7.11 Other External Combustion - Natural Gas

Categories 307, 968, 1590, and 1591

7.11.1 Introduction

Categories 307, 968, 1590, and 1591 account for criteria pollutant emissions (particulate, organic, NOx, SOx, and CO) from natural gas (NG) combustion sources in the Bay Area.

Categories 307, 1590, and 1591 estimate emissions from natural gas combustion sources, such as industrial, commercial, and institutional water and steam boilers, furnaces, space heaters, process heaters, pre-heaters, ovens, and after burners, etc. Category 968 accounts for emissions from the natural gas fired glass melting furnaces at glass manufacturing facilities.

7.11.2 Methodology

Point Sources

Point Sources are operations that emit air pollution into the atmosphere at a fixed location within a facility, for which the Air District has issued a permit to operate, e.g., refinery cooling towers. These could also be a collection of similar equipment / sources located across multiple facilities, e.g., reciprocating engines.

During the permit to operate (PTO) issuance process, the BAAQMD collects information from the operating facility and/or determines from published literature, e.g., EPA’s document AP-42, characteristics of a source including maximum throughput, emission factors for emitted pollutants, and control factors associated with downstream abatement devices. These characteristics are then stored for future use in the BAAQMD’s internal database. Facilities that hold a permit to operate are required to renew this permit periodically (this period varies based on facility and source type). Upon renewal, the facilities are requested to provide any updates to source characteristics as well as the source throughput for the last 12 months. This throughput, in combination with the emission factors and controls factors stored in the internal database, are used to estimate annual emissions at the source level. These source level emissions are then sorted and aggregated into categories.

Categories 307 and 968 are point source categories and follow the above methodology for emissions estimates.

Area Sources

Categories 1590 and 1591 are considered an area source category since they cover facilities / emission sources that are not directly permitted by the District, and hence not systematically cataloged. Emissions for area source categories are determined using the formula:

Current Year Emissions = Base Year Emission X Growth Profile, and,

Base Year Emission = Throughput X Control Factor X Emission Factor

where,

  • throughput or activity data for applicable base year(s) is determined using a top-down approach (e.g. state-, national-level data);
  • emission factor is derived from general literature, specific literature and reports, and/or source testing results provided by Air District staff;
  • control factor (if applicable) is determined by District and state rules and regulations in effect;
  • and, historical backcasting and forecasting of emissions is based on growth profiles as outlined in the Trends section of this chapter

More details on throughput, county distribution, emission factors and controls is provided in the following subsections.

The District Non-Residential Area source NG throughputs are derived by subtracting the total District Point Source NG throughput (stated in the section above) from the total Non-Residential NG usage throughput published by the California Energy Commission (CEC) .

\[ \begin{eqnarray}   \text{[CEC Non-Residential usage]} - \text{[District Data Bank Point source usage]} \\ = \text{[District Non-Residential Area source usage]} \end{eqnarray} \]

Detailed steps taken in development of natural gas throughput estimates for the Area source Category 1590 - Industrial sources; and Category 1591 - Commercial sources are as follow:

Step 1: Collect Natural Gas Usage Data

Natural gas (NG) usage (in millions of therms, Mtherms) is obtained from the California Energy Commission (CEC) website on a county-by-county basis. This publicly available dataset contains Residential and Non-Residential NG usage for years 1990 to 2020. It should be noted that this dataset does not include electricity generation NG usage (consumption at power plants), consistent with CEC categorization protocol. The NG data is first converted to millions of standard cubic feet (mil. ft3) using a conversion factor of 97.02 (mil. ft3/ therm). For the current base year Emissions Inventory (EI), only the data available for years 2011-2015 are used for calculations and projections (both backcasting and forecasting; see Step 5). Going forward this dataset is referred to as NG_nonres_res_public_CEC.

For further resolution into different sectors (e.g., industrial, commercial etc.), county-by-county annual NG consumption data (in therms) is also additionally obtained from CEC staff directly (internal communication). This supplemental data set provides NG usage at a sector-level for all Bay Area counties for years 1990-2019. Once again, only the sector-speciated data for years 2011-2015 are used for calculations and projections (see Step 5). This dataset, consistent with CEC categorization schemes, does not include NG consumption for electricity generation. The supplemental NG data set is referred to as NG_sector_internal_CEC.

