Second report of the National Pollutant Release Inventory working group 2001 to 2002: chapter 2

2. Criteria Air Contaminants: Speciation

• 2.1 Background
• 2.1.1 Introduction
• 2.1.2 CAC Speciation and Regional Air-Quality Modelling
• 2.2 Framework for VOC Speciation
• 2.2.1 A List of VOC Species for Speciation
• 2.2.2 A De-minimus for Reporting of Species
• 2.3 Parameters for VOC Speciation
• 2.3.1 The VOC Species Listing
• 2.3.2 The Reporting De-minimus
• 2.3.3 Treatment of Mixtures and Isomers
• 2.3.4 Break-out of Speciation Reporting to Parallel Break-Out of Total VOC Reporting
• 2.3.5 Speciation by Reporting Facilities, Supported by Ongoing Efforts to Broaden and Improve Speciation Tools
• 2.4 Additional Issues and Ongoing Work Requirements Related to Speciation
• 2.4.1 Bridging Unspeciated VOC Emissions
• 2.4.2 Speciation of Other CAC
• 2.4.3 Reporting Out of CAC Information

2.1 Background

2.1.1 Introduction

The following CACs were added to the NPRI for the 2002-reporting year: oxides of nitrogen, sulphur dioxide, carbon monoxide, volatile organic compounds (VOC), and particulate matter. Particulate matter includes separate listings for total particulate matter (TPM), particulate matter less than 10 micrometers (PM₁₀) and particulate matter less than 2.5 micrometers (PM_2.5).

In adding CACs to the NPRI, EC took into account the views presented in the First Report of the WG (October 31, 2001). That report addressed substance-specific definitions and thresholds, as well as a number of other data requirements, such as temporal variation and stack reporting parameters. Additional data elements - related in particular to speciation of VOC and other CACs - were also to have been considered by the WG in 2001, for inclusion in the 2002 reporting year. However, these additional elements were deferred to the 2003 reporting year, to allow more time for research and analysis of possible approaches. A Speciation Subgroup (SG) of the WG was formed in September 2001, to assist EC in further developing options for consideration by the full WG (refer to Annex B for more details).

As a starting point, the WG recognizes the rationale for speciation, with regional air quality modelling being a key driver. There was also recognition of the need to balance improved information against additional burdens.

Additional drivers for speciation include domestic and international programs that require information on emissions, trends and forecasts. Key amongst these are:

• Canada-Wide Standard for Particulate Matter and Ozone
• Canada Wide Acid Rain Strategy
• Ozone Annex of the 1991 Canada-US Air Quality Agreement
• Convention on the Long Range Transport of Air Pollutants
• Development of Ambient Air Quality Objectives
• The International Joint Commission
• The United Nations Economic Commission for Europe.

The WG agreed that VOC -- which for the current reporting year is to be reported only as a total, although in any given facility there may be tens or hundreds of constituent species -- is the priority focus for speciation. VOC emissions influence atmospheric chemistry and air quality in a number of ways, including the following:

• They contribute to the formation of secondary pollutants (e.g., ozone, sulphates, nitrates, and secondary organic aerosol)
• They contribute directly and indirectly to organic aerosol mass
• Some individual VOC species affect human health.

2.1.2 CAC Speciation and Regional Air-Quality Modelling

A regional air quality model (RAQM) is a computer code that numerically solves a comprehensive, time-dependent mathematical model of air quality processes in the atmosphere. The model is a set of equations and relationships that describe the set of atmospheric dynamic, physical, and chemical processes that govern air quality.

To date, emissions processing systems have been used in conjunction with process-level reporting of total VOC emissions to estimate speciated VOC emissions for input into RAQMs. Reporting of total VOC emissions at the facility level does not provide sufficient information for use by RAQMs. Where more detailed information is not available, assumptions are made.

It is important to note that air quality models and monitoring are complementary processes. RAQMs support the following initiatives:

• The building of linkages between emissions to the atmosphere and ambient air concentrations/ atmospheric deposition
• Synthesizing best understanding of all processes relevant to acid deposition
• Determining the importance of various processes
• Testing scientific hypotheses (e.g., linearity assumption)
• Filling in gaps in monitoring networks
• Designing/optimizing monitoring networks
• Determining source attribution
• Simulating "what if" scenarios.

RAQMs require speciated VOC information because different VOC species have different:

• reactivities (RAQMs require speciated information to determine gas-phase and aqueous-phase atmospheric reactivities);
• ozone-forming potentials;
• volatilities; and
• aerosol-forming potentials.

