Loading...
HomeMy WebLinkAboutDAQ-2025-000773DRAFT Northern Wasatch Front Serious State Implementation Plan Area Source Inventory Technical Supporting Document (2017, 2023 and 2026) Background and Overview: Area source emissions are smaller pervasive emissions sources that do not qualify as point sources under the relevant emissions cutoffs (>= 50 tpy of NOx or VOC). Area sources encompass more widespread sources that may be abundant, but that, individually, release small amounts of a given pollutant. These are sources for which emissions are estimated as a group rather than individually. Examples typically include dry cleaners, residential wood heating, auto body painting, and consumer solvent use. With the exception of certain oil and gas industry sources, area sources generally are not required to submit individual emissions estimates. EPA’s oil and gas industry inventory methods were used for the Northern Wasatch Front (NWF) Serious State Implementation Plan (SIP) for the 2015 ozone NAAQS and are explained in EPA’s 2020 NEI TSD. Though they are part of the area source inventory, they will not be mentioned extensively in this document. The main distinction between point and area sources is the methodology used to estimate emissions. Point sources are inventoried individually while area sources are inventoried collectively and given as county totals. The only exception to this distinction is based on Utah Administrative Code R307-150 which requires a triennial emissions inventory submission from sources with Standard Industrial Classification codes in the major group 13 that have uncontrolled actual emissions greater than one ton per year for a single pollutant of PM10, PM2.5, oxides of nitrogen (NOx), oxides of sulfur, carbon monoxide or volatile organic compounds (VOC). These sources include but are not limited to industries involved in oil and natural gas exploration, production, and transmission operations; well production facilities; natural gas compressor stations; and natural gas processing plants and commercial oil and gas disposal wells, and ponds. The term “process” is used here to name an operation or activity that produces emissions. Area sources include broad groups of processes. Several examples are provided below. The following examples include a brief discussion of how emissions are generally calculated for each category, where the activity data came from, where the projection data came from, and how the controls are applied. 1. Commercial and consumer solvent usage. This generally employs population or employment census data multiplied by a VCPy1 (see “Calculation Methodology for Solvent Emissions” for more information) emission factor and reducing emissions by an applicable control (such as 10% reduction). Projections are based on census population and employment growth estimates and future control measures are also applied. Point source emissions are also subtracted, where applicable, to avoid potential double counting of emissions. 2. Stationary fuel combustion. In general, Energy Information Agency (EIA) data are used and adjusted by subtracting point source fuel use and applying an emission factor to the remaining fuel use. This is also projected using EIA future growth/decline estimates. 1 Seltzer, K., E. Pennington, V. Rao, B. Murphy, M. Strum, K. Isaacs, AND H. Pye. Atmospheric Chemistry and Physics Peer-Review: Reactive Organic Carbon Emissions from Volatile Chemical Products. Atmospheric Chemistry and Physics. Copernicus Publications, Katlenburg-Lindau, Germany, 21(6):5079-5100, (2021). DRAFT 3. Livestock. Activity data (animal population) is typically derived from agricultural census figures and entered into EPA’s livestock emissions tool. Emissions outputs are then projected from the base year based on agricultural employment growth estimates. 4. Waste treatment and disposal. This is generally based on waste collection estimates and multiplied by an emission factor derived from points source reports which accounts for control measures as well as EPA-developed emission factors (such as for composting). Emissions are projected based on population growth. 5. Miscellaneous industrial manufacturing operations. Often surrogate information such as product per individual consumed or employment census figures are multiplied by a relevant emission factor and adjusted for appropriate controls. Projections are typically based on population or employment growth. 6. Fuel distribution. Fuel dispensed from Utah Tax Commission or EPA modeled estimates serve as the source of activity data which is then multiplied by the applicable emission factor for each sub-category. When applicable, emissions are then adjusted by a control factor and emissions are projected based on EIA fuel consumption estimates. 