Previously, an assumed split of 55% Industrial to 45% Commercial, which remained constant for all years, was applied to the first dataset, NG_nonres_res_public_CEC. With the availability of the supplemental dataset NG_sector_internal_CEC, the actual split of Industrial to Commercial can now be obtained down to a county level on a yearly basis.

NG-consumption data is available by year and by county in both datasets mentioned above. However, to ensure consistency between past Base Year EIs and the current Base Year EI (base year of 2015), NG usage totals are derived from the CEC-published NG_nonres_res_public_CEC dataset. The NG_sector_internal_CEC dataset is only used to determine sector-specific splits on a yearly basis.

Step 2: Determine Industrial and Commercial Split for Natural Gas Consumption and Apply the Split

The NG_sector_internal_CEC dataset has NG-usage divided by sector, but there are several sectors in addition to Industrial and Commercial, such as, Mining, Transportation, Communications, and Utilities (TCU), Agricultural Water Pump, and Unclassified. The data in these respective sectors are assumed to be included in either the Industrial or Commercial sector. The apportionment of these other sectors into either Industrial or Commercial is as follow:

  1. TCU (Transportation, Communications, and Utilities) sector is assumed to be 100% Commercial

  2. Mining sector is assumed to be 100% Industrial

  3. Agriculture with Water Pump sector is assumed to be 100% Industrial

  4. Unclassified are split into Industrial and Commercial based on based on split of Industrial and Commercial sectors for a given year (after the addition of data in Mining and Ag Water Pump to Industrial sector)

After applying the above categorization, a percent split is determined on a by year, by county basis.

The revised and newly derived splits is then applied to the NG_nonres_res_public_CEC dataset to determine the Total (Area Sources + Point Sources) natural gas usage in the Bay Area, by county, by year, and by sector (Industrial and Commercial; in mil. ft3).

Step 3: Determine Point Source Natural Gas Consumption

The Air District’s Point Source (permitted source) NG usage is subtracted from the total NG usage in the Air District to determine the Area Source throughput for categories 1590 - Industrial, and 1591 - Commercial. The point source NG usage includes both combustion and non-combustion sources. The regional consumption of NG also includes the use of NG as the main feedstock for hydrogen production, which in turn is used in the fractional distillation of crude oil at refineries. Shown below are all Point source categories in the Base Year 2015 emissions inventory that use natural gas.

Other Point Source NG Categories:

Category 15 - Industrial/Commercial Processes/Facilities- Petroleum Refining Facilities- Flares And Blowdown Systems

Category 16 - Industrial/Commercial Processes/Facilities- Petroleum Refining Facilities- Other Refining Processes

Category 49 - Industrial/Commercial Processes/Facilities- Other Industrial/Commercial Processes- Gas Production Fields

Category 54 - Industrial/Commercial Processes/Facilities- Other Industrial/Commercial Processes- Other Industrial Commercial

Category 275 - Petroleum Product/Solvent Evaporation- Other Organic Compounds Evaporation Industrial/Commercial- Coating Auto Refinishing

Category 299 - Combustion - Stationary Sources- Fuels Combustion- Oil Refineries External Combustion- Natural Gas & Other Gases

Category 303 - Combustion - Stationary Sources Fuels Combustion- Reciprocating Engines- Gas Fuel

Category 305 - Combustion - Stationary Sources Fuels Combustion- Turbines Gas Fuel

Category 307 - Combustion - Stationary Sources Fuels Combustion- Other External Combustion- Natural Gas

Category 968 - Combustion - Stationary Sources Fuels Combustion- Glass Melting Furnaces- Natural Gas

Category 1687 - Combustion - Stationary Sources Fuels Combustion- Landfills- Internal Combustion Engines

Category 1749 - Combustion - Stationary Sources Fuels Combustion- Cement Plant Combustion- Natural Gas

Power Plant Point Source NG Categories:

Category 290 - Combustion - Stationary Sources Fuels Combustion- Cogeneration- Boilers

Category 291 - Combustion - Stationary Sources Fuels Combustion- Cogeneration- Turbines

Category 292 - Combustion - Stationary Sources Fuels Combustion- Cogeneration- Reciprocating Engines

Category 293 - Combustion - Stationary Sources Fuels Combustion- Power Plants- Boilers

