quinet, E. (2000). Evaluation methodologies of transportation projects in France.

Transport Policy, 7(1), 27-34.

Reggiani, A. (1998). Accessibility, trade and location behaviour, Aldershot, Ashgate.

Rehfeld, C. (1998). Transport infrastructure investments and decision support

systems., Doctoral Dissertation. Institute of Planning, Transport Studies: Technical

University of Denmark, Lyngby.

Rietveld, P. and Nijkamp, P. (1993). Transport and Regional Development. In

European Transport Economics, J. Polak and A. Heertje, (Eds.), European

Conference of Ministers of Transport (ECMT). Blackwell Publishers, Oxford.

Riitters, K., Wickham, J., and Coulston, J. (2004). Use of road maps in national

assessments of forest fragmentation in the United States. Ecology and Society,

9(2), 13.

Robbins, C. S., Dawson, D. K., and Dowell, B. A. (1989). Habitat area requirements

of breeding forest birds of the middle Atlantic States. Wildlife Monographs 103. The

Wildlife Society, Bethesda, MD.

Rothengatter, W. (2005). National systems of transport infrastructure planning: the

case of Germany. In Round Table 128: National Systems of Transport

Infrastructure Planning, ECMT, (Eds.), ECMT, Paris.

Roy, R. (2003). European versus National-Level Evaluation: The Case of the PBKAL

High-Speed Rail Project. In Transport Projects, Programmes and Policies:

Evaluation needs and capabilities, A. Pearman, P. Mackie, and J. Nellthorp, (Eds.),

Ashgate, Aldershot.

Saaty, T. L. (1990). Multicriteria Decision Making: The Analytic Hierarchy Process,

RWS Publications, Pittsburgh, PA.

SACTRA (1999). Transport and the economy. Standing Advisory Committee on

Trunk Road Assessment. The Stationery Office, London.


- 200 -

Sadek, S., Bedran, M., and Kaysi, I. (1999). GIS Platform for multicriteria

evaluation of route alignments. Journal of Transportation Engineering, 125(2), 144-

151.

Salling.K.A., Leleur, S., and Jensen, A. V. (in press). Modelling decision support and

uncertainty for large transport infrastructure projects: The CLG-DSS model of the

Øresund Fixed Link. Decision Support Systems.

Savelberg, F. and Vogelaar, H. (1987). Determinants of a northern high-speed

railway. Transportation, 14(2), 97-111.

Sayers, T. M., Jessop, A. T., and Hills, P. J. (2003). Multi-criteria evaluation of

transport options-flexible, transparent and user-friendly?. Transport Policy, 10(2),

95-105.

Sánchez, A. and Aparicio, A. (2004). Escenarios de futuro en Pirineos. VI Congreso

de Ingeniería de los Transportes, Ciudad Real.

Sánchez, A. and Zamorano, C. (2006). Geographic Information System analysis of

accessibility as an indicator of potential land use changes and of induced impacts on

the environment. Application to the Pyrenean area of France and Spain.

Transportation Research Record: Journal of the Transportation Research Board, No

1983, 24-32.

Schade, W., Mackie, P., Nellthorp, J., Burgess, A., and Renes, G. (2004).

Methodological advances in Project Assessment within a European context.

Deliverable D5 of the IASON (Integrated appraisal of Spatial Economic and Network

Effects of transport investments and policies) project of the 5th Framework

Programme of the European Commission, TNO INRO, Delft, The Netherlands.

Schürmann, C., Spiekermann, K., and Wegener, M. (1997). Accessibility

indicators.Berichte aus dem Institüt for Raumplanung 39, IRPUD, Dortmund.

Schürmann, C., Spiekermann, K., and Wegener, M. (2004). Spatial impacts of the

Trans-European Networks. 2004 WCTR Proceedings, WCTR Society, Istanbul.

Scottish Executive (2000). Guidance on Accessibility Measuring Techniques. Derek

Halden Consultancy for the Scottish Executive Central Research Unit, Edinburgh.


- 201 -

Seligmann, B. (2005). National systems of transport infrastructure planning: the

case of France. In Round Table 128: National Systems of Transport Infrastructure

Planning, ECMT, (Eds.), ECMT, Paris.

