Traems: the Transport Planning Add-on Environmental Modelling System

This paper describes the advanced development, at Griffith University, of a GIS-based system that combines outputs from transport scenario testing methods with land use information for modelling the environmental impacts from road traffic. This system, known as “the TRansport Add-on Environmental Modelling System (TRAEMS)” was developed using MapInfo GIS and is intended for use by transport planners as an add-on module to existing transport planning models. The model incorporates environmental factors such as noise, air pollution, energy consumption and pollution in stormwater run-off in a system that will fit almost seamlessly onto the current range of transport models in use. By incorporating these factors into a userfriendly computer package, the transport planning process and environmental implications can be considered simultaneously, creating the potential for a more efficient and environmentally sensitive planning process. The outputs from the system provide planners with immediate information on the environmental effects of any transport proposal being considered and thereby aid in the selection of a preferred alternative. This paper reports progress in the development of the noise, air pollution and water quality modules of this system and demonstrates the maximum utility of this form of modelling in the testing of transport planning scenarios. Proceedings of the 19 ARRB Transport Research Conference, Section C, pp114-138 INTRODUCTION Transportation plays a vital role in contributing to economic development and in enhancing the quality of life. In particular, road investments are one of the most significant determinants of urban growth and form of settlements. However, the use of the road infrastructure by vehicular traffic causes adverse environmental changes which are detrimental to the community as a whole. Environmental damage is seen in the area of atmospheric air pollution, pollution of the natural drainage system, noise generation and disturbance to human settlements, as well as safety and accessibility problems. There is increased community concern with these matters to the extent that, environmental consequences of transport are becoming a contributory factor in the shaping of urban transport policies. The last thirty years has seen the development of transport planning processes (Travel Demand Modelling or Travel Forecasting) to predict demand for travel. This transport planning process aims to provide information for the design of transport networks that allow an optimum and efficient movement of traffic. It defines future road and public transport network scenarios in terms of traffic flow efficiencies (usually as volume/capacity ratios), but generally stops short of considering the environmental impacts of such scenarios – at least at the time the transport scenarios are being modelled. The current research is based on the premise that the evaluation of environmental impacts of transport scenarios should be undertaken at the transport network modelling and testing stage. This research has demonstrated that this is possible because of the data overlap between travel forecasting and environmental modelling. Figure1 illustrates the relationship between the data required for environmental modelling and the data used or output from the travel demand modelling (TDM). The concept of evaluating environmental impacts of transport during the planning stage, using data available from the travel demand modelling process, underlies this research program being conducted by Griffith University. The genesis of the research was early work by Brown and Patterson (1990) who combined noise prediction models and traffic planning to predict noise impacts from road traffic throughout Logan City. Building on this work a theoretical shell has been developed in which modules have been incorporated for other environmental impacts of transport such as air pollution, energy and greenhouse impacts, stormwater pollution and visual impacts. By incorporating these factors into a user-friendly GIS-based computer package, the transport planning process and environmental implications can be considered simultaneously, creating a more efficient environmentally conscious transport planning process. The schematic architecture of the model adopted shown in Figure 2 provides the structure of the system being developed. This system is now known as the TRansport Add-on Environmental Modelling System (TRAEMS). The outputs provide planners with immediate information on the environmental effects of any transport proposal being considered and thereby aid in the selection of a preferred alternative. This paper reports on the progress of the development of the noise, air pollution and water quality modules of this system and demonstrates the usefulness of this form of modelling in the testing of transport planning scenarios. INTERFACE BETWEEN TRANSPORT MODELS AND TRAEMS TRAEMS is designed to provide a practical tool for estimating and evaluating the environmental impacts of road transport proposals within MapInfo GIS. For the system to serve as an add-on program to travel demand models for use by a whole range of users, its user interface should be designed in a way as to be able to fit almost seamlessly onto the current range of transport models in use. There are a range of integrated environmental and transport models being developed in Australia and elsewhere (Nielsen and Jocabsen 1996, Taylor 1996). The approach taken in the current research has been that models tend not to be used if they interfere too much with the way people (in this case transport planners) work. Much effort has been expended to ensure that TRAEMS can fit on to any TDM output and can provide an Proceedings of the 19 ARRB Transport Research Conference, Section C, pp114-138 extension to the models already being used – hence the important inclusion of the term “Addon” in the acronym for the system. To achieve this, the interface approach to the integration