7. Miscellaneous sources (agricultural/forest burning, structure fires, mining, and construction). These employ a variety of surrogate (census, survey), reported (acres treated) or throughput activity data and have emission factors and controls (where applicable) applied. Projections are also based on various methods including those described above. Each of these broad groups of processes contains a number of more specific groups or categories that share similar emission processes and emission estimation methods. Activity data is used to calculate area source categories. This data includes population, employment, Vehicle Miles Travelled, fuel usage, animal, crop, oil and gas industry throughput submissions, and other estimates. A list of the individual data tables and sources of the data used in the calculation processes is included in the associated R2 script and accompanying Excel spreadsheets. The activity data workbook and R script contain annual activity data by county, and detailed emissions summations by category from database queries or applicable EPA/NOMAD (Nonpoint Methods Advisory group) tool outputs and additional emissions estimates (“gap filling”). All databases, workbooks, and scripts are available upon request. Area sources were adjusted for potential overlaps and double counts with point sources. Adjusted categories include bakeries, mining and quarrying, fuel combustion, degreasing solvents, publicly owned treatment works (POTW), and municipal landfills. Adjustments typically involve subtracting point source activity data from total estimated statewide or countywide activity data prior to calculating emissions or, less frequently, subtracting point source emissions from overall emissions estimates for a particular category. Emissions data from the area source inventory, EPA/NOMAD tool outputs and gap filling emissions are, after being compiled, processed through an open source emissions modeling platform called SMOKE3 3 https://www.cmascenter.org/smoke/ 2 https://www.r-project.org/ DRAFT (Sparse Matrix Operator Kernel Emissions) which adjusts the data for the desired episode and applies additional relevant controls. The base year inventory is the primary inventory from which other inventories are derived. Thus, all inventories are consistent with data provided in the base year inventory. The 2017 base year was selected as the foundation for the serious SIP for the NWF and the 2023 and 2026 projection years were projected based on this year. The projection year inventories (2023 and 2026) project future air pollution emissions. The goal in developing emission projections is to attempt to account for as many of the important variables that affect future year emissions as possible. Emission projections provide a basis for developing control strategies for this State Implementation Plan (SIP), conducting attainment analyses, and tracking progress towards meeting air quality standards. Emission projections are a function of change in activity (growth or decline) combined with changes in the emission rate or controls applicable to the source. To a large extent, projection inventories are based on forecasts of industrial growth, population growth, changes in land use patterns, and transportation growth. The way the area source categories were projected is explained in “Overview of Projection Methods”. Baseline and Projected Year Inventories A base year inventory is comprised of a comprehensive, accurate, current inventory of actual emissions from sources of VOCs and NOX emitted within the boundaries of the NAA. The base year for this SIP submittal is 2017. Although 2020 is the most recent calendar year for which a complete triennial inventory was submitted to the EPA, it was determined that 2020 emissions were not representative of the emissions in a typical year in the state due to people staying home during the COVID-19 pandemic. As a result, 2020 emissions were projected back to 2017 proportionally using relevant 2017 to 2020 activity data ratios and to take advantage of new methods and emission factors available in 2020. Moreover, UDAQ has developed projected emission inventories for 2023 and 2026 based on the base year inventory described above. For both the 2023 and 2026 projected emissions inventories, 2020 emissions were projected forward respectively to 2023 and 2026 proportionally using relevant 2020 to 2023 and 2020 to 2026 activity data ratios. Both the base year and projected year emissions inventories were adjusted to account for local administrative rules, including gas pump emissions in anticipation of the passage of proposed rule R307-328. Because R307-328 was not implemented until 2023, gas pump emissions for base year 2017 and projection year 2023 are uncontrolled. For 2026, it was assumed that R307-328 was fully phased in and gas pump emissions were adjusted accordingly. Gas pump emissions for 2026 are subject to change depending on whether or not the rule is passed. Counties Included in Base and Projected Year Inventories Data from all counties in Utah were utilized to calculate emissions for the base and projected year inventories. These include Beaver, Box Elder, Cache, Carbon, Daggett, Davis, Duchesne, Emery, Garfield, Grand, Iron, Juab, Kane, Millard, Morgan, Piute, Rich, Salt Lake, San Juan, Sanpete, Sevier, Summit, Tooele, Uintah, Utah, Wasatch, Washington, Wayne, and Weber Counties. DRAFT Summary of Pre-SMOKE Emissions of Primary CAP Pollutants The baseline R script is included in this submission along with several other spreadsheets and source files. These include spreadsheets with activity and emissions data used in the emissions calculation process and source files for the activity and emissions data. These files were used to generate outputs of pre-SMOKE emissions data of primary CAPs in R. The R script and associated output spreadsheets are available upon request. Appendices 1, 3 and 5 (attached at the end of this document) show a summary of emissions by county of CO, NOx, NH3, SO2, PM10, PM2.5, and VOCs for the 2017 base year and 2023 and 2026 projection year inventories. Appendices 2, 4 and 6 (also