Air Toxics Emissions Inventory Protocol for the Great Lakes States

3.0 Source Identification and Categorization

This section discusses the target compounds and associated facility and area emission sources that should be included in the regional air toxics inventory. Procedures for adding or deleting compounds from the target compounds list are also outlined. Literature references to use in determining the number and location of facilities in a given state are also provided. Standard Industrial Classification (SIC) codes, Source Classification Codes (SCC,) and Area and Mobile Source (AMS) codes, which link sources, devices, and processes to emission factors for the target compounds, are presented in this section.

3.1 Compounds of Interest

The following discussion presents the list of target compounds and the procedure for revising this list.

3.1.1 Identification of Target Compounds

The list of target compounds is now considered the complete list of compounds of interest for use in the air toxics emissions inventory. All compounds listed in this table are to be included in each state's air toxics inventory and uploaded to the regional database of emissions data and estimates. It also specifies if a given compound is included in the following lists of toxic chemicals:

• U. S. Environmental Protection Agency's (U.S. EPA) Great Waters Program;
• The list of compounds initially targeted by the Great Lakes Commission (GLC); and
• Title III of the 1990 Clean Air Act Amendments (CAAA).

The table also indicates the Chemical Abstract Service (CAS) number of each compound, when available, and whether the compound is considered a carcinogen by the U.S. EPA.

3.1.2 Procedure to Revise List of Target Compounds

Using the process described in Section 2.2, the states developed the following procedure to add or delete compounds from the list. Compounds eligible to be added will come from the following information sources:

• The Great Waters List of toxic air pollutants under Section 112(m) of the 1990 CAAA;
• The Hazardous Air Pollutants (HAPs) List under Section 112(b) and 112(c)(6) of the 1990 CAAA;
• The Urban Air Toxics List developed under Section 112(k) of the 1990 CAAA;
• The list of toxic air pollutants in the Great Lakes National Program Office (GLNPO) Great Lakes Water Quality Initiative;
• Any Great Lakes State toxic air pollutant list established as part of a Great Lakes State regulatory proceeding;
• The Toxic Release Inventory System (TRIS), Superfund Amendments and Reauthorization Act (SARA) 313 database for any of the Great Lakes Basin states;
• The U.S. EPA Integrated Risk Information System (IRIS) database; and
• The International Joint Commission (IJC) List of toxic air pollutants.

Any Great Lakes State may submit a proposal to the Technical Steering Committee to list or delist a compound. The proposal for listing or delisting a compound must include:

• The toxic air pollutant name (including all common synonyms) and the CAS Number, where available; and
• The reason for listing or delisting the compound, indicating:
  • The information source (see above) from which the compound was obtained;
  • The carcinogenicity, mutagenicity, teratogenicity, and/or bio-accumulation status of the candidate compound; and
  • Any additional or new information regarding the listing status of the candidate compound.

The Technical Steering Committee will participate in each proposal and comment on the advisability of listing or delisting compounds. The states will act on the proposed changes to the compound list following the procedures specified in Section 2.0 The Technical Steering Committee will submit an updated list of target compounds for consideration by GLC and the Great Lakes States no less than once every two years.

3.2 Source Coding Systems

The most important building block of any database of emissions data and estimates is the method of characterizing the sources, devices and processes that can create or control emissions. When a relational database management system (such as ORACLE®) is used to manage these emissions data and estimates, one or more codes are typically used to uniquely identify a given source, device, or process, and also to identify valid combinations of those sources, devices and processes.

The sections below describe these coding systems and their use in preparing the regional inventory.

3.2.1 SIC Codes

3.2.1.1 Description

The SIC codes were developed to classify establishments by the type of activity in which they are engaged. Establishments are economic entities (generally at a single physical location) that conduct business, perform services, or engage in industrial operations. Examples of SIC establishments include:

• Factories;
• Mills;
• Stores;
• Hotels;
• Movie theaters;
• Mines;
• Farms;
• Ranches;
• Railroad depots;
• Airline terminals;
• Sales offices;
• Warehouses; and
• Central administrative offices.

The SIC is intended to cover the entire range of economic activities at an establishment. The economic activities covered are:

• Agriculture;
• Forestry;
• Fishing;
• Hunting and trapping;
• Mining;
• Construction;
• Manufacturing;
• Transportation;
• Communications;
• Electric, gas and sanitary services;
• Wholesale trade or retail trade;
• Finance, insurance and real estate;
• Personal, business, professional, repair, recreation and other services; and
• Public administration.