of GIS and models was adopted. Under this method, routines are developed which behave like modules resident within the GIS and operates via a user interface normally developed using the programming language of the GIS. It does depend on the GIS for data input, output and display capabilities. The flow of data between the GIS and the model are depicted in Figure 3. It shows the GIS receiving TDM data and output which are used to generate the various network map layers in the GIS. In addition to the TDM output data, the GIS also receives land use information, which include building setbacks, water catchment details, zonal data and demographic data. From the GIS map layers created, the GIS then generates process data as input for the model and then receives model output for display. The entire process runs transparently to the user via the user interface. The user interface was developed using the MapInfo programming language MapBasic. It operates entirely within MapInfo medium in the form of pull-down menus. The entire system is user-friendly providing a step by step guide to all operations. It is thus simple to use requiring no prior knowledge of MapInfo. A special tool is developed for creating a map display (GIS spatial layer) of the transport modelling network by importing the output data sets and the node coordinates used in the transport models. The node coordinates can be in any of the following six coordinate systems: longitude/latitude; longitude/latitude (AGD 66); longitude/latitude (AGD 84); Australian Map Grid (AGD 66); Australian Map Grid (AGD 84); Non Earth (or local) coordinate system. The output data set is assumed to be in ASCII format. Currently, the delimiter parameters allowed are comma, single space and tab, hence TDM outputs generated in any other formats are not catered for. The above tool and each of the environmental modules in TRAEMS are developed as separate modules within the system. The environmental modules operates independently of each other. They however make use of the same underlying base data comprising the network details and output from the transport demand model. During the analysis stage, the program uses the data integration capabilities of the GIS to integrate the above base data sets and the land use information required depending on the model being considered to generate the required data sets needed. THE TRAFFIC NOISE MODULE The traffic noise model The noise module was designed based on a concept of noise immission modelling first proposed by Brown and Patterson (1990). The emphasis here is on the estimation of the noise immission and not the noise emission. Noise immission is defined as the impact or exposure of the emitted traffic noise on the adjacent land use. As pointed out by the above authors, noise is only a problem if it affect noise sensitive sources. Immissions modelling is thus a preferred approach to environmental impact measurement as it provides an unambiguous measure of impact of the pollutant being measured. This, together with the use of GIS and the introduction of land use information, differentiates the approach adopted here from other noise prediction models. The noise prediction procedure used is based on the Calculation of Road Traffic Noise (CoRTN) procedures developed by the UK Department of Transport (UK DoT 1988). A review of the literature has shown that this model can accurately be used to predict traffic noise on Australian roads (Saunders et al 1983). This is supported by the several computerised noise prediction programs that have been developed using this method, for example NETNOISE (Woolley 1994, Woolley and Taylor 1998) and POLDIF (Taylor and Anderson 1988). The developed version of the above method provides options for predicting the L10, 18h or L10, 1h dB(A) scale. Proceedings of the 19 ARRB Transport Research Conference, Section C, pp114-138 Land use information Because of the focus on noise immissions, the noise module requires land use information as input. This information is not available from the transport demand models. The collection of this information for a large City, such as Brisbane, is a major task and is yet to be completed. At the moment an approach which uses heads-on digitising (that is digitising based on screen image) has been developed to capture the locational details of noise sensitive buildings. It uses aerial photograph (raster image) of the area, road network layer, and a comprehensive cadastral database containing detailed information on each property in the area. The method is however, slow and therefore suited to small size study areas only, hence the need for more systematic and efficient land use information collection methods. In this regard two line of research are currently being investigated: pattern recognition using raster images of the city, and interpolation based on town planning data. Output The output from the noise module includes noise emission and noise immission thematic maps and a State of Environment Report chart. The noise emission levels along each link is computed at a distance of 10 m from the edge of the nearside roadway, and displayed in the form of a link emission map. This map does not directly depict the existence of noise problems and their magnitude. It however, has the ability to isolate transport corridors that generates high noise emissions which may be helpful in the planning and developments of the adjacent land use by provided prior information on the noise emissions. The immission maps depicts the noise levels computed in front of the facade of all noise sensitive buildings along the roadway. Here the network links are shaded according to the maximum noise immission at the facade of any dwelling located along the link. In some cases, some dwellings along the link may experience a lower noise level than those depicted in the figure depending on the distance of the dwelling from the road. This information provides a different illustration of the noise environment of the area by clearly depicting noise problem areas and their