SENDA 3 (1986). Informe metodológico sobre impactos socioeconómicos a nivel

local, interregional y de interés comunitario del enlace entre Irurzun y Andoain.

SENDA 3. Planificación y transportes integrados, S.A.

Serageldin, I. (1996). Sustainability and the Wealth of Nations. First Steps in an

Ongoing Journey. Environmentally Sustainable Development Studies and

Monographs Series No. 5, World Bank, Washington, D.C.

Sheate, W. R. (1992). Strategic environmental assessment in the transport sector.

Project Appraisal, 7(3), 170-174.

Small, K. A. (1999). Project Evaluation. In Essays in Transportation Economics and

Policy, J. A. Gómez-Ibáñez, W. B. Tye, and C. Winston, (Eds.), The Brookings

Institution, Washington, D.C.

Spiekermann, K. and Neubauer, J. (2002). European Accessibility and

Peripherality: Concepts, Models and Indicators. Nordregio Working Paper 2002:9,

Stockholm.

Spiekermann, K. and Wegener, M. (1994). The shrinking continent: new time-space

maps of Europe. Environment and Planning B: Planning and Design, 21(6), 653-

673.

Stead, D. (2001). Transport intensity in Europe -indicators and trends. Transport

Policy, 8(1), 29-46.

Steer Davies Gleave (2004). High Speed Rail: International Comparisons.

Commission for Integrated Transport (CFIT), UK, London.

Stoelinga, A. and Luikens, H. (2005). National systems of transport infrastructure

planning: the case of The Netherlands. In Round Table 128: National Systems of

Transport Infrastructure Planning, ECMT, (Eds.), ECMT, Paris.

Sudgen, R. and Williams, A. (1978). The principles of practical cost-benefit analysis,

Oxford University Press, Oxford.


- 202 -

Talen, E. (1998). Visualizing fairness: Equity maps for planners. Journal of the

American Planning Association, 64(1), 22-49.

Talen, E. and Anselin, L. (1996). Assessing spatial equity: an evaluation of

measures of accessibility to public playgrounds. Environment and Planning A, 30,

595-613.

Tavasszy, L. A., Burgess, A., and Renes, G. (2004). Conclusions and

recommendations for the assessment of economic impacts of transport projects and

policies. Final Report of the IASON (Integrated appraisal of Spatial Economic and

Network Effects of transport investments and policies) project of the 5th Framework

Programme of the European Commission, TNO Inro, Delft, The Netherlands.

Taylor, M. A. P., Woolley, J. E., and Zito, R. (2000). Integration of the global

positioning system and geographical information systems for traffic congestion

studies. Transportation Research C, 8, 257-285.

Törnqvist, G. (1970). Contact systems and regional development, C.W.K. Gleerop,

Lund.

Transport & Mobility Leuven and K.U.Leuven (2006). Final Report on TREMOVE

model v 2.44. Service contract for the European Commission, DG Environment,

Brussels.

Tsamboulas, D., Beuthe, M., Grant-Muller, S., Leleur, S., Nellthorp, J., Panou, K.,

Pearman, A., and Rehfeld, C. (1998). Innovations in decision analysis (for

transport initiatives' evaluation).Deliverable D10, Volume 1: Main text, of the

EUNET (Socio-Economic and Spatial Impacts of Transport) project of the 5th

Framework Programme of the European Commission.

Turró, M. (1999). Going trans-European. Planning and financing transport networks

for Europe, Pergamon, Oxford.

UNFCCC (1997). Kyoto Protocol to the United Nations Convention on Climate

Change. UNFCCC Secretariat, Bonn, Germany.

van Exel, J., Rienstra, S., Gommers, M., Pearman, A., and Tsamboulas, D. (2002).

EU involvement in TEN development: network effects and European value added.

Transport Policy, 9(4), 299-311.


- 203 -

van Wee, B. (2002). Land use and transport: research and policy challenges.

Journal of Transport Geography, 10, 259-271.

van Wee, B., Hagoort, M., and Annema, J. A. (2001). Accessibility measures with

competition . Journal of Transport Geography, 9, 199-208.

Vickerman, R. (2000). Evaluation methodologies for transport projects in the United

Kingdom. Transport Policy , 7(1), 7-16.

