2012 and 2018 Emissions Updates for the CAPCOG Region and Milam Counties

Transcription

1 2012 and 2018 Emissions Updates for the CAPCOG Region and Milam Counties Prepared by the Capital Area Council of Governments, December 2013 FY Rider 8 Near-Nonattainment Grant, Task 3.1 Prepared in Cooperation with the Texas Commission on Environmental Quality The preparation of this report was financed through grants from the State of Texas through the Texas Commission on Environmental Quality

2 Table of Contents Section 1 Introduction... 9 Section 2 Point Source Updates Section 3 Area Source Updates Section 3.1 Industrial Fuel Combustion Section Manufacturing Fuel Consumption Section Non-Manufacturing Industrial Fuel Consumption Section Emissions Calculations and Summaries Section Industrial Fuel Combustion Spatial Allocation Improvements Section 3.2 Commercial Fuel Combustion Section Statewide Commercial Fuel Consumption Totals Section Mobile Source Fuel Consumption Subtractions Section Growth Factors Section Allocation of Statewide Consumption to County Levels Section Point Source Subtractions Section Annual Fuel Consumption Totals for 2012 and Section Temporal Allocation Section Emissions Totals Section 3.3 Selected Oil and Gas Production Equipment Section Artificial Lift/Pumpjack Equipment Section Compressor Engines Section Heaters Section Growth Projections to Section Emission Summaries Section Spatial Allocation Section 4 On-Road Source Updates Section 4.1 Link-Based Updates for Austin Round Rock MSA Section 4.2 Extended Idling Section Frontage Road Extended Idling Section Total Extended Idling Hours Section Emissions Estimates Section Hourly Profiles Section Spatial Allocation Section 5 Non-Road Source Updates Section 5.1 Agricultural Equipment Section Equipment Population, Fuel Type, and Horsepower Updates Section Activity Updates Section Growth and Scrappage Updates Section Emissions Modeling and Calculations Section Agricultural Equipment Spatial Allocation Improvements Section 5.2 Construction and Mining Equipment Section Mining and Quarry Operations DCE Subsector Section Heavy-Highway Construction DCE Subsector Section Landfill Operations DCE Subsector Section Remaining Construction Equipment Subsectors Section Notes on Problems with TexN for Construction and Mining Equipment Section 5.3 Industrial Equipment

3 Section Population and Engine Profile Updates for Aerial Lifts, Forklifts, and Sweepers/Scrubbers Section Transportation Refrigeration Unit Population and Engine Profile Updates Section Terminal Tractors Population Updates Section Growth Factors Section Annual Activity Updates Section Updated Temporal Allocation Profile for LPG Forklifts Section 5.4 Residential Lawn and Garden Equipment Section Equipment Population Updates Section Annual Activity Updates Section Monthly Allocation of Residential Lawn and Garden Activity Section Weekday/Weekend Allocation of Residential Lawn and Garden Activity Section 5.5 Aviation: Austin-Bergstrom International Airport Appendix A: Electronic Files Submitted Appendix B: Oil and Gas Active Well Counts and Production Data Appendix C: CAMPO Contract for EPS3 File Preparation for On-Road Link-Based Emissions Inventories Appendix D: Supplemental Data for Mine and Quarry DCE Subsector Emissions Inventory Figure 1: Default and Updated Point Source NOX Emissions by County Figure 2: Comparison of Nationwide "Industrial" and Manufacturing Natural Gas Consumption Figure 3: Spatial Allocation of Industrial Natural Gas Combustion Figure 4: Spatial Allocation of Industrial Distillate Fuel Oil Combustion Figure 5: Spatial Allocation of Industrial LPG Combustion Figure 6: Spatial Allocation of Residual Fuel Oil Consumption Figure 7: PAVE Plot of Industrial Fuel Combustion NO X Emissions, Figure 8: Comparison of Growth of Oil and Gas Production in Caldwell and Fayette Counties, Figure 9: PAVE Plot of Oil and Gas Production NOX Emissions, Figure 10: Default and Updated On-Road Start and Running NO X Emissions for 2012 and 2018 by County (tpd) Figure 11: PAVE Plot of On-Road Exhaust NO X Emissions, Figure 12: PAVE Plot of Off-Network Start NO X Emissions, Figure 13: PAVE Plot for Extended Idling NO X Emissions, Figure 14: Default and Updated Agricultural Equipment NO X Emissions by County Figure 15: 2012 and 2018 Default and Updated Region-Wide Agricultural Tractor Equipment Populations Figure 163: 2012 and 2018 Default and Updated Region-Wide Combine Equipment Populations Figure 17: Engine-Powered and Electric-Powered Irrigation Pumps in Texas by Year Figure 18: 2012 and 2018 Default and Updated Irrigation Set Equipment Populations Figure 19: Cattle Inventories in CAPCOG Program Area as of January 1st, Figure 20: 2012 and Wheel Tractor, Baler, Mower, Sprayer, Tiller, and Swather Equipment Populations Figure 21: Head of Cattle as of January 1, Figure 22: Acres of Cotton Harvested by Year Figure 23: Comparison of Spark-Ignition (S.I.) and Compression-Ignition (C.I.) Annual Activity Estimates (hours/year) Figure 24: Activity Levels for "Other Agricultural Equipment"

4 Figure 25: Default TexN and Updated Irrigation Set Activity Levels (hours/year) Figure 26: Cumulative Age Distribution of HP Tractors by Model Year Figure 27: Default and Updated Agricultural Equipment NO X Emissions by County (tons per day) Figure 28: CropScape Land Cover Data for Travis County, Figure 29: Example of Attribute Table with Gridcode Counts From Isectpolyrst Script. (Only three columns of gridcodes are shown) Figure 30: PAVE Plot of Agricultural Tractor NO X Emissions, Figure 31: Default and Updated Region-Wide Construction and Mining Equipment Emissions Figure 32: Default and Updated Annual Hours of Use for Mine and Quarry Operations DCE Figure 33: Default and Updated Average Horsepower Ratings for Mine and Quarry Operations DCE Figure 34: PAVE Plot for Mine and Quarry Operations DCE Subsector NO X Emissions, Figure 35: PAVE Plot for Heavy Highway Construction DCE Subsector NO X Emissions, Figure 36: PAVE Plot for Landfill Operations DCE Subsector NO X Emissions, Figure 37: Non-Zero NR.BMX files Produced from Default 2012 Heavy Highway Construction DCE (#9) Subsector Run Figure 38: Non-Zero NR.BMX files Produced from Default 2012 Mine and Quarry Operations DCE (#23) Subsector Run Figure 39: Non-Zero NR.BMX files Produced from Default 2012 Landfill Operations DCE (#10) Subsector Run Figure 40: Screenshot of MySQL TexN Database Table "SCCDCEXWALK" Figure 41: Default and Updated Industrial Equipment Ozone Season Weekday Emissions for 2012 and 2018 (tons per day) Figure 42: Default and Updated Aerial Lift, Forklift, and Sweeper Equipment Populations for the Austin- Round Rock MSA, Figure 43: Default and Updated Engine Type Distribution for Aerial Lifts, Figure 44: Default and Updated Residential Lawn and Garden Equipment Ozone Season Weekday Emissions for 2012 and 2018 (tons per day) Figure 45: Default and Updated Lawn and Garden Equipment Annual Activity (hours per year) Figure 46: Default and Updated Monthly Allocation of Residential Lawn and Garden Equipment Activity (except Chainsaws) Figure 47: Default and Updated Weekday/Weekend Allocation of Residential Lawn and Garden Activity Table 1: CAPCOG Emissions Inventory Updates for 2012 and 2018 Modeling Scenarios... 9 Table 2: Updated Point Source Emissions by County for 2012 and 2018 (tons per day) Table 3: Comparison of Assumptions for Default and Updated Point Source Emissions Estimates Table 4: Example of Calculation of Industrial Fuel Consumption Rates using MECS and CBP Data Table 5: Comparison of Reported and Calculated Nationwide Fuel Consumption Totals (trillion BTU) Table 6: Errors in Calculated County-Level Employment at the 6-Digit NAICS Code Level Table 7: Estimated Point Source Manufacturing Employees Table 8: 2011 Industrial Kerosene Consumption by County Table 9: Estimated Area Source Manufacturing Fuel Consumption by County, 2011 (MMBTU) Table 10: Estimated Nationwide Natural Gas Consumption and Employment in Selected Mining Subsectors, Table 11: Expenses and Consumption of Natural Gas for Agricultural Irrigation, Table 12: Agricultural Area Source Industrial Natural Gas Combustion Table 13: Agricultural Area Source Industrial Natural Gas Combustion

5 Table 14: Projected Industrial Fuel Consumption for the West South Central Census Region (quadrillion BTU) Table 15: Texas Industrial Gas Consumed by Month, 2012 (MMCF) Table 16: Default and Updated Weekly Allocation of Activity Table 17: Emission Factors and Heat Conversion Rates Table 18: 2012 Industrial Fuel Combustion Ozone Season Weekday Emissions by Fuel Type (tons per day) Table 19: 2018 Industrial Fuel Combustion Ozone Season Weekday Emissions by Fuel Type (tons per day) Table 20: Updated Industrial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) Table 21: Default Industrial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) Table 22: Difference for Industrial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) Table 23: Statewide Fuel Consumption Estimates for the "Commercial" Sector Table 24: EPA Nationwide Non-Road Fuel Consumption Subtractions Table 25: Forklift Sales by SIC Code for Austin-Round Rock MSA Table 26: Statewide Fuel Consumption Estimates for the "Commercial" Sector with Non-Road Adjustments Table 27: EIA Commercial Sector Fuel Consumption Estimates for 2011, 2012, and 2018 (quadrillion BTU) Table 28: Statewide Fuel Consumption Estimates for the "Commercial" Sector with Non-Road Adjustments Table 29: Commercial Sector Employment by County, Table 30: Commercial and Institutional Point Source Employment Estimates Table 31: Commercial Sector Employment by County with Point Source Subtractions, Table 32: 2012 Commercial and Institutional Fuel Consumption Totals Table 33: 2018 Commercial and Institutional Fuel Consumption Totals Table 34: Monthly Consumption of Commercial Sector Natural Gas in Texas, Table 35: Commercial Sector Weekly Allocation Factors Table 36: EPA Commercial and Institutional Fuel Combustion Emissions Factors (lbs per unit of fuel consumed) Table 37: Industrial Fuel Combustion Ozone Season Weekday Emissions by Fuel Type (tons per day) Table 38: Updated Commercial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) Table 39: Default Commercial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) Table 40: Difference for Commercial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) Table 41: Source Classification Codes for Targeted Oil & Gas Production Equipment Emissions Updates 45 Table 42: Regular Producing Oil Wells by County, September Table 43: Drilling Completions for Oil Wells September August Table 44: Fractions of Oil Wells Requiring Artificial Lift in RRC Districts 1 and 3, Table 45: Artificial Lift/Pumpjack Engine Profile and Associated Emission Factors Table 46: 2012 Annual and Ozone Season Day Artificial Lift/Pumpjack Emissions (tons per day) Table 47: 2012 Gas Production Totals, Including Casinghead Gas and Gas Well Gas (MCF) Table 48: Drilling Completions for Gas Wells September August

6 Table 49: Gas Wells Completed in 2012 and Active September 2012 for RRC Districts 1 and Table 50: Compression Requirements from Previous Studies by Region of Texas Table 51: Compressor Engine Makes and Models with SCCs, HP Ratings, % of Load, and NO X, CO, and VOC Factors (g/hp-hr) Table 52: Compressor Engine Makes/Models with Multiple SCC Listings Table 53: Compressor Engine Emissions by County, Table 54: Compressor Engine Emissions by SCC, Table 55: Heater/Treater Emissions Factors Table 56: CENRAP Weighted Averages of Texas Basin Level Data Table 57: Active Oil and Gas Wells by County, September Table 58: 2012 Ozone Season Day Heater-Treater Emissions at Oil and Gas Wells by County (tons per day) Table 59: Oil and Gas Production Equipment Emissions Growth Factors Table 60: Updated Pumpjack, Compressor, and Heater Ozone Season Weekday Emissions by County (tons per day) Table 61: Default Pumpjack, Compressor, and Heater Ozone Season Weekday Emissions by County (tons per day) Table 62: Difference for Pumpjack, Compressor, and Heater Ozone Season Weekday Emissions by County (tons per day) Table 63: Austin-Round Rock MSA Weekday Start, Running, and Evaporative CO, NO X, and VOC for 2012 and 2018 (tons per day) Table 64: Comparison of Basis Used for Default and Updated Emissions Estimates Table 65: Truck Parking Capacity by County, Table 66: Extended Idling Rates for Truck Stops and Other Facilities (idling hours per parking space) Table 67: Frontage Road Idling Rates by Day Type (Idling Hours per Mile of Interstate Frontage Road).. 69 Table 68: Two-Way Interstate Frontage Road Mileage Suitable for Idling Table 69: Extended Idling Hours per Day by Day Type and County for 2012 and Growth Factor for Table 70: Extended Idling Emissions Rates for CO, NO X, NO, NO 2, HONO, and VOC (grams per hour) Table 71: 2012 and 2018 Summer Weekday (Monday Thursday) Extended Idling Emissions by County (tpd) Table 72: Hourly Distribution of Idling by Day Type Table 73: Summary of Agricultural Equipment Updates Table 74: Updated Weekday Agricultural Equipment Emissions by County for 2012 and 2018 (tons per day) Table 75: Default Weekday Agricultural Equipment Emissions by County for 2012 and 2018 (tons per day) Table 76: Difference Between Default and Updated Weekday Agricultural Equipment Emissions (from Default to Updated Rates) by County for 2012 and 2018 (tons per day) Table 77: 2002, 2007, 2012 and 2018 Tractor Populations for <40 HP Tractors Table 78: 2002, 2007, 2012 and 2018 Tractor Populations for HP Tractors Table 79: 2002, 2007, 2012 and 2018 Tractor Populations for 100+ HP Tractors Table 80: Allocation of Tractor Populations to Fuel Types Table 81: Default and Updated Allocations of <40 HP, HP, and 100+ HP Tractor Populations to Horsepower Bins Table 82: Default and Surveyed Average Horsepower Ratings Table 83: 2002, 2007, 2012, and 2018 Combine Equipment Populations Table 84: Statewide Distribution of Combines by Fuel Type in

7 Table 85: Statewide Counts of Farm and Ranch Irrigation Pumps by Fuel Type: 1988, 1994, 1998, 2003, and Table 86: Projected Statewide Irrigation Set Populations by Fuel Type, 2012 and Table 87: Allocation of Statewide Irrigation Set Populations (all fuel types) to Counties Table 88: 2008 Allocation of Irrigation Sets by Fuel Type Table 89: Pechan Equipment Ratios for Two-Wheel Tractors, Balers, Mowers, Sprayers, Tillers, and Swathers by Farm Type Table 90: Farms, Forage Acreage, and Cattle and Calves Inventories for Selected Farm Types, Statewide Table 91: Pechan Mower Equipment Ratios and Derived Statewide Equipment Populations for 1987 and 1992 for Texas Table 92: Comparison of Survey-Derived Statewide Mower Populations and Census of Agriculture Populations Table 93: Comparison of Hay Baler and Agricultural Tractor Fuel Type Distributions in Pechan Study Table 94: Baseline and Projected Cattle Populations by Year, 2007, 2012, and Table 95: Forage Harvester Equipment Population by County and Year Table 96: 2007, 2012, and 2018 Cotton Picker and Stripper Populations by County Table 97: 2012 "Other Agricultural Equipment" Populations Table 98: 2018 "Other Agricultural Equipment" Populations Table 99: Texas Irrigation Fuel Expenses, Prices, and Quantity Consumed, Table 100: Comparison of Updated and 8760 Hour/Year Irrigation Fuel Consumption Rates (gallons per piece per year) Table 101: Growth Factors Used in Agricultural Equipment Emissions Modeling Table 102: Chi-Squared Test of Independence for HP Tractor Age Distributions Table 103: 2012 and 2018 Emissions by County for Typical Ozone Season Weekday (tons per day) Table 104: 2012 and 2018 Emissions by Equipment Type for Typical Ozone Season Weekday (tons per day) Table 105: Agricultural Equipment Land Use Assignments for Spatial Allocation Table 106: Summary of DCE Subsector Emissions Estimates and Spatial Allocation Updates Table 107: Default and Updated Region-Wide Construction and Mining Equipment Emissions (tons per ozone season day) Table 108: Mine and Quarry Diesel Construction Equipment Population Ratios (Pieces per 10,000 Annual Labor-Hours) Table 109: Active Mines and Quarries and Pit Labor-Hours, Table 110: Mine and Quarry Growth Factors by County Table 111: 2012 Mine and Quarry Equipment Populations by County Table 112: 2018 Mine and Quarry Equipment Populations by County Table 113: Default Mine and Quarry Operations DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) Table 114: Updated Mine and Quarry Operations DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) Table 115: Difference Between Updated and Default Mine and Quarry Operations DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) Table 116: Heavy Highway Construction Equipment Profiles (Piece-Days per $M or Lane-Mile of Construction) Table 117: Default Heavy Highway Construction DCE Summer Weekday Emissions, 2012 and 2018 (tons per day)