The SIC codes are described in the Standard Industrial Classification Manual--1987 published by the federal Office of Management and Budget (OMB, 1987). The SIC Manual contains a detailed description of each classification and assigns it a four-digit, numeric SIC code. The SIC code contains:

• A four-digit industry code (the most detailed, e.g., "1382" is "oil and gas field exploration services");
• A three-digit industry group code (e.g., "138x" is "oil and gas field services");
• A two-digit major group code (e.g., "13xx" is "oil and gas extraction"); and
• A one-digit division code for specific major groups (e.g., Division B is "mining").

By identifying emission sources by their establishment type, SIC codes help to organize and report emissions data by industry type (e.g., rubber manufacturing). The SIC codes can be used with other source coding systems to organize and report emissions data by type of device (e.g., boiler) or process (e.g., fuel combustion). More importantly, the use of SIC codes makes it possible to access other databases that contain SIC information when identifying sources to include in the inventory, and when estimating their emissions.

3.2.1.2 Requirements for Use

The SIC codes identify the types of industries that generate emissions and are used extensively for reporting and for matching to other databases, which requires a full four-digit SIC code. Therefore, each facility source uploaded to the regional repository shall be coded with a full four-digit SIC code, called the primary SIC code. If a four-digit SIC code is too detailed for a facility (i.e., the facility has more operations than can be identified by a single four-digit SIC code), secondary and tertiary four-digit SIC codes may be entered. The option also exists to enter a four-digit SIC code for any device within a facility source.

A set of valid SIC codes is maintained as a reference table in the regional inventory maintained by GLNPO. Any SIC codes for state emissions databases that are not valid for GLC use will need to be converted prior to uploading.

3.2.2 SCCs and AMS Codes

The SCCs and AMS Codes are the building blocks upon which most emission inventory systems are structured. These systems were developed and are maintained by the U.S. EPA. The U.S. EPA has also developed a breakdown of likely SIC/SCC combinations.

The U.S. EPA has developed a tier system to consistently categorize and summarize emissions data that use SCC and AMS codes. The system (with slight modifications to be more applicable to air toxics emissions) groups these codes into useful reporting categories for criteria air pollutants. The mapping of tier categories to SCC and AMS codes is contained in a reference file for the Regional Air Pollutant Inventory Development System (RAPIDS).

In a number of situations, SCC and AMS codes overlap (i.e., a SCC code will define the same type of source as an AMS code). Reconciliation of data using these two coding systems is discussed in Section 4.0.

The following discussion describes the SCC and AMS coding systems and how they are to be used in developing the state portions of the regional inventory.

3.2.2.1 SCCs

The SCC codes describe a source, along with its devices and processes, and link those processes to an "approved" emission factor. An approved emission factor is either contained in the Factor Information Retrieval System (FIRE) or FIRE/GLC, or is a "source-specific" factor derived using the procedures and criteria described in Section 4.0. The SCC characterizes air polluting processes by source type; device type; any related raw material, fuel, or product; any control devices; and release location and type.

The SCC is an eight-digit, numeric code divided into four fields (A-Bx-Cxx-Dx) representing:

• Field A - the category process;
• Field B - the major industry;
• Field C - the major product; and
• Field D - different operations for the category process.

Each SCC code has a corresponding description.

The SCC codes vary in detail from process to process, and there are some inconsistencies in the assignment of SCCs. For some processes, there are SCCs for individual release points or unit operations within the process. In other cases, an entire process may be represented by a single SCC. In addition, some processes may be represented both individually and as part of a larger general category. For example, the manufacture of polyethylene is represented by several SCCs: 3-01-018-60 through 3-01-018-66. The fields include:

• Manufacturing operations (3);
• Chemical manufacturing (-01);
• Plastics production (-018); and
• Recovery system (polyethylene) (-60), purification system (polyethylene) (-61), extruder (-65), and packing/shipping (-66).

However, the manufacture of epoxy resins is represented in the SCC system by only one SCC, 3-01-018-47, with the fields:

• Manufacturing operations (3);
• Chemical manufacturing (-01);
• Plastics production (-018); and
• Epoxy resins (-47).

An SCC typically had a single "unit" of activity (SCC unit). However, as more data became available, the requirement of a single SCC unit combination was changed. Some activities now have multiple units that are appropriate. For example, pounds per ton and pounds per million British Thermal Unit (lb/MMBtu) of heat input are both used for coal combustion. Emission factors contained in FIRE (and FIRE/GLC) are structured in this manner. Therefore, each state will treat SCCs as being unitless and the unit of activity will be associated with the emission factor instead of with a particular SCC.