Category 1595 - Combustion - Stationary Sources Fuels Combustion -Power Plants- Turbines

Accounting for point source NG consumed in hydrogen (H2) production:

Category 16, Other Refining Processes, represents emissions from H2 production. It has been observed that many refineries have been grouping the NG used as a feedstock in H2 production into the refinery fuel gas (RFG) usage when they report data to the Air District. This grouping process makes the NG used for H2 production indistinguishable when looking at Category 16 permit data. Therefore, to correctly account for the amount of NG being used for H2 production in the Bay Area, an independent approach has been adopted which includes the following datasets:

  1. Annual amount of NG used for H2 production feedstock for the Petroleum Administration for Defense District (PADD) 5, West Coast region, is obtained from the Energy Information Administration (EIA) website. EIA archives contain NG usage data only back to year 2010.

  2. The West Coast NG throughput data above is allocated to the San Francisco Bay Area by using the proportion of CO2e emissions for all refineries and H2 production plants in the West Coast PADD5 region compared to CO2e emissions reported for Bay Area refineries and H2 production plants.

This ratio-based apportionment is used to derive H2 production related NG throughput for counties where refineries and H2 plants are located in the Bay Area, for years 2011 to 2015. This usage is given a pseudo category id 5000 for accounting purposes. NG throughputs for H2 production in 2015 for Contra Costa and Solano counties were 24,055 and 4,678 (mil. ft3) respectively.

Apportioning Category 307

Category 307 is the only point source category that has a mix of both Commercial and Industrial sources. All other point source categories (15, 16, 49, 54, 299, 303, 305, 968, 1687, 1784, and pseudo category 5000) are assumed to be completely Industrial. Category 307, thus, accounts for all of (100%) the Commercial point source NG usage. Category 307 throughput is divided into Industrial and Commercial sectors based on Standard Industrial Codes (SIC) at the source level.

To obtain the final Area Source throughputs for categories 1590 (Industrial) and 1591 (Commercial), the total point source NG usage (by Industrial and Commercial sector) is negated from the NG_nonres_res_public_CEC usage data set.

Step 4: Deal with Negative Throughputs

The Area Source NG throughputs are derived by subtracting the total Point Source NG throughputs from the total Non-Residential NG data. For some counties, the Industrial : Commercial split is highly skewed towards one sector (e.g., Contra Costa county, Marin county etc.). For these counties, the uncertainty itself in the Point Source throughput data of the dominant sector (obtained from BAAQMD permit data) is larger than the absolute magnitude of Point Source throughput for the minor sector. At the same time, owing to uncertainties in self-reporting mechanisms, the confidence interval of the total Point Source throughput of such a county most likely overlaps the absolute value of the total Non-Residential throughput (obtained from CEC publication). In a county with skewed distribution of sector contributions, it is quite possible that the subtraction of the Point Source data from the Total NG data can yield a small negative quantity as the Area Source throughput. As these two datasets are independently sourced, they cannot be expected to be completely consistent with each other in terms of scope and inclusion of all sources. All negative values are observed to be well within the % certainty of the total throughputs. However, it is not feasible to assume all natural gas usage for commercial units is being utilized in excess by larger permitted sources and therefore fully covered by permitted categories. In those cases, where the Area Source throughput for a county is statistically negligible (or negative), the throughputs for such a year would be replaced with an averaged throughput over the five year time-period (between years 2011-15) using values for those years where the throughputs are non-zero (even if very small).

Step 5: Determining Forecast and Backcast

The EI team has the most confidence in Point Source data reported for years 2011-2015. The datasets are not only more recent but there is a better historical knowledge of emission factors and throughputs used to derive these data as per the internal Reporting Year (RY) data pipeline. The BAAQMD archives also have missing point source data for years 1991 and 1992. Additionally, there is a limitation of producing NG-throughput data used for H2-production due to lack of temporal coverage in the input datasets from EIA and EPA prior to year 2010 (see section above).

Hence, it has been decided that NG-throughput will be calculated for Categories 1590 and 1591 using the classical Area Source calculation approach that has been applied to several other Non-Point Source categories in the current base year emissions inventory. In this approach, the throughput data and subsequently emissions are calculated for years 2011-2015, and appropriate growth profiles are used to backcast the historical data and make future projections. In this specific case, NG-data throughput data is calculated for each year between years 2011-2015 at the county level, so the county distribution is unique for each year (unlike traditional Area Sources where county fractions typically remain constant). As such the data calculations are performed for Categories 1590 and 1591 separately as “Special Cases” (but still using the classic area source calculation approach).