Vickerman, R., Spiekermann, K., and Wegener, M. (1999). Accessibility and

Economic Development in Europe. Regional Studies, 33(1), 1-15.

Voogd, H. (1997). The changing role of evaluation methods in a changing planning

environment: some Dutch experiences. European Planning Studies, 5(2), 257-266.

Voogd, H. and Woltjer, J. (1999). The communicative ideology in spatial planning:

some critical reflections based on the Dutch experiences. Environment and Planning

B, 26, 835-854.

Vreeker, R., Nijkamp, P., and Ter Welle, C. (2002). A multicriteria decision support

methodology for evaluating airport expansion plans. Transportation Research D,

7(1), 27-47.

Wegener, M. (2001). New spatial impact models. International Journal of Applied

Earth Observation and Geoinformation, 3(3), 224-237.

Wegener, M., Eskelinen, H., Fürst, F., Schürmann, C., and Spiekermann, K. (2000).

Indicators of Geographical Position. Final Report of the Working Group

'Geographical Position' of the Study Programme on European Spatial Planning

(SPESP). IRPUD, Dortmund.

Williams, H. C. W. L. (1976). Travel demand models, duality relations and user

benefit analysis. Journal of Regional Science, 16(2), 147-166.

Wilson, A. G. (1971). A family of spatial interaction models and associated

developments. Environment and Planning, 3 (1), 1-32.

Yao, E. and Morikawa, T. (2005). A study of an integrated intercity travel demand

model. Transportation Research A, 39, 367-381.


- 204 -

Zhu, X. and Liu, S. (2004). Analysis of the impact of the MRT system on

accessibility in Singapore using an integrated GIS tool. Journal of Transport

Geography, 12, 89-101.

Appendix A

- 205 -

APPENDIX A: DEFINITION OF CRITERIA WEIGHTS

A.1. Questionnaire

In cooperation with other four universities, TRANSyT (Centre for Transport

Research, UPM) is collaborating in a research project whose objective is the

strategic assessment of transport infrastructure plans at a national level. The

project is focused in the extension of the high capacity road and rail networks for

interurban passenger travel. For that purpose, an assessment methodology based

in a multicriteria analysis is being developed.

Given the strategic nature of the assessment, the analysis is focused in

criteria such as regional or social cohesion or the strategic environmental

assessment. One of the key aspects of the methodology is the definition of the

weights to be assigned to each of these criteria.

This is why we would be grateful if you could spend a few minutes filling the

questionnaire included in the back part of this sheet. If you are interested in

receiving information about the results of this questionnaire, please include your e-

mail address in the lower part of this sheet.

Many thanks for your cooperation,

Elena López Suárez [email protected]

TRANSyT, Centre for Transport Research

ETSI Caminos, Canales y Puertos, Universidad Politécnica de Madrid

RESPONDENT INFORMATION

Mark with an X in your sector:

SECTOR

Civil servant

Academic

Consultant

Construction

Other (indicate)………………………………………….

E-mail (if you want to receive the results of the questionnaire):……………………


- 206 -

RANK THE FOLLOWING CRITERIA IN DESCENDING ORDER OF IMPORTANCE:

CRITERION DESCRIPTION IMPORTANCE

Network efficiency Improvement of the quality of the surface transport system

Cross-border integration Improvement of network efficiency in cross-border regions

Social cohesion Improvement of the potential for interaction of structurally lagging and/or inaccessible regions

Regional cohesion Reduction in the regional differences in accessibility

Environmental impact Minimization of strategic environmental impacts

MAKE PARIWISE COMPARISONS BETWEEN CRITERIA, MARKING WITH AN X IN THE FOLLOWING TABLE:

Network efficiency Regional cohesionCross-border integration Social cohesionStrategic environmental impact Cross-border integrationSocial cohesion Network efficiencyRegional cohesion Strategic environmental impactCross-border integration Network efficiencySocial cohesion Regional cohesion Network efficiency Strategic environmental impactCross-border integration Regional cohesion Strategic environmental impact Social cohesion

is absolutelymore important

that

is slightly more important that

is equallyimportant that

is stronglymore important

that

is very stronglymore important

that

is slightly lessimportant that

is strongly lessimportant that

is very stronglyless important

thatis absolutelyless important

that

Network efficiency Regional cohesionCross-border integration Social cohesionStrategic environmental impact Cross-border integrationSocial cohesion Network efficiencyRegional cohesion Strategic environmental impactCross-border integration Network efficiencySocial cohesion Regional cohesion Network efficiency Strategic environmental impactCross-border integration Regional cohesion Strategic environmental impact Social cohesion

is absolutelymore important

that

is slightly more important that

is equallyimportant that

is stronglymore important

that

is very stronglymore important

that

is slightly lessimportant that

is strongly lessimportant that

is very stronglyless important

thatis absolutelyless important

that

Appendix A

- 207 -

A.2. Criteria weights

The above questionnaire was distributed in two engineering-transportation events:

the VI Transportation Engineering Congress 2004 (CIT2004), held in Zaragoza

between 23-25 June, 2004, and the Doctorate Course ‘Transport Policy in the

European Union’, held in November 2004 in the Civil Engineering School of the

Polithecnic University of Madrid. By the time that the questionnaire was designed,

the environmental criterion joined all environmental impacts in a single criterion.

Later, this criterion was split into global warming and habitat fragmentation. This

motivated that the weight attached to each of the two above mentioned

environmental criteria has been computed as half the total environmental weight.

The answers obtained are believed to constitute a valid sample to represent

the preferences of Spanish DMs, transport planners and transport engineering

researchers.

The REMBRANDT procedure (Lootsma, 1982) requires the respondents to

express their preferences via pairwise comparisons on a qualitative scale, as

represented in the Figure included in the questionnaire. These qualitative answers

are given their corresponding numerical values in a +8/-8 interval. These values

are subsequently transformed, using a logarithmic scale, in order to derive each

criterion’ weight.

TABLE A.1. summarizes the results of the weights attached to each of the

criteria, after applying the REMBRANDT procedure to the information contained in a

sample of 38 questionnaires. For practical reasons, weights have been normalized

so that they sum up to 1.

Table A.1.: Base weight profile

CRITERION WEIGHT

Efficiency

Network efficiency 0.261

Cross-border integration 0.130

Cohesion

Regional cohesion 0.212

Social cohesion 0.193

Environment

Global warming 0.102

Habitat fragmentation 0.102


- 208 -

Appendix B

- 209 -

APPENDIX B: CASE STUDY APPLICATION OF THE

ACCESSIBILITY MODEL

This Appendix includes further information on the transport network used for the

calculations of travel times needed for the accessibility model and detailed NUTS-5

results of the network efficiency accessibility indicator.

Do-nothing and PEIT transport networks for the road and rail modes are

included in Figures B.1. to B.6.

Detailed accessibility results are included in Table B.1., which is included

only as a .pdf file in the CD (Table1.pdf), in which each term means:

� CMUNI: NUTS-5 code

� ACE_0: Value of the network efficiency accessibility indicator in the do-

nothing alternative. Road mode

� ACE_P: Value of the network efficiency accessibility indicator in the PEIT

alternative. Road mode

� ARE_0: Value of the network efficiency accessibility indicator in the do-

nothing alternative. Rail mode

� ARE_P: Value of the network efficiency accessibility indicator in the PEIT

alternative. Rail mode


- 210 -

Figure B.1. Spanish road network of the do-nothing alternative (A0)

Appendix B

- 211 -

Figure B.2. Spanish road network of the PEIT alternative (APEIT)


- 212 -

Figure B.3. French road network

Appendix B

- 213 -

Figure B.4. Portuguese road network


- 214 -

Figure B.5. Study area rail network of the do-nothing alternative (A0)

Appendix B

- 215 -

Figure B.6. Study area rail network of the PEIT alternative (APEIT)