To maintain consistency in calculating emissions and reporting emission estimates, states will not assign their own SCCs, but will use only those SCCs issued by the U.S. EPA and selected for use in this project. In those instances when any state disagrees with an SCC issued by the U.S. EPA, or if an appropriate SCC has not been developed by the U.S. EPA, the procedures outlined in Section 2.0 will be used to address the issue.

The Tier Organization System, proposed for use by U.S. EPA, groups SCC and AMS codes (discussed below) into three tiers of source categories (i.e, Tier 1, Tier 2 and Tier 3). The Tier Organization System should be used to allow direct comparison of the state emission estimates and to prepare regional emission summaries. A table of tier codes and their relationship to SCC and AMS codes is included in the RAPIDS reference tables (see Section 5.3).

3.2.2.2 AMS Codes

The SCC codes do not include area and mobile sources. For area and mobile sources, emissions are not calculated for each individual source, but, instead, are calculated as an aggregation of individual sources (e.g., architectural coating, pesticide application, and on-road motor vehicles). The U.S. EPA extended the SCC system to include area and mobile sources developing a separate coding system called AMS codes. The AMS codes follow the same general structure as SCC codes, but instead use a ten-digit code patterned Ax-Bx-Cxx-Dxx. Otherwise, the SCC and AMS coding systems are equivalent.

Examples of AMS codes include the following:

• Mobile sources (-22);
• Industrial processes (-23);
• Construction (-11);
• General building construction (-010); and
• Blasting (-030).

3.2.2.3 Requirements for Use

Each process uploaded to the regional repository must be coded with a single eight-digit SCC code or a ten-digit AMS code, as appropriate, issued by the U.S. EPA and identified as valid for GLC use in the SCC-AMS reference table in the regional database at the GLNPO. Any SCC or AMS codes that may be in use in state emissions databases that are not U.S.EPA issued and are not valid for GLC use will need to be converted by that state prior to uploading.

Avoid using non-specific SCC and AMS codes (i.e., those that are listed as general, other, not classified). If an identified process has no specific SCC or AMS code, a new code should be obtained using the procedures described in Section 3.2.3. Any requests for new codes should be made known to the other states through the Technical Steering Committee.

The following steps are to be taken by the states to remove generic SCC-AMS codes before each state's data are loaded into the regional inventory:

Step 1: Determine if a non-generic SCC-AMS code is available.
The U.S. EPA may have assigned an SCC-AMS code to cover the subject process since the generic SCC-AMS code was entered for the subject process. If no SCC-AMS code exists, go to Step 2.

Step 2: Disaggregate the subject process. The process coded into the state's database may be an aggregation of several processes. Determine if SCC-AMS codes are available for the process if it is entered in the database in a more aggregated form. If the process cannot be disaggregated, go to Step 3.

Step 3: Request a new SCC-AMS code. Request a new SCC-AMS code from the U.S. EPA using the procedure described in Section 3.2.3. A copy of the request should be distributed to all states through the Technical Steering Committee. If no SCC-AMS code was assigned, go to Step 4.

Step 4: Document. Document the steps taken to obtain a non-generic SCC-AMS code and reason for using a generic SCC-AMS code. A generic SCC-AMS code will be considered acceptable only for those cases where the U.S. EPA considers the process to be too unique to warrant having a new SCC-AMS code assigned to it.

3.2.3 Obtaining SCC and AMS Codes

The SCC and AMS codes are frequently used in permits, inventories, and databases (such as AIRS and FIRE) to identify emissions data and other emissions-related information. Emission/device/process data may be stored, manipulated, and retrieved using the SCC or AMS code assigned to the source rather than the description of the source. The complete, current version of the SCC table can be obtained from the U.S. EPA. Some emission sources have no SCC or AMS code and corresponding description. In such cases,the requestor should contact the U.S. EPA Office of Air Quality Planning and Sampling (OAQPS), Emissions Inventory Branch (EIB) to have an SCC assigned to the source. The requestor should supply information about the source to the U.S. EPA to aid them in developing and assigning an SCC or AMS code. This information should include:

A description of the emission source for which an SCC or AMS is requested. If possible, the requestor should provide a diagram of the overall process showing related emission sources. If SCC or AMS codes exist for the related emission sources, they should be indicated on the diagram.

Control status of the emission source. The requestor should clearly indicate if the emission point is controlled or uncontrolled. If the emission source is controlled, either by add-on equipment or process modifications, the control should be described. If an SCC-AMS code exists for the uncontrolled process, it should be indicated. That code can usually be applied to the controlled process, in conjunction with a control device code.