8 Table 118: Updated Heavy Highway Construction DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) Table 119: Difference Heavy Highway Construction DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) Table 120: Landfill Equipment DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) Table 121: Landfill Equipment DCE Spatial Allocation for 2012 and Table 122: Remaining Construction Subsector Summer Weekday Emissions, 2012 and 2018 (tons per day) Table 123: Industrial Equipment Updates Table 124: Aerial Lifts, Forklifts, and Sweepers/Scrubbers Engine Type Distributions Table 125: Default and ENVIRON Aerial Lift HP Bin Distributions by Engine Type Table 126: ERG-Updated 2001 Statewide Transportation Refrigeration Unit Population by HP Range Table 127: Industrial Equipment Annual Activity by SCC and Fuel Type Table 128: Industrial Equipment Weekdays and Weekend Day Activity Allocations in Season.dat file Table 129: Residential Lawn and Garden Equipment Ratios (pieces/single-family detached unit household) Table 130: Base Year Housing and Population Data Table 131: Estimated Growth of SFDUs by County from Base to 2012 and Table 132: Single Family Detached Units by County in 2012 and Table 133: Engine Type and Horsepower Allocation of Residential Lawn and Garden Equipment, 2012 and Table 134: Landings and Take-Offs at Austin-Bergstrom International Airport by Activity Type Table 135: 2012 Daily ABIA Activity (LTO) and Emissions (tons per day) Table 136: 2018 Daily ABIA Activity (LTO) and Emissions (tons per day) Table 137: Oil and Gas Well Counts and Production Data by County, Table 138: OIl and Gas Well Counts and Production Data by County, Table 139: Oil and Gas Well Counts and Production Data by County, Table 140: OIl and Gas Well Counts and Production Data by County, Table 141: Oil and Gas Well Counts and Production Data by County, Table 142: Oil and Gas Well Counts and Production Data by County, Table 143: Oil and Gas Well Counts and Production Data by County, Table 144: Mine and Quarry Diesel Construction Equipment Average Hours Per Year Table 145: Mine and Quarry Diesel Construction Equipment Average Horsepower Ratings Table 146: Mine and Quarry-Level Production Data for

9 Section 1 Introduction The Capital Area Council of Governments (CAPCOG) has developed new emissions estimates for 2012 and 2018 and spatial allocation data that can be used as part of a photochemical modeling project that is jointly being undertaken by CAPCOG and the Alamo Area Council of Governments. This project involves using the June 2006 photochemical modeling episode 1 created by the Texas Commission on Environmental Quality (TCEQ) with emissions inventory updates to represent 2012 and 2018 ozone season scenarios. CAPCOG submitted updated ozone season day estimates of carbon monoxide (CO), nitrogen oxides (NO X ), and volatile organic compounds (VOC) from point, area, on-road, and non-road inventories for all eleven counties in its Rider 8 program area. CAPCOG s Rider 8 program area includes the ten counties in the CAPCOG planning region Bastrop, Blanco, Burnet, Caldwell, Fayette, Hays, Lee, Llano, Travis, and Williamson Counties and Milam County. CAPCOG also submitted updated spatial allocation data for a number of sources. These updates take advantage of the most recent research, models, and activity data available for each source category. The table below summarizes the updates that were updated for this analysis. The table shows whether a category was updated for the Austin- Round Rock Metropolitan Statistical Area (MSA) and the other six counties in the region for 2012 and 2018, and whether any updates to the spatial surrogates were made. These updates replaced the emissions inventory data AACOG has already loaded into the 2012 and 2018 modeling files for preliminary modeling runs. A complete description of those updates is available from AACOG. This report describes the updates CAPCOG has made for each source category and provides comparisons to the data it is replacing. Table 1: CAPCOG Emissions Inventory Updates for 2012 and 2018 Modeling Scenarios Source Category 2012 MSA 2012 Other 2018 MSA 2018 Other Spatial Updates Point Sources Area Source Industrial Fuel Combustion Area Source Commercial Fuel Combustion Area Source Oil and Gas Production Equipment On-Road Start, Running, and Evaporative Emissions On-Road Extended Idling Non-Road Agricultural Equipment Non-Road Construction and Mining Equipment Mine and Quarry Equipment Non-Road Construction and Mining Equipment Heavy Highway Construction Non-Road Construction and Mining Equipment Landfill Equipment Non-Road Industrial Equipment Non-Road Residential Lawn and Garden Equipment Non-Road Aviation and Ground Support Equipment Austin-Bergstrom International Airport (ABIA) 1 ftp://amdaftp.tceq.texas.gov/pub/rider8/ 9

10 Section 2 Point Source Updates CAPCOG used TCEQ s 2011 point source emissions inventory data for ozone season day emissions estimates for all point sources in the CAPCOG region and Milam County other than Electrical Generating Units (EGUs) that report to the U.S. Environmental Protection Agency s (EPA s) Air Markets Program Data (AMPD) 2. For these EGUs, CAPCOG used the average daily NO X emissions reported for June, July, and August, 2012, and multiplied the ratios of each source s NO X to VOC and CO emissions as reported in TCEQ s 2011 point source emissions inventory. There was one unit (Decker Creek GT-3A) that did not have reported ozone season day emissions for 2011, but did have NO X emissions reported for the 2012 ozone season. For this source, CAPCOG multiplied the NO X emissions by the ratio of VOC and CO to NO X for its companion unit, GT-3B. Since these sources have data reported at the hourly level, CAPCOG calculated the average 24-hour diurnal profile for each unit based on its average NO X emissions each hour in June, July, and August, TCEQ staff confirmed that this was the most appropriate approach to establishing 2012 baseline emissions estimates for each point source. 3 One point source that is often identified as an EGU but which does not report to EPA s Acid Rain Database is the University of Texas at Austin s Hal Weaver Power plant. For this source, 2011 emissions were available from TCEQ s point source emissions inventory. Since this source does report fuel consumption and electrical generation data to the Energy Information Administration (EIA) under form EIA-923 4, CAPCOG calculated the 2012 emissions by multiplying the 2011 emissions by the ratio of reported fuel consumption in June, July, and August, 2012, to the reported fuel consumption in June, July, and August, For 2018 estimates, CAPCOG assumed that there would be no change in point source emissions. CAPCOG determined that, given the available data and time constraints for this project, that this would be the most appropriate assumption for 2018 point source emissions in the region. TCEQ staff also recommended this approach in a phone conversation prior to CAPCOG submitting the data to AACOG. 5 The table below shows the updated total daily NO X, VOC, and CO emissions from point sources in each of the 11 counties that are part of this study area. Table 2: Updated Point Source Emissions by County for 2012 and 2018 (tons per day) County CO NO X VOC Bastrop Blanco Burnet Caldwell Fayette Hays EPA. Air Markets Program Data. Last accessed July 11, Phone conference with Ron Thomas and Miranda Kosty, Texas Commission on Environmental Quality. May 13, EIA. Form EIA-923 Detailed Data with previous form data (EIA-906/920). Last accessed July 11, Phone conversation with Miranda Kosty, Texas Commission on Environmental Quality. June 28,

11 County CO NO X VOC Lee Llano Milam Travis Williamson TOTAL The existing point source emissions estimates are based on data used for the attainment demonstrations incorporated into the State Implementation Plan (SIP) for the Dallas-Fort Worth (DFW) 6 and the Houston-Galveston-Brazoria (HGB) ozone nonattainment areas. The DFW SIP revision was used for 2012 point source emissions, while the HGB SIP revision was used for the 2018 point source emissions. The DFW SIP revision was adopted by TCEQ on December 7, 2011, while the HGB SIP revision was adopted by TCEQ on March 10, The table below shows the differences in the basis of the emissions estimates for electricity generating units (EGUs) and non-electricity generating units (NEGUs). Table 3: Comparison of Assumptions for Default and Updated Point Source Emissions Estimates Sources 2012 DFW SIP HGB SIP and 2018 CAPCOG Update ARD EGUs Average of each hour for 3 rd Quarter 2008 Average of each hour for 3 rd Quarter 2007 Average of each hour for June August 2012 Non-ARD EGUs 2008 Point Source Emissions Inventory; Post-2008 EGUs with approved permits 2006 Point Source Emissions Inventory; Post EGUs with approved permits 2011 Point Source Emissions Inventory Non-EGUs 2008 Point Source Emissions Inventory Projected to Point Source Emissions Inventory Projected to Point Source Emissions Inventory TCEQ. Appendix B: Emissions Modeling for the DFW Attainment Demonstration SIP Revision for the 1997 Eight- Hour Ozone Standard. December 7, TCEQ. Appendix B: Emissions Modeling for the HGB Attainment Demonstration SIP Revision for the 1997 Eight- Hour Ozone Standard. March 10, Using the higher growth rate from of the Dallas Federal Reserve Bank s Texas Industrial Production Index (TIPI) and EPA s Economic Growth Analysis System version 5.0 (EGAS5). 10 Using the Dallas Federal Reserve Bank s Texas Industrial Production Index (TIPI) and EPA s Economic Growth Analysis System version 5.0 (EGAS5) for industries not covered by the TIPI. 11

12 NO X Emissions (tons per day) 2012 and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties The chart below shows the difference between the default and updated point source NO X emissions estimates for 2012 and 2018 for each county. Figure 1: Default and Updated Point Source NOX Emissions by County Default 2018 Default 2012 and 2018 Updated 12

13 Section 3 Area Source Updates CAPCOG updated the 2012 and 2018 emissions estimates for three categories of area source emissions: 1) industrial fuel combustion, 2) commercial fuel combustion, and 3) oil and gas production equipment. These source categories account for the vast majority of NO X emissions among the area sources in the region. Section 3.1 Industrial Fuel Combustion CAPCOG updated the 2012 and 2018 industrial fuel combustion estimates for all 11 counties in the program area. These updates rely on industry-specific fuel consumption factors derived from Energy Information Administration (EIA) and county-level manufacturing data from the County Business Patterns (CBP). The default method used by EPA for the 2008 and 2011 National Emissions Inventory (NEI) relies on allocating statewide consumption of natural gas, distillate fuel oil, residual fuel oil, and liquefied petroleum gas (LPG) down to the county level based on total manufacturing employment, after subtracting for point source fuel consumption, consumption for feedstock, and non-road fuel consumption. This approach does not account for substantial variations in the fuel intensity of manufacturing processes within the manufacturing sector, and it does not specifically account for nonmanufacturing fuel consumption that the EIA classifies as industrial. Under EIA s classification scheme, the industrial sector includes industries classified under the North American Industrial Classification System (NAICS) Codes 11 (Agriculture, Forestry, Fishing, and Hunting), 21 (Mining, Quarrying, and Oil and Gas Extraction), 23 (Construction), and (Manufacturing). CAPCOG s estimates include two segments: manufacturing establishments and agricultural establishments. CAPCOG separately estimated the fuel consumed in each county for each establishment type and then applied standard emissions factors from AP-42 to these activity estimates to obtain emissions estimates. Section Manufacturing Fuel Consumption According to the definition used by the EIA, manufacturing includes all establishments engaged in the mechanical or chemical transformation of materials or substances into new products. The EPA s method for estimating industrial fuel combustion implicitly assumes that all of the area source fuel consumption in the industrial sector was in the manufacturing sector. EPA s method allocates activity to each county based on total manufacturing employment; however, it does not capture the significant variations in fuel consumption rates for each type of manufacturing. CAPCOG s method for estimating emissions from this category seeks to account for these variations by developing NAICS-specific fuel consumption factors. Section Manufacturing Fuel Consumption and Employment by NAICS Code In order to estimate a baseline for manufacturing fuel combustion activity, CAPCOG used the Energy Information Administration s (EIA s) 2010 Manufacturing Energy Consumption Survey 11 (MECS) and employment data from the Census Bureau s County Business Patterns (CBP). 12 Since MECS reports fuel consumption for industries at a much more refined classification level than the statewide reporting of 11 EIA Manufacturing Energy Consumption Survey. Last accessed July 11, U.S. Census Bureau. County Business Patterns. Last accessed 13

14 fuel consumption for the industrial category under the State Energy Data System, it provides a method of estimating fuel consumption that accounts for the substantial variations in fuel consumption based on the mix of industrial sources within a region. CAPCOG estimated county-level activity by calculating national employment-based fuel consumption rates for each NAICS code reported by EIA. CAPCOG divided the fuel consumption reported for each NAICS code in Table 3.2: Energy Consumption as a Fuel by Manufacturing Industry & Region (trillion Btu) by the number of employees nationwide for 2010 in the CBP. Section Calculation of National Employment Figures for Undisclosed NAICS Codes For four industries (NAICS codes , , and , and ), the exact number of employees was not disclosed in the CBP based on the Census Bureau s rules for disclosure. For two of these (NAICS Codes and ), it was possible to calculate the total by subtracting the number of employees at the same level from the next-higher level NAICS code: For the two other NAICS codes, both shared the same classification at the next highest level, making it impossible to perform the above calculation. Instead, CAPCOG estimated the number of employees based on what information was provided for these industries. CBP reports the number of establishments for each NAICS code within the following ranges:,,. 1-4 employees, 5-9 employees, employees, employees, employees, employees, employees, employees, and 1,000 or more employees. CAPCOG calculated the maximum and minimum number of employees for NAICS code and for each of the three subordinate NAICS codes (325181, , and ). CAPCOG first multiplied the minimum value of each range by the number of establishments and then added each of those values together to obtain a minimum value for the whole NAICS code and did the same for the maximum values. Since there is no maximum for the 1,000 or more employees grouping, CAPCOG used 2,000 as the maximum value. The following series of equations shows the calculations for NAICS Code

15 ( ) The values of the midpoints for these four NAICS code were the following: CAPCOG then adjusted the midpoints for the two undisclosed NAICS codes ( and ) to reconcile them with the total number of employees reported in NAICS codes and CAPCOG did this by first subtracting the number of employees in NAICS code from the number in NAICS code to get the total sum of employees in NAICS codes and In order to estimate the number of employees for each of those NAICS codes, CAPCOG then multiplied each midpoint by the ratio of total employees for both NAICS codes to the sum of the midpoints for both NAICS codes. This procedure provided an estimate of the number of employees in each NAICS code that was consistent with the total number of employees in the higher-order NAICS code while reflecting what information was known about the relative number of employees between these two undisclosed NAICS codes. Section Calculation of Manufacturing Fuel Consumption Factors other than Kerosene For many of the industry types, more detailed data at the 4-digit, 5-digit, or 6-digit level were provided in the MECS. In order to accurately represent fuel consumption rates for a particular industry, it was necessary to subtract out the fuel consumption and employment data for lower-level NAICS codes in order to obtain the fuel consumption rates for the other industries within the higher-level NAICS code. For example, the table below shows the initial adjustment that was made to the natural gas fuel consumption rate for Computer and Electronic Products Manufacturing (NAICS Code 334) to account for out the more specific information available for Semiconductors and Related Devices Manufacturing (NAICS Code ). Table 4: Example of Calculation of Industrial Fuel Consumption Rates using MECS and CBP Data NAICS Code Natural Gas Consumed (trillion BTU) Employees, Fuel Consumption Rate (MMBTU/employee) , ,