For both backcasting and forecasting throughputs (and subsequently emissions), historical CEC sector-specific NG-use data is combined with California Air Resources Board (CARB) NG-data that includes NG-use projections based on adopted Rules and State Laws, to generate appropriate normalized county-specific growth profiles. These growth profiles are applied to year 2011 NG-throughput data (for backcasting), and to year-2015 NG throughput data (for forecasting), to create a complete NG-throughput (and subsequently emissions) timeline for years 1990-2040.

Step 6: Calculating emissions using Area Source calculation approach

The NG throughputs for Categories 1590 (Commercial) and 1591 (Industrial) are then utilized to derive emission estimates for various pollutants using the current inventory Area Source calculation approach. The calculation approach uses the Area Source formula to derive emissions for different pollutants. Emissions are derived for the following pollutants: Total Organic Gases (TOG), Methane (CH4), Particulate Matter (PM), Nitrogen Oxides (NOx), Sulfur Dioxide (SO2), Carbon Monoxide (CO), Carbon Dioxide (CO2), and, Nitrous Oxide (N2O), using the respective emission factors (ef_qty). The regulatory and technology controls implemented for specific pollutants are incorporated in the control factors (cf_qty). The emissions of Reactive Organic Gases (ROG) are derived by subtracting CH4 emissions from TOG emissions. The PM emissions are further classified into PM2.5 and PM10 emissions based on a speciation profile for gaseous combustion obtained from CARB’s list of PM speciation profiles.

In this manner, a complete multi-pollutant emissions dataset for various counties for NG Commercial and Industrial Area Sources is put together from years 1990 to 2040. This dataset is then input back into the Special Cases emissions master spreadsheet for further post-processing and data visualization with other base year inventory data.

(a) Activity Data / Throughput

For point sources categories 307 and 968, process material throughputs are as reported by the plants and contained in the Air District internal database.

For area source categories 1590 and 1591, throughput information is obtained as explained in steps one through six above.

(b) County Distribution / Fractions

Point source categories 307 and 968: The District’s point source database contains county distribution information for each facility based on its geographical location.

Area source categories 1590 and 1591: County distribution is based on natural gas usage in each county.

(c) Emission Factors

Point Sources Emission Factors

Point source emission factors may be source specific factors reported by the plants through source test results or, applicable general factors, i.e., from the U.S. Environmental Protection Agency (EPA).

Area Source Emission Factors

Emission factor information for natural gas combustion was obtained from the U.S. Environmental Protection Agency (EPA) document AP-42317, the Energy Information Administration (EIA)318 and the California Energy Commission (CEC)319.

(d) Control Factors

The Air District Regulation 9, Rule 7 controls emissions of NOx and CO from industrial and large commercial NG combustion sources such as boilers, steam generators, and process heater320. The controls implemented as a part of this regulation are applied to category 1590, 307, and 968.

The Air District Regulation 9, Rule 6 controls emissions of NOx and CO from residential and small commercial NG combustion sources such as boilers, steam generators, and process heater321. Controls implemented as a part this regulation are applied to category 1591.

(e) Speciation

The ROG/TOG ratios applied to this category, or this group of related categories are based on an Air District internal speciation profile. Multiple data sources have been used for developing speciation profiles, such as Air District-approved source tests, TOG speciation ratios used by other regional air quality agencies, and relevant literature including latest speciation profiles developed by CARB322 and the US Environmental Protection Agency323.

Organic speciation is based on the methane emissions relative to the TOG emissions, such that ROG emissions are the result of subtracting methane from TOG emissions.

Organic speciation fractions are shown below:

Cat_ID TOG ROG/TOG Fraction
C1590 1.00 0.791
C1591 1.00 0.791
C307 1.00 0.4223
C968 1.00 0.4223

PM speciation: The PM2.5/PM and the PM10/PM ratios applied to these categories are consistent with size fractions of speciation profiles developed by the California Air Resources Board (CARB) and published on their emissions inventory webpage324.