Industry description. The requestor should describe the industry in which the emission source is found and applicable SIC code, if known. For example, the emission source might be a waste combustor in the waste disposal industry, which equates to SIC 4952 Refuse Systems. If the requestor is aware of other industries that use this operation, they should provide this information also.

Pollutants emitted. The requestor should provide a list of pollutants potentially emitted by the source. Also, the requestor should indicate pollutants that definitely are not emitted from the source, if known.

Related process parameters. The requestor should provide process parameters to which emissions can be related. For example, a related process parameter for a fuel combustion emission source might be weight of fuel consumed.

Related or similar processes in other industries with existing SCCs.

The requestor may provide information about the emission source to U.S. EPA using the SCC Request Form. Some sections of the form are to be completed by U.S. EPA. This form is designed to illustrate the type of information that the requestor should consider before submitting the request. Not all items apply to all source types, and the information can be submitted in other formats (e.g., tables) that may be more applicable or convenient. This would be particularly true when requesting several related SCCs.

Figure 3.2 illustrates the SCC assignment procedure. If an applicable SCC does not exist in the U.S. EPA's Aeromatric Information Retrieval System (AIRS) for the requestor's described process, an eight-digit SCC, a 70-character SCC description, and an associated process throughput unit will be assigned; the process description and units provided must be suitable for these fields. The assigned code, description, and units will be transmitted to the requestor and to U.S. EPA's National Air Data Branch (NADB) for addition to the AIRS database. The U.S. EPA will also provide the FIRE contractor with a hard copy or electronic version of new SCCs and descriptions to be added to the FIRE database. The FIRE database is to be updated several times a year and uploaded to U.S. EPA's Clearinghouse for Inventory and Emission Factors (CHIEF) Bulletin Board. A list of the newly assigned SCCs (since the previous update) found in the FIRE database will be available on the CHIEF Bulletin Board at that time.

If U.S. EPA receives an SCC request for a source considered to be an area source rather than a point source, that information will be transferred to NADB for assignment of an AMS code.

3.3 Identifying Possible Target Compounds Emission Sources

The Source Summary Database (SSD) lists emission sources, devices, and processes that may discharge one or more of the target compounds, and will be used by each state to identify source categories. The SSD lists these sources by SICs, SIC/SCC combinations, and SCC/AMS codes.

SSD data were obtained from the following sources:

• The National Air Toxics Information Clearinghouse (NATICH);
• The Crosswalk (XWALK)/Air Toxics Emission Factor Database (XATEF) (which is being replaced by the FIRE System);
• Data provided by the states of Indiana, Illinois, Michigan and Wisconsin;
• FIRE;
• Data extracted from the TRIS for the eight Great Lakes states (TRIS is a database containing facilities in the United States, identified by SIC only, that have emissions of certain chemicals above a particular usage or production rate); and
• SPECIATE (a database of speciation factors that are weight percentages of the volatile organic compounds [VOCs] or total particulate matter [PM], presented by source categories for approximately 300 source types).

The latest available version of the SPECIATE database was used to identify sources (by SCC only) that might emit one or more of the target compounds. A number of sources were identified in SPECIATE using the default speciation profile ("Overall Average Profile," designated with a "0000" profile number). However, this default profile is not a true overall average, but only indicates that no valid profiles of source associates are found in SPECIATE for the specified pollutant. Given the low probability that such associations will identify sources that emit compounds of interest, these associations are not included in the SSD and should not be used to estimate air toxic emissions.

Area sources listed in the AIRS AMS subsystem database were associated with VOC and PM speciation profiles from SPECIATE. Similar facility and area source descriptions were compared with previously assigned profiles to make these associations. Speciation profile ratings and VOC percentages were also considered in assigning profiles. Area and mobile sources that did not emit VOCs were not assigned profiles. TheseAMS/SPECIATE VOC/PM association data were used to identify possible pollutant/area source associations and are included in the SSD.

The XWALK database does not contain any emission factors, but lists the chemicals that could be emitted from a given air pollution source. The data derived from this XWALK database should be used with caution because it has not been updated recently and many source categories may no longer emit the compounds listed.

The next table summarizes the possible pollutant/source associations included in the SSD. Whenever possible, sources are represented by both SCC and SIC. However, in some instances, only SIC associations were possible. In most cases, either SCC or AMS codes are matched to possible emission sources.

The GLC column in the pollutant/source associations table includes unique occurrences of a compound from either FIRE, state information, or SPECIATE. The SCC column under GLC and FIRE also includes AMS codes. The pollutant list in the summary table is longer than the list in of 49 target compounds because many of the various dioxin/furan congeners are listed separately in the former table. The SCCs in the SSD will vary from one digit to the full eight digits, due to the SCC variations found in the FIRE repository database. The SICs in the SSD also vary from two to four digits because of variations in the different data sources. Where used, the AMS codes contain all ten digits.