attached at the end of this document) include summaries of VOC emissions by county associated with solvent emissions. Please note that hose pump emissions are only available for applicable counties at this time including Salt Lake, Davis, Weber, Tooele, and Utah counties. Note that any missing Federal Information Processing Standards (FIPS) codes—numbers that uniquely identify geographic areas, such as countries, states, and counties—in the output used for SMOKE mean that there were zero emissions for any pollutants in the county represented by the missing FIPS code, and any blank rows/columns indicate that there either are no emissions for the process represented by the respective Source Classification Code or that emissions data is unavailable for the pollutant associated with that process. Tools/Software Used to Calculate Emissions R version 4.4.0 was used to calculate the majority of emissions for the base and projected year area source inventories. Excel 365 was used to calculate emissions for the base and projected year inventories, and Excel workbooks were also prepared using Excel 365 to read data into R. Emissions Calculation Methodology, Activity Data, Controls, and Emission Factors Emissions from area sources are nearly always estimated employing a surrogate calculation procedure (i.e. a proxy). Direct measurement of area source emissions is hardly ever practical because of technical and cost considerations given the scope and scale of these aggregated emission sources. It is important to note that evolving methodologies for the calculation of area source emissions results in ever present changes in estimated emissions. Therefore, comparison between past or future SIPs or National Emission Inventory datasets should be made with caution, particularly in nonpoint inventories, given the high degree of improvement including changes in calculation methodologies, emission factors, and apportionment methods developed in a triennial cycle. There are four basic approaches for developing an area source emission estimate: 1) extrapolation from a sample set of the sources (surveys, permit files, or other databases); 2) material balance methods; 3) mathematical models; and 4) emission factors applied to activity data. The calculation procedures determine what data is used to estimate the area source emissions. A list of the individual data tables and sources of the data used in the calculation processes are included in the attached R script. As previously noted, activity data examples include census-based population estimates, employment figures (specified to relevant NAICS codes), livestock census counts, and fuel and chemical consumption estimates. These data are used in the calculations to estimate emissions for area sources. DRAFT Control strategy projections are estimates (see section describing controls applied for more details) of future year emissions that also include the expected impact of modified or additional control regulations. State and local planners should determine if any future scheduled regulations, whether at the federal, state, or local level, apply to sources in their area. Future year emissions may also be affected by fuel switching, fuel efficiency improvements, improvements in performance due to economic influences, or any occurrence that alters the emission producing process. Programs other than those aimed at reducing the emissions of the criteria pollutants of interest may affect the future year emissions. These may include energy efficiency programs and pollution prevention programs. These should all be reflected in the projections through the future year control factor, emission factor, or in some cases, by adjusting the activity growth forecast. Control factors and emission factors vary by source category and are continuously being revised and improved based on field and laboratory measurements. States also examine the future year control factors or emission factors in relation to the base year value to ensure any existing controls are not double-counted by taking additional credit in the future year. Overview of Projection Methods Emission projections for area sources depend upon the change in source level activity and changes in the emission factor applicable to the source. For area sources, the most appropriate equation used to project emissions is: Efy = Eby * G * C Where: Efy = projection year emissions Eby = base year emissions G = growth factor C = control factor, accounting for changes in emission factors or controls The base year activity (fuel use, employment, population) will vary depending on the source category. The growth activity type should align with the base year activity type as closely as possible. The above equation is only an example of the necessary calculation for emission projections; further complicating factors required for an accurate projection may require the development of a more vigorous equation. As with point sources, area source projections can be made using local studies or surveys or through surrogate growth indicators, to approximate the rise and fall in expected activity. The most commonly used surrogate growth indicators are those parameters typically projected by various census data sources such as population, housing, land use, and employment. Area sources rarely have detailed information based on surveys of individual emitters. Generally surrogate growth rates, as characterized by source type, must be