VIEWS: 1 PAGES: 13 POSTED ON: 10/14/2012
The Economic and Environmental Consequences of Regulating Traffic Integrated Traffic, Regional Economic, Impact and Assessment Models – MERGE (Model for Exchanging Regionalised Geographic Entities) Jacob Kronbak & Steen Leleur Centre for Traffic and Transport (CTT) Technical University of Denmark (DTU) Abstract The prime aim of the TRIP project “Integrated Traffic, Regional Economic and Impact Models” is to create a decision-making tool for use with transport policy initiatives and transport sector projects, which takes into account both the economic and the environmental consequences of alternative decisions. The decision-making tool is constructed through a GIS1-based linkage of different types of models; a transport model, an interregional economic model, an environmental impact model and an assessment model. This paper presents the chosen framework for the modelling tool (MERGE - Model for Exchanging Regionalised Geographic Entities) and describes the perspectives for qualifying the decision- making process by integrating existing models. The above four mentioned models are applied to develop MERGE in its first version. It should be noted that MERGE is given a modular structure that allows other models – some are briefly described in the paper – to enter the framework. In this way it becomes possible to tailor specific studies in accordance with specific demands and budget means. The paper also includes a description of how the framework will be used for evaluating a transport policy initiative in the form of a road-pricing scenario in Denmark and finally the paper describes conclusions and research perspective. 1 Geographical Informations System Introduction The purpose of the paper is to give a presentation of the current status of the Danish TRIP project 2.3 which overall has been described as follows [TRIP (2000)]: “The prime aim of the project is to create a decision-making tool for use with transport policy initiatives and transport sector projects, which takes into account both the economic and the environmental consequences of alternative decisions. The decision-making tool will be tested in relation to road pricing in Denmark.” More detailed the following elements have been identified in the modelling process: Development of a system of data exchange between transport models, regional economic models and impact models, specifically designed to facilitate environmental impact evaluation. Extension and improvement of existing models through increasing the quality of data relating to exogenous variables. Assessment of the predictive power of transport models and regional economic models. Development of a research framework for improvement of the existing models based on the experience of the project. Identification of the environmental consequences, economic efficiency and distributional implications of different policy options. The paper sets off with an introduction to MERGE and the chosen transport, interregional economic, environmental impact and assessment models. This is followed by a description of how the framework will be used for evaluating a transport policy initiative in the form of a road-pricing scenario in Denmark and finally the paper describes conclusions and research perspective. MERGE The keyword in MERGE (Model for Exchanging Regionalised Geographic Entities) is model integration. First of all MERGE has to link the transport, interregional economic, environmental impact and assessment models together into the desired decision-making tool by making procedures for transferring in- and output data between the models. For each type of models an existing model has been chosen for the decision-making tool: MERGE module A. The National Transport Model, LTM, developed by the Danish National Road Directorate and others. MERGE module B. The environmental and economic impact assessment tool TicMap, developed at the Technical University of Denmark. MERGE module C. The interregional economic model LINE, covering all Danish regions, developed by the Local Government Research Institute (AKF) in Copenhagen. MERGE module D. The Composite Model for Assessment (COSIMA) developed at the Technical University of Denmark. The integration of these models addresses some questions that are not only relevant for the specific models in question but also for a more general application. This means that some of the procedures developed in MERGE are quite universal for model integration. Instead of making MERGE a specific tool for integrating the LINE, LTM, TicMap and COSIMA models, the starting point has been a more general approach to model integration (incl. data transfer) and from there on to focus on the specific models. This approach has the advantage that if other (or better) models become available they can be utilised with less effort than if MERGE was specifically designed for the mentioned models. This can be referred to as model modularity and leads to a structure of MERGE as illustrated in Figure 1. Figure 1: Schematic structure of the MERGE modular structure. The principle is that other models can be ”snapped” on to MERGE and