Communication of the impacts of air pollution from industrial sources which may cause cancer and acute and chronic health effects is one of the greatest challenges faced by a risk manager. How the information is presented to the affected public is as important as the discussions on the health impact of the chemicals released routinely or accidentally into the environment. This paper will deal with the integration of GIS with a computer dispersion-modeling program to communicate effectively the magnitude and extent of the impact of air pollution on the community from industrial sources.

Introduction

On December 3, 1984, at the Union Carbide plant in Bhopal, India, a cloud of methyl isocyanate (MIC) gas, an extremely toxic gas, was released from their storage tanks, resulting in exposures to the community downwind from the plant with dire consequences. A significant number of the residents of Bhopal who lived downwind from the plant died and continue to do so and there are also a significant number still suffering from the aftermath of the incident.

The disastrous incident set the impetus for the development of a number of community-right-to-know regulations in this country at the federal, state and local government levels and the reassessment by industrial facilities of their operations. In addition, citizens' groups joined forces to pass legislation (e.g., Proposition 65 in California) mandating that industries disclose the hazards they pose and to communicate the risks of excess cancer and non-cancer health effects to the affected communities. The ensuing regulations required public communication of the adverse health impact and risks from routine or daily operation of the industrial facilities neighborhood by neighborhood. They were also required to anticipate worst-case (Bhopal type) scenario or a potential accidental event.

In California, the State legislature passed the Air Toxics "Hot Spots" Information and Assessment Act of 1987¹ (Assembly Bill 2588). This established a statewide program for the inventory of air toxics emissions from individual facilities as well as requirements for risk assessment and public notification of potential health risks. The state Legislature found and declared (California Health and Safety Code 44300 et.seq) that:

• Significant uncertainty exists about the amounts of potentially hazardous air pollutants which are released , the location of those releases, and the concentration to which the public is exposed; and,
• These releases may create localized concentrations of air toxics "hot spots" where emissions from specific sources may expose individuals and population groups to elevated risks of adverse health effects, including, but not limited to, cancer and contribute to the cumulative health risks of emissions from other sources in the area.

In November 1986, California voters approved an initiative to address growing concerns about exposures to toxic chemicals. That initiative became The Safe Drinking Water and Toxic Enforcement Act of 1986, better known by its original name: Proposition 65. One of the requirements of the Proposition 65 is the posting of warning in either written, printed or in a graphic manner that indicates the "affected area" from the release of the chemicals above the threshold levels identified in the Act. A typical posting for Proposition 65 is shown in Figure 1.

The Emergency Planning and Community Right to Know Act, also known as the "Bhopal" legislation was passed by Congress in 1986 and was designed to inform communities about the potential threats in their midst.

The Federal Accidental Release Prevention Program [Title 40, Code of Federal Regulations (CFR) part 68] was developed to require facilities that handle regulated substances in quantities exceeding specified thresholds, to prepare a hazard assessment and submit a risk management plan (RMP). The RMP is part of the their risk management programs to prevent the accidental release of regulated substances. One of the requirements of the hazard assessment is the identification of geographic areas most vulnerable to hazardous release events (worst-case accidental releases, potential offsite consequences of releases or other alternate release scenario) so that the potential impacts can be quantified. The RMP becomes a public document once it is submitted to the federal, state and local administering agencies and will be used for communicating the offsite consequences to the public.

A requirement in all these federal, state and local regulations is risk communication. A number of computer models and programs are available as tools to assist a risk manager in communicating risk to the public. This paper discusses the integration of computer dispersion modeling software with ArcView with the Spatial Analyst extension to overlay risk isopleths to show cumulative impacts on the community. The steps involved in the integration of GIS with computer dispersion model for risk communication is shown in Figure 2.

Modeling Program

The preferred dispersion modeling program recommended by the Environmental Protection Agency (EPA) for industrial sources is Industrial Source Complex Short Term (ISCST3)². This dispersion model is used to calculate concentration or deposition values for different averaging time periods at pre-defined receptors areas in the ambient air from the release of pollutants from a single or a multiple of sources (point, volume, area) with actual site specific pre-processed meteorological information. The dispersion model is a Gaussian model that assumes the releases from the sources occur in a steady state manner from the time of release until it reaches the receptors.