used. While surrogate growth indicators such as employment and population are reasonable estimators of future air pollution-generating activity for traditional area source emitters (manufacturing, population-based activities), other indicators may be more appropriate for non-traditional emitters. Policy changes which DRAFT may lead to increased or decreased activity in a category must also be considered. For example, future emissions from agricultural tilling will be affected by trends towards conservation tillage as well as total acres tilled. Projections of total acres tilled may not trend with agricultural earnings as operations due to changes in crop yields. The amount of prescribed burning which takes place each year is driven by the policy of federal and state forests and land management agencies. The projection year control factor for area sources should account for both changes in emissions due to new levels of control required by federal, state, and local regulations and process modifications or technology improvements. Emitters in the manufacturing sector, such as industrial, commercial, and institutional fuel combustion, may be assigned a traditional control measure to limit emissions. However, for many area sources, conventional control methods are often inapplicable; instead, control of area source emissions may involve process modifications such as limiting agricultural burning practices, paving with emulsified asphalt or concrete, or stabilizing dirt roads. The control factors should also account for market-driven process changes, such as the move toward lower-solvent or water-based paints (this can be both market and regulatory-driven) and conservation tillage. It is noteworthy that spatial issues may also impact area source projections. Urban sprawl may result in decreases in area source emissions related to farming, such as agricultural tillage and managed burning. Conversely, urban sprawl may then result in increases in other area source emissions associated with residential areas, such as dry cleaning and consumer solvent use. Notes explaining how the area source emissions were calculated are included in the attached R script. Typically, activity data is multiplied by a relevant emission factor for a given pollutant to estimate the tons emitted per year for that pollutant. Controls measures may also be applied to the emission factor itself or by reducing the emissions by the relevant percent control factor. The R script calculates the emissions and contains a list of assumptions, emission factors, controls, equations and references for specific emissions categories and is included in this submission. The notes also explain the activity data that was used for the emissions calculations and how the data was projected. As mentioned above, the R script is available upon request. Calculation Methodology for Solvent Emissions Emissions associated with solvent categories were calculated and input into SMOKE separately from the rest of the area source inventory. This was done because solvent processes generate a large portion of the total VOC emissions in the area source inventory, and EPA has in recent years shifted to using a new python-based framework for Volatile Chemical Product emissions called VCPy. This new framework combined in EPA’s Wagon Wheel tool performs point source subtractions for solvents on behalf of the states whereas before EPA would leave it to the states to perform point source subtractions for solvent emissions themselves. As with other area source categories, emissions for solvents were calculated by multiplying population or employment data by relevant emissions factors, subtracting point source emissions, and apply relevant controls. Baseline emissions calculations and point source subtractions for most Utah solvent categories were performed by EPA through the VCPy framework in the Wagon Wheel tool. UDAQ then manually applied relevant controls to solvent categories to adjust emissions accordingly with state programs and regulations. DRAFT Summary of Controls Applied Over time, UDAQ has adopted and implemented various programs and regulations to reduce emissions associated with specific emissions categories. These programs and regulations, referred to as “rules,” have been adopted to reduce emissions that previously exceeded Clean Air Act (CAA) thresholds. Rules are generally applied to specific counties that exceed CAA thresholds rather than the state as a whole. Each rule specifies a target by which UDAQ would like to reduce emissions associated with the relevant emissions category, which is used to derive a control factor for use in the area source inventory. Uncontrolled emissions are multiplied by the relevant control proportions (which are computed by subtracting the control factor from one). Controls are generally phased in over time by a specific percent per year until they are 100% phased in. A controls spreadsheet was developed in Excel that was used to apply specific controls in R. The controls spreadsheet and associated R script describing specific controls are available upon request. Please see Appendix 7 at the end of this document for a snapshot of some of the controls that were applied in the development of the area source inventory. Methodology Changes/Updates Since the NWF moderate SIP and 2017 NEI, the EPA has made changes to the calculation methodologies for the emissions of several area source categories. These include