CARB Particulate Matter (PM) Speciation Profile number 120, “Gaseous Material Combustion”, is used to speciate total PM emissions into PM10 and PM2.5 emissions. Particle size distribution is shown below:

Size Fraction of Total Particulate (TSP)
10 micron 1.000
2.5 micron 1.000

(f) Sample Calculations

The following equation can be used to calculate any criteria pollutant emissions for any county within the Bay Area Air District:

TOG Emissions (tons/ year) = Throughput (millions ft3/ year) x TOG Emission Factor (lbs/ millions ft3) x County Fraction x Growth Factor x Control Factor x (ton/2000 lbs)

7.11.3 Changes in Methodology

To improve NG-related combustion emissions for Area Source categories 1590 and 1591, the following changes were applied to Base Year 2015 methodology:

  1. Apportioning CEC’s NG-data for minor sectors (including Ag pump, TCU, Mining and Unclassified) into Industrial and Commercial sectors to enable more accurate accounting for Bay Area’s county specific NG throughputs. This step was ignored and skipped in previous base year emissions inventory methodology.

  2. Using data-derived sector splits across years 2011-2015 in the the current base year emissions inventory as compared to a fixed split applied in the earlier emissions inventories. Since the growth profiles are also derived from sector-specific and county-specific CEC and CARB data, the time-varying sector splits for historical and future years are also indirectly included in the entire timeline for the current emissions inventory NG-throughputs.

  3. Accounting for NG consumed in Bay Area for H2 production and removing this throughput quantity from the total Industrial NG throughput before estimating Area Source NG throughput. This step was missing from the previous emissions inventories leading to erroneous Area Source throughputs.

  4. Using a refined Standard Industrial Classification (SIC) system to appropriately apportion Point Source NG usage into Industrial and Commercial sectors.

7.11.4 Emissions

A summary of emissions by category, county, and year are available via the associated data dashboard for this inventory publication.

7.11.6 Future Improvements

The following improvements will be instituted in the next base year emissions inventory iteration to further improve the accuracy of the estimates produced in the Base Year 2015:

Get a full set of data from EIA and EPA for corresponding data sets used to derive NG throughput related to H~2` production so that this methodology can be applied to all years instead of having to use growth profiles for backcasting prior to year 2011.

Regulatory control factors will be improved to reflect current standards and fleet turnover.

7.11.7 Uncertainties

Uncertainties exist due to not having a complete data set for all years to derive NG throughputs related to H2 production in the Bay Area. A further discussion of uncertainties built into the throughput calculation approach for Categories 1590 and 1591 is included in the ‘Methodology’ section above.

7.11.8 Contact

Authors: Abhinav Guha, Sukarn Claire

Reviewer: Ariana Husain, Abhinav Guha, Song Bai

Last Update: November 06, 2023

7.11.9 References & Footnotes

  1. USEPA. AP-42. Compilation of Air Pollutant Emissions Factors. https://www.epa.gov/sites/default/files/2020-09/documents/1.4_natural_gas_combustion.pdf↩︎

  2. Energy Information Administration (EIA). https://www.eia.gov/energyexplained/natural-gas/use-of-natural-gas.php↩︎

  3. California Energy Commission (CEC). https://www.energy.ca.gov/↩︎

  4. Air District Regulation 9 - Rule 7. https://www.baaqmd.gov/rules-and-compliance/rules/reg-9-rule-7-nitrogen-oxides-and-carbon-monoxide-from-industrial-institutional-and-commercial-boiler↩︎

  5. Air District Regulation 9 - Rule 6. https://www.baaqmd.gov/rules-and-compliance/rules/reg-9-rule-6-nitrogen-oxides-emissions-from-natural-gasfired-water-heaters?rule_version=Adopted↩︎

  6. ORGPROF. CARB. 2022. https://ww2.arb.ca.gov/speciation-profiles-used-carb-modeling↩︎

  7. SPECIATE. USEPA. 2022. https://www.epa.gov/air-emissions-modeling/speciate↩︎

  8. PMSIZE. CARB. 2022. https://ww2.arb.ca.gov/speciation-profiles-used-carb-modeling↩︎

  9. Association of Bay Area Government (ABAG). Projections. https://abag.ca.gov/our-work/land-use/forecasts-projections↩︎