The reliability of the different data sources used to identify pollutant/source associations varies. For example, the existence of an emission factor in FIRE indicates a high probability that a given source actually emits a particular compound of interest. However, an association in XWALK may only indicate some previous evidence that a pollutant was emitted by a given source. Therefore, the order (from FIRE to SPECIATE) of the data sources in the summary table are indicative of the reliability of the pollutant/source associations. Although the reliability of SPECIATE and XWALK are essentially the same, the data in XWALK are more current; SPECIATE is, therefore, last in the summary table.

Over 300,000 pollutant/profile/SCC and AMS associations are represented in the SSD database. In addition to the possible pollutant/source associations table, the following source summary files are available (.dbf) from GLC in electronic format to allow the states to further analyze and summarize these data:

FIRE.DBF (5,654 records) - This file contains the SCC/SIC/pollutant associations from all of the available modules in FIRE. Emission factors exist for all of these associations.

GLC4.DBF (11,261 records) - This file contains the SCC/SIC/pollutant associations derived from all of the information sources discussed above.

GLC4SPEC.DBF (27,782 records) - This file contains the SCC and AMS/profile/pollutant associations derived from SPECIATE; the AS associations were obtained from a study that Radian performed previously that mapped AMS codes to speciation profiles for VOC sources.

GLC4AREA.DBF (1,756 records) - This file contains area source/profile/ pollutant associations derived from SPECIATE. The area sources in this file use an "old" three-digit code that preceded the development of AMS codes. These associations were not incorporated into GLC4SPEC.DBF due to possible duplication with AMS associations.

GLC4SUM.DBF/GLC4SUM.FRX/GLC4SUM.FRT - Contains the database and report files for Table 2, which provides a summary of the number of possible pollutant/source associations found in the expanded SSD.

GLC4SCC.DBF/GLC4SCC.FRX/GLC4SCC.FRT - Contains the database and report files for the pollutant/SCC associations (this table has 546 pages).

GLC4SIC.DBF/GLC4SIC.FRX/GLC4SIC.FRT - Contains the database and report files for the pollutant/SIC associations (this table has 125 pages).

GLC4CMB.DBF/GLC4CMB.FRX/GLC4CMB.FRT - Contains the database and report files for the pollutant/SCC/SIC associations (this table has 89 pages).

GLC4POL..DBF/GLC4POL.FRX/GLC4POL.FRT - Contains all the pollutants for each SCC and a memo file called GLC4POL.FPT (this table has 635 pages).

The main data elements in the files listed above include:

• CAS number;
• Pollutant name;
• SIC Code;
• SIC Code description;
• SCC;
• SCC description;
• AMS Code; and
• AMS Code description.

Many of the pollutant/source category associations do not have emission factors in FIRE; determining emission estimates for certain facilities and source types may, therefore, be somewhat difficult. Where appropriate, the Protocol contains alternative emission estimation procedures for some pollutant/source category associations not included in FIRE (see Section 4.0). Alternative emission estimation techniques include: source-specific emission factors, source-specific models (e.g., TANKS), mass balance procedures, and speciation profiles. Combining speciation profiles with AP-42 criteria pollutant emission factors is not recommended except in the absence of any other information. Emission estimates derived using speciation profiles should be used with extreme caution; such estimates are quite uncertain.

Sources of information that will assist each state in identifying the number and location of specific facilities of interest are presented in Section 3.4. Each state will review these information sources to identify facilities to be included in its component of the regional inventory.

3.4 Information Sources for Identifying Facilities

Each state will consult the following information sources to identify specific facilities that may emit compounds of interest:

• Existing state VOC and PM or PM10 emission inventories;
• "Locating and Estimating" documents (see Section 8.0);
• The Toxic Release Inventory (TRI) for each state;
• State water pollution and hazardous waste files;
• Local industrial directories available from local development authorities (if such directories are organized by SIC);
• Some equipment manufacturers and suppliers (e.g., incinerators, cooling towers, plating tanks, storage tanks, solvents, control equipment) may have good contacts within the industry to locate sources of interest;
• The Dun and Bradstreet "Million Dollar Directory" listing companies with sales over $1,000,000 per year by SIC and county;
• The Dun and Bradstreet "Middle Market Directory" listing companies with sales between $50,000 and $1,000,000 per year by SIC and county; and
• The National Business Lists of companies (by SIC and county), with information on financial strength and number of employees.