16 NAICS Code Natural Gas Consumed (trillion BTU) Employees, 2010 Fuel Consumption Rate (MMBTU/employee) 334 minus , For some NAICS codes, the total quantity of fuel consumed was not reported, either because the reported value was less than 0.5 trillion BTU or because the Relative Standard Error for that NAICS/fuel type combination was greater than 50%. For these, CAPCOG set the fuel consumption value to zero. This assumption should not have any significant impact on emissions totals, since when those values are assumed to be zero, the calculated sum of each major area source fuel is more than 98% of the reported nation-wide total consumption of that fuel in the manufacturing sector. Table 5: Comparison of Reported and Calculated Nationwide Fuel Consumption Totals (trillion BTU) Fuel Reported Total Calculated Total Calculated % of Total Residual Fuel Oil % Distillate Fuel Oil % Natural Gas 5,211 5, % NGL and LPG % Other 5,059 5, % CAPCOG used the data in Table 3-6: Selected Wood and Wood-Related Products to separate out the energy consumption from biomass and pulping/black liquor from the other category. These two fuel types make up 630 and 824 trillion BTU of fuel consumption, respectively, accounting for a total of 1,454 trillion BTU of fuel consumption, primarily in the wood products and paper manufacturing sectors, with the rest in food, beverage, and furniture manufacturing. That still leaves 3,605 trillion BTU in other fuel, which includes net steam and other energy used to produce power and heat. A portion of that fuel was kerosene, which is reported statewide for the industrial sector in EIA s Adjusted Sales of Fuel Oil and Kerosene. Since there is more specific information available for that fuel type, CAPCOG addressed it differently, but CAPCOG assumed that all of the remaining fuel consumed was consumed by point sources or did not result in ozone-forming emissions (such as steam). Since the MECS also provides the fuel consumption by end use 13, it is also possible to subtract from the fuel consumption totals the amount of fuel consumed for non-road mobile equipment. The total reported distillate fuel oil consumption for onsite transportation was 39 trillion BTU, while reported NGL and LPG onsite transportation consumption was 17 trillion BTU. 14 CAPCOG subtracted the total fuel consumed for onsite transportation from each NAICS code s total fuel consumption in order to obtain the estimated fuel consumption in stationary equipment. While other equipment classified as 13 EIA Manufacturing Energy Consumption Survey. Table 5.2: Energy Consumed as a Fuel by End Use by Mfg. Industry with Net Electricity (trillion BTU). Last accessed July 15, EIA defines onsite transportation as, The direct nonprocess end use that includes energy used in vehicles and transportation equipment that primarily consume energy within the boundaries of the establishment. Energy used in vehicles that are found primarily offsite, such as delivery trucks, is not measured by the MECS. 16

17 non-road could be included among the other uses reported, onsite transportation was the only end use that was clearly mobile and would not be part of the area source emissions inventory. Similar to the treatment of unreported fuel consumption totals for total fuel consumption, CAPCOG assumed that any fuel consumption reported as less than 0.5 trillion BTU or with a RSE value of over 50% was zero. This assumption resulted in no difference between the calculated and reported total of distillate consumed for onsite transportation, and resulted in a calculated total for NGL/LPG of 15 trillion BTU compared to the reported total of 17 trillion BTU (88%). This process yielded nation-wide fuel consumption data and employment data for 81 distinct industry groupings. CAPCOG divided the 2010 fuel consumption by the 2010 employment to produce fuel consumption factors for each of the 81 industry groupings and each fuel type. Section County-Level Manufacturing Employment for 2011 In order to estimate county-level fuel consumption, CAPCOG applied the 2010 fuel consumption factors to 2011 manufacturing employment data for each of the 11 counties in CAPCOG s program area using the 2011 CBP. In many cases, the exact number of employees was not disclosed, but ranges were provided and the numbers of establishments in various employment ranges were provided. CAPCOG used the procedures described in Section : Calculation of National Employment Figures for Undisclosed NAICS Codes of this report in order to estimate county-level employment for each NAICS code or grouping that fuel consumption factors had been calculated for. One additional constraint was added to the minimum and maximum employment figures for each NAICS code: the employment ranges expressed in the CBP were added to the analysis, thereby further narrowing the maximum and minimum employment ranges. These ranges were designated by letters, which corresponded to the following values: a employees, b employees, c employees, d. N/A e employees, f employees, g. 1,000 2,499 employees, h. 2,500 4,999 employees, i. 5,000 9,999 employees, j. 10,000 24,999 employees, k. 25,000 49,999 employees, l. 50,000 99,999 employees, and m. 1,000,000 or more employees. As a quality check, CAPCOG compared the estimated totals back to the ranges of maximum and minimum values. This check revealed some errors, although it appears that at least some of the errors are due to inconsistencies between the CBP s reported employment ranges and the calculated ranges using the number establishments with varying levels of employment. These errors result in only very 17

18 minor differences with the total manufacturing employment for each county and could not be otherwise reconciled, so CAPCOG left them as calculated. Table 6: Errors in Calculated County-Level Employment at the 6-Digit NAICS Code Level County 6-Digit NAICS Codes Reported Errors at 6-digit level Total Employees Net Error Value Bastrop Blanco Burnet Caldwell <-1 Fayette Hays , Lee Llano <-1 Milam Travis , Williamson , Region Subtotal , For the whole region, there were errors in 7.7% of the 6-digit NAICS code/county records, accounting for a net difference of employees, which represent less than 1% of the total manufacturing employment in the region. This means that these errors whatever the source would not likely change the results by any more than 1% if resolved. In order to not double-count fuel consumed by industrial point sources, CAPCOG used the 2011 point source emissions inventory to identify all of the manufacturing point sources in the region. CAPCOG matched the Standard Industrial Classification (SIC) codes in the point source inventory to corresponding NAICS codes using the Census Bureau s NAICS/SIC crosswalk, and then assumed that the point source had the largest number of employees for that specific 6-digit NAICS code and subtracted the estimated employment for each point source from its county s employment in the corresponding NAICS code. For some cases, the point source was the only establishment in that NAICS code, in which case, CAPCOG set the employment to zero. For the others, CAPCOG used the midpoint and adjustment technique described in Section to estimate the employment for each facility. Table 7: Estimated Point Source Manufacturing Employees RN Site County SIC NAICS Estimated Employees RN Hanson Brick Bastrop RN Acme Brick Bastrop RN Texas Lehigh Hays RN LF Manufacturing Main Plant Lee RN LF Manufacturing QFRP Plant Lee RN Alcoa Inc. Milam RN APAC Texas Inc. Travis RN Austin Counter Tops Inc. Travis

19 RN Site County SIC NAICS Estimated Employees RN Austin White Lime Inc. Travis RN Freescale-Ed Bluestein Travis RN Spansion LLC Travis RN Freescale-Oak Hill Travis RN Samsung Travis ,750 CAPCOG multiplied the adjusted 2011 county-level employment data by the 2010 fuel consumption rates for each of the fuels reported in the MECS except for coal, petroleum coke and breeze, and nonbiomass other. CAPCOG assumed that only point sources consumed these other fuels, and as a result, their emissions would not be in the area source inventory. Section Manufacturing Kerosene Fuel Consumption While the MECS provides detailed data on consumption of distillate fuel oil, it does not specifically provide data for kerosene. The EIA s Adjusted Annual Sales of Fuel Oil and Kerosene, however, do provide state-level consumption of industrial kerosene consumption. In 2011, there were 45,145 thousand gallons of kerosene consumed in the industrial sector, according to the EIA s Adjusted Sales of Fuel Oil and Kerosene report. This report, unlike some others, does not include farm consumption. The industrial consumption translates into trillion BTU, making up only a small fraction of the other fuel reported consumed besides the selected wood and biomass fuels reported by EIA. This assumed a MMBTU/barrel thermal conversion rate, as reported in EIA s Conversion factors for the 2011 Annual Energy Review. 15 In Texas, non-farm industrial sources consumed, 22,557 thousand gallons of kerosene. CAPCOG allocated this statewide total to industrial area sources in each county based on the total number of employees in manufacturing in each county minus the estimated number of employees in industrial point sources in the county. Table 8: 2011 Industrial Kerosene Consumption by County County Manufacturing Employment 2011 Minus Point Sources Industrial Kerosene (1,000 gallons) Bastrop Blanco Burnet Caldwell Fayette Hays 3, Lee Llano 75 2 Milam U.S. Energy Information Administration. Annual Energy Review Appendix A: British Thermal Unit Conversion Factors. Table A1: Approximate Heat Content of Petroleum Products. 19

20 County Manufacturing Employment 2011 Minus Point Sources Industrial Kerosene (1,000 gallons) Travis 21, Williamson 6, SUBTOTAL 35,231 1,083 Section Summary of Manufacturing Area Source Fuel Consumption by County In order to calculate the total emissions from area source industrial fuel combustion in the manufacturing sector, it is necessary to add all of the fuel consumed for each NAICS code in each county. The following table shows the estimated consumption of fuel from area sources in the manufacturing sector in Table 9: Estimated Area Source Manufacturing Fuel Consumption by County, 2011 (MMBTU) County Distillate Oil Residual Oil Natural Gas LPG Kerosene Pulping Liquor Biomass Bastrop 4,455 1, ,772 1,816 3, ,951 Blanco 1, , ,198 Burnet 29,817 1, ,428 6,831 3, ,656 Caldwell 2, ,903 1,597 2, ,139 Fayette 4,776 2, ,527 2,156 3, ,626 Hays 62,661 11, ,884 7,207 13, ,004 Lee 1, ,488 1, Llano 1, , Milam 1, , , ,733 Travis 95,508 17,529 3,063,773 27,899 90, ,274 Williamson 36,870 5,175 1,220,261 9,886 26, ,817 TOTAL 241,767 41,542 5,974,652 59, , ,027 Section Non-Manufacturing Industrial Fuel Consumption The industrial sector as defined by the EIA encompasses not only manufacturing, but also farming, mining, and construction. If a farming, mining, or construction establishment operates permanent, stationary combustion equipment, assuming it was not a point source or accounted for in the oil and gas emissions inventory, it would be considered part of the area source industrial fuel combustion inventory. If the MECS data is compared to other data on industrial fuel consumption reported by the EIA, it becomes apparent that there are significant differences. For natural gas and kerosene, statewide data are available from the EIA that allow these emissions to be calculated apart from the fuel consumption in the manufacturing sector. Data is also available on distillate consumption in the farming, construction, and mining sectors, but there is not enough information available to determine whether any of the reported statewide consumption of distillate in these sectors includes combustion in stationary equipment or whether it is exclusively for non-road 20

21 Billion Cubic Feet 2012 and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties equipment use. For the purposes of this inventory, CAPCOG assumes that 100% of the distillate fuel consumed in these sectors is for non-road use. The chart below shows the total nation-wide quantity of natural gas consumed in the industrial sector as reported by EIA 16 and the total quantity of gas consumed in the manufacturing sector as reported in the EIA s MECS from , , , and In 2010, non-manufacturing consumption accounted for 18.49% of all industrial natural gas consumption nationwide. Figure 2: Comparison of Nationwide "Industrial" and Manufacturing Natural Gas Consumption 9,000 8,000 7,000 6,000 5,000 4,000 Industrial Manufacturing 3,000 2,000 1, Within the South Census Region, which Texas is in, there was a difference of 454,125 MMCF between the sum of all of the states reported industrial natural gas consumption for 2010 and the natural gas consumption reported for the South Census Region in the MECS, which amounts to 12.91% of all of the 16 U.S. Energy Information Administration. U.S. Natural Gas Industrial Consumption. Last accessed July 17, U.S. Energy Information Administration Manufacturing Energy Consumption Survey. Table 1-1: First Use of Energy for All Purposes (Fuel and Nonfuel). Last accessed July 17, U.S. Energy Information Administration Manufacturing Energy Consumption Survey. Table 1-1: Consumption of Energy for All Purposes (First Use) by Mfg. Industry & Region (physical units). Last accessed July 17, U.S. Energy Information Administration Manufacturing Energy Consumption Survey. Table 1-1: Consumption of Energy for All Purposes (First Use) by Mfg. Industry & Region (physical units). Last accessed July 17, U.S. Energy Information Administration Manufacturing Energy Consumption Survey. Table 1-1: Consumption of Energy for All Purposes (First Use) by Mfg. Industry & Region (physical units). Last accessed July 17,

22 natural gas consumed in the industrial sector for the region. Presumably, this natural gas is being consumed by other subsectors within the broader industrial classification, including farming, mining, and construction. Lease fuel and plant fuel are not counted in the annual reported industrial fuel consumption and are instead reported separately, and natural gas consumed in oil and gas production is accounted for in separate SCC codes. CAPCOG decided to assume that 100% of the residual natural gas fuel consumption occurred on farms. As described below, there is a basis for estimating natural gas consumption in the mining sector, but the total amount of natural gas consumed in the sector is probably only about billion cubic feet annually, out of a difference of 782 1,262 billion. Construction, by its very nature, is a temporary activity and would not be expected to operate permanent, stationary natural gas or kerosene combustion equipment. Section Natural Gas Consumption at Mines and Quarries The 2007 Economic Census contains data on nationwide expenditures on fuel in the mining sector that can be used to estimate fuel consumption data. 21 Since lease and plant fuel are already accounted for in the oil and gas emissions inventory, the relevant data for industrial fuel combustion in the mining sector are for NAICS codes 212 and 213. The table below shows the reported consumption of natural gas, along with the calculated consumption (using the EIA s reported price of natural gas for industrial customers in 2007 of $7.62 per MCF 22 and a heat value of 1022 BTU/CF). Table 10: Estimated Nationwide Natural Gas Consumption and Employment in Selected Mining Subsectors, 2007 NAICS Code Consumption ($1,000) Consumption (MMBTU) 22 Employees Consumption Rate (MMBTU/Employee) ,220 2,025,370 38, , ,065 39, ,528 24,289,533 5,044 4, , , , ,260 3, ,832 4, ,707 3,553,978 33, ,476 8, ,947 3,186,697 8, ,297 2,301,762 29, ,943 4,916,113 3, , ,741 2,760,065 2, ,329 12,951,854 2,014 6, , ,192 1, ,417 2,317,731 4, , ,281 6, U.S. Census Bureau Economic Census. Table EC0721I3: Mining: Industry Series: Selected Supplies, Minerals Received for Preparation, Purchased Machinery, and Fuels Consumed by Type for the United States U.S. Energy Information Administration. Natural Gas Prices. Last Accessed July 17, 2013.

23 NAICS Code Consumption ($1,000) Consumption (MMBTU) Employees Consumption Rate (MMBTU/Employee) SUBTOTAL 622,980 82,901, , Where exact employment figures were withheld, CAPCOG used the midpoint of the range reported by the CBP. The consumption data were not disclosed for NAICS codes , , , , , and If the total employment for these industries (34,143.5) is multiplied by the total fuel consumption rate for those subsectors that had consumption data disclosed, it is possible to estimate the fuel consumed in these sectors too (5,467,292 MMBTU). CAPCOG calculated what the 2011 NO X emissions would be at quarries in the Austin-Round Rock Metropolitan Statistical Area (MSA) using the fuel consumption rates and CBP employment data in conjunction with the emissions factor for small boilers in AP The total for the entire MSA would only amount to 4.38 tons per year, or only about 0.01 ton per day for the entire MSA. Since these emissions are so small and diffuse across the region, CAPCOG decided not to directly estimate the small amount of natural gas consumption at area mines and quarries, and instead assigned all of the nonmanufacturing natural gas consumption to the agricultural sector. Section Natural Gas Consumption at Farms Natural gas can be used on farms in both non-road and stationary equipment. In order to isolate the stationary equipment activity, CAPCOG used the United States Department of Agriculture s (USDA s) 2008 Farm and Ranch Irrigation Survey, which provides statewide data on fuel expenditures for irrigation. 24 Using the 2008 prices of natural gas for industrial consumers 25, it is possible to calculate the natural gas consumed by irrigation equipment in each state in the South Census Region. The expenses for the states marked with an asterisk has the total expenses withheld, so expenses were imputed based on the number of farms reporting natural gas use in that state and the average expense per farm for all states that had expenses withheld (this residual amount was calculated by subtracting the reported totals from the nation-wide total). Table 11: Expenses and Consumption of Natural Gas for Agricultural Irrigation, 2008 State Expenses ($1,000) Price MMCF Consumed Alabama $0 $ Arkansas $3,918 $ Delaware* $5 $12.54 <0.5 District of Columbia $0 $ U.S. Environmental Protection Agency. AP-42, 5th Edition, Volume I, Chapter 1: External Combustion Sources, Section 4: Natural Gas Combustion. July U.S. Department of Agriculture, National Agricultural Statistics Service Farm and Ranch Irrigation Survey. Table 20: Energy Expenses for On-Farm Pumping of Irrigation Water by Water Source and Type of Energy: 2008 and _20.pdf. Last Accessed July 17, U.S. Energy Information Administration. Natural Gas Prices. 23