- 216 -

CMUNI ACE_0 ACE_P ARE_0 ACE_P CMUNI ACE_0 ACE_P ARE_0 ACE_P1001 1.36 1.32 4.93 3.35 2011 1.46 1.43 4.90 3.221002 1.33 1.30 4.95 3.32 2012 1.38 1.32 4.23 2.731003 1.37 1.34 6.13 4.03 2013 1.44 1.39 5.47 3.281004 1.37 1.34 5.83 3.63 2014 1.44 1.37 4.62 3.131006 1.29 1.26 4.52 2.54 2015 1.36 1.32 4.23 2.781008 1.33 1.29 4.66 2.57 2016 1.45 1.39 5.09 3.431009 1.37 1.32 4.77 3.22 2017 1.51 1.47 5.15 3.451010 1.34 1.31 5.01 3.38 2018 1.33 1.29 4.31 2.811011 1.37 1.33 4.66 2.76 2019 1.46 1.40 4.46 3.041013 1.36 1.32 4.96 3.38 2020 1.43 1.39 4.65 3.211014 1.31 1.28 4.58 2.59 2021 1.40 1.35 4.29 2.851016 1.40 1.36 4.70 2.88 2022 1.42 1.37 4.46 2.961017 1.41 1.37 5.23 3.63 2023 1.40 1.36 5.40 3.171018 1.34 1.30 4.76 2.67 2024 1.37 1.33 4.58 3.121019 1.41 1.36 4.60 2.78 2025 1.33 1.29 4.25 2.661020 1.30 1.28 4.68 2.72 2026 1.40 1.36 4.51 3.051021 1.35 1.32 4.69 2.61 2027 1.35 1.31 4.41 2.981022 1.39 1.33 4.60 2.69 2028 1.49 1.44 5.43 3.711023 1.40 1.35 4.79 2.89 2029 1.33 1.28 4.16 2.711027 1.37 1.34 4.90 3.32 2030 1.45 1.42 4.70 3.031028 1.35 1.31 4.61 3.28 2031 1.45 1.42 4.74 3.071030 1.41 1.37 5.04 3.72 2032 1.35 1.31 4.20 2.741031 1.39 1.33 4.70 2.79 2033 1.38 1.34 4.88 3.211032 1.41 1.36 4.53 2.72 2034 1.38 1.33 4.47 3.011033 1.39 1.32 4.71 2.80 2035 1.33 1.28 4.02 2.551034 1.38 1.34 4.69 2.79 2036 1.38 1.32 4.35 2.891036 1.32 1.30 4.95 3.17 2037 1.35 1.32 4.25 2.591037 1.39 1.35 4.91 2.81 2038 1.38 1.33 4.26 2.781039 1.40 1.33 4.46 2.65 2039 1.38 1.34 4.45 3.011041 1.40 1.34 4.65 2.74 2040 1.36 1.32 4.27 2.831042 1.36 1.33 5.11 3.33 2041 1.42 1.36 4.57 3.071043 1.39 1.32 4.40 2.59 2042 1.49 1.45 4.87 3.191044 1.37 1.33 4.82 3.50 2043 1.41 1.36 4.52 3.071046 1.30 1.27 4.60 2.63 2044 1.45 1.41 4.68 3.001047 1.29 1.26 4.49 2.51 2045 1.37 1.32 4.33 2.861049 1.33 1.29 4.69 2.72 2046 1.37 1.31 4.25 2.791051 1.36 1.31 4.71 3.14 2047 1.46 1.41 4.64 3.141052 1.37 1.32 4.79 3.45 2048 1.35 1.29 4.23 2.761053 1.37 1.33 4.74 3.18 2049 1.47 1.43 4.98 3.281054 1.31 1.28 5.05 3.64 2050 1.35 1.31 4.15 2.661055 1.35 1.31 4.75 2.79 2051 1.35 1.31 4.40 2.891056 1.41 1.37 4.98 3.41 2052 1.37 1.31 4.21 2.761057 1.38 1.34 4.70 2.79 2053 1.41 1.36 4.30 2.881058 1.33 1.29 4.77 2.67 2054 1.39 1.34 4.40 2.941059 1.32 1.29 4.54 2.46 2055 1.58 1.55 5.54 4.351060 1.41 1.35 4.51 2.70 2056 1.37 1.34 4.68 3.021061 1.37 1.33 4.79 3.24 2057 1.41 1.36 4.42 2.991062 1.30 1.27 4.56 2.57 2058 1.51 1.46 5.37 3.661063 1.32 1.29 5.17 3.75 2059 1.46 1.40 4.82 3.241901 1.31 1.28 4.72 2.62 2060 1.41 1.36 4.86 3.171902 1.30 1.26 4.52 2.54 2061 1.35 1.32 4.35 2.912001 1.40 1.35 4.51 3.06 2062 1.43 1.38 5.13 3.322002 1.42 1.38 4.56 3.12 2063 1.39 1.35 4.75 3.072003 1.33 1.29 3.96 2.52 2064 1.41 1.36 4.37 2.932004 1.37 1.34 4.58 2.92 2065 1.41 1.36 4.37 2.922005 1.38 1.34 5.38 3.14 2066 1.42 1.37 4.62 3.112006 1.43 1.39 4.83 3.15 2067 1.45 1.41 5.20 3.492007 1.41 1.37 4.68 3.24 2068 1.43 1.36 4.65 3.062008 1.43 1.36 4.77 3.16 2069 1.33 1.29 4.03 2.572009 1.31 1.28 4.03 2.54 2070 1.42 1.36 4.92 3.29