used in coherence with the other models in the modelling framework (examples of other present and future models are listed in Table 1). Besides the modularity of MERGE another important factor in the model integration is consistency. One of the objectives of MERGE is to provide data exchange between the transport, regional economic, impact and assessment models. All these types of models rely to a certain extent on spatially distributed data but not necessarily on data with the same spatial distribution. When integrating the models into a decision support system, it is important to ensure a common consistent basis so that results and conditions to some extent are identical or at least consistent. It is especially important to be able to reproduce results and datasets. It might sound simple, but if e.g. a model requires 50 zones, MERGE has to be able to generate this number of zones, preferably from any base dataset, under a number of different conditions e.g. equal number of inhabitants within the zone; equal area of the zones etc. At the same time it has to be possible to keep track of where data originated from and to give some estimates of the accuracy of not only the original data, but also the generated data. This is commonly known as metadata (or data on data). Generation of new datasets from an existing dataset is where GIS have been proven as a very powerful tool. Results and input data all have some kind of spatial attributes e.g. population data can be on a municipal or a parish level. It is not necessary that all the integrated models actually use the spatial reference but the spatial reference can be used in MERGE to generate and exchange datasets. LTM – MERGE module A The Danish Road Directorate (VD), the Danish National Rail Company (DSB), the Danish National Rail Agency, and the Danish Ministry of Transport have jointly developed the Danish National Transport Model (LTM). The model provides an assessment of the national and regional passenger transport by car, rail, bus, ferry and aeroplane in Denmark. The model assesses both the development of the total transport effort in Denmark, as well as the link flows. The zoning structure of the model is based on Danish parishes (sogne), making the zoning structure quite detailed with approximately 1,100 zones. In Copenhagen the zoning structure is based on the Greater Copenhagen Transport Model (HTM). All border crossings to Germany and all ferry connections abroad are linked to each of their port zones. The road network consists of approximately 10,000 links with information on travel speed, link length and traffic counts. The rail network is based on information on routes, travel times, frequencies, cost structure and passenger counts. The model also contains a bus network of approximately 1,100 lines with information on routes and travel costs. Finally, the model contains a network of domestic air routes. The socio-economic data is obtained from the national travel habit surveys. The present version of the model is based on interviews from this survey of October 1994 to September 1995, involving approximately 14,000 people. The model is maintained in EMME/2 that calculates supply data (travel time, distance, travel cost etc.) between each zone pair. The resistance matrices produced are then used in the hierarchical logit-model FREDERIK, consisting of four sub-models: 1. Model of car ownership based by zone 2. Model of trip frequency by zone 3. Model of destination choice 4. Modal split model (car, bus, train, aeroplane, bicycle/walk). The sub-models of FREDERIK are applied once for each of 9 trip purposes: 1. Business trips 2. Private long distance trips 3. Home-work trips 4. Leisure trips, weekday 5. Shopping trips, weekday 6. Other trips, weekday 7. Leisure trips, weekend 8. Shopping trips, weekend 9. Other trips, weekend. FREDERIK undertakes the analyses based on the information on supply and socio-economic data on population and workplaces by zone as well as assumptions on the taxation and travel cost structure and development in the economy. TICMAP – MERGE module B The TicMap (Traffic Impact and Cost MAPping) model [Wass-Nielsen, M. & Hviid Steen, C. (2001)] is a MapInfo tool for traffic impact calculation. The basis of TicMap is four impact models for: 1. Accidents 2. Noise 3. Emission 4. Severance and perceived risk Each of these impact models are based on Danish impact assessment models and the can be run individually from within MapInfo. The results from the impact models are used to make a assessment of the impacts on each road segment in the network. For a closer description of the impact models see e.g. [Wass-Nielsen, M. & Hviid Steen, C. (2001)], [Clausen, F., Wätjen, W. & Leleur, S. (1991)] or [Leleur, S. (2000)] LINE – MERGE module C LINE is an interregional general equilibrium model that has been developed at The Local Government Research Institute (AKF) in Copenhagen [Madsen, B.; Jensen-Butler, C. & Dam, P.U. (2001)]. In relation to the project proposal two characteristics of LINE are important. First the model incorporates the consequences of changes in transport costs for regional competitiveness and