24 State Expenses ($1,000) Price MMCF Consumed Florida* $5 $11.72 <0.5 Georgia $423 $ Kentucky* $19 $ Louisiana $646 $ Maryland $7 $ Mississippi $160 $ North Carolina $0 $ Oklahoma $21,851 $ ,677 South Carolina $0 $ Tennessee $0 $ Texas $194,499 $ ,707 Virginia* $21 $ West Virginia $0 $ TOTAL $221,553 $ ,884 The calculated total natural gas consumption for irrigation in Texas represented 1.6% of all industrial natural gas consumed in Texas in CAPCOG applied the region-wide difference of 12.91% between the sum of the industrial natural gas consumption and the region s 2010 manufacturing natural consumption to Texas s 2010 industrial fuel consumption total to estimate the amount of natural gas consumed by farms in Texas. This produced a total of 183,190 MMCF statewide. CAPCOG then subtracted from that total the 21,707 MMCF consumed in irrigation equipment in 2008 which CAPCOG assumed was a non-road source - from this statewide total in order to obtain an estimate of the natural gas consumed by area sources at farms (161,483 MMCF). There is also an alternative method of estimating Texas s farm area source natural gas consumption that involves allocating the 454,125 MMCF consumed for purposes other than manufacturing in the south region to each state based on agricultural production data. Using land in farms from the 2007 Census of Agriculture 26 (CAPCOG s basis for allocating statewide totals to the county level) would increase Texas s non-manufacturing natural gas consumption from 183 billion cubic feet to 211 billion cubic feet, while using fuel expenses would decrease the consumption to 120 billion cubic feet. Given the range of consumption estimates using this method, it isn t clear that this provide a superior method of estimating statewide fuel consumption. CAPCOG allocated the 161,483 MMCF of area source natural gas combustion to the each of the 11 counties in its Rider 8 program area based on the total land in farms. This surrogate has the advantage of being encompassing of all agricultural production, although it does not necessarily reflect the intensity of natural gas usage. The data on expenses for fuel was another option for allocating, but since 26 U.S. Department of Agriculture, National Agricultural Statistics Service. Volume 1, Chapter 2: State Level Data. Table 1: State Summary Highlights. 1_001.pdf. Last Accessed July 17,

25 diesel dominates this category, it wouldn t necessarily provide a clearer picture of natural gas consumption either. The table below shows the allocation to each county in the program area for Table 12: Agricultural Area Source Industrial Natural Gas Combustion 2010 County Land in Farms (acres) NG (MMCF) Bastrop 402, Blanco 395, Burnet 482, Caldwell 304, Fayette 565, Hays 235, Lee 325, Llano 538, Milam 538, Travis 262, Williamson 541, Texas Total 130,398, ,483 Section Kerosene Consumption at Farms Statewide data on consumption of kerosene at farms is available from EIA s Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. 27 For 2011, the total was 142,000 gallons. CAPCOG allocated this consumption to each county based on land in farms from the 2007 Census of Agriculture. The table below shows the allocation of the 2011 consumption to each county. Table 13: Agricultural Area Source Industrial Natural Gas Combustion 2010 County Land in Farms (acres) Kerosene (1,000 gallons) Bastrop 402, Blanco 395, Burnet 482, Caldwell 304, Fayette 565, Hays 235, Lee 325, Llano 538, Milam 538, Travis 262, Williamson 541, Texas Total 130,398, U.S. Energy Information Administration. Texas Kerosene Adj. Sales/ Deliveries to Farm Customers. Last Accessed July 18,

26 Section Emissions Calculations and Summaries Once CAPCOG calculated the baseline fuel consumption activity, it needed to apply growth factors, temporal profiles, and emissions factors in order to estimate ozone-season day emissions of CO, NO X, and VOC for 2012 and Section Growth Factors CAPCOG projected the 2011 and 2010 baseline activity to 2012 and 2018 using the EIA s reference case in the Annual Energy Outlook 2013 projections for industrial consumption of various fuels for the West- South Central census region, which consists of Arkansas, Louisiana, Oklahoma, and Texas. 28 CAPCOG divided the 2012 and 2018 consumption of each fuel type by the baseline activity (2011 for manufacturing, and farm kerosene, 2010 for farm natural gas) in order to calculate growth factors. CAPCOG assigned kerosene to the other petroleum category and assigned pulping liquor and biomass to the biofuels category. The table below reflects the total energy consumption in the industrial sector for each fuel. Table 14: Projected Industrial Fuel Consumption for the West South Central Census Region (quadrillion BTU) Fuel LPG Distillate Fuel Oil Residual Fuel Oil Other Petroleum Natural Gas Biofuels Heat and Co-products Section Temporal Profiles CAPCOG calculated the monthly profile of industrial fuel combustion based on the natural gas consumption reported in the industrial sector in Texas for 2012 in EIA s Texas Natural Gas Industrial Consumption. 29 Table 15: Texas Industrial Gas Consumed by Month, 2012 (MMCF) Month Gas Consumed Fraction January 132, February 123, March 127, April 123, May 122, June 121, July 127, U.S. Energy Information Administration. Annual Energy Outlook Table 2-7: Energy Consumption by Sector and Source, West South Central, Reference Case. May U.S. Energy Information Administration. Texas Natural Gas Industrial Consumption. 26

27 Month Gas Consumed Fraction August 129, September 123, October 124, November 125, December 130, The summer season (June August) accounted for 25.03% of annual natural gas consumption; therefore, CAPCOG assumed that 25.03% of all industrial fuels were consumed in summer months. CAPCOG adjusted the EPA s default weekly temporal profile for industrial fuel combustion as indicated in its temporal profiles for the CAIR Platform in order to make Monday Friday activity uniform. 30 The following table shows the default and updated allocations of weekly activity. Table 16: Default and Updated Weekly Allocation of Activity Day Default Fraction Updated Fraction Monday Tuesday Wednesday Thursday Friday Saturday Sunday In order to calculate daily activity, then, annual activity is multiplied by (summer fraction), multiplied by 7 days per week /92 days (allocation to a given week in the summer), and multiplied by (allocation to a given weekday). Section Emissions Factors CAPCOG used the following heat conversion factors and emissions factors to calculate the annual emissions for each fuel and county. Table 17: Emission Factors and Heat Conversion Rates Fuel CO Factor NO X Factor VOC Factor Heat Conversion Rate Distillate Engines 0.95 lbs/mmbtu 4.41 lbs/mmbtu 0.36 lbs/mmbtu n/a 30 U.S. Environmental Protection Agency. Temporal Profile CAIR Platform. Temporal Cross-Reference File for CAIR Platform. February

28 Fuel CO Factor NO X Factor VOC Factor Distillate Boilers Residual 5 lbs/1,000 gallons 5 lbs/1,000 gallons 20 lbs/1,000 gallons 55 lbs/1,000 gallons 0.2 lbs/1,000 gallons 0.28 lbs/1,000 gallons Natural Gas 84 lbs/mmcf 100 lbs/mmcf 5.5 lbs/mmcf Butane Propane 8.4 lbs/1,000 gallons 7.5 lbs/1,000 gallons 15 lbs/1,000 gallons 13 lbs/1,000 gallons Wet Wood 0.6 lbs/mmbtu 0.22 lbs/mmbtu Dry Wood 0.6 lbs/mmbtu 0.49 lbs/mmbtu Kerosene 4.82 lbs/1,000 gallons lbs/1,000 gallons 0.9 lbs/1,000 gallons 0.8 lbs/1,000 gallons lbs/mmbtu lbs/mmbtu 0.33 lbs/1,000 gallons Heat Conversion Rate MMBTU/1,000 gallons MMBTU/1,000 gallons 1,025 MMBTU/MMCF 103 MMBTU/1,000 gallons 91.33/1,000 gallons n/a n/a 135 MMBTU/1,000 gallons External combustion rates for distillate fuel oil and residual oil external emission rates were based on AP-42, Table for boilers <100 MMBTU/hour for CO and NO X, and Table for industrial boilers (using non-methane total organic compounds or NMTOC). 31 The emissions rates for distillate consumed in internal combustion engines were based on AP-42, Table The emissions rates for natural gas were based on AP-42, Table for CO and NO X (small, uncontrolled boilers) and Table for VOC. 33 The emissions rates for LPG combustion were based on Table (VOC emissions rate calculated as difference between total organic compounds and methane). 34 Emission factors for wet wood and dry 31 U.S. Environmental Protection Agency. AP-42, 5th Edition, Volume I, Chapter 1: External Combustion Sources, Section 3: Fuel Oil Combustion. September 1999, Corrected May U.S. Environmental Protection Agency. AP-42, 5th Edition, Volume I, Chapter 3: Internal Combustion Sources, Section 3: Gasoline and Diesel Industrial Engines. August U.S. Environmental Protection Agency. AP-42, 5th Edition, Volume I, Chapter 1: External Combustion Sources, Section 4: Natural Gas Combustion. July U.S. Environmental Protection Agency. AP-42, 5th Edition, Volume I, Chapter 1: External Combustion Sources, Section 5: Liquefied Petroleum Gas Combustion. July

29 wood were based on AP-42, Table The emission rates for kerosene were based on the emission rates listed in EPA s ici_fuel_combustion_2011_tx_rev file for the 2011 NEI. Since there isn t a strict 1:1 correspondence between each of the emission rates and the fuel consumption figures obtained by CAPCOG, it was necessary to make some assumptions in assigning activity to each emissions rate. CAPCOG assumed that 50% of distillate fuel is consumed in internal combustion devices and 50% is consumed in external combustion devices. This assumption matches EPA s assumption in the documentation for the 2011 National Emissions Inventory 36. CAPCOG contacted Greg Lauderdale at TCEQ to determine whether TCEQ used any different assumption, and he indicated that they used the tool CenSARA developed through a contract with ERG. CAPCOG contacted the author of the report, and she stated that they used the EPA default, although she was not aware of any data they used as the basis for that assumption. For LPG, CAPCOG assumed a 60% Butane/40% Propane mixture, corresponding to the composition identified by EIA in its Annual Energy Review for The compound emissions rates are equal (albeit with fewer significant figures) than EPA s emissions rate for the 2011 NEI. CAPCOG assigned the pulping liquor to the wet wood emissions factors and all other biomass to the dry wood emissions factors. Since many of the emission rates were expressed as pounds of emissions per physical quantity, it was necessary to convert the fuel consumed in MMBTU to physical quantities. For distillate fuel oil, residual fuel oil, butane, propane, and kerosene, CAPCOG used the Energy Information Administration s estimated heat content of petroleum products as reported in Appendix A to the Annual Energy Review The heat content of natural gas was based on the natural gas delivered to Texas consumers in 2011 in EIA s Heat Content of Natural Gas consumed. 38 Section Emission Totals for 2012 and 2018 Once growth factors, temporal profiles, and emission factors were applied to the county-level activity, the following emission totals were produced for 2012 and 2018 for the region at large. Table 18: 2012 Industrial Fuel Combustion Ozone Season Weekday Emissions by Fuel Type (tons per day) Source SCC CO NO X VOC Manufacturing Distillate Oil Combustion Manufacturing Residual Oil Combustion Manufacturing Natural Gas Combustion U.S. Environmental Protection Agency. AP-42, 5th Edition, Volume I, Chapter 1: External Combustion Sources, Section 6: Wood Residue Combustion in Boilers. Update 2003, September U.S. Environmental Protection Agency. Documentation of Industrial, Commercial, and Institutional Fuel Combustion Emissions Estimates for ftp://ftp.epa.gov/emisinventory/2011nei/doc/ici_fuel_combustion_by_state/ici_fuel_combustion_2011_tx_rev.zi p 37 U.S. Energy Information Administration. Annual Energy Review Appendix A: British Thermal Unit Conversion Factors. Table A1: Approximate Heat Content of Petroleum Products U.S. Energy Information Administration. Heat Content of Natural Gas Consumed. 29

30 Source SCC CO NO X VOC Manufacturing LPG Combustion Manufacturing Wood Combustion Manufacturing Kerosene Combustion Farm Natural Gas Combustion Farm Kerosene Combustion TOTAL Table 19: 2018 Industrial Fuel Combustion Ozone Season Weekday Emissions by Fuel Type (tons per day) Source SCC CO NO X VOC Manufacturing Distillate Oil Combustion Manufacturing Residual Oil Combustion Manufacturing Natural Gas Combustion Manufacturing LPG Combustion Manufacturing Wood Combustion Manufacturing Kerosene Combustion Farm Natural Gas Combustion Farm Kerosene Combustion TOTAL The following tables show the total industrial fuel combustion emissions by county. Table 20: Updated Industrial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) County 2012 CO 2012 NOX 2012 VOC 2018 CO 2018 NOX 2018 VOC Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson TOTAL Table 21: Default Industrial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) County 2012 CO 2012 NO X 2012 VOC 2018 CO 2018 NO X 2018 VOC Bastrop

31 County 2012 CO 2012 NO X 2012 VOC 2018 CO 2018 NO X 2018 VOC Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson TOTAL Table 22: Difference for Industrial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) County 2012 CO 2012 NO X 2012 VOC 2018 CO 2018 NO X 2018 VOC Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson TOTAL Section Industrial Fuel Combustion Spatial Allocation Improvements CAPCOG used the industrial fuel combustion emissions spatial allocation surrogates previously developed for Bastrop, Caldwell, Hays, Travis, and Williamson Counties based on establishment-specific employment counts in September 2005 in order to allocate distillate fuel oil, residual fuel oil, natural gas, and LPG combustion for manufacturing sources within those counties. This is documented in a report produced in 2013 by CAPCOG. 39 The process involved multiplying the employment at a specific facility by fuel consumption figures developed similar to those described above for Since different industries have different fuel intensities and different demand for each fuel type, the allocation varies 39 Capital Area Council of Governments. Spatial Allocation Surrogate Updates for Selected Area and Non-Road Sources in the Austin-Round Rock-San Marcos Metropolitan Statistical Area. July

32 by fuel type and more accurately reflects the fuel consumption patterns for particular manufacturing industries. The figures below show the spatial distribution of the estimated 2005/2006 area source industrial fuel consumption at manufacturing establishments within the MSA. Figure 3: Spatial Allocation of Industrial Natural Gas Combustion 32

33 Figure 4: Spatial Allocation of Industrial Distillate Fuel Oil Combustion Figure 5: Spatial Allocation of Industrial LPG Combustion 33

34 Figure 6: Spatial Allocation of Residual Fuel Oil Consumption For kerosene and wood fuel combustion, CAPCOG used the default industrial fuel combustion spatial allocation. For farm industrial fuel combustion, CAPCOG used the spatial allocation factors developed for agricultural tractors that is described further in Section The following PAVE plot shows the aggregated spatial allocation of industrial fuel combustion NO X emissions in

35 Figure 7: PAVE Plot of Industrial Fuel Combustion NO X Emissions, 2012 Section 3.2 Commercial Fuel Combustion CAPCOG updated the 2012 and 2018 commercial fuel combustion estimates for all 11 counties in the program area. These updates rely on the most recent state-level fuel consumption data and projections from EIA and county-level employment in NAICS Codes from the CBP and the Texas Workforce Commission (TWC). This method closely follows the methods EPA has used for the 2008 and 2011 NEI with modifications in the method used for point source subtractions and assumptions used for non-road subtractions. EIA defines the commercial sector as an energy-consuming sector that consists of service-providing facilities and equipment of businesses; Federal, State, and local governments; and other private and public organizations, such as religious, social, or fraternal groups. While the EIA does have a Commercial Building Energy Consumption Survey (CBECS) that provides more detailed data on differences in fuel consumption patterns for different types of commercial buildings, CAPCOG determined that further research would be needed before using the data collected from that survey for inventory development. Future research projects by CAPCOG may include data from this survey. Section Statewide Commercial Fuel Consumption Totals CAPCOG obtained updated statewide fuel consumption totals from EIA. The following table shows the most recent statewide fuel consumption data for the commercial sector. 35

36 Table 23: Statewide Fuel Consumption Estimates for the "Commercial" Sector Fuel Year Quantity Units Natural Gas ,297 MMCF Distillate Fuel Oil ,691 1,000 gallons Residual Fuel Oil ,850 1,000 gallons Kerosene ,000 gallons LPG ,851 1,000 barrels Since the emission rates for LPG combustion are reported for 1,000 gallons of fuel consumed, CAPCOG converted the LPG consumption in barrels to 77,742 gallons (42 gallons/barrel). Section Mobile Source Fuel Consumption Subtractions In order to avoid double-counting emissions from fuel consumed in this sector from on-road and nonroad sources used by commercial establishments, it is necessary to subtract any mobile source fuel consumption from fuel consumption totals. The focus of these adjustments is the distillate fuel oil and LPG consumption from mobile sources. EPA s methodology for estimating commercial distillate fuel oil emissions in the 2011 NEI relies on an assumption from a regulatory impact analysis for low-sulfur fuels that assumes that 100% of the No. 2 fuel oil (which makes up 100% of the distillate reported for 2011) is consumed by mobile sources. 45 While CAPCOG believes that this estimate fails to account for distillate oil used for emergency generators, in order to maintain consistency with EPA estimates and methods, CAPCOG chose not to attempt to modify this assumption. ERG provided an estimate of emissions from emergency generators in the Dallas-Fort Worth Area in its 2005 report on selected industrial and commercial equipment. 46 This estimate can be roughly scaled by population to this region to obtain a rough estimate of the emissions from such generators in this region; about 0.13 tons per day of NO X emissions and 0.01 tons per day of VOC emissions. 40 Energy Information Administration. Texas Natural Gas Consumption by End Use. Last accessed August 9, Energy Information Administration. Texas Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. Last accessed August 9, Energy Information Administration. Texas Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. Last accessed August 9, Energy Information Administration. Texas Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. Last accessed August 9, Energy Information Administration. State Energy Data System. Table F12: Liquefied Petroleum Gases Consumption Estimates, Last accessed August 9, Environmental Protection Agency. Draft Regulatory Impact Analysis, Chapter 7: Estimated Costs of Low-Sulfur Fuels. Last accessed August 9, Wells, Sam; Eastern Research Group, Inc. Data Collection, Sampling, and Emissions Inventory Preparation Plan for Selected Commercial and Industrial Equipment: Phase II. August 31, Last Accessed July 25,