Page 1 of 66


Page 2 of 66


Page 3 of 66


Page 4 of 66


Page 5 of 66


Page 6 of 66


Page 7 of 66


Page 8 of 66


Page 9 of 66


Page 10 of 66


Page 11 of 66


Page 12 of 66

CMUNI ACE_0 ACE_P ARE_0 ACE_P CMUNI ACE_0 ACE_P ARE_0 ACE_P9398 1.31 1.27 5.78 3.03 9482 1.36 1.32 5.21 2.749400 1.34 1.29 4.77 3.85 9483 1.35 1.31 4.84 3.939403 1.36 1.33 4.93 4.01 9485 1.28 1.25 4.78 3.439405 1.37 1.31 4.96 4.01 9901 1.33 1.30 4.83 3.569406 1.28 1.25 4.81 3.55 9902 1.32 1.29 4.48 2.599407 1.38 1.33 4.90 2.92 9903 1.33 1.27 6.26 4.079408 1.30 1.27 4.84 3.48 9904 1.30 1.27 4.47 2.559409 1.37 1.32 6.51 4.31 9905 1.32 1.29 4.67 2.789410 1.37 1.33 5.86 3.73 9906 1.28 1.24 4.39 2.469411 1.32 1.27 4.78 2.74 9907 1.28 1.25 4.34 2.439412 1.35 1.30 5.04 2.77 9908 1.37 1.34 6.32 4.149413 1.34 1.28 5.76 2.92 10001 1.43 1.36 5.04 3.83