through this, the level of regional economic activity and relocation of economic activity and households. Second, LINE can model changes in interaction and thereby changes in origins and destinations for different types of traffic. Two basic economic processes form the skeleton of LINE. These two processes are denoted the real circle and the cost-price circle. The real circle is the demand-production-demand process and is an iterative or a round-by-round process: Regional demand determines regional production, which generates regional income, which in turn influences regional demand. This forms a circle, where round-by-round impacts add up to a total impact. In the cost-price circle, regional level of wages and prices of intermediate consumption commodities determine the cost of production. Regional production cost is transformed into regional prices of commodities (basic prices), which in turn determine regional prices in markets and thereby intermediate consumption cost and final demand prices (buyers prices). This cost-price circle also includes a round-by-round process, where cost-price changes are passed into the regional and interregional production system [Madsen, B.; Jensen-Butler, C. & Dam, P.U. (2001)]. The LINE real circle A simplified illustration of the LINE real circle for modelling the consequences for production, income and employment are show in Figure 2. Figure 2: The LINE real circle [Madsen, B.; Jensen-Butler, C. & Dam, P.U. (2001)]. The regional economic impacts calculated in LINE can be summed up like this: Reductions in the price of commodities influence the disposable income for households. An increase in the regional disposable income influences the private consumption. An increase in private consumption influences the amount of commodities purchased by the households under the assumption that the composition of commodities is assumed unchanged. The increase in private consumption again influences the supply of commodities to an area either by an increase in the Danish production (by place of production) or an increased import. An increase in production leads to an increase in disposal income for the households (by place of household). An increase in production will at the same time lead to an increase in demand for raw material (by place of demand) that again will lead to an increase in production of raw material (by place of production) etc. (under the assumption that the composition of raw material for each industry is unchanged). This demand circle shows in other words how changes in the economic activity in one area influence the surrounding areas through a series of mechanism e.g. household consumption, commodity trade and commuting [Madsen, B.; Jensen-Butler, C. & Dam, P.U. (2001)]. The LINE cost-price circle The prices of commodities are first of all determined by the cost of production. But the price of commodities is also influenced by changes in transport cost both when raw material is transported to the place of production, when commodities are transported from the place of production to the market and when the commodities are transported from the market to the consumer. Besides that the disposal income of the households is also influenced by the cost of commuting. The influences on prices are shown on Figure 3. The figure shows (in a simplified way) how cost and price changes in the production influence the cost of commodities on the market, which again influences cost and prices in production etc. At the same time the figure shows how LINE calculates the changes in regional production cost and the price of commodities. Figure 3: The LINE cost-price circle [Madsen, B.; Jensen-Butler, C. & Dam, P.U. (2001)]. As it can be seen on Figure 3, the production prices of specific commodities are determined by the sum of production cost meaning cost of row material, wages and profit. The costs of transport are added to the basic cost when commodities are transported from the producer to the market and thereby the production cost is made into the selling price. The commodities are purchased by either the households (private consumption on address) or by other industries (raw material after place of production). The price of the commodities – from the households point of view – are hereby added the cost of the transport from the marked to the household. The cost of transport thereby influences the disposable income of the household. The price – from the industries point of view – includes the cost of transport of raw material to the place of production. The commodities from these industries are then again added the cost of transport – meaning that the households or industries in the next change of the production chain will have these costs added to the price etc. The cost of transport also directly influences the disposable income and wages. If the cost of commuting is lowered then the disposable income – all other things equalled – will go up [Madsen, B.; Jensen-Butler, C. & Dam, P.U. (2001)]. COSIMA – MERGE module D The COSIMA model (composite model for assessment) has been worked out to provide a more comprehensive assessment of transport initiatives