37 For LPG fuel, EPA used 2006 nationwide fuel consumption estimates from NONROAD equipment it considered part of the commercial sector, as defined by EPA, and divided that total by the quantity of LPG consumed in the commercial sector reported in the SEDS for The following equipment types were used by EPA for this calculation: Table 24: EPA Nationwide Non-Road Fuel Consumption Subtractions SCC Class Description Fuel (1,000 bbl/year) Lawn and Garden Equipment Chippers and Stump Grinders (Commercial) Commercial Equipment Generator Sets 1, Commercial Equipment Pumps Commercial Equipment Air Compressors Commercial Equipment Welders Commercial Equipment Pressure Washers Commercial Equipment Other Commercial Equipment Airport Ground Support Airport Ground Support Equipment Equipment 59 SUBTOTAL n/a n/a 3,971 Research conducted by ENVIRON for CAPCOG in early 2013 on industrial forklifts showed that 61% of the forklift sales in the Austin-Round Rock Metropolitan Statistical Area (MSA) have been to commercial establishments (SIC Codes 42-99). 48 Table 25: Forklift Sales by SIC Code for Austin-Round Rock MSA SIC Codes Description Forklift Sales Agriculture Mineral Extraction Construction Manufacturing Transportation, Communications, Gas, and Sanitary Services Wholesale Trade Retail Remaining Commercial Sector Subtotal 424 Total n/a 694 Since the commercial sector made up 61% of forklift sales, CAPCOG decided to allocate 61% of forklift fuel consumption to the commercial sector. CAPCOG adjusted the EPA s estimated LPG forklift fuel consumption for 2006 based on previous research by ERG in 2005, which indicated that LPG forklifts are 47 U.S. Environmental Protection Agency. ftp://ftp.epa.gov/emisinventory/2011nei/doc/ici_fuel_combustion_by_state/ici_fuel_combustion_2011_tx_rev.zi p. Last access August 9, [ENVIRON REPORT] 37

38 used 1,270 hours per year. Therefore, CAPCOG adjusted the fuel consumption total by 1270/1800 in order to reflect this reduced usage. The adjusted allocation of LPG forklift fuel consumption to the commercial sector was 21,400 thousand barrels for When added to the previously identified commercial fuel consumption totals, the new estimate for nationwide LPG non-road fuel consumption for the commercial sector was 25,371 thousand barrels. In order to adjust the 2011 Texas LPG consumption total, CAPCOG divided this amount by 32,109 thousand barrels of fuel consumed by the commercial sector as reported in the SEDS. For reasons unknown, the data provided by EPA resulted in a lower total for commercial LPG consumption (22,411 barrels) that does not match the data currently in SEDS. 49 This calculation showed that 79% of 2006 LPG fuel consumption in the commercial sector was for non-road use. Therefore, CAPCOG reduced the 2011 commercial LPG fuel consumption by 79% in order to reflect this non-road adjustment. The adjusted statewide totals are presented below. Table 26: Statewide Fuel Consumption Estimates for the "Commercial" Sector with Non-Road Adjustments Fuel Year Quantity Units Natural Gas ,297 MMCF Distillate Fuel Oil ,000 gallons Residual Fuel Oil ,850 1,000 gallons Kerosene ,000 gallons LPG , ,000 barrels Section Growth Factors In order to estimate 2012 and 2018 fuel consumption totals, CAPCOG used the projected growth in commercial sector fuel consumption from the EIA s 2013 Annual Energy Outlook for the West South Central region. 55 CAPCOG used the ratios of projected fuel consumed in 2012 and 2018 to 2011 fuel consumption in order to project the statewide consumption for each year for petroleum fuels. For Energy Information Administration. Texas Natural Gas Consumption by End Use. Last accessed August 9, Energy Information Administration. Texas Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. Last accessed August 9, Energy Information Administration. Texas Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. Last accessed August 9, Energy Information Administration. Texas Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. Last accessed August 9, Energy Information Administration. State Energy Data System. Table F12: Liquefied Petroleum Gases Consumption Estimates, Last accessed August 9, U.S. Energy Information Administration. Annual Energy Outlook Table 2-7: Energy Consumption by Sector and Source, West South Central, Reference Case. May

39 natural gas, CAPCOG used the ratio of 2018 to 2012 consumption. The table below shows the fuel consumption totals and factors CAPCOG used for these projections. Table 27: EIA Commercial Sector Fuel Consumption Estimates for 2011, 2012, and 2018 (quadrillion BTU) Fuel / /2012 Propane Kerosene Residual Fuel Natural Gas CAPCOG then multiplied the baseline 2011 petroleum fuel consumption totals by the 2012/2011 ratios, and then multiplied the 2012 fuel consumption totals by the 2018/2012 ratios. This produced the following estimates of statewide fuel consumption. Table 28: Statewide Fuel Consumption Estimates for the "Commercial" Sector with Non-Road Adjustments Fuel Units Natural Gas , ,156 MMCF Distillate Fuel Oil ,000 gallons Residual Fuel Oil 58 1,711 10,360 1,000 gallons Kerosene ,000 gallons LPG ,290 1,000 gallons Section Allocation of Statewide Consumption to County Levels In order to allocate statewide fuel consumption totals to the county level, CAPCOG used the same method used by EPA in its documentation for the 2008 and 2011 NEI. This entails using the ratio of county-level employment in NAICS codes to statewide employment in these sectors. CAPCOG used the Census Bureau s 2011 employment data reported in the CBP for the employment data for all of these NAICS codes other than public administration (NAICS code 92, which is not reported in CBP). In order to account for employment in the Public Administration sector, CAPCOG used the TWC s Datalink, which does provide county-level employment data in the public administration sector. Both the CBP and 56 Energy Information Administration. Texas Natural Gas Consumption by End Use. Last accessed August 9, Energy Information Administration. Texas Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. Last accessed August 9, Energy Information Administration. Texas Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. Last accessed August 9, Energy Information Administration. Texas Adjusted Distillate Fuel Oil and Kerosene Sales by End Use. Last accessed August 9, Energy Information Administration. State Energy Data System. Table F12: Liquefied Petroleum Gases Consumption Estimates, Last accessed August 9,

40 the TWC s Datalink use the quarterly census of employment and wages to compile employment totals. CAPCOG used the data for March 2011 in order to match the timeframe reflected in the CBP. For several counties, employment totals were not disclosed for every NAICS code for CAPCOG used a procedure similar to the procedure described for manufacturing establishments described in the Industrial Fuel Combustion section of this report (Section 3.1) in order to estimate employment in these sectors. CAPCOG subtracted the reported employment totals for every NAICS code that had a reported total from the county-wide employment total in order to obtain the total employment for the undisclosed NAICS codes. CAPCOG then used data on the employment ranges and the number of establishments in various employment ranges in order to calculate employment midpoints for each undisclosed NAICS code. CAPCOG then adjusted these midpoints in order to ensure that their sum equaled the total employment for undisclosed NAICS codes. For example, for Bastrop County, there were 11,436 total employees reported in the CBP, of which 11,052 were accounted for directly in the employment totals for fully disclosed NAICS codes. The difference between these two figures is 384 employees, which were in the agriculture, forestry, fishing, and hunting sectors (NAICS code 11), the utilities sector (NAICS code 22), the real estate and rental leasing sector (NAICS code 53), management of companies and enterprises (NAICS code 55), educational services (NAICS code 61), and industries not classified (NAICS code 99). CAPCOG calculated midpoints of possible ranges for each of those NAICS codes using the information available, and then added the values to yield a total of employees. CAPCOG then adjusted the values by multiplying them by 81% to ensure that the sum equaled 384. Finally, CAPCOG added the public administration employees from the TWC Datalink. The table below shows the total number of commercial sector employees in 2011 and the corresponding percentage of the statewide total. Table 29: Commercial Sector Employment by County, 2011 Area Employees % Bastrop 10, % Blanco 1, % Burnet 8, % Caldwell 4, % Fayette 5, % Hays 35, % Lee 3, % Llano 2, % Milam 3, % Travis 477, % Williamson 103, % Texas 7,926, % 40

41 Section Point Source Subtractions In order to avoid double-counting point source fuel consumption, CAPCOG subtracted the estimated employment at the commercial and institutional point sources from each county s employment total. CAPCOG used data on the employment range for corresponding NAICS codes within a given county, as well as reviews of data obtained through internet searches, in order to estimate each point source s employment. The table below shows the estimated employment for each commercial/institutional point source in the region that was used for these point source subtractions. Table 30: Commercial and Institutional Point Source Employment Estimates Facility Name NAICS County Employees Austin-American Statesman Travis Coupland Pump Station Williamson 2.5 Center Union Gas Compressor Station Bastrop 2.5 Flint Hills Austin Terminal Travis 74.5 Flint Hills Mustang Ridge Terminal Caldwell 7 Texas State University Hays 4,000 3M Company-River Place Travis 1,500 UT-Austin Hal Weaver Power Plant Travis 24,000 For Austin-American Statesman, the Coupland Pump Station, the Center Union Gas Compressor Station, Flint Hills Austin Terminal, and Flint Hills Mustang Ridge Terminal, CAPCOG cross-referenced the reported SIC code with the appropriate NAICS code, then reviewed the county CBP data and identified the largest employer within that NAICS code, and used the midpoint of the employment range. For Texas State and 3M Company, CAPCOG used information from the Austin Chamber of Commerce on the region s top employers. 61 Another list of top employers in the region indicated that UT-Austin employs 24,000 employees, so CAPCOG used that figure for the Hal Weaver Power Plant. UT uses the Hal Weaver Power plant to supply its electrical, heating, and chilled water needs, through district electricity, heat, and chilled water systems. As such, CAPCOG believes it is appropriate to subtract all of the employees at UT-Austin s main campus from Travis County s commercial employment total. While there are employees on another campus the Pickle Research Campus there is no data CAPCOG was able to find that would enable it to estimate only the employees at UT s main campus. The table below shows the adjusted county employment totals that were used for the adjusted county-level employment figures. Table 31: Commercial Sector Employment by County with Point Source Subtractions, 2011 Area Employees % Statewide Total Bastrop 10, % Blanco 1, % Burnet 8, % Caldwell 4, % Fayette 5, % Hays 31, %

42 Area Employees % Statewide Total Lee 3, % Llano 2, % Milam 3, % Travis 451, % Williamson 103, % Section Annual Fuel Consumption Totals for 2012 and 2018 After the non-road and point source subtractions were accounted for, CAPCOG then allocated the statewide fuel totals to each county in the region. The tables below show the annual consumption estimates for each county for 2012 and Table 32: 2012 Commercial and Institutional Fuel Consumption Totals County Natural Gas (MMCF) Distillate Oil (1,000 gallons) Residual Oil (1,000 gallons) Kerosene (1,000 gallons) LPG (1,000 gallons) Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis 9, ,032 Williamson 2, Total 13, ,434 Table 33: 2018 Commercial and Institutional Fuel Consumption Totals County Natural Gas (MMCF) Distillate Oil (1,000 gallons) Residual Oil (1,000 gallons) Kerosene (1,000 gallons) LPG (1,000 gallons) Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis 10, ,099 Williamson 2,

43 County Natural Gas (MMCF) Distillate Oil (1,000 gallons) Residual Oil (1,000 gallons) Kerosene (1,000 gallons) LPG (1,000 gallons) Total 14, ,527 Section Temporal Allocation CAPCOG used the monthly allocation of natural gas consumed in 2012 in order to allocate all commercial fuel consumption to each season. Table 34: Monthly Consumption of Commercial Sector Natural Gas in Texas, 2012 Month Fuel Consumed (MMCF) Fraction January 22, February 20, March 14, April 12, May 10, June 9, July 9, August 9, September 9, October 12, November 14, December 20, TOTAL 165, June August accounted for % of all commercial sector fuel consumption in This is slightly higher than the 14% in EPA s default temporal allocation file 62. CAPCOG then allocated activity to each week based on the number of weeks over the 92 days between June 1 and August 31 (92 days/7 days a week = weeks). CAPCOG then allocated the weekly emissions to weekdays, Saturdays, and Sundays based on EPA s default weekly allocation factors, except that it adjusted the weekday allocations to have uniform allocations for Monday Friday. The table below shows the allocation factors. Table 35: Commercial Sector Weekly Allocation Factors Day Type Fraction Weekday Saturday 0.13 Sunday U.S. Environmental Protection Agency. Temporal Profile CAIR Platform. Last Accessed 8/26/

44 Section Emissions Totals CAPCOG calculated the emissions totals using the emissions factors provided by EPA in its 2011 National Emissions Inventory Documentation. The following table summarizes these factors. Table 36: EPA Commercial and Institutional Fuel Combustion Emissions Factors (lbs per unit of fuel consumed) Fuel CO NO X VOC Unit Source Residual Oil ,000 gallons ERTAC 2009 Natural Gas MMCF ERTAC 2009 LPG ,000 gallons ERTAC 2009 Kerosene ,000 gallons ERTAC 2009 Table 37: Industrial Fuel Combustion Ozone Season Weekday Emissions by Fuel Type (tons per day) Fuel Type CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Residual Oil Natural Gas LPG Kerosene Total Table 38: Updated Commercial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Table 39: Default Commercial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette

45 County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Hays Lee Llano Milam Travis Williamson Total Table 40: Difference for Commercial Fuel Combustion Ozone Season Weekday Emissions by County (tons per day) County CO 2012 NOX 2012 VOC 2012 CO 2018 NOX 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Section 3.3 Selected Oil and Gas Production Equipment CAPCOG updated the emissions estimates for three types of oil and gas production equipment that produce NO X emissions artificial lift/pumpjacks, and compressor engines, and heater-treaters. CAPCOG based these estimates on methods and calculation developed by Eastern Research Group (ERG) previously conducted research on behalf of TCEQ 63. The EPA Source Classification Codes (SCCs) for each of these equipment types are presented in the table below. Table 41: Source Classification Codes for Targeted Oil & Gas Production Equipment Emissions Updates SCC Short Description OIL AND GAS EXPLORATION AND PRODUCTION ARTIFICIAL LIFT ON SHORE OIL PRODUCTION HEATER TREATER GAS PRODUCTION COMPRESSOR ENGINES ON-SHORE GAS PRODUCTION HEATERS NATURAL GAS FIRED 2-CYCLE LEAN BURN COMPRESSOR ENGINES <50 HP NATURAL GAS FIRED 2-CYCLE LEAN BURN COMPRESSOR ENGINES 50 TO 499 HP

46 SCC Short Description NATURAL GAS FIRED 4-CYCLE LEAN BURN COMPRESSOR ENGINES 500+ HP NATURAL GAS FIRED 4-CYCLE RICH BURN COMPRESSOR ENGINES <50 HP NATURAL GAS FIRED 4-CYCLE RICH BURN COMPRESSOR ENGINES 50 TO 499 HP ON-SHORE GAS PRODUCTION DEHYDRATORS NATURAL GAS FIRED 4-CYCLE RICH BURN COMPRESSOR ENG HP W/ NON SPECIFIC CATALYTIC REDUCTION CAPCOG obtained updated activity data for the CAPCOG counties and Milam County for 2012 from the Texas Railroad Commission (RRC) in conjunction with the assumptions and formulae developed by ERG for its 2008 estimates in order to establish updated estimates for CAPCOG assumed that all oil and gas production equipment emissions were uniformly distributed across each day of the year. Section Artificial Lift/Pumpjack Equipment Artificial lift refers to the creation of below-ground pressure to extract oil where bottom-hole pressure does not naturally occur sufficient to displace the oil to the well surface. A pumpjack is a specific artificial lift technique that uses a mechanical pump to create bottom-hole pressure. These pumpjacks are sometimes powered by natural gas, resulting in emissions. ERG s calculated the pumpjack emissions based on a report by ENVIRON International Corporation for the Central States Regional Air Partnership (CENRAP) 64. The formula ERG used to estimate the emissions of each criteria pollutant for a given county is shown below. ( ) ( ) Where: E ik = the emissions for county i, and pollutant k (tons/yr.) W i = the total number of active oil wells in county i f pumpjack = fraction of oil wells with artificial lift engines e pumpjack = fraction of pumpjack engines that are electrically operated EF k = emission factor for pollutant k (g/hp-hr) HP = horsepower of the engine (HP) LF = load factor of the engine while operating T annual = annual number of hours engine is in use (hr/yr) 907,180 is the conversion factor from grams to tons of emissions The total number of active oil wells within a given county (W i) for 2012 was taken from Texas Railroad Commission s (RRC) online database of regular producing wells per county 65. Well counts are provided at the county level in February and September of each year. CAPCOG used the number of regular producing oil wells for each county reported for September 2012 as the basis for the value ofw i