changes to the solvents, agriculture field burning, agricultural silage, livestock waste, and crops and livestock dust categories. These changes are explained in detail in EPA’s 2020 NEI TSD. In addition, UDAQ decided to adopt a new calculation methodology for commercial/institutional and industrial natural gas combustion emissions (the calculation spreadsheet which was incorporated into the R script is available upon request). Previously, UDAQ calculated commercial/institutional and industrial emissions resulting from natural gas combustion by using emissions factors and the emissions calculation methodology based on NOMAD’s 2014 ICI Combustion Tool and using natural gas consumption estimates provided by the natural gas distributor. The NOMAD method is explained in detail in EPA’s 2020 NEI TSD. Now, UDAQ estimates emissions using a new method shown in Equations 1 and 2 (which use NOx as an example but other relevant CAP pollutants were also calculated by the same methods employing their specific emission factor). This methodology was previously implemented by the South Coast and San Joaquin Air Districts and uses boiler data provided by the Utah Labor Commission which was then analyzed to categorize boilers by capacity and application (commercial/institutional or industrial). A new emissions factor representing current emissions per unit was derived by dividing an assumed uncontrolled emissions threshold of 80 ppmv by 833 (ppm-mmbtu/lbs). Equation 1: NOx emissions per day = Emission Factor (80 ppmv NOx/833 ppm-mmbtu/lbs) x 24 hours x Total Boiler Capacity (7053 MMBtu/hr) x Operating Capacity (0.75)/2000 lbs = 6.10 tons NOx/day; Equation 2: NOx emissions per year = 6.10 tons NOx/day x 365 days = 2,225.18 tons NOx/year. DRAFT Development and Result of Additional Products of Serious SIP Inventory Aside from the methodology changes described above, UDAQ decided to implement a new methodology to account for emissions resulting from the operation of regular and low permeation gasoline hoses in Utah. These emissions were previously unaccounted for in the area source inventory, and EPA has not formally assigned a Source Classification Code to this category. Data on the number of pumps in the Utah Non-Attainment Areas as well as emissions factors for regular hoses and low permeation hoses were provided to UDAQ by the Utah Department of Agriculture. The analysis was only performed for Non-Attainment Area counties as data on the number of pumps not within the Non-Attainment Areas is not available at this time. To calculate emissions for each gas pump, the number of gas pump hoses in each Non-Attainment Area was multiplied by the appropriate emissions factor for the analysis (i.e., VOC Emissions Per Year = 1,734 Hoses in Davis County x 5.48 lbs VOC/hose/year). It was also assumed that there were 2 hoses per gas pump. For 2017 and 2023 emissions, it was assumed that controls from the proposed gas rule had not yet been implemented, so an emissions factor for regular gas pump hoses was used. For 2026 emissions, it was assumed that the proposed gas rule had been adopted and implemented by the Utah Air Quality Board, so an emissions factor for low permeation hoses was used. Although the analysis assumed controls would be implemented in 2026, an uncontrolled emissions scenario for 2026 was also calculated in case the proposed rule is not passed. The analysis was performed in Excel 365, and the Excel workbook is available upon request. Quality Control Methods and Results The data distributions for the 2017 base year as well as the 2023 and 2026 projection year area source inventories from the Serious SIP were compared against the data distributions from the 2017 base year and 2023 projection year from the Moderate SIP as well as 2017 NEI data and 2022 EPA Emissions Modeling Platform emissions estimates. This was done using a separate R script that had been designed to check specifically for any significant differences between the data distributions of the aforementioned data sets and the Serious SIP area source data. Also, bar charts were created in R showing the total emissions of each pollutant by Source Classification Code to check for any egregious data outliers. Overall, the quality control process showed that emissions in the Serious SIP inventories were similar to what was shown in the Moderate SIP inventories, differences between various data sets could be accounted for, and that the distribution of emissions was consistent with what was expected. DRAFT Appendices DRAFT Appendix 1: 2017 Base Year Emissions in Tons/Year (Not Including Solvent Emissions) *Gas hose emissions only include NWF Non-Attainment Counties and Utah county; data for counties outside these are not available right now. DRAFT Appendix 2: 2017 Base Year Solvent Emissions in Tons/Year DRAFT Appendix 3: 2023 Projection Year Emissions in Tons/Year (Not Including Solvent Emissions) *Gas hose emissions only include NWF Non-Attainment Counties and Utah county; data for counties outside these are not available right now. DRAFT Appendix 4: 2023 Projection Year Solvent Emissions in Tons/Year DRAFT Appendix 5: 2026 Projection Year Emissions in Tons/Year (Not Including Solvent Emissions) DRAFT *Gas hose emissions only include NWF Non-Attainment Counties and Utah county; data for counties outside these are not available right now. DRAFT Appendix 6: 2026 Projection Year Solvent Emissions in Tons/Year DRAFT Appendix 7: Controls Applied