10002 1.42 1.36 4.90 4.15 10063 1.46 1.38 5.02 3.8610003 1.41 1.33 4.97 3.13 10064 1.40 1.32 4.95 3.1910004 1.41 1.34 4.65 2.89 10065 1.47 1.41 4.97 3.0410005 1.44 1.36 4.90 3.79 10066 1.45 1.39 5.17 4.4010006 1.44 1.35 4.87 3.72 10067 1.38 1.28 4.71 2.9210007 1.42 1.34 4.83 3.05 10068 1.47 1.42 5.03 3.1010008 1.43 1.36 4.95 3.49 10069 1.41 1.34 5.05 3.1410009 1.45 1.39 4.93 4.18 10070 1.43 1.37 4.77 2.9410010 1.40 1.32 4.80 3.15 10071 1.48 1.41 5.11 3.2010011 1.50 1.44 5.08 3.25 10072 1.45 1.37 5.91 4.3910012 1.39 1.31 4.55 2.80 10073 1.40 1.33 4.68 3.9410013 1.39 1.33 4.80 3.01 10075 1.44 1.39 4.79 2.9310014 1.47 1.42 4.87 2.99 10076 1.37 1.27 4.63 3.5010015 1.42 1.35 5.04 3.82 10077 1.53 1.47 5.44 4.6610016 1.40 1.30 4.68 3.57 10078 1.46 1.39 5.31 4.0710017 1.48 1.43 5.28 3.46 10079 1.47 1.42 5.01 3.0910018 1.40 1.30 4.64 3.17 10080 1.43 1.35 5.04 3.8210019 1.36 1.30 4.68 2.78 10081 1.45 1.37 4.98 3.0510020 1.41 1.34 4.72 3.98 10082 1.39 1.32 4.63 2.8710021 1.39 1.31 4.58 3.08 10083 1.47 1.42 4.95 3.0410022 1.45 1.38 5.10 3.15 10084 1.45 1.37 5.15 3.2710023 1.41 1.33 4.93 3.27 10085 1.37 1.33 4.97 3.1110024 1.43 1.36 5.18 3.95 10086 1.42 1.34 4.95 3.7510025 1.46 1.38 5.03 3.87 10087 1.50 1.44 5.26 3.3810026 1.38 1.32 4.53 2.67 10088 1.42 1.34 4.87 3.7810027 1.46 1.39 4.93 3.13 10089 1.42 1.33 4.82 3.7210028 1.40 1.36 5.04 3.20 10090 1.44 1.36 4.91 3.7610029 1.55 1.49 5.34 3.48 10091 1.48 1.43 4.88 2.9810030 1.40 1.35 4.65 2.77 10092 1.43 1.37 5.16 4.3910031 1.43 1.36 4.96 3.16 10093 1.45 1.38 5.02 3.9210032 1.42 1.35 4.82 3.34 10094 1.45 1.39 5.21 4.0410033 1.53 1.46 5.11 3.25 10095 1.41 1.33 4.97 3.5710034 1.44 1.36 4.89 3.72 10096 1.45 1.37 5.14 3.9210035 1.45 1.39 5.13 3.94 10097 1.40 1.35 4.65 2.8110036 1.46 1.38 4.97 3.81 10098 1.43 1.29 4.70 2.8110037 1.37 1.29 4.37 2.62 10099 1.37 1.28 4.62 2.7910038 1.42 1.33 4.75 2.99 10100 1.40 1.32 5.00 3.1610039 1.45 1.38 5.08 3.97 10101 1.38 1.30 4.82 3.0310040 1.41 1.31 4.85 3.04 10102 1.40 1.34 4.98 4.2310041 1.48 1.42 5.18 4.02 10103 1.37 1.31 4.65 2.8410042 1.44 1.39 4.76 2.90 10104 1.44 1.38 4.89 2.9810043 1.44 1.37 4.87 4.14 10105 1.46 1.41 4.82 2.9210044 1.51 1.45 5.66 4.86 10106 1.43 1.34 4.88 3.7010045 1.36 1.27 4.43 2.62 10107 1.45 1.40 5.23 4.0310046 1.45 1.39 5.00 3.70 10108 1.50 1.45 5.23 3.2610047 1.40 1.31 4.63 3.52 10109 1.51 1.45 5.43 4.6510048 1.49 1.45 4.94 3.10 10110 1.47 1.42 4.84 2.9410049 1.39 1.29 4.47 2.72 10111 1.45 1.42 4.94 3.06