47 Table 42: Regular Producing Oil Wells by County, September 2012 Data Wells Bastrop 294 Blanco 0 Burnet 0 Caldwell 3,236 Fayette 549 Hays 0 Lee 794 Llano 0 Milam 1,158 Travis 21 Williamson 45 TOTAL 6,097 ERG assumed that the fraction of wells that require artificial lift engines (F pumpjack) was equal to the fraction of all active wells greater than one year old. ERG s reasoning behind this estimation was that wells that had been completed within a year would still retain sufficient naturally-occurring bottom-hole pressure to drive crude oil from the well bore without the necessity of artificial lift, while older wells would require artificial lift to bring oil to the surface. This fraction was calculated using the following formula: ( ) In order to calculate this fraction, ERG (and CAPCOG) used the RRC s drilling completion data. The RRC provides data on the number of oil wells completed by district by month. 66 CAPCOG obtained the data for Districts 1 and 3 (D1 and D3) for September 2011 August 2012, which are presented below. 67 Table 43: Drilling Completions for Oil Wells September August 2012 Month D1 D3 Sep Oct Nov Dec Jan Feb Mar Apr May Note that the drilling data presented here is the total oil well drillings completed. The RRC data also breaks down the number of new, re-enter, and recompletion drilling completions, but since ERG used total completions, CAPCOG did as well in order to maintain consistency. 47

48 Month D1 D3 Jun Jul Aug Total 1, CAPCOG then summed the total active wells for each county in each region as of September 6, 2012, in order to obtain the total active wells within a given year for each region. The table below summarizes these data, and the calculated fraction of oil wells requiring artificial lift. Table 44: Fractions of Oil Wells Requiring Artificial Lift in RRC Districts 1 and 3, 2012 RRC District Wells Completed in 2012 Active Wells September 2012 Fraction Requiring Artificial Lift District 1 1,162 16, District , CAPCOG applied each district s fraction requiring lift to the corresponding counties within this region: District 1: Bastrop, Burnet, Blanco, Caldwell, Hays, Llano, Milam, Travis, and Williamson Counties; and District 3: Fayette and Lee Counties. Since this approach includes all oil wells, including those that were hydraulically fractured (which would not likely use a pumpjack), it may over-estimate the pumpjack emissions, although to the extent that this activity is captured in the previous year s drilling (and therefore not part of the fraction requiring artificial lift), it should not cause any significant double-counting. CAPCOG used the following default assumptions from ERG s report: E pumpjack is 70% for all counties, based on a survey of four major engineering consultants that specialize in oil and gas production; LF is 71%, based on an average value of a power function relationship given a range of values from 10% - 100% 68 ; T annual equals 4,380 hours (1/2 of the year). CAPCOG also used the same horsepower ratings (HP) and emissions factors (EF) as ERG did for this source. ERG averaged the emissions factors and horsepower ratings across four models of Arrow-C model pumpjacks as shown below. ERG found these to be the most common fuel-powered engine models in the state, based on survey feedback from operators, leasers, and energy consultants

49 Table 45: Artificial Lift/Pumpjack Engine Profile and Associated Emission Factors Arrow Series C Model Horsepower Fuel Consumption (MMBtu/Hp-Hr) NO X (g/hp-hr) CO (g/hp-hr) VOC (g/hp-hr) C C C C Average ERG assumed that none of the pumpjacks in operation in 2008 were subject to EPA s New Source Pollution Standards (NSPS) for Oil and Gas Production, which applied to pumpjack engines manufactured on or after August 1, In the absence of any clear data indicating the age distribution of the pumpjacks in operation in this area, CAPCOG preserved these assumptions for the 2012 estimates. The following sample calculation shows how these data were used to calculate NO X emissions for pumjacks in Caldwell County in ( ) ( ) Where, E CALDWELL,NOX = the NO X emissions for Caldwell County (tons/yr.) W CALDWELL= active oil wells in county Caldwell County Sept = 3,236 f pumpjack = fraction of oil wells with artificial lift engines = e pumpjack = fraction of pumpjack engines that are electrically operated = 0.70 EF NOX = emission factor for NO X (g/hp-hr) = HP = average horsepower of the engine = LF = load factor of the engine while operating = 0.71 T annual = annual number of hours engine is in use (hr/yr) = 4, ,180 is the conversion factor from grams to tons of emissions Given these values, the equation yields the following value: ( ) ( ) CAPCOG calculated 2012 annual NO X, CO and VOC emissions and divided by 366 days/year (for 2012 because of the leap year) in order to produce ozone season day emissions. While oil production is available on a monthly basis, it is not possible to tell what percentage of these data are from newly completed wells and from older wells, therefore, a uniform temporal distribution is most appropriate. These estimates are presented below CFR

50 Table 46: 2012 Annual and Ozone Season Day Artificial Lift/Pumpjack Emissions (tons per day) Pollutant CO (tpy) NO X (tpy) VOC (tpy) CO (tpd) NO X (tpd) VOC (tpd) Bastrop Blanco Burnet Caldwell Hays Fayette Lee Llano Milam Travis Williamson TOTAL , Section Compressor Engines Compressor engines are used at oil and gas wellheads to increase the pressure of natural gas that arrives at a well s surface so it can be transported through gathering lines further down the production processing line. These compressors are powered by natural gas, though their report does not detail the extent to which gas produced on sight is used to power the compression engines themselves. CAPCOG s updates for the compressor engine estimates replace all of the existing estimates, including those from AACOG s Eagle Ford Shale inventory for the counties where overlap occurred in order to avoid double-counting. CAPCOG s updates are based on ERG s report, which used a 2005 study for the Houston Advanced Research Council (HARC) 70. This study gathered data on about 1,300 compressor engines and field survey data from an additional 64 engines. For a given criteria pollutant in a given county, the emissions from a particular compressor engine model are calculated as follows: ( ) Where: E ikj = emissions for county i, and pollutant k (tons/yr), for engine make/model j TGP i= total gas production in county i, (MCF/yr) F 1i = fraction of wells requiring compression in county i F 2j= fraction of compression load represented by engines of type j EF jk = emission factor for engine type j, and pollutant k (g/hp-hr) C i = compression requirements for county i, (Hp-hr/MCF) 907,180 is the conversion factor from grams to tons of emissions

51 The amount of gas produced in a given county in 2012 (TGP i) was tabulated from an RRC query-based platform that links to a database, which provides comprehensive information on the amount of gas and oil produced at the county, district, and offshore area or specific gas or oil field over specified monthly and annual date ranges 71. Natural gas production in particular is further delineated by whether the gas was produced directly from a RRC-classified gas well or from the casinghead of a RRC-classified oil well Total gas production in 2012 (the sum of casinghead gas and gas from gas wells) is shown in the table below. Table 47: 2012 Gas Production Totals, Including Casinghead Gas and Gas Well Gas (MCF) 74 County Gas Well Casinghead Total Bastrop 99, , ,196 Blanco Burnet Caldwell 353,124 11, ,217 Hays Fayette 6,875,974 8,908,255 15,784,229 Lee 6,823,309 1,337,519 8,160,828 Llano Milam 359,286 99, ,973 Travis Williamson TOTAL 10,473,053 14,511,390 24,984,443 The formula used for calculating the fraction of gas wells requiring compression (F 1i) is similar to the formula used for the fraction of oil wells requiring artificial lift: ( ) Similar to the method used to calculate the fraction of oil wells not requiring artificial lift, CAPCOG obtained the monthly gas well drilling completions by district for September 2011 August 2012 in order to estimate the number of gas wells not requiring compression Casinghead gas refers to natural gas that is gathered at the wellhead during the production of oil from a designated oil well. 73 Because a given well can produce both oil and gas, as well as gas condensate, oil and gas wells cannot be classified strictly according to the function or purpose of the well. Therefore, the RRC uses a ratio to determine whether a given well should be categorized as either an oil or gas well. Specifically, after a well begins production, RRC examines the production data provided to them by the well operator. Given production data, a well is designated as a gas well if it produces more than 100,000 scft. of natural gas per barrel of oil. If the ratio is any less than 100,000 scft. per barrel of oil, then the well is classified as an oil well. 74 Data downloaded from Texas Railroad Commission Oil and Gas Production Data Query on May 28, Note that more recent queries have provided slightly different results than the May 28 th query. 51

52 Table 48: Drilling Completions for Gas Wells September August 2012 Month D1 D3 Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Total CAPCOG subtracted these totals from the district-wide active well totals as of September 2012 in order to obtain the total number of wells requiring compression (those drilled over one year before). CAPCOG used the oil and gas well data in Region 1 for Bastrop, Burnet, Blanco, Caldwell, Hays, Llano, Milam, Travis, and Williamson Counties, and Region 3 data for Fayette and Lee Counties. Table 49: Gas Wells Completed in 2012 and Active September 2012 for RRC Districts 1 and 3 RRC District Wells Completed in 2012 Active Wells September 2012 Fraction Requiring Compression District , District , Compression Requirement per County (C i) depends upon the compression ratio or the ratio of suction pressure at the wellhead site to output pressure into the gas gathering lines. The lower that this ratio is, the more mechanical energy in horsepower (HP) is needed from a compressor engine to generate adequate pressure to move the gas through the gathering lines. The volume of gas needed to fuel this compression is directly proportional to the power required from the compressor. A compression requirement as measured in HP-hours per MCF represents the conversion of inlet pressure to outlet pressure. The average requirement ERG calculated from previous studies was 3.21 hp-hrs/mcf. CAPCOG decided to use this value for all counties, including those in District 3, because of the variation across studies within a given district. 52

53 Table 50: Compression Requirements from Previous Studies by Region of Texas 75 Study TRC District 2 TRC District 3 TRC District 6 All Other Texas Areas HARC n/a 2005 and 2008 Pollution Solutions n/a n/a 2.97 n/a Final In order to calculate the emissions for each SCC code, CAPCOG allocated compression activity to each make and model based on the fractions of compression load (F 2j) identified in a previous ERG study in Each make and model s source classification code, horsepower rating, percentage of engine load, and emissions factors are presented in the table below. 77 Table 51: Compressor Engine Makes and Models with SCCs, HP Ratings, % of Load, and NO X, CO, and VOC Factors (g/hp-hr) Engine Make/Model SCC HP % of Load NO X CO VOC AJAX C % AJAX DPC % AJAX DPC % AJAX DPC % AJAX DPC % AJAX DPC % AJAX DPC % AJAX DPC % AJAX DPC % CAT G3516 TALE % Wauk L7042 GL % CAT G3512 TALE % Wauk VRG % GEMINI G % CAT G3306 NA % CAT G3304 NA % Wauk VRG % CAT G3306 TA % Wauk F817 G % Based on Table 4-6 from ERG s 2010 report for TCEQ page Based on Table 4-3 from ERG s 2010 study for TCEQ. 53

54 Engine Make/Model SCC HP % of Load NO X CO VOC Wauk F1197 G % CAT G3406 NA % CAT G3306 NA HCR % CAT G342 NA % CAT G342 TA % Wauk VRG % CAT G3406 TA % CAT G % CAT G399 TA % Wauk L7042 GSI % CAT G398 TA % Wauk L7042 G % For four makes/models, there were multiple SCC codes listed, one more specific (such as cycle lean burn compressor engines <50 HP), and one more general (gas production compressor engines). Note that for several makes and models, two SCCs were listed. All other makes/models were either exclusively lean-burn or exclusively rich-burn, so it was not immediately evident how to address these. Since only one emissions rate was provided and the vast majority of the engines for each of the four makes/models were lean-burn, CAPCOG decided to assign these models to the more specific SCC codes. This may result in the lean-burn SCC codes having slightly higher emissions and the rich-burn SCC codes having slightly lower emissions than if the activity was split between rich burn and lean burn SCC codes. However, since the same emissions rate is used, it would not have any impact on the emissions estimates for compressor engines at large. Collectively, these makes/models account for 6.21% of the compression load, with lean-burn engines accounting for 95% of all of the engines in the survey, meaning that this assumption would only have a de minimis impact on any SCC-specific analysis. Table 52: Compressor Engine Makes/Models with Multiple SCC Listings Engine Make & Model Lean-Burn Rich-Burn SCC 1 SCC 2 AJAX DPC AJAX DPC AJAX C AJAX DPC The following shows an example of these calculations for NO X emissions from model number CAT G3306 TA in Fayette County for ( ) 54

55 Where: E FAYETTE,NOX,CAT3306TA = NO X emissions for Fayette County from CAT G3306 TA engines (tons/yr) TGP FAYETTE = total gas production in Fayette County in 2012 (MCF/yr) = 15,784,229 F 1FAYETTE = fraction of gas wells requiring compression in Fayette County = F 2CATG3306TA= fraction of compression load represented by CAT G3306 TA engines = EF CATG3306TA,FAYETTE = emission factor for NO X for CAT G3306 TA engines (g/hp-hr) = C FAYETTE = compression requirements for Fayette County (Hp-hr/MCF) = ,180 is the conversion factor from grams to tons of emissions ( ) For each county, CAPCOG calculated the emissions for each make and model, and then aggregated them by SCC code. The following tables show the total compressor engine emissions by county and by SCC. Table 53: Compressor Engine Emissions by County, 2012 County CO (tpy) NO X (tpy) VOC (tpy) CO (tpd) NO X (tpd) VOC (tpd) Bastrop Blanco Burnet Caldwell Hays Fayette Lee Llano Milam Travis Williamson TOTAL Table 54: Compressor Engine Emissions by SCC, 2012 SCC Code CO (tpy) NO X (tpy) VOC (tpy) CO (tpy) NO X (tpy) VOC (tpy) TOTAL

56 Section Heaters Oil and gas producers use heaters (also referred to as boilers) at oil and gas well sites for three separate functions: 1) to provide heat for separating crude oil into petroleum products before it is taken farther along the production line for further processing, 2) to maintain the proper temperature for oil and gas while they remain in temporary storage tanks, and 3) to maintain temperature for "inline pipes (also called gathering lines) that move petroleum products farther down the production line. Heaters typically use natural gas to power their activity and therefore can be significant sources of NO X, CO and VOC in oil-producing and gas-producing areas. ERG s method for estimated heater emissions is based on the 2008 CENRAP study previously described under Section Emissions for a single heater are based on the following equation: Where: ( ) E fp = emissions of pollutant p from heater type f (tons/yr) EF fp = emission factor for pollutant p (lbs/mmcf) Q = heater rating (MMBTU/hr) HV f = the local heating value for natural gas for heater type f (MMBTU/MMCF) T annual = annual hours of operation (hr/yr) hc = heater cycling fraction to account for the fraction of operating hours that the heater is firing 2,000 = conversion factor from lbs. to tons Emissions factors (EF fp) are based on EPA AP-42 s compilation of emission factors 78 for external natural gas combustion sources. The table below summarizes AP-42 emission factors for NO X, CO, and VOC. Table 55: Heater/Treater Emissions Factors Pollutant Factor (lbs/mmcf) NO X 100 CO 84 VOC 5.5 ERG used the heater firing rating (Q), annual activity (T annual), the local heating values (HV f), and the heating cycle (hc) from the CENRAP study, based on survey data at the Texas basin level. 78 U.S. Environmental Protection Agency. AP-42, 5th Edition, Volume I, Chapter 1: External Combustion Sources, Section 4: Natural Gas Combustion. July