than made possible by applying a conventional cost-benefit analysis (CBA). Thereby COSIMA deals with a mix of CBA effects and non-CBA effects. Typically the non-CBA effects - when seen from a modelling viewpoint - are more difficult to handle as compared to the CBA effects where handbook approaches (pricing and procedures) are available for many transport planning problems. In brief we will refer to the CBA effects as effects where pricing manuals and procedures exist and to the non-CBA effects as multi-criteria analysis (MCA) effects as this type of analysis, stemming from operations research, becomes relevant for the extension of the conventional CBA. The idea of COSIMA can briefly be described by the following seven steps as formulated for the assessment of a number of alternative by-pass projects for a Danish town currently in need of relieving of the through traffic [Leleur, S. (2001)]: 1. The first task is to determine the CBA effects being relevant for the concrete appraisal study. In the by-pass example, following the Danish Road Directorate’s standard model, the effects are: travelling time, vehicle operating costs, accidents, maintenance costs, noise, air pollution and severance & perceived risk. The investment enters the analysis denominated as the construction costs. 2. The next task is to determine the MCA effects of relevance. These may be measured either in some type of quantitative unit, an example could be changes in strategic mobility see [Kronbak, J. (1998)], or by judgement using a +5, .., 0, .., -5 scale. In the example below three MCA effects are taken into account by using such a point scale: network accessibility, urban planning and landscape. 3. With CBA and MCA effects laid down the so-called “anchoring” part of the model formulation can take place. Hereby is meant determining the importance of the MCA effects against the CBA effects and in-between each other. Several MCA techniques are relevant here: direct weights, pairwise comparison, swing weights, etc. see [Leleur, S. (2000)]. Criteria importance is denominated by weights on the individual criteria adding up to 100%. 4. At this stage the base case scenario termed A is modelled and presented to the decision- makers together with the assumed interesting assessment questions; these concern issues that may have a principal influence on the decisions to be made from the study. The decision-maker involvement may lead to revision of both the kind of MCA impacts included and their weights in scenario A. Part of this exchange with decision-makers is also to formulate suitable additional scenarios; in the by-pass example these are the scenarios B, C and D. 5. Afterwards COSIMA is run for all the scenarios and the assessment questions are scrutinised and related to possible sources of uncertainty. This “deterministic run” of the model and the identification of “varying levels” of information give intermediate results that in the following model step are examined as concerns their “feasibility risks” which indicate that a result in the deterministic run may be associated with such uncertainty that precaution is needed. 6. Next the so-called “stochastic run” of COSIMA is undertaken. This in fact is a Monte Carlo simulation where parameters and data have been replaced with suitable probability distributions that can represent the actual information level. In the by-pass example used as background for this overview of steps in COSIMA both distributions arrived at empirically by data fitting and distributions set on the basis of reasoning are made use of. One could refer to those as “objective” and “subjective” probability assessments. 7. At this stage the assessment questions are addressed on the basis of the model results and the assumptions behind and a second exchange with the decision-makers is carried out. With the layout of the COSIMA model as a transparent tool box two principal possibilities are available now. The study may simply end here if the decision-makers are confident about the model outcome, or the decision-makers may want to feed back in the process and re- address some of the previous model settings to shed light on some issues. It should be noted that COSIMA is more or less tailored dependent on the concrete application. As should appear from the overview of methodological steps above, features for applying both scenarios and risk examinations are available. When incorporating COSIMA in the MERGE model software, the way the other three model categories are set for the actual planning problem will influence the possibilities for the assessment analyses to be carried out in the COSIMA module. Other models The following is a listing of other models of interest for the MERGE development. Table 1. Other models with present or future possibility of use in connection with MERGE. Name Description Reference Transport Models HH/KR traffic The Copenhagen/Ringsted (K/R) model. The model has been developed at the national model Danish railway agency in cooperation with the Centre for Traffic and Transport – Technical University of Denmark (CTT-DTU). COMPAS model The COMPAS (Copenhagen Model for Passenger Activity The model are to be developed at The Danish Scheduling) model is planned as an activity-based model Transport Research Institute (DTF). for modelling of individual transport demand based on individual choices of activities in various locations. OK traffic model The OK (Oslo - København) traffic model Principles have been worked out by TetraPlan and CTT-DTU. TSM The TSM (Traffic Sketch Model) is a simple traffic model The model has been used at the Centre for designed only to sketch out the future traffic. Traffic and Transport – Technical University of Denmark in connection with student theses and simple (sketch) traffic modelling. Regional Economic Models ASTRID New regional economic model The model is under development at the Institute of Local Government Studies (AKF). POINTER The POINTER (Potential Interaction) model describes the The model has been developed at the Centre potential for interaction, based on the transport system and for Traffic and Transport – Technical the spatial distribution of the population. University of Denmark and has been used in several theses and EU projects. JOINTER The JOINTER (Joining of Territories) takes its offspring in The model has been developed at the Centre the POINTER model but describes the potential for for Traffic and Transport – Technical economic interaction. University of Denmark in cooperation with the Transport Studies Unit – University of Oxford. Impact Models COPE The COPE (Corridor Planning and Evaluation) model is a The model has been developed at the Centre GIS based model for planning and evaluation transport for Traffic and Transport – Technical corridors. University of Denmark and has been used in several theses. SEAM The SEAM (Scenario-based appraisal methodology) The model has been developed at the Centre model can make scenario based project evaluation. for Traffic and Transport – Technical University of Denmark and has been used in several theses. Assessment models VD-model The Vejdirektorat (Danish Road Directorate) has a well- Danish Road Directorate proven assessment framework for CBA of new road infrastructure investments. EUNET As part of the EUNET project a comprehensive framework The EUNET project had among others as its for innovative cost-benefit/multicriteria decision analysis main objectives to develop an innovative cost- methodology has been developed. benefit/multicriteria decision analysis methodology and to develop a methodology for measurement and valuation of socio- economic development effects. PlanTIS The PlanTIS (Transportation Information System) model is The PlanTIS model has been developed by a GIS-based project database and presentation tool. It PLANCO Consult as a part of the TEN-Invest forms a strong base for implementing assessment tools. project for the European union. Road Pricing Scenario The description for the TRIP project “Integrated Traffic, Regional Economic and Impact Models” states the following [TRIP (2000)]: A choice will be made of a suitable empirical example for application of the modelling framework. A relevant case is introduction of road pricing in Denmark. This change in paradigm from single car and gasoline taxes to advanced user charging clearly influences emissions, the regional economy, the traffic system, accessibility and their interrelations. The model complex will permit evaluation of the environmental and economic consequences of the policy instrument.” In Denmark different schemes for road pricing has been dealt with in detail in the FORTRIN research programme [Kildebogaard, J. & Ildensborg-Hansen, J. (2001)]. The main focus of the TRIP Road Pricing Scenario (TRIP-RPS) is, however, to demonstrate the framework and not to make a comprehensive analysis of road pricing scenarios in Denmark. Due to that the case study will be based on one of the scenarios from the FORTRIN project but it will in no other way be related to the FORTRIN project. One of the main points in using GIS as the base for MERGE is the possibility to include the spatial dimension within the modelling framework. This indicates, that the TRIP-RPS should differ spatially within Denmark. There are many other - spatial and non-spatial - parameters that would be interesting to change in a road pricing scenario, but the TRIP-RPS aims at being relative simple in order to focus on the advantage and weakness of the model framework and not necessarily on the results. This means that the following limitations have been included in the TRIP-RPS. – There are no redistribution effects. This means that possible revenue from RP will not be used to e.g. subsidy public transport. – The RPS should only be link based – meaning that it will only influence the network impedance (cost) and not (in short term) change the car ownership (as modelled in the LTM- FREDERIK model). – It would be an advantage if the exchange of data to the LTM could be tabular (as with the LINE model). One of the RP schemes in the FORTRIN project is the “Goal Orientation Management Scenario”. In this scheme the tariff structure is drawn up in order to support two independent objectives [Kildebogaard, J. & Ildensborg-Hansen, J. (2001)]: The total CO2-emission from the transport sector should be reduced, and Traffic in the urban and residential areas should be reduced. The road users should be motivated to follow the intentions of the local traffic plans, ie. use the trunk roads as far as possible. The FORTRIN “Goal Orientation Management Scenario” has, however, a quite elaborated differentiation between: 3 vehicle types, 2 time periods, 3 road types and 3 area types. This gives a total combination of 54 different road prices. It is, however, for the TRIP-RPS, impractical to differentiate between so many vehicle types, time periods, road types and area types. Instead the TRIP-RPS will focus on 2 vehicle types: Car and Truck and only differentiate between two types of areas: Urban and Rural. This gives a TRIP-RPS tariff structure with 4 different road prices as shown in Table 2. Table 2: The tariff structure for the TRIP-RPS. Road Price Urban area Rural area Car X X Truck X X The TRIP-RPC will be calculated for one full “circle” in the MERGE model structure as illustrated in Figure 4. Figure 4: The dataflow in one circle of the TRIP-PRS. The sequence in the TRIP-RPS for one full circle can be illustrated as in Figure 4: 1. Network, Traffic Analyse Zones (TAZ) and link traffic are sent from the LTM to MERGE. 2. MERGE sends the link traffic to TicMap for impact calculation. 3. TicMap returns the calculated environmental impacts to MERGE. 4. MERGE calculates the resistance matrices based on the TRIP-RPS and the LTM network and sends the matrices to LINE. 5. LINE returns the modelled regional economic development to MERGE. MERGE transforms the results from municipalities to parishes. 6. MERGE sends the socio-economic data from LINE and the new transport prices from the TRIP-RPS to LTM. 7. LTM calculates new link traffic based on the output from LINE and the TRIP-RPS and sends the result to MERGE. 8. MERGE sends the new link traffic to TicMap for impact calculation. 9. TicMap returns the environmental impacts based on the TRIP-RPS to MERGE and MERGES calculates the changes in environmental impact for the TRIP-RPS. 10. MERGE sends all results to COSIMA for assessment. Conclusion and research perspective The paper has presented the MERGE modelling structure with emphasis on integration and modularity. Major results to be obtained with the project can be listed as follows: Development of a system of data exchange between transport models, regional economic models and impact models, specifically designed to facilitate environmental impact evaluation. Extension and improvement of existing models through increasing the quality of data relating to exogenous variables. Development of a research framework for improvement of the existing models based on the experience of the project. Assessment of the predictive power of transport models and regional economic models and of the enhanced applicability of the impact and assessment models supported by case calculations. Currently, the different modules are tested on case calculations. In the remaining part of the TRIP project the full MERGE model will be tested and the different findings reported. Even as this should be considered as only a preliminary conclusion the results obtained so far indicate that the MERGE software will make it possible to undertake a range of interesting examinations within practical transport planning that have so far not been dealt with due to shortcomings of existing software tools and not the least their practical interlinking. Literature Clausen, F., Wätjen, W. & Leleur, S. (1991): ”The Danish Highway Priority Model, EURET, Concerted Action 1.1. Kildebogaard, J. & Ildensborg-Hansen, J. (2001): ”FORTRIN programmet: Et variabel kørselsafgiftssystem – tekstscenarier og økonomi”, Center for Trafik og Transport, Danmarks Tekniske Universitet. Kronbak, J. (1998). Trafikplanlægning og GIS-baserede konsekvensberegninger, Ph.D. dissertation, Department of Planning, Technical University of Denmark. Kronbak, J. (2002): “Data exchange between transport models, regional economic models and impact models, specifically designed to facilitate environmental impact evaluation”, Paper at “Trafikdage in Aalborg”. Leleur, S. (2000): “Road Infrastructure Planning – A Decision-Oriented Approach”, Polyteknisk Forlag, ISBN 87-502-0824-1. Leleur, S. (2001). Transport Infrastructure Planning: Modelling of Socio-Economic Feasibility Risks, Proceedings of the EUROSIM Congress, Delft, 2001. Madsen, B.; Jensen-Butler, C. & Dam, P.U. (2001): ” The LINE-Model” AKF Forlaget, December 2001 (http://www.akf.dk/eng2001/line_model.htm). TRIP (2000): “TRIP 2.3: Integrated Traffic, Regional Economic and Impact Models”, Project description, TRIP 2000. Vejdirektoratet (2002): ”Trafikøkonomiske enhedspriser 2000”, The Danish Road Directorate, Ministry of Transport, Denmark 2002. Wass-Nielsen, M. & Hviid Steen, C. (2001): ”Development of a MapInfo tool for mapping and analysis of traffic impacts – in Danish (Udvikling af et værktøj i MapInfo til kortlægning og analyse af trafikale effekter). Centre for Traffic & Transport, Technical University of Denmark.
"Paper CTT Danmarks Tekniske Universitet"