Page 13 of 66


Page 14 of 66


Page 15 of 66


Page 16 of 66


Page 17 of 66


Page 18 of 66


Page 19 of 66


Page 20 of 66


Page 21 of 66


Page 22 of 66


Page 23 of 66


Page 24 of 66


Page 25 of 66


Page 26 of 66


Page 27 of 66


Page 28 of 66


Page 29 of 66


Page 30 of 66


Page 31 of 66


Page 32 of 66


Page 33 of 66


Page 34 of 66


Page 35 of 66


Page 36 of 66


Page 37 of 66


Page 38 of 66


Page 39 of 66


Page 40 of 66


Page 41 of 66


Page 42 of 66


Page 43 of 66


Page 44 of 66


Page 45 of 66


Page 46 of 66


Page 47 of 66


Page 48 of 66


Page 49 of 66


Page 50 of 66


Page 51 of 66


Page 52 of 66


Page 53 of 66


Page 54 of 66


Page 55 of 66


Page 56 of 66


Page 57 of 66


Page 58 of 66


Page 59 of 66


Page 60 of 66


Page 61 of 66


Page 62 of 66


Page 63 of 66


Page 64 of 66


Page 65 of 66

CMUNI ACE_0 ACE_P ARE_0 ACE_P CMUNI ACE_0 ACE_P ARE_0 ACE_P50148 1.48 1.41 5.22 3.80 50273 1.44 1.39 5.10 4.0850149 1.47 1.39 4.78 3.02 50274 1.44 1.39 5.11 4.0950150 1.35 1.30 4.03 2.85 50275 1.47 1.41 4.44 3.2550151 1.43 1.37 5.16 3.76 50276 1.43 1.36 4.94 3.5450214 1.46 1.41 4.38 3.10 50277 1.35 1.31 4.35 3.1050215 1.37 1.33 4.80 3.57 50210 1.52 1.45 5.68 4.3450216 1.36 1.31 4.24 2.70 50211 1.35 1.30 4.20 3.0350217 1.40 1.34 4.10 2.57 50212 1.34 1.29 4.03 2.9650218 1.46 1.41 4.77 3.14 50213 1.56 1.50 5.06 4.2250219 1.34 1.28 4.18 2.70 50278 1.40 1.34 4.26 3.1350220 1.43 1.35 4.83 3.45 50279 1.35 1.31 3.80 2.5250221 1.48 1.43 4.72 3.55 50280 1.41 1.33 4.45 2.8750222 1.36 1.29 4.14 3.01 50281 1.40 1.32 4.32 3.4250223 1.35 1.30 3.96 2.89 50282 1.36 1.32 4.40 3.1550224 1.40 1.34 4.58 2.91 50283 1.42 1.33 4.52 2.7850225 1.33 1.28 4.21 3.05 50284 1.39 1.34 3.96 2.6650227 1.44 1.34 4.65 3.92 50285 1.34 1.28 4.17 2.7050228 1.35 1.30 4.08 2.91 50286 1.35 1.32 3.77 2.5150229 1.40 1.36 3.91 2.65 50287 1.39 1.35 4.48 3.2350230 1.45 1.39 4.84 3.09 50288 1.33 1.27 3.95 2.4850231 1.36 1.31 4.13 2.96 50289 1.42 1.36 4.94 3.9350232 1.48 1.35 5.54 4.10 50290 1.41 1.35 4.42 2.7950233 1.48 1.43 4.85 3.70 50291 1.50 1.43 4.81 3.1850234 1.43 1.36 4.64 3.06 50292 1.40 1.34 4.48 2.8150235 1.36 1.31 4.07 2.61 50293 1.39 1.35 3.95 2.6850236 1.37 1.33 4.29 3.13 50294 1.42 1.36 4.57 2.9050237 1.41 1.33 4.45 2.95 50295 1.44 1.37 4.44 2.8150238 1.46 1.38 4.67 3.29 50296 1.41 1.35 4.27 3.1450239 1.44 1.40 5.13 4.11 50297 1.31 1.26 3.36 2.1950240 1.41 1.35 4.37 3.24 50298 1.35 1.30 4.00 2.5550241 1.36 1.33 4.25 3.08 50901 1.46 1.40 4.98 3.5950242 1.36 1.33 3.80 2.54 50902 1.44 1.37 5.01 3.6250243 1.39 1.35 4.27 3.1150244 1.40 1.34 4.74 3.1150245 1.47 1.37 5.50 4.1450247 1.32 1.27 3.92 2.8250248 1.46 1.39 4.93 3.0050249 1.41 1.36 4.39 2.8550250 1.51 1.44 4.65 3.0750251 1.39 1.31 4.37 2.8750252 1.37 1.32 4.01 2.4950253 1.34 1.30 4.37 3.1250254 1.41 1.37 4.38 3.2250255 1.39 1.35 4.51 2.8750256 1.51 1.46 5.42 4.4050257 1.37 1.33 3.84 2.5850258 1.43 1.36 4.53 2.8650259 1.38 1.34 4.97 3.8250260 1.41 1.37 4.19 2.9150261 1.40 1.32 4.45 2.9450262 1.33 1.27 3.99 2.9250263 1.40 1.36 4.55 3.3050264 1.43 1.36 4.41 2.7850265 1.43 1.35 4.52 2.9350266 1.44 1.40 4.47 3.3050267 1.48 1.41 4.98 3.2550268 1.49 1.42 5.01 3.0850269 1.35 1.30 4.15 2.9750271 1.48 1.44 5.31 4.2950272 1.32 1.27 3.99 2.89

Page 66 of 66

DOCTORAL THESIS - oa.upm.esoa.upm.es/373/1/ELENA_LOPEZ_SUAREZ.pdf · sus siempre acertados...

Documents

Transcript of DOCTORAL THESIS - oa.upm.esoa.upm.es/373/1/ELENA_LOPEZ_SUAREZ.pdf · sus siempre acertados...