Home Articles Air pollution modelling for Chennai city using GIS as a tool

Air pollution modelling for Chennai city using GIS as a tool

Air pollution modelling for Chennai city using GIS as a tool

H. M. G. Charlot
Teaching Research Associate CES
[email protected]

M. Kanagaraj
Research Associate
[email protected]

R. Narayanan
Research Scholar, DUSD
Anna University, Chennai
[email protected]

Introduction
If the London fog disaster in the early fifties had not killed thousands of people over a short period, few would have bothered about acute pollution disasters. The air that man breathes is polluted by industrial and automobile emission bringing into the atmosphere Suspended particular matter (SPM), Oxides of Sulfur and nitrogen, Carbon monoxide, photochemical oxidants and Hydrocarbons. These pollutants, individually and collectively, have teratogenic, carcinogenic or mutagenic effects and can also cause respiratory ailments, the physiological barriers being ineffective against them.

In the 1980s, the polluted air in the cities could be traced to the chimneys of factories. But by the 1990s it was more than apparent that the major contributor to the haze and the poisons in the air was not factories but automobile – cars, buses, trucks, three-wheelers and two-wheelers. For over a decade now, there is no dispute over the fact that more than half the pollution load in our cities is due to automobile exhaust.

In the light of these concerns, there is clearly a need for improved information on levels of traffic-related air pollution. This information is required for a wide range of purposes: to help investigate the relationship involved as inputs to health risk assessment, to assist in establishing and monitoring air quality standards, and to help evaluate and compare transport policies and plans. For all these purposes information is required not only on temporal trends in air pollution but also on geographic variations. GIS based vehicular pollution models are needed to identify pollution hot spots, to define at risk groups, to show changes in spatial patterns of pollution resulting from policy or other interventions and to provide improved estimates of exposure for epidemiological studies.

This paper presents a methodology to develop a GIS based vehicular pollution model using a coordinated approach, taking into consideration of all the parameters influencing vehicle emissions, unlike other conventional approaches which have used GIS as a preprocessing/post processing tool.

Geographic Information Systems
A geographic information system (GIS) is a computer-based information system that enables capture, modeling, manipulation, retrieval, analysis and presentation of geographically referenced data. The rise of GIS technology and its use in a wide range of disciplines provides transportation and air quality modelers with a powerful tool for developing new analysis capability. The organization of data by location allows data from a variety of sources to be easily combined in a uniform framework.

Another important feature of GIS is its ability to bridge the technical gap between the need of analysts and decision-makers for easy understanding of the information. The user friendliness of GIS is a feature that has made GIS one of the most used platforms for planning all over the world. The ability of GIS to answer technical questions also makes GIS an excellent tool. Literature on GIS data structures, applications, and vendor products are substantial.

Air pollution modelling for Chennai city using GIS as a tool

The following section will briefly cover the past use of GIS in transportation and air quality analysis and the issue of spatial data quality.

Earlier applications of GIS in mobile emission modeling

Emission inventories
Models have been developed to estimate hourly estimates of emissions, which utilises GIS in developing mobile source estimates for input into photochemical models. The main function of the GIS in such model was the spatial aggregation of travel demand forecasting model features into a grid. Spatially defined vehicle mixes by trip purpose, temporal factors, hourly temperatures, trip volumes, trip speeds, and modal percentages are used as inputs.

Zonal estimates were allocated to traffic analysis zone centroids that were re-allocated to grid cells. Link estimates were allocated to nodes and re-allocated to cells. The use of points to represent these features did not take full advantage of the spatial structure provided by the original input data. Traffic Analysis Zones (TAZ) falling along grid cell boundaries should have their portions divided.

This strategy would limit grid cell sizes to those significantly larger than TAZs, which can be quite large (30-40 square km) for some metropolitan areas. Also, no mention is made of strategies for identifying the confidence ranges of the estimates.

The model supports the use of GIS, but did not take full advantage of the research value of GIS. Further, the model did not have the flexibility to answer the diverse impact or mitigation questions that arise from estimating emissions.

GIS for transportation planning and air quality analysis
Researchers used GIS as a preprocessor and postprocessor to mobile emission modeling. Although they relied on existing models to estimate emissions, they showed how GIS could be valuable in the management of emission related data. They made the connection between the needs of transportation planners and decision-makers and the spatial tools and features of GIS.

Microscale analysis
Researchers at Utah State University used GIS in developing microscale analyses of a small group of intersections. They linked a GIS with CALINE3 and CAL3QHC to predict pollutant concentration levels. The value of GIS (outside of spatial data storage and data visualization) was its ability to compare concentration results to other non-related data. The contribution is significant to this research because it provides a foundation for the argument that a GIS approach is not restricted to developing emission inventories, but can be easily expanded to a number of other related issues.