57 Table 56: CENRAP Weighted Averages of Texas Basin Level Data Data Weighted Average Heater Firing Rating (MMBtu/hr) Annual Activity (hr/yr) Gas Well Heating Value (MMBtu/MMCF) Gas Well Heating Value (MMBtu/MMCF) Heater Cycling ,076 1,209 1,655 1 To expand these estimates to county-wide inventories, ERG used the following equation: Where: E fpj = the total heater emissions of pollutant p in county j and well type f (ton/yr) E fp = the total emissions of pollutant p from a single heater at well type f (ton/yr) W fj = the total number of wells of type f in county j N = the typical number of heaters per well in the county 2,000 = the conversion factor from pounds to tons of emissions CAPCOG used W fj taken from the RRC s list of active wells for each county 79. The following table summarizes the total number of active oil and gas wells in each county for September Table 57: Active Oil and Gas Wells by County, September 2012 County Gas Oil Bastrop Blanco 0 0 Burnet 0 0 Caldwell 6 3,236 Fayette Hays 0 0 Lee Llano 0 0 Milam 11 1,158 Travis 0 21 Williamson 0 45 TOTAL 400 6,097 The typical number of heaters per well (N) was taken from the CENRAP study s weighted average of Texas basin-level survey responses

58 The following example shows the calculation of an oil heater. ( ) ( ) This emissions rate is then applied to the estimated number of oil heaters in the county with the following equation. Using these equations, CAPCOG calculated the annual emissions for each county for CAPCOG then divided the annual totals by 366 in order to obtain the ozone season day emissions. Table 58: 2012 Ozone Season Day Heater-Treater Emissions at Oil and Gas Wells by County (tons per day) County CO Oil NO X - Oil VOC - Oil CO Gas NO X - Gas VOC - Gas Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson TOTAL Section Growth Projections to 2018 CAPCOG developed growth estimates for the industry in the region based on an exploratory data analysis (EDA) approach to a review of oil and gas production and active well counts in each county from , combined with a review of local conditions in order to estimate the most appropriate growth factors for each type in each county. In reviewing the data, CAPCOG found that trends are specific to each county and each energy production type. The graph below shows how idiosyncratic the trends were. It shows the ratio of oil and gas production to a 2006 base year in Caldwell County and Fayette County the counties with the most oil and gas production emissions in the region for each year from 2006 to

59 Ratio of Production to 2006 Base Year 2012 and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties Figure 8: Comparison of Growth of Oil and Gas Production in Caldwell and Fayette Counties, Caldwell - Oil Production Fayette - Oil Production Caldwell - Gas Production Fayette - Gas Production In the chart above, it is evident that there is significant growth in oil production in Caldwell County from , although it looks like growth tapers off starting in Over this same period, there is also a steady, apparently linear decline in gas production in Caldwell County. Fayette County oil production, on the other hand, declined from 2006 to 2009, but then started to climb again starting in Growth slowed year over year from , however. Gas production in Fayette County, on the other hand, steadily declined over the same period as well. These data show the difficulty in identifying appropriate growth rates for this source category. CAPCOG staff collectively reviewed the data and developed agreed-upon assumptions used for each county and calculated the results, as shown in the table below. Table 59: Oil and Gas Production Equipment Emissions Growth Factors County Oil Sources Oil Growth Basis Gas Sources Gas Growth Basis Bastrop Power curve using annual production Power curve using annual production Blanco No production No production Burnet No production No production Average of growth rate Caldwell for wells and production Linear curve, using from , linear 2012 production growth to 2018 Fayette Linear projection based on growth Linear projection based on growth 59

60 County Oil Sources Oil Growth Basis Gas Sources Gas Growth Basis Hays No production No production Lee Linear projection using Power curve, using production 2012 Production Llano No production No production Milam Linear growth of oil well Linear growth of gas well count count Travis Erratic changes No production Williamson Erratic changes No production Appendix B contains charts with the detailed well count and production data used to calculate the projections. Section Emission Summaries The table below shows the aggregate CO, NO X, and VOC emissions for all three source types (pumpjacks, compressor engines, and heater-treaters). Table 60: Updated Pumpjack, Compressor, and Heater Ozone Season Weekday Emissions by County (tons per day) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Table 61: Default Pumpjack, Compressor, and Heater Ozone Season Weekday Emissions by County (tons per day) 80 County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Based on modeling files. There appears to have been no changed between 2012 and 2018 in the default files. 60

61 County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Milam Travis Williamson Total Table 62: Difference for Pumpjack, Compressor, and Heater Ozone Season Weekday Emissions by County (tons per day) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Section Spatial Allocation Oil and gas wells were spatially allocated using a well location shapefile purchased from the RRC for Bastrop, Caldwell, Lee, Fayette, and Milam Counties, which together represent the vast majority of oil and gas production for the CAPCOG region ( downloaded March 10, 2013). Travis and Williamson County emissions were allocated based on the default spatial surrogates. The figure below shows the PAVE plot for the oil and gas emissions. 61

62 Figure 9: PAVE Plot of Oil and Gas Production NOX Emissions, 2012 A subset of wells coded as active (those with SYMNUM codes of (2, 4, 5, and 6) were selected from this shapefile, and a new shapefile was made from the selection. This shapefile was brought into ArcGIS along with the 4-km grid. Counties were processed separately in order to account for multiple clipped cells with matching ID values because the original cell crosses a county boundary. The Isectpntpoly script in Geospatial Modeling Environment was used to acquire well counts for each 4- km grid cell. The settings used were as follows: In: County-clipped active wells shapefile Poly: 4-km grid county-clipped shapefile Field: AMDA_id, center_x, center_y The output of this script is different than that of Isectpolyrst (the script used for agricultural equipment), in that, rather than giving grid cells with counts, it appends the fields selected (see above settings) to each point feature in the active wells shapefile. To obtain counts of wells in each grid cell, the attribute table was opened for each county, and the AMDA_id field was summarized (by right-clicking on the field name and choosing Summarize from the context menu). The boxes were checked for center_x and center_y to be included in the output table, and Maximum was chosen as the statistic ( Minimum would be equally acceptable, since these are unique coordinates associated with a grid cell, the result would be the same this step is used only to append the x-y coordinates to the output table). The output table contains active well counts for each grid cell, from which ratios can be calculated in Excel. 62

63 CAPCOG initially embarked on an ambitious plan to allocate oil and gas production and drilling equipment based on levels of production for each well and drilling activity contained in database records obtained from RRC. The data acquired consisted of the Drilling Permit Master and Trailer, Oil Production Data, and Gas Production Data (hereafter collectively referred to as RRC tabular data). This proved to be a difficult task because of incompleteness of the RRC tabular data and its difficulty of use. The RRC tabular data are only available in extended binary decimal interchange (EBCDIC) format, an archaic encoding system rare in use. This required conversion by an outside party (Digital Data Services of Lakewood, CO) to comma separated values (CSV) format to be human-readable and able to be joined to other datasets, such as the GIS data. Production data obtained by CAPCOG in CSV format uses RRC lease numbers for oil and RRC ID numbers for gas as unique identifiers instead of API numbers that correspond to the spatial data, making the joining of these data difficult. Development of crosswalk tables with all three identifiers (lease number, RRC ID, and API number) would greatly assist in joining these unique identifiers to a well location. Widespread mismatches between API Numbers in drilling permit data and those in the spatial data, as well as missing or incomplete API numbers in the attributes of the spatial data made accurate allocation of drilling activity difficult as well. Matches between data downloaded from the RRC s Drilling Permit (W- 1) Query tx.us/ewa/drillingpermitsqueryaction.do) were more complete and could provide the basis for future allocation of drilling activity. The reason that the Drilling Permit Master and Trailer was obtained rather than just using the W-1 query results is that the W-1 query results only include permit approval date, and have no information on completion date. Because of the volume of data, it was not possible to predict the level of mismatch between the spatial datasets and the RRC tabular data until the data was acquired and an attempt was made. It is possible that acquisition of wellbore datasets from RRC might provide the necessary crosswalk tables to improve the match rate between production/drilling data and spatial data (these datasets are described at This would require further study and possible consultation with the RRC. 63

64 Section 4 On-Road Source Updates CAPCOG made two sets of updates to the default 2012 and 2018 on-road emissions inventories for the program area. First, CAPCOG obtained hourly, link-based emissions inventories produced by CAMPO to replace the existing, non-link-based inventories for running, start, and evaporative emissions. CAPCOG also updated the extended idling emissions for all eleven counties based on research conducted by CAPCOG and TTI in 2011, updated to reflect the locations and capacities of truck idling locations in 2012, and updated to reflect the idling emissions rates and growth factors in CAMPO s 2012 and 2018 linkbased emissions inventories. Section 4.1 Link-Based Updates for Austin Round Rock MSA CAPCOG used updated, link-based on-road emissions data supplied by the Capital Area Metropolitan Planning Organization (CAMPO) for Bastrop, Caldwell, Hays, Travis, and Williamson Counties to replace the existing non-link-based inventories for those counties. The updated inventories included updated hourly emissions generated using MOVES2010b for Weekdays (Monday Thursday), Fridays, Saturdays, and Sundays. CAPCOG used the updated data for all processes except for extended idling, which is described in the next section. This report is simply providing a summary of the changes in emissions that occurred as a result of this update; for a full description of the development of these inventories, please review the report that the Texas Transportation Institute (TTI) prepared for CAMPO documenting their work. 81 The following table provides a summary of the average weekday start, running, and evaporative on-road emissions of CO, NO X and VOC for 2012 and Table 63: Austin-Round Rock MSA Weekday Start, Running, and Evaporative CO, NO X, and VOC for 2012 and 2018 (tons per day) County 2012 CO 2012 NO X 2012 VOC 2018 CO 2018 NO X 2018 VOC Bastrop Caldwell Hays Travis Williamson TOTAL These estimates increase the NO X emissions from these processes by 8.04 tons per day (tpd) in 2012 and 4.60 tpd in 2018, representing 17% and 18% increases, respectively. These updates result in an additional 3.43 tpd in NO X reductions over this time period beyond what was already projected. The figure below compares the default and updated NO X emissions for each county. The NO X emissions were divided into three species in its submission to AACOG: NO, NO 2, and HONO. The modeled emissions of these species were directly used for modeling purposes rather than relying on default speciation profiles. 81 Texas Transportation Institute. Austin Five-County Region MOVES-Based On-Road Mobile Source Modeling Emissions Inventories for 2012 and June

65 NO X Emissions (tons per day) 2012 and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties Figure 10: Default and Updated On-Road Start and Running NO X Emissions for 2012 and 2018 by County (tpd) Default 2012 Updated 2012 Default 2018 Updated Bastrop County Caldwell County Hays County Travis County Williamson County TTI used a number of updated parameters for the 2012 and 2018 inventories. The following table shows some key assumptions for each inventory. Table 64: Comparison of Basis Used for Default and Updated Emissions Estimates Parameter Default Basis 82 Updated Basis Emissions Model MOVES2010a MOVES2010b Activity 2008 Highway Performance Travel Demand Model 2005 Management System Data validation and 2015 analysis year Meteorology June August 2008 June August 2011 Pollutants CO, VOC, NO X, NO, NO 2 CO, VOC, NO X, NO, NO 2, HONO Population and Age Distribution July 2010 July 2012 Gasoline RVP (psi) Gasoline Sulfur (ppm) ETOHVolume (volume %) 10% 10% MTBEVolume (volume %) 0% 0% ETBEVolume (volume %) 0% 0% TAMEVolume (volume %) 0% 0% Aromatic (volume%) Olefin (volume%) Benzene (volume%)

66 Parameter Default Basis 82 Updated Basis E200 (vapor% at 200 degrees F) E300 (vapor% at 300 degrees F) TxLED Adjustment July 2010 July 2012 The link-based emissions were spatially allocated by Providence Engineering using tools created by TCEQ and TTI. The figures below show the distribution of start and exhaust NO X emissions for The contract used by CAMPO that describes the requirements for this work is included in Appendix C since Providence was not required by CAMPO to produce a report for this project. Figure 11: PAVE Plot of On-Road Exhaust NO X Emissions,

67 Figure 12: PAVE Plot of Off-Network Start NO X Emissions, 2012 Section 4.2 Extended Idling CAPCOG updated the 2012 and 2018 estimate of extended idling emissions by combination long-haul trucks using data collected in 2011 at area truck stops and a comprehensive and updated list of locations where idling would be expected to occur within the region. The activity estimates were based on activity rates and parking capacity estimates developed in CAPCOG s report on extended idling 83. A handful of new locations were added to the list for 2012, including the following: Buc-ee s In Bastrop, Buc-ee s in Luling, Texaco in Giddings. The combined parking capacities for all facilities in each county in 2012 are listed below. 83 Capital Area Council of Governments. Extended Idling Activity Estimates for Combination Long-Haul Trucks in Central Texas for the Years 2006, 2008, and November

68 Table 65: Truck Parking Capacity by County, 2012 County Spaces Bastrop 74 Blanco 93 Burnet 16 Caldwell 106 Fayette 68 Hays 62 Lee 23 Llano 0 Milam 45 Travis 90 Williamson 212 TOTAL 789 CAPCOG then multiplied the parking capacities at each location by an activity rate based on observational data. CAPCOG had developed a separate activity estimate for the Mustang Ridge truck stop based on its anomalous activity collected in State Highway 130 was under construction at the time, which may have contributed to the difference in the data. Since the highway opened in 2012, CAPCOG instead used the activity rates (extended idling hours per day per parking space) observed for the other locations and applied them to the capacity at the Mustang Ridge truck stop. Table 66: Extended Idling Rates for Truck Stops and Other Facilities (idling hours per parking space) Day Type Rate Monday Thursday 9.67 Friday 7.20 Saturday 6.23 Sunday 4.73 CAPCOG calculated the activity rates for 2012 for each facility (i) and day type (j) using the following equation: CAPCOG then looked up the corresponding grid cell location for each facility and listed the X and Y coordinates for each location. Section Frontage Road Extended Idling In addition to idling at truck stops and other locations with defined parking areas, extended idling also occurs along frontage roads for interstate highways. During the 2011 data collection, CAPCOG collected data on idling along frontage roads on Interstate Highway (IH) 35, which resulted in the following set of activity rates: 68

69 Table 67: Frontage Road Idling Rates by Day Type (Idling Hours per Mile of Interstate Frontage Road) Day Type Rate Monday Thursday 1.04 Friday 0.77 Saturday 0.67 Sunday 0.51 CAPCOG used Google Earth and a KMZ file with the 4km x 4km modeling domain grid cells in order to obtain the number of frontage road miles along interstate highways in each grid cell. There are only two such highways in the region: Interstate 35, which crosses Hays, Travis, and Williamson Counties; and Interstate 10, which crosses Caldwell and Fayette Counties. Upon examination of the aerial photography for Fayette County, CAPCOG determined that they do not provide the same kind of access to the interstate highway that the frontage roads in the other counties did, and therefore assumed that no idling occurred along frontage roads in Fayette County. The following table shows the total frontage road mileage for each county used in the calculations. In the activity calculations, the mileage listed below was multiplied by two to account for the frontage road going in either direction. Table 68: Two-Way Interstate Frontage Road Mileage Suitable for Idling County Miles Bastrop 0 Blanco 0 Burnet 0 Caldwell 4.71 Fayette 0 Hays Lee 0 Llano 0 Milam 0 Travis Williamson TOTAL CAPCOG then used the following equation to calculate the idling hours for link (k) on day type (l). Once the 2012 idling hours were calculated, CAPCOG then projected the activity to 2012 using the ratio of idling hours in 2018 to 2012 from the link-based inventories for Bastrop, Caldwell, Hays, Travis, and Williamson Counties prepared by TTI in 2013 and the non-link based inventories for Blanco, Burnet, Fayette, Lee, and Milam Counties prepared by TTI in These projections are based on projected growth in Combination Long-Haul Truck VMT for each county. 69

70 Section Total Extended Idling Hours The total idling hours (facility plus frontage road) for each day type for 2012 are provided below, and the growth factor used to project to Table 69: Extended Idling Hours per Day by Day Type and County for 2012 and Growth Factor for 2018 County Mon.-Thu. Fri. Sat. Sun Factor Bastrop Blanco Burnet Caldwell 1, Fayette Hays Lee Llano Milam Travis Williamson 2, , , , TOTAL 7, , , , Section Emissions Estimates CAPCOG obtained emissions rates for CO, NO X, NO, NO 2, HONO, and VOC from TTI s link-based inventories. These used 2012 age distributions, providing for more up-to-date emissions rates than the non-link based inventories would. For non-txled Counties (Blanco and Burnet), CAPCOG divided the Austin-Round Rock MSA emission rates for NO X, NO, NO 2, and HONO by the TxLED adjustment factors. The TxLED adjustment factor for 2012 was , and the factor for 2018 was The table below shows the emission rates used for these updates. Table 70: Extended Idling Emissions Rates for CO, NO X, NO, NO 2, HONO, and VOC (grams per hour) Year Area CO NO X NO NO 2 HONO VOC 2012 TxLED Non-TxLED TxLED Non-TxLED Applying these rates to the activity levels described above produces the following emissions totals for 2012 and Table 71: 2012 and 2018 Summer Weekday (Monday Thursday) Extended Idling Emissions by County (tpd) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet

71 County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Section Hourly Profiles CAPCOG calculated an updated hourly profile based on the hourly distribution provided by TTI in its linkbased inventory for the region. The table below shows the fraction for each hour. Table 72: Hourly Distribution of Idling by Day Type Hour Mon-Thu Fri Sat Sun TOTAL

72 Section Spatial Allocation CAPCOG allocated extended idling emissions directly to the truck stops, frontage roads, and other idling locations in proportion to the estimated daily hours of extended idling. The PAVE plot below shows the spatial allocation of NO X emissions from extended idling for More detailed spatial allocation data is available in the appendices. Figure 13: PAVE Plot for Extended Idling NO X Emissions,

73 Section 5 Non-Road Source Updates CAPCOG made significant updates to the non-road emissions inventories for 2012 and These included updates to agricultural equipment, industrial equipment, construction and mining equipment, residential lawn and garden equipment, and aviation, auxiliary power units, and ground support equipment at Austin-Bergstrom International Airport (ABIA). CAPCOG made no changes to the emissions estimates for commercial equipment, commercial lawn and garden equipment, railway maintenance equipment, recreational equipment, recreational marine equipment, locomotives, oil and gas drill rigs, or aviation sources other than ABIA. For the agricultural equipment, industrial equipment, construction and mining equipment, and residential equipment sources, CAPCOG obtained updated equipment population and/or activity data from research that was more specific to the region or more current than what was currently being used; CAPCOG then used these data with the Texas NONROAD (TexN) model version 1.6 to estimate ozone season day and weekend day emissions. For all TexN runs, CAPCOG used typical year meteorology with all rules enabled, and all post-processing adjustments (TxLED, altitude, soil, climate, ground cover, and reformulate gasoline) enabled. For the ABIA emissions estimates, CAPCOG used data from a report prepared by ERG for TCEQ in Section 5.1 Agricultural Equipment CAPCOG made extensive improvements to the representation of 2012 and 2018 non-road agricultural equipment emissions based on original research. These included updated county-level emissions estimates for all equipment types and the development of highly detailed spatial allocation surrogates for each equipment type based on the actual 2012 agricultural land use. These improvements build on research conducted by CAPCOG in 2012, including updated 2006 tractor emissions estimates 85 and updated spatial surrogates based on 2008 agricultural land use. 86 The default emissions estimates for agricultural equipment are based on default 2012 and 2018 ozone season day runs of TexN. The default equipment populations, annual activity estimates, seasonal activity profile, and weekday/weekend profile were all based on the statewide agricultural equipment emissions inventory prepared by E.H. Pechan and Associates for TCEQ in Eastern Research Group, Inc. Development of Statewide Annual Emissions Inventory and Activity Data for Airports. July 15, ftp://amdaftp.tceq.texas.gov/pub/offroad_ei/airports/tex/erg_statewide_airport_ei_report_july_2011.pdf. Last Accessed July 25, Capital Area Council of Governments. Agricultural Tractor 2006 Ozone Season Weekday Emission Inventory for the CAPCOG Program Area. Austin, TX; August ed.pdf. 86 Capital Area Council of Governments. Spatial Allocation Surrogate Updates for Selected Area and Non-Road Sources in the Austin-Round Rock Metropolitan Statistical Area. Austin, TX; August ed.pdf. _Development_of_Updated_Spatial_Surrogates_for_Selected_Area_and_Non-Road_Sources_Final.pdf 87 E.H. Pechan and Associates. Development of Emissions Inventory of Agricultural Equipment in All Texas Counties Part Two. Durham, NC; August 15,

74 CAPCOG s updates include a combination of top-down and bottom up improvements, and take advantage of a wide range of data sources, including the U.S. Department of Agriculture (USDA s) Census of Agriculture, 88 the USDA s Farm and Ranch Irrigation Survey, 89 a regional survey of tractor operators conducted by ERG for CAPCOG in 2012, 90 Pechan s 2009 study, annual agricultural production statistics from USDA s QuickStats system, 91 the USDA s CropScape data, 92 USDA reports on fuel prices for farmers, 93 and used equipment listings. The table below summarizes CAPCOG s updates. Table 73: Summary of Agricultural Equipment Updates SCC SCC Description Population HP 22xx Wheel Tractors 22xx Agricultural Tractors 22xx Combines 22xx Balers 22xx Mowers 22xx Sprayers 22xx Tillers > 6 HP 22xx Swathers 22xx Other Agricultural Equipment 22xx Irrigation Sets Annual Activity Spatial Allocation The figure and tables below summarize the updated emissions estimates U.S. Department of Agriculture Census of Agriculture. Volume 1, Chapter 2: Texas State and County Data. Issued February 2009, Updated December Last Accessed September 23, U.S. Department of Agriculture Farm and Ranch Irrigation Survey. Volume 3. Special Studies. Part 1. Issued November 2009, Updated February 2010 and July hp. Last accessed September 23, Eastern Research Group, Inc. Agricultural Equipment Emission Inventory Survey Final Report. Prepared for the Capital Area Council of Governments. Austin, TX; December ed.pdf U.S. Department of Agriculture. CropScape 2012 Cropland Data Layer. Last Accessed September 23,

75 Weekday NO X Emissions (tons per day) 2012 and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties Figure 14: Default and Updated Agricultural Equipment NO X Emissions by County Default 2012 Updated 2018 Default 2018 Updated Table 74: Updated Weekday Agricultural Equipment Emissions by County for 2012 and 2018 (tons per day) County 2012 CO 2012 NO X 2012 VOC 2018 CO 2018 NO X 2018 VOC Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total

76 Table 75: Default Weekday Agricultural Equipment Emissions by County for 2012 and 2018 (tons per day) County 2012 CO 2012 NO X 2012 VOC 2018 CO 2018 NO X 2018 VOC Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Table 76: Difference Between Default and Updated Weekday Agricultural Equipment Emissions (from Default to Updated Rates) by County for 2012 and 2018 (tons per day) County 2012 CO 2012 NO X 2012 VOC 2018 CO 2018 NO X 2018 VOC Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Section Equipment Population, Fuel Type, and Horsepower Updates CAPCOG s updates to equipment populations and horsepower characteristics focused on leveraging existing, high-quality data from the Census of Agriculture, the 2012 survey conducted by ERG for CAPCOG, and updated growth projections reflecting more up-to-date agricultural production data. Section Agricultural Tractors CAPCOG used the county-level tractor inventories of tractors reported in the 2002 and 2007 Censuses of Agriculture as the basis for 2012 and 2018 tractor populations. In both 2002 and 2007, the Census of Agriculture reported the total number of agricultural tractors in three HP bins: <40 HP, HP, and 100+ HP. CAPCOG calculated growth projections by county and tractor size, using exponential growth factors rather than linear growth to avoid producing negative values in some of the estimates. The annual exponential growth factors were calculated using the following equation: 76

77 Where: G ij = the annual exponential growth factor for HP bin i in county j, and P ijt = the number of tractors in HP bin i in county j in year t. CAPCOG consulted with Rick Baker at ERG as to whether exponential or linear growth would be more appropriate for these projections. He indicated that while exponential growth likely underestimates the reduction in agricultural production in the region from 2007 to 2012, it provides the best way to take advantage of all of the data and avoid negative values for certain HP bins in certain counties. Equipment populations for 2012 and 2018 were calculated as follows: The tables below show the equipment populations for each county reported in the 2002 Census of Agriculture 94 and 2007 Census of Agriculture 95, the annual growth factor (GF), and the projections for 2012 and Table 77: 2002, 2007, 2012 and 2018 Tractor Populations for <40 HP Tractors County % of 2002 GF Bastrop 1,453 1, % Blanco % Burnet 1, % Caldwell 1, % Fayette 2,044 1, % Hays % Lee 910 1, % Llano % Milam 1,133 1, % Travis % Williamson 1,775 1, % TOTAL 12,334 9, % , , National Agricultural Statistics Service Census of Agriculture. Volume 1, Part 43A: Texas State and County Data. Table 38: Machinery and Equipment on Operation: 2002 and U.S. Department of Agriculture, June df. Last accessed August 7, National Agricultural Statistics Service Census of Agriculture. Volume 1, Part 43A: Texas State and County Data. Table 41: Machinery and Equipment on Operation: 2007 and U.S. Department of Agriculture, Issued February 2009, Updated December _041_041.pdf. Last accessed August 7,

78 Table 78: 2002, 2007, 2012 and 2018 Tractor Populations for HP Tractors County % of 2002 GF Bastrop 1,658 1, % Blanco % Burnet 839 1, % Caldwell 1,036 1, % Fayette 2,387 2, % Hays % Lee 1,580 1, % Llano % Milam 1,934 1, % Travis % Williamson 2,029 1, % TOTAL 13,731 12, % , ,612.4 Table 79: 2002, 2007, 2012 and 2018 Tractor Populations for 100+ HP Tractors County % of 2002 GF Bastrop % Blanco % Burnet % Caldwell % Fayette % Hays % Lee % Llano % Milam % Travis % Williamson % TOTAL 3,686 3, % , ,379.6 The updated equipment populations result in much higher estimates of the number of tractors in the <40 HP and HP tractor groupings, but not much of a difference in the estimates of emissions for the 100+ HP group. 78

79 Figure 15: 2012 and 2018 Default and Updated Region-Wide Agricultural Tractor Equipment Populations 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 <40 HP HP 100+ HP 2012 Default 2012 Updated 2018 Default 2018 Updated CAPCOG allocated the tractor equipment populations to each fuel type based on survey data collected by ERG within the 11-county area in 2012 as documented in CAPCOG s 2013 report on agricultural tractor emissions. 96 This resulted in many more small (<100 HP), gasoline-powered tractors, which are included in TexN for some (but not all) HP bins, and LPG tractors, which are not included in TexN at all. The table below shows the fuel type allocations for each HP bin. Table 80: Allocation of Tractor Populations to Fuel Types HP Range Diesel Allocation Gasoline Allocation LPG Allocation <40 HP 55.5% 42.5% 2.5% HP 89.4% 7.6% 3.0% 100+ HP 100.0% 0.0% 0.0% CAPCOG also used ERG s survey results to update the allocation of equipment populations for each of the three horsepower bins used by the Census of Agriculture to the TexN horsepower bins. Each county s tractor populations within the <40 HP, HP, and 100+ HP ranges were allocated to each TexN HP bin based on the survey results relative distribution of tractors within the Census of Agriculture HP bins. The table below shows the default and updated TexN HP bin allocations for each of the Census of Agriculture HP groupings. 96 Capital Area Council of Governments. Agricultural Tractor 2006 Ozone Season Weekday Emission Inventory for the CAPCOG Program Area. Austin, TX; August ed.pdf. 79

80 Table 81: Default and Updated Allocations of <40 HP, HP, and 100+ HP Tractor Populations to Horsepower Bins HP Min HP Max Tractors Allocated Default % Updated % <40 HP <0.1% 7.5% <40 HP 27.3% 27.5% <40 HP 72.7% 65.0% HP 25.1% 16.7% HP 39.7% 47.0% HP 35.2% 36.4% HP 43.3% 87.1% HP 37.3% 11.9% HP 19.3% 1.0% HP <0.1% 0.0% Finally, CAPCOG also used ERG s survey data to update the average HP ratings for several TexN HP bins, shown in the table below marked with an asterisk. These results were statistically significantly different from the average horsepower ratings at a 95% confidence level. Table 82: Default and Surveyed Average Horsepower Ratings HP Bin Default Avg. HP Survey Avg. HP HP* HP HP HP* HP* HP HP* HP HP Section Combines CAPCOG used county-level equipment population data from the 2002 and 2007 Censuses of Agriculture to update the combine equipment populations for 2012 and Since Blanco County had two farms that reported owning a combine in 2007, but the total number of combines was not disclosed, CAPCOG assumed that there two combines for that county in CAPCOG calculated a region-wide exponential growth factor to apply to 2007 equipment populations since combines can often be used across county lines for harvesting, and since small changes in the populations for counties with few combines could skew the results for those counties. From 2002 to 2007, the number of combines reported within the region declined from 498 to 395, a decrease of 21%. This translates to an annual exponential decrease of 4.5% per year, resulting in a region-wide population of 313 combines in 2012 and 237 combines in This trend results in somewhat higher populations than the trends in crop production data from grain harvesting would produce. 80

81 Table 83: 2002, 2007, 2012, and 2018 Combine Equipment Populations County Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Region Subtotal CAPCOG allocated the 2012 and 2018 equipment populations to gasoline and diesel fuel types based on the default ratios in TCEQ s 2007 emissions inventory, prepared by Pechan 97 : Table 84: Statewide Distribution of Combines by Fuel Type in 2007 Fuel Type Number % Diesel 8, % Gasoline % Total 9, % 97 Thesing, Kirsten. Agricultural Equipment in All Texas Counties Part Two. E.H. Pechan & Associates. August 15, Last accessed August 5,

82 Figure 163: 2012 and 2018 Default and Updated Region-Wide Combine Equipment Populations Default Updated Section Irrigation Sets CAPCOG updated the irrigation set equipment populations in each county based on statewide equipment population data reported in the USDA s Farm and Ranch Irrigation Survey and the amount of irrigated land in each county as reported in the 2007 Census of Agriculture. The Farm and Ranch Irrigation Survey provides statewide counts of the number of irrigation pumps by fuel type used on farms and ranches. The following table shows the statewide number of irrigation pumps reported in the survey from 1988 through Table 85: Statewide Counts of Farm and Ranch Irrigation Pumps by Fuel Type: 1988, 1994, 1998, 2003, and 2008 Year Electric Natural Gas LPG Diesel Gasoline Total Pumps ,750 22,025 2,090 1, , ,855 20, , , ,700 20,924 1,028 5, , ,713 15, , , ,437 12, , ,187 These data show a clear decline in engine-powered irrigation pumps over this time period. The following figure shows the total number of engine-powered irrigation sets and electric-powered irrigation sets. By applying a linear best-fit line to these data, CAPCOG obtained a slope of -441, meaning that the number of engine-powered irrigation sets can be expected to decrease statewide by 441 per year. 82

83 Figure 17: Engine-Powered and Electric-Powered Irrigation Pumps in Texas by Year 70,000 60,000 50,000 40,000 30,000 Engine-Powered Electric-Powered Linear (Engine-Powered) 20,000 10,000 y = x + 903, R² = CAPCOG projected the total statewide 2008 engine-powered irrigation set population from 2008 to 2012 and 2018 and then allocated the population to each fuel type based on the each fuel type s fraction of internal combustion irrigation pumps reported for While the downward trend for internal combustion irrigation sets appears to be primarily driven by the decrease in the dominant fuel type natural gas there are not clearly discernible patterns for any of the other fuel types that would allow CAPCOG to confidently make separate projections by fuel type. Table 86: Projected Statewide Irrigation Set Populations by Fuel Type, 2012 and 2018 Year Total IC Pumps CNG LPG Diesel Gasoline ,986 11, , ,340 9, , CAPCOG then allocated these statewide totals to each county based on the county s fraction of the statewide total of irrigated land as reported in the 2007 Census of Agriculture, as the table below shows. Table 87: Allocation of Statewide Irrigation Set Populations (all fuel types) to Counties Geography Irrigated Land (acres) % Statewide 2012 Pop Pop. Bastrop County 2, % Blanco County 1, % Burnet County 1, % Caldwell County % Fayette County 1, % Hays County %

84 Geography Irrigated Land (acres) % Statewide 2012 Pop Pop. Lee County 3, % Llano County % Milam County 2, % Travis County 1, % Williamson County % Region Subtotal 16, % Texas Total 5,010, % 14,986 12,340 Each county s irrigation set population was then allocated to each fuel type based on the following ratios from the 2008 survey, as the table below shows. Table 88: 2008 Allocation of Irrigation Sets by Fuel Type Fuel Type Number Percentage Natural Gas 12,998 78% LPG 417 2% Diesel 2,428 14% Gasoline 907 5% These updates result in quite dramatic changes from the default population estimates, as the figure below shows. Figure 18: 2012 and 2018 Default and Updated Irrigation Set Equipment Populations Natural Gas LPG Diesel Gasoline 2012 Default 2012 Updated 2018 Default 2018 Updated 84

85 Section Two-Wheel Tractors, Balers, Mowers, Sprayers, Tillers, and Swathers For two-wheeled tractors, hay balers, mowers, sprayers, tillers and swathers, the most recent data on equipment populations and activity are from Pechan s 2007 statewide survey for TCEQ. Since the 2007 survey coincided with the 2007 Census of Agriculture, it provides the best existing representation of equipment populations