2.2 Framework for VOC Speciation

The WG accepted the approach, proposed by the Speciation SG, of establishing a specified list of species, and a reporting de-minimus.

2.2.1 A List of VOC Species for Speciation

Any species on the list would be reportable assuming the threshold trigger for total VOC (10 tonnes released to air) and the species de-minimus (see below) are met. WG members concurred on this "specified list" approach to speciation after considering an "open list" approach - i.e., one where facilities would be required to speciate based on the full range of over 800 VOC species. This latter was deemed to be unnecessary and overly burdensome.

It was agreed that the list should focus on VOC species of relative significance in relation to the formation of ground level ozone and particulate matter. The WG’s recommended approach to accomplishing this is as follows:

• The VOC species listing should be identified based on specific criteria. (ENGO members of the WG support 90% target coverage for each criterion, while industry members support an "80/20" approach to the criteria as a whole.)
• The specific criteria include: mass, reactivity, solubility, and condensability. (See Annex C for more detail.)
• Cross-checks (e.g., based on formulations/solvents, industry-specific information, and area source and regional specificity analyses) should be performed to ensure comprehensive identification of VOC species.

Industry members of the WG also recommend that consideration be given to removing species from the list if they have gone unreported for a number of years.

2.2.2 A De-minimus for Reporting of Species

The de-minimus is the threshold established for determining when individual species on the species list must be reported. Thus, once total VOC reporting is triggered, reporting will also be required for any species on the specified list that is released in quantities equal to or greater than the de-minimus.

The following are foundations for identifying an appropriate de-minimus:

• NPRI reporting thresholds should continue to apply. In other words, the threshold for total VOC remains at 10 tonnes release to air, and individual VOC that are currently listed on the NPRI must meet their respective thresholds (either the standard 10-tonnes manufacture, process, otherwise use (MPO) threshold, or the applicable alternate threshold (ATH)).
• In keeping with the total VOC threshold, the de-minimus should be a release-based threshold.

2.3 Parameters for VOC Speciation

Based in part on application of the preceding framework, the WG recommends the following parameters for VOC speciation:

2.3.1 The VOC Species Listing

Master List of Candidates

The WG agreed on the development of a "master-list" of candidates for speciation, based on the following:

• The VOC species already individually listed on the NPRI
  
• As shown in Table D-1 of Annex D, there are currently 101 VOC substances/mixtures/isomers individually listed on the NPRI.
  
• Additional species identified from national analysis that meet the criteria established in subsection 2.2.1
  
• Tables D-2 and D-3 of Annex D list the species identified in this manner. The list was developed using information in the Residual Discharge Information System (RDIS) for 1995, and the US EPA SPECIATE v3.0 database. At the 90% coverage level for the criteria in subsection 2.2.1, there are 64 individual species and 18 mixtures. There are also 12 isomer groupings. Of these 94 substances, 35 are already individually listed on the NPRI, hence show up in Table D-1. Thus, a total of 59 species would be added to the master list.
  
• Additional species identified from regional analysis that meet the criteria established in subsection 2.2.1
  
• As per Table D-4 in Annex D, a total of twelve additional species (i.e., not already on the NPRI and not identified through the preceding national analysis) were identified from a regional analysis of the Windsor-Quebec corridor. This analysis was performed to determine if the national list was relevant to the regional analyses.
  

EC staff will consolidate the tables in Annex D into a single master list, undertaking additional reconciliation and eliminating any remaining duplication.

Final List of VOC Species

In principle, the WG agrees that the existing individually listed NPRI substances should be included in the final specified list. Differing views remain on whether the 80% or 90% coverage target should apply for the species identified through application of the criteria at the national level. WG members have also not come to a recommendation on which, if any, of the species identified from regional analysis should be included. EC will undertake additional analysis on these twelve substances, to present to the WG at its September meeting. Information and perspectives from stakeholders are requested.

2.3.2 The Reporting De-minimus

Based on an analysis of percent coverage at different de-minimus levels, as shown in Annex E, industry members of the WG believe that a de-minimus of 1 tonne is appropriate. (Given this de-minimus, once total VOC reporting is triggered, reporting would also be required for any species on the specified list that is released in quantities of 1 tonne or greater.) ENGO members of the WG agree that 1 tonne is an appropriate de-minimus only if it is applied in conjunction with a specified list that is based on a target coverage of 90% for the four criteria.

It is important to reiterate that no change is implied to the overall thresholds that trigger reporting. In other words, total VOC continue to be reportable by a facility only when at least 10 tonnes total are emitted, as per the Canada Gazette Notice for the 2002 NPRI. (Similarly, emissions of any VOC species individually listed on the NPRI continue to be reportable when the threshold for that VOC species is exceeded.)