Influencing decision-makers
Othofer developed an interesting approach to predicting location specific emission production estimates for changing control strategies. Instead of developing estimates using detailed location-specific emission producing activities and emission rates, they disaggregated large zonal estimates using emission-producing activities. The advantage of this approach is its simplicity and its straightforward recognition that the data needed to predict emissions at smaller levels does not exist or the relationships are undefined. The disadvantage is that the ability to predict changes among the disaggregated levels is a function only of the change of the overall larger units. Thus, the true effects of activity changes on emissions cannot be measured. The project produced high-quality graphics that indicated locational variation in emission-producing activities. The project was successful because elected officials could ‘see’ areas that have potentially high emissions and therefore had evidence for developing actions for those specific areas. Although, the modeling capability of the project is limited, its ability to influence action through spatial communication is a noteworthy contribution to the use of GIS in this arena.

Lacunae in air pollution modelling for Chennai city
Chennai City is the fourth largest metropolis in India. The Chennai metropolitan area covers an extent of 1172 Sq.km of which the corporation area, which is identified as the city extends over 172 Sq.km. As per 2001 census the population of Chennai City is 42.16 lakhs. The vehicle population during 1999-2000 is around 11.15 lakhs.

The ambient air quality of Chennai has deteriorated with an increase in the number of vehicles and industrial pollution. A recent study by the State Pollution Control Board (PCB) found that the levels of suspended particular matter (SPM) ranged from 274 to a mind-boggling 1,470 micrograms/cubic meter (mg/m3) at several areas, which was much higher than the WHO prescribed limit of 200 mg/m3. The level of carbon monoxide ranged from 12 to 70 parts per million (ppm) as against the permitted 35-ppm. The study also showed that emission from nearly 50 percent of the vehicles in the city exceeded the permitted levels and the pollution load in the atmosphere increased by 3.5 percent annually.

The entire city has got only 6 ambient air quality monitoring stations. With this limited number of stations, to represent the air quality in chennai city spatially is a difficult task. Hence a co-ordinated methodology to map the air quality in Chennai City using GIS is explained in the following paragraphs.

Air pollution modelling for Chennai city using GIS as a tool

Model design parameters
The following parameters have been chosen for the mobile emission model.

The parameters are:

  • Develop estimates of the production of
    automobile exhaust pollutants in space and time
    A more accurate, verifiable, estimate of the pollutants may prove more useful in predicting the impact of motor vehicles.
  • Comprehensive representation of vehicle technologies

Differences in vehicle technologies / characteristics have been shown to significantly affect vehicle emission rates. The list of desired vehicle characteristics are model year, engine size, weight (or mass), emission control type(s), fuel delivery type, transmission type, cross-sectional area, and number of cylinders.

  • Separate and quantify high-emitting vehicle emissions

    A small percentage of the fleet disproportionally contributes to total mobile source emissions. By separating this small high-emitting portion of the operating fleet, it will be easier to predict the impacts of control strategies that may target high emitters.

  • Separate start, hot-stabilized, and enrichment emission quantities and locations

    By separating estimates into specific emission modes, mode-specific impact strategies can be more efficiently evaluated. Further, emission rates for each mode are predicted using different variables. Engine starts are primarily influenced by vehicle characteristics and engine temperature. Hot stabilized and enrichment emissions are primarily influenced by vehicle characteristics and operating condition.

  • Include Speed related factors

    The relation between speed and emission levels has been well established various.

  • Include emission rates from the statistical approach

    Emission rates from the statistical approach need to be included because the research indicates that modal parameters better characterize accurate emission rate estimation. Because the modal emission rates models are available, they can be immediately integrated into the model framework. The approach also produces separate start and running exhaust emission estimates, addressing one of the previously defined model design parameters.

  • Include activity measures from travel demand forecasting models

    Travel demand forecasting models are the primary predictive tools for regional level vehicle activity. Despite their well-documented problems, they have characteristics that make them very attractive for a spatially-resolved model. First of all, they have a defined structure and connectivity that translates into a spatial form (zones, links, and nodes). Second, they develop estimates using socioeconomic information, allowing the model to be indirectly affected by social and economic changes.

  • Use of Geographic Information Systems

Using GIS is important because it is designed to handle the spatial data management and modeling functions key to the research goals. Without GIS, complex spatial analysis and manipulation algorithms would have to be re-created. Its widespread use and popularity among planning agencies is significant enough to warrant its use.