The output from the dispersion model is ground level concentration values at the desired or specified receptors for the averaging hours (1 hour, 6 hours, 24 hours, annual, etc.) specified. From this data, a concentration isopleth can be drawn for sources that have the same concentrations. Most of the software packages available either in the public domain or from the private sectors create a plot file for concentrations. However, to calculate the risk from the pollutants emitted from the source, the concentration values would have to be multiplied with the potency factors for the pollutants emitted and the risk adjusted if the pollutants have multiple routes of exposures.

The ISC3 model is a steady-state Gaussian plume model which can be used to assess pollutant concentrations from a wide variety of sources associated with an industrial source complex. The model is recommended for:

• Rural or urban areas
• Flat or rolling terrain
• Transport distances of less than 50 kilometers
• 1-hour to annual averaging times; and,
• Continuous toxic air emissions.

The input requirements for the source data are the following:

• Location.
• Emission rate
• Physical stack height
• Stack gas exit velocity
• Stack inside diameter
• Stack gas temperature

Optional inputs include:

• Source elevation
• Building dimensions
• Particle size distribution with corresponding settling velocities; and
• Surface reflection coefficients

Meteorological data.

ISCS3 requires pre-processed hourly weather data that includes stability class, wind direction, wind speed, temperature, and mixing height.

The output file gives the concentration or deposition values for each identified receptor for the specified averaging hours within the specified meteorological period. The output for plot files and partial contribution files can also be specified in the program. The partial contribution file (pcon.bin) contains the relative contribution for the concentration for the averaging time from each source based on an emission rate of 1 gm/sec.

Health Risk Assessment

The health risk assessment is a detailed comprehensive analysis prepared to evaluate and predict the dispersion of hazardous substances in the environment and the potential for exposure to human populations and to assess and quantify both the individual and population-wide health risks associated with those levels of exposure (H&S 44306). The risk assessment shall consider the potency, toxicity, quantity, and volume of hazardous materials released from the facility, the proximity of the facility to potential receptors, including, but not limited to, hospitals, schools, day care centers, worksites, and residences. Figure 3 shows a Summary Health Risk Assessment Form that shows the maximum risk, the pollutants that contribute to the risk, and the location of the maximum risk receptor.

The California Air Pollution Control Officers Association (CAPCOA) published the Air Toxics "Hot Spots" Program Risk Assessment Guidelines³ in 1991 and was revised in 1993. A program, Assessment of Chemical Exposure for AB 2588 (ACE2588) was designed for AB2588 risk assessments, and it incorporates the algorithms and recommendations in the CAPCOA's Risk Assessment Guidelines.

The ACE2588 model interfaces with ISCST3 to predict ambient air concentrations, and to assess the exposure to and associated risks of toxic pollutants emitted from a single or multiple sources in a facility. The program also calculates risk from other exposure pathways such as dermal, soil, water, food and mother's milk.

Integrated Models

Currently the California Air Resources Board is developing an integrated program Hot Spot Analysis and Reporting Program (HARP) that will integrate the modeling program (ISCST), the health risk assessment program (HRA) and GIS. A number of consulting companies⁴ have developed modeling programs which have integrated GIS with the output of dispersion modeling.

Company Name _Carthage Anodizing Inc______________________
Facility Name _ __________________________
Facility Address _400 E. Cloaks St____________________________________
_Compton, CA 90222___________________________________
AQMD ID Numbers _A49384___

A. CANCER RISK
1. Inventory Report Basis 1989 Curent emissions
(circle one only)

2. Maximum Cancer Risk to Receptors (based on Tables III-5* and III-6* substances)
a. Max Offsite _______35.6 x 10^-6___ location:_387,548E; 3,752,639N

b. Residence _______35.6 x 10^-6___ location:_387,548E; _3,752,639N

c. Worker ________8.8 x 10^-6_ location:_387,348E; 3,752,739N

3. Substances Accounting for 90% of Cancer Risk _hexavalent chromium______
Processes Accounting for 90% of Cancer Risk __chrome anodizing__________

4. Population Exposed to Specific Risk Levels (including worker population)
a. >1x10-6 __15,820________
b. 1x10-6 to 1x10-5 __ 15,370________
c. 1x10-5 to 1x10-4 ___ 450_______
d. 1x10-4 to 1x10-3 __________
e. >1x10-3 __________

5. Cancer Burden __0.046________ (including worker population)

6. Maximum Distance to Edge of 1 x 10-6 Cancer Risk Isopleth (meters) 1500______

7. Screening Cancer Risk to Most Exposed Individual (based on Table III-7* substances)
a. Residence (without silica) ____________________
b. Residence (silica only) ____________________