magnitude. It provides important quantitative information on the extent and intensity of the exposure, in terms of the number of buildings experiencing unacceptable noise levels, in addition to the location of the problems (Affum and Brown 1997). An optional graphical output produced from the noise module shows the distribution of traffic noise levels at the facades of the buildings in the study area. While not initially designed for this purpose, by using a calibrated set of traffic volumes reflecting the existing traffic situation, this output chart from TRAEMS can effectively be used to provide a State of Environment Report for road traffic noise. The chart could also be used to monitor the changes, over time, in noise impacts. THE AIR POLLUTION MODULE Air emissions from road traffic Air quality is one of the key issues confronting policy makers involved in environmental matters. In urban areas, air pollutants consist of primary emissions such as: carbon monoxides (CO), hydrocarbons (HC), volatile organic compounds (VOC), oxides of nitrogen, (NOx), oxides of sulphur (SO2), and fine particulate (such as dust, soot and lead). In addition carbon dioxide (CO2) is also produced in great quantities from the combustion of fossil fuel (Hickman and Colwill 1982). Road transport is a major contributor of these air pollutants in urban areas (ATSE 1997). The main culprits are CO, HC, VOC, NOx and lead, resulting mainly from the combustion of the fossil fuel used to propel vehicle engines. As a result, increase in the vehiclekilometres travelled is accompanied by increase in the road transport related air pollutants in urban areas. The future of clean air in major cities is therefore linked to future transportation Proceedings of the 19 ARRB Transport Research Conference, Section C, pp114-138 plans for those cities. Effective air pollution management relies on both good land usetransportation integration plans and polices directed at controlling transport emissions. The air pollution emissions prediction model Several air pollution models have been developed using vehicle and traffic flow data and local emission factors (EPA 1991, USEPA 1985, Nguyen 1995, etc.). These studies have one important similarity in that the estimated emission is based on an emission rate or factor computed as pollutant mass per kilometre travelled per vehicle. In general it is assumed that the air pollution emissions of vehicles in a given traffic stream is given by: E N P i i i i n = ⋅ = ∑ 1 Eq. 1 where P = emission factor (in g/km) for vehicle of a particular type/fuel i, Ni = total volume of vehicle type/fuel i, and n = total number of vehicle types into which the vehicle fleet is classified The air emission model selected for use in TRAEMS was developed based on traffic data (vehicle kilometres travelled and speed) and emission factors obtained from the US and NSW (USEPA 1985 and Pengilly 1989). In developing the model the emissions from the vehicle fleet in Australia were classified into three main types: • Light duty petrol vehicles comprising passenger cars and light commercial vehicles (LDPV) • Heavy duty petrol vehicles comparing medium, and heavy CV and buses (HDPV) • Heavy duty diesel vehicles comparing medium, and heavy CV and buses (HDDV) The air pollution emissions from road traffic considered in TRAEMS are CO, HC and NOx. They are determined from the following equations (USEPA 1985 and Pengilly 1989). For all vehicle types CO and HC emissions are determine from equation 2. Ε=P·Exp(A + BS + CS + DS + ES + FS) Eq. 2 NOx emissions for vehicle types LDPV and HDPV are determined using equation 3. Ε=P·(A + BS + CS + DS + ES) Eq. 3 For HDDV vehicles NOx emissions are calculated using equation 4 below. Ε=P·Exp(A + BS + CS + DS + ES) Eq. 4 where Ε = estimated speed corrected emission in g/km (valid for 8-88km/h) S = average speed on the road link in miles/h P = city-cycle emission factor in g/km unique for each pollutant A, B, C, D, E are the speed correction factor coefficients which are unique for each pollutant and each vehicle type. These models were selected because they have been found to perform reasonably well under Australian conditions (RTA 1994). The form of the model was developed using US data sets, and NSW data sets have been used to developed emission factors and coefficients suitable for Australian conditions (RTA 1994). The above models also take into account the differences in speeds on the different segments of the network, the variations in vehicles types and fuel used. The disaggregation permit it to be applied at the network disaggretgate levels, normally employed in TDM. Proceedings of the 19 ARRB Transport Research Conference, Section C, pp114-138 Implementation of the air pollution emission models in TRAEMS In TRAEMS, each road link or segment is considered as an independent line source of pollution with constant traffic variables. The link-based output from the TDM (speeds, traffic volume, and vehicle fleet composition, if given) are supplemented by vehicle air emission factors and used in estimating the emission levels. The implemented components of the air pollution module to date include: • predicting the quantity of CO, HC , and NOx emissions levels from each link; and • division of the study area into a grid specified by the user, and the computation of the total emission levels in each grid. Further development of the air pollution module will deal with the dispersion of the estimated link-based emissions and hence, the determination of the final concentration levels of each pollutant, taken into consideration the topological and meteorological conditions of the area. When this is in place, TRAEMS will have the potential to assess air pollution immissions: for example, the proportion of the community experiencing critical air pollution levels. The difficulty in using this method in transport planning lies with the determination of the average emission factors (P) for each vehicle class and fuel used. As part of the Brisbane Integrated Transport Study, city-wide