Model approach
The conceptual design of the proposed research model

The following sections describe the five major tiers of the model design.

Spatial environment
The objective of the spatial environment tier is to unify input data under a common zonal and lineal structure. The size and scope of the zones and lines depend on the users and their specific needs.

Zonal data
The zonal module defines the polygon structure used to represent data and emission estimates for engine starts. It is the role of the zonal module to combine the polygons of various input data (i.e. socioeconomic, land use, TAZ) into a single polygon dataset.

Lineal data
The road module defines the lineal data used for predicting running exhaust emissions.

Conflation
Conflation is the blending of two line databases. Conflating the abstract travel demand forecasting network and a spatially accurate comprehensive road database is needed to improve the spatial accuracy of the travel model results.

Air pollution modelling for Chennai city using GIS as a tool

Fleet characteristics
An improved capability to identify the emission significant components of the operating fleet is important to emission rate accuracy. Spatially variant emission estimates are needed, requiring spatially resolved sub-fleet characterization. Therefore, there is a need for identifying procedures that can accurately predict spatially resolved vehicle characteristics for urban areas.

Vehicle characteristics
The first vehicle characteristic module has two major tasks: determine individual vehicle location parameters and emission-specific characteristics.

Vehicle geocoding
Address Geocoding is a process whereby standard address fields of road name, road type, and ZIP code are used to identify corresponding lines in a road database.

Decoding
Raw registration data can usually provide a few important vehicle characteristics (VIN, make, model, model year, and number of cylinders), but more information can be developed from the vehicle identification number (VIN). These files should represent a comprehensive description of the region’s fleet characteristics. These files can be further processed to develop the emission-rate specific fleet distributions.

High emitting vehicles
A high emitting vehicle is one that has malfunctioning or tampered with emission control systems causing higher than normal emissions. It is expected that a small percentage of high emitting vehicles account for a large percentage of total emissions. High emitter determination is an important model design parameter and therefore it is appropriate to characterize these vehicles differently.

Technology grouping

Once vehicles are identified as high or normal emitters, they are characterized into technology groups. Technology grouping is the process of combining vehicles according to the emission standards of the make.

Vehicle activity
The emission-important vehicle activity estimates provided by the regional travel models are: the number and location of peak hour (or daily) trip origins, road segment volumes, and road segment average speeds. Temporal travel behavior and modal (idle, cruise, acceleration, and deceleration) operations. This

  • Engine Start Activity
  • Intra-zonal Running Exhaust Activity
  • Modal Activity

Road grade
The impacts of road grade on emissions are included in the model design. Road grade affects vehicle emissions by impacting the load on the engine. Gravity exerts a force on a vehicle that must be counteracted to maintain a constant speed. Road grade is not included in mandated emission models because tests on the actual effects have not been completed and because metropolitan areas do not maintain spatially defined road grade estimates. Including vehicle activity impacts resulting from road grade provides an important step in emission model development.

Conceptual design of the proposed research model


Conceptual design of the proposed research model

Air pollution modelling for Chennai city using GIS as a tool

Temporal variability
The temporal variability also plays a major role in. Although average speed cannot be predicted to determine LOS F during off-peak hours, volume-to-capacity ratios provide sufficient information for selection of appropriate speed and acceleration profiles.

Facility emissions
Facilities are divided into zones and lines corresponding to the previously mentioned emission modes of engine starts and running exhaust (respectively). Facility estimates are used to allocate emission production to those vector spatial data structures currently used by transportation planners. By tying emission production estimates to facilities, tasks regarding research, reporting, validation, or control strategy development are made easier.

Engine start zonal facility estimates
Zonal facilities include the zonal representations of land use, and Census blocks. The model design allows for other zonal designations to be included, but only the three mentioned have been required. The zones have been included in the definition of facilities because planners to aggregate socioeconomic information use them. While running exhaust emissions occur within zones, they are better tied to modal activity that occurs on the road. Engine starts, however, occur at trip origins, generally characterized with point or zonal information.

Management of data dictionary
The directory structure may include the following data

Directories

  • zone: stores all zonal data and coverages
  • road: stores all lineal data and coverages
  • grid: stores all vector grid data
  • emest: stores all emission estimate data
  • tech: stores all technology group data
  • grade: stores all road grade related data
  • raster: stores all raster data
  • raw: place to store backup copies of data and programs
  • aml: stores all AML code
  • code: stores all C code
  • templates: stores a number of INFO file templates
  • spac: stores all speed / acceleration profiles
  • modmat: stores all modal matrices
  • lookup: stores all ASCII lookup files
  • temp: stores temporary files used during
    program runs

Contribution of Geographic Information Systems in the proposed Mobile emission Model
Geographic information systems can be advantageously used in the proposed model for the following activities

Spatial data organization
Data in the model can be organized based on their spatial character. Structuring the multiple layers of data in this manner provides data connectivity that would be difficult without GIS and topology.