B. HAZARD INDICES **
1. Highest Chronic Hazard Indices (based on Tables III-8* and III-10* substances)
Residential chronic HI: Liver__ toxicological endpoint:_0.2534
Worker chronic HI: _______ toxicological endpoint:_______________

2. Substances Accounting for 90% of Chronic Hazard Index: Cr +6, 1-1-1 TCA

Processes Accounting for 90% of Chronic Hazard Index _chrome anodizing, degreaser

3. Highest Acute Hazard Indices (based on Table III-9* substances)
Residential acute HI: _CNS_____ toxicological endpoint: _0.0028_______
Worker acute HI: _________ toxicological endpoint: ______________

4. Substances Accounting for 90% of Acute Hazard Index _111-TCA_____________
Processes Accounting for 90% of Acute Hazard Index __Degreaser__________

Methodology

The following steps were taken to create the risk isopleths to show cumulative impacts.

• Step 1 Prepare and run dispersion model, ISCST3
• Step 2 Prepare a health risk assessment program- ACE2588 model
• Step 3- Prepare output file in Excel for import into ArcView
• Step 4- Connect Excel file to ArcView with SQL Connect and ODBC
• Step 5 Grid receptor locations in ArcView
• Step 6- Create risk isopleths with Spatial Analyst
• Step 7 Project views from UTM to decimal degrees
• Step8 Overlay risk isopleths, demographic data and street data

To show the cumulative effects of two nearby sources on the same receptor locations, two modeling runs and health risk assessment were conducted. The input data for the sources were assumed to be identical, with the same source parameters and emission rates, except for their location. The model runs were based on five sources: 3 points and 2 volume sources, 5 pollutants, 112 receptors, and meteorological data for the City of Compton, California, for 365 days. The concentrations for 1-hour maximum and the annual average were generated for each receptor.

The input file was prepared in the format required to run ISCST3 by using a text editor. Programs with front-end interfaces create the input file in the required format for the model. One such program developed by EPA is EXINTER ⁵ (expert interface). Programs in the public domain all have user friendly front-end interfaces which prepare the input file to run the model.

To run the model, the source data, source location, emission data, receptor data, site specific meteorological data from the closest air monitoring station, receptor data, and emission data are needed. Figure 4 shows the various inputs needed for the model run.

The input file that was generated to run ISCST3 for one of the sources is as shown in Table 1. A partial listing of the output file from ISCST3 is as shown in Table 2. The output file generates the partial contribution file (pcon.bin) for further processing in ACE 2588. Table 2 shows the some receptors points with their concentration levels. The ACE2588 reads the pcon.bin file to calculate the cancer risk, cancer burden and non-cancer health effects.

To run ACE2588, the input file as shown in Table 3 was prepared in the required format. This input file identifies the number of receptors, number of sources, hours of model runs and number of days. In addition, the pollutant IDs' are identified as well as the actual emission rates and the receptor locations where the risk values are to be calculated.

The output from the modeling program is either a concentration or a deposition rate based on the emission rate entered. The concentration values have to be multiplied with the unit risk factor, or potency factor to obtain the cancer and non-cancer risk values and also have to be adjusted for different routes of exposures. For a single pollutant, this can be done easily in a spreadsheet program. However, the complexity adds up when there are multiple pollutants with different emission rates with different routes of exposures. Programs such as ACE 2588, the Health Risk Assessment (HRA) program or the Hot Spot Analysis and Reporting Program (HARP) available from the California Air Resources Board make the task much simpler.

The ACE2588 program is used to calculate the cancer risk from both inhalation and non-inhalation pathways, cancer burden, health hazard index for acute and chronic effects for multiple pollutants with different emission rates at each receptor location and to take into account multiple exposure routes. Samples of input and output files from ACE2588 are shown in Table 3 and 4. The data from Tables 4 and 5 are combined later in Excel and connected to ArcView to grid the receptor areas and plot the risk contours.

ACE2588 RUN FOR Carthage Anodizing Inc, Compton(7/28/99)

(This procedure was obtained from Esri's Robert Burke⁶ and their technical support team).