2.3.3 Treatment of Mixtures and Isomers

With respect to mixtures, the WG recommends that reporting be either by individual VOC species (if known), or by mixture. With respect to isomers, the WG recommends that reporting be based on the aggregated total, except where individual isomers have significantly different ozone and/or PM forming properties. In this case, the individual isomers are identified in the VOC species listing and will be reported individually to the NPRI; but the remainder of the isomers may be reported as a group that excludes the individual isomer. The WG recommends that EC develop appropriate guidance for dealing with both mixtures and isomers. The WG also requests that EC determine the legal requirements for due diligence, and develop appropriate guidance, for reporting mixtures versus the individual substances that comprise the mixture, and for instances when those components are listed individually on the NPRI with MPO thresholds.

2.3.4 Break-out of Speciation Reporting to Parallel Break-Out of Total VOC Reporting

Based on the Canada Gazette Notice for the 2002 NPRI, VOC (and other CAC substances) must be reported not only as a total on a facility basis; but also, where major stacks exist, separate numbers must be reported for ground-level emissions and for emissions from each major stack. ("Major stacks" refers to stacks with a height equal to or greater than 50 m that emit a specified minimum quantity of a CAC substance - 5 tonnes for VOCs.) The WG agreed that speciation of VOC should be on the same basis.

2.3.5 Speciation by Reporting Facilities, Supported by Ongoing Efforts to Broaden and Improve Speciation Tools

In keeping with the NPRI tradition of self-reporting of emissions, facilities should undertake their own speciation estimates to provide to EC. One tool to assist facilities is the US EPA database SPECIATE 3.1, which contains standard speciation profiles for about 600 processes. In some cases, facilities may have speciation profiles that are better tailored to their processes. Other existing tools must be made available as applicable; and ongoing development of tools is important. WG members acknowledge that initial limitations in data quality are likely. Improvements are expected as reporters become familiar with requirements and procedures, and progress is made in guidance and methodologies.

2.4 Additional Issues and Ongoing Work Requirements Related to Speciation

2.4.1 Bridging Unspeciated VOC Emissions

Some means of representing the unspeciated VOC emissions is required for modelling purposes. The WG identified the following possible approaches, based on optional questions to reporters:

• The NPRI reporting software could prompt reporters to voluntarily provide information on the range (e.g., high, medium, low) in relation to key properties (i.e., reactivity, solubility, condensability) for the unspeciated portion of total VOC releases to air.
• The NPRI software could prompt reporters to voluntarily provide information (including CAS numbers and air emission values) that they may have access to about additional VOC species that are not on the specified list. Some measure of judgement by reporters is implied in determining what additional species would warrant being reported; but of particular interest to modellers would be information on the largest contributors by mass to the unspeciated VOC mass emissions.

2.4.2 Speciation of Other CAC

Since CO and SO₂ are individual substances, and speciation of NOx is already effectively handled by modellers, the WG agrees that PM should be the focus for further speciation work. Given the lead-time needed to develop recommendations, the WG agrees that the target for application should be the 2004 reporting year.

PM mass emissions consist mostly of the following six components: elemental carbon, sulphates (SO4), ammonia (NH4), crystal material, organic carbon, and NO₃. PM sources include combustion and process-related activities, and fugitive emissions. The WG recommends that EC perform additional analysis, consult with major emission source sectors, and prepare a proposal for consideration in the next round of consultations. The following framework was established:

• Identify the major source sectors for PM emissions
• Determine the percentage breakdown of PM components per category
• Adopt a similar approach to exploring options for PM speciation as that used for VOC speciation.

2.4.3 Reporting Out of CAC Information

A single VOC species may be reported individually if listed as such on the NPRI, and also as part of the total VOC listing. Given this, WG members have expressed concerns about the potential for users of NPRI data to double-count the emissions. All WG members support EC’s stated intention to publish the following on an annual basis, starting with the 2002 reporting year:

• Appropriate context for the CAC information contained in the NPRI, at the time that the NPRI data is made public.
• Summary tables of the national, comprehensive CAC emissions inventory, following the NPRI data release.
• A national summary report of the CAC emissions inventory, following the summary tables.

Building on the concept of providing context for interpreting the data, industry members of the WG request EC to ensure that clear direction (e.g., a protocol and guidance) to reduce the risks of double counting is brought to the attention of data users in a range of situations; for example, when querying the NPRI data-base. EC will also explore possibility of grouping certain substances (e.g., ozone precursors) for purposes of reporting out.