Framework of the GIS based Mobile Emission Model


Framework of the GIS based Mobile Emission Model

Air pollution modelling for Chennai city using GIS as a tool

Spatial data joining
Data sets of different characteristics can be merged together to form a single entity. Specifically, GIS allows improved spatial resolution of the travel demand forecasting model network by conflating to a spatially accurate road database. This capability of GIS also allows linkages to occur between the various area sources of information (land use, Population, etc.).

Spatial query
GIS provides the ability to search data by locational parameters. Specifically, the technique can be used to predict the on-road fleet distribution required for the identification of the fleet registered within a certain distance from the individual road segments.

Spatial aggregation
GIS provides the ability to aggregate irregular polygon data and line data into regular user-defined grid cells. This capability makes GIS vital for efficiently developing mobile emission inventories, regardless of the modeling approach used.

Spatial data visualization
The map-making and graphic display capabilities found in most GISs are extremely useful in communicating model results to individuals from various technical backgrounds. Given the importance of mobile emissions in determining transportation improvements, this feature has significant value.

Conclusion
This paper has made an attempt towards an integrated mobile emission modeling. Further work in this direction is on the way and further improvements and changes in the proposed methodology will be incorporated. There is always a scope for further refinement of the methodology with appropriate changes and additions.

References

  • [Bachman 1998] Bachman, William, Jessica Granell, Randall Guensler, and John
  • Leonard; “Research Needs In Determining Spatially-Resolved Sub-fleet Characteristics”; Transportation Research Record; Transportation Research Board, Washington, DC, 1998.
  • [Bachman 1996] Bachman, William, Wayne A. Sarasua, and Randall Guensler; “A
  • GIS Framework for Mobile Source Emissions Modeling”; Transportation Research
  • Record No. 1551. National Academy Press; January, 1996.
  • [Benson 1989] Benson, Paul; “CALINE4 – A Dispersion Model for Predicting
  • Pollutant Concentrations Near Roadways” (FHWA/CA/TL-84/15); State of
  • California Department of Transportation, Division of New Technology and
  • Research; Sacramento, CA; November, 1984, Revised June, 1989.
  • [Calspan 1973] Calspan Corporation; “Automobile Exhaust Emission Surveillance”;
  • Environmental Protection Agency #APTD-1544; NTIS PB220-775; Office of
  • Mobile Source Air Pollution Control; Ann Arbor, MI; May, 1973.
  • [CARB 1990] California Air Resources Board; Emission Inventory 1987; Technical
  • Support Division; Sacramento, CA; March, 1990.
  • [Carlock 1993] Carlock, Mark; “An Analysis of High Emitting Vehicles in the On-
  • Road Motor Vehicle Fleet”; The Emission Inventory: Perception and Reality,
  • Proceedings of an International Specialty Conference; Air and Waste Management
  • Association; Pittsburgh, PA; October, 1993; pp 207-228.
  • [Guensler 1993a] Guensler, Randall; “Data Needs for Evolving Motor Vehicle
  • Emission Modeling Approaches”; Transportation Planning and Air Quality II; Paul
  • Benson, Ed.; American Society of Civil Engineers; New York, NY; 1993.
  • [Guensler 1994] Guensler, Randall; “Vehicle Emission Rates and Average Vehicle
  • Operating Speeds”; Institute of Transportation Studies, University of California,
  • Davis, CA; 1994.
  • [Hallmark 1996] Hallmark, Shauna and Wende O’Neill; “Integrating GIS-T and Air
  • Quality Models for Microscale Analysis”; Transportation Research Record No.
  • 1551; National Academy Press; 1996.
  • [Harvey 1991] Harvey, G. and E. Deakin; “Toward Improved Regional
  • Transportation Modeling Practice”; National Association of Regional Councils;
  • Washington DC; December, 1991.
  • [Heywood 1988] Heywood, John B.; Internal Combustion Engine Fundamentals;
  • McGraw-Hill Publishing Company; New York, NY; 1988.
  • [Kunselman 1974] Kunselman, P., H.T McAdams, C.J. Domke, and M.E. Williams;
  • “Automobile Exhaust Emission Modal Analysis Model”; Calspan Corporation;
  • Buffalo, NY; Environmental Protection Agency (EPA-460/3-74-005); Office of
  • Mobile Source Air Pollution Control; Ann
    Arbor, MI; January, 1974.