To extract the output data and to create a table for creating the contours is a challenge because the output includes text description and data. However data in Excel can be assessed directly using a Microsoft Excel ODBC driver and ArcView's ODBC connectivity. The data in the MS Excel file must be first divided into tables by assigning a name to a block of cells which are to represent a table. In MS Excel Version 7.0, this was done in the following manner:

1. Highlight the block of cells representing a single table in ArcView.
2. In MS Excel, click on the menu options Insert/Name/Define. This will bring up a menu where the name of the block can be defined.
3. Type in the name to be assigned to the highlighted cells. This is the name that will appear in the tables list on the SQL connect window in ArcView. Several tables may be defined in a single worksheet. A drop down list of the tables defined in a single worksheet is provided above the upper left corner of the worksheet.
4. Using the ODBC Administrator in the control panel, configure the Excel ODBC driver to access the MS Excel workbook file that contains the defined tables. Connect.
5. In ArcView, go to menu options Project/SQL Connect. Select Excel file from the Connection drop down options, and Connect. Once the connection is established, select the table name from the Table Column. From the Columns section select the desired columns. Once the selection is completed, press the Query button to extract the table and columns identified. Provide a different name or select the default name. When the query is completed, the table is available in ArcView.

The tables which were created in steps 3&4 include the receptor information with the X and Y coordinates and the risk information are used to create the isopleths. Two isopleths are generated, one for the initial source and the other for another identical source receptor located nearby but affecting the same receptors. The tables containing the receptor and risk data is added as an Event Theme Table where the X and Y field are the X and Y coordinates of the receptors. This table is added to the View and converted to a shape file for processing in Spatial Analyst. Spatial Analyst, an extension for ArcView, was used to create the risk contours. In Spatial Analyst, the contours were created using the options Surface/Create Contours, accepting the screen defaults and choosing the IDW method for creating the contours. The X, Y, and Z columns were assigned to the X-coordinates , Y coordinates for the receptors and risk, respectively. In addition a script, logcon.apr, courtesy Robert Burke from Esri, provided the option to specify the contour intervals. In this case, risk intervals of 10 were chosen. This helped reduced the number of risk contours and focus the impact from the higher risk levels.

Since the data for the receptors were in Universe Transverse Mercator (UTM) coordinates, the view had to be projected to decimal degrees to enable street and other demographic data to be displayed. The projector script, prjctr.avx, a sample script with ArcView was used to project the views from UTM to decimal degrees.

The results of the risk isopleths are as shown in figures 5,6, and 7. To communicate the risk impacts, various demographic layers were added to the view to demonstrate the impacts on the community.

Figure 7

Summary and Conclusion

The questions asked most frequently in risk communication meetings are from individuals wanting to know what their own family is exposed to and if they are safe living in the risk isopleths displayed and what the regulatory agency or the industry is doing to address the problems. Questions such as these are to dealt with delicately and are also extremely difficult to answer. An inappropriate response could lead a complete breakdown of the communication and trust. The communication fails if the risk communicator resorts to technical jargon and downplays the risk of exposure or compares it to involuntary risk probabilities. The use of GIS is a tool among many others that can be utilized by the risk communicator to address public concerns and fears.

With GIS, the street layer and demographic information can be layered with the risk isopleths, to show the cumulative impacts, if any, on the community, and also the affected highest risk areas as shown in Figure 8. In comparison, quite often what is available to the risk manager is the static map as shown in Figure 1. With a GIS system, the risk manager can effectively answer the community questions and concerns about their concerns in an interactive manner. This paper hopefully demonstrates that GIS can be used to integrate dispersion models and can be a useful tool for risk communication.

References

1. California's Health and Safety Code, Division 26, Section 44300 et.seq.)
2. Guidelines of Air Quality Models, USEPA, 40 CFR Part 51, Appendix W.
3. User's Guide to the Assessment of Chemical Exposure for AB 2588 (ACE2588) Computer Model, March 1991.
4. Environmental Manager, Air and Waste Management Association's Magazine for Environmental Managers, 1999 Environmental Computer Software Buyers Guide, Page 77, July 1999.
5. EPA's EXINTER Program for Dispersion modeling, available at http://www.epa.gov/ttn/scram
6. Robert Burke, Esri, rburke@Esri.com, E-mail communications.

Mohan Balagopalan
Senior Engineer
South Coast Air Quality Management District
21865 E.Copley Drive, Diamond Bar, CA 91765
909-396-27049(phone)
909-396-2999 (fax)

mbalagopalan@aqmd.gov