emissions factors for South East Queensland were developed for various vehicle types and pollutants (ARUP 1995). These values are site specific and change over time. The changes in vehicle emission factors are accommodated in the air pollution module by allowing the user to specify them using the menu form shown in Figure 4. This form is activated by selecting the “parameter values” button from the main “input data” form (shown in Figure 5) which is displayed when the “air emission module” in TRAEMS is activated. This figure is used to select the MapInfo TABLES containing the relevant data items from the TDM and is available to all the environmental modules. The output from this module are a series of maps showing the various pollutants levels emitted per unit length on each link and the total emissions level for each pollutant in each grid cell. THE WATER QUALITY MODULE Water quality pollutants from road traffic A critical review by Tomerini (1997) has shown that the same pollutants emitted by the road and road traffic are found in urban stormwater runoff and receiving water bodies. This indicates that transport activities are a contributing source to overall urban water quality pollutants. The major pollutants in water bodies that are generated by roadway activities are Polycyclic Aromatic Hydrocarbons (PAHs) and the heavy metals: lead, zinc and copper (Peterson and Batley 1992). These pollutants are a product of exhaust emissions, tyre wear, brake linings, motor parts, vehicle fuel and lubrication system losses, litter and spillage, and the wear of road surfaces. The water quality model A literature search failed to find a reliable model for prediction of quantitative water quality impacts from road transport. This is due, in part, to a lack of understanding of the principles and the complexities of the processes involved in road traffic impacts on water quality. The traditional method of considering road runoff as an indicator of the influence of roadwayrelated activities on urban stormwater quality is flawed. This is because pollutants produced from roadway-related activities are distributed across the entire catchment, not only on road surfaces, through the processes of interception, dry deposition, translocation and resuspension. Proceedings of the 19 ARRB Transport Research Conference, Section C, pp114-138 A detailed account of the stormwater pollution process is presented in Tomerini (1997) but, briefly, the process is as follows. When a pollutant is discharged from motor vehicle it may initially be emitted to the atmosphere or deposited onto impervious or pervious surfaces. The pollutants in the atmosphere may fall and accumulate onto plants and buildings via interception processes or onto impervious and pervious surfaces through the mechanism of dry deposition. These pollutants move between impervious and pervious surfaces via translocation processes, while resuspension processes also take place by the movement of pollutants from the catchment surfaces back to the atmosphere. During a rainfall event, the pollutants that have accumulated in the atmosphere during the antecedent dry period may be washed out of the atmosphere via wet deposition. Those that accumulated on surfaces throughout the catchment including the roadways are removed via runoff. Those accumulated on unstable pervious surfaces may be removed via erosion. The accumulated pollutants from all the above processes may all find their way into the receiving water bodies which is often a natural creek or stream. This means that, overall, water quality impacts from road traffic result from runoff from the entire catchment. It should also be noted that, in addition to the above processes, the pollutants may also undergo changes by chemical and biological processes. The quantity and quality of the receiving water in the creek or stream also influence the final concentration and fate of pollutants. TRAEMS has therefore proposed a model for traffic related stormwater pollution based on the traffic load intensity within each water catchment. The water quality module uses: • the total vehicle kilometres travelled (VKT) on roadways within a catchment (or subcatchment) as a surrogate measure of substance which may pollute water bodies; • the assumption that roadway emissions within a particular catchment will largely be washed off within that catchment; • the notion that even if we are unable to predict absolute levels of stormwater pollution from roadways at present, that knowledge of the relative potential for traffic related water pollution of different catchments is useful. The output is a relative pollutant load in the receiving water bodies. This variable is referred to as the Relative Potential Pollution (RPP). The output from TRAEMS is displayed as RPP (a relative variable) across all subcatchments in the study area. Improvement of the water quality module will involve the development of submodels to describe the pollutant build-up, wash-off and chemical change components of the stormwater pollution system. These could transform the model from one that estimates the RPP between subcatchments to one that predicts absolute water pollution concentrations. APPLICATION TO CASE STUDY Study area and data use This section reports on the application of TRAEMS to Brisbane City data to provide preliminary NOx emission and water quality information for the city. The data sets used in this application are network and traffic data, water catchment boundaries and air pollution emission factors. The network and traffic data used were obtained from Brisbane City Land Use Transport Study (LUTS) conducted by Veitch Lister Consulting Pty Ltd. A digital version of waterway catchment maps produced in the “Strategy Plan for the Management of Brisbane Waterways data sets” were accessed (Snowy Mountains Engineering Corporation, 1990) and air pollution emission factors were those developed as part of the Brisbane Integrated Transport Study (ARUP 1995) Proceedings of the 19 ARRB Transport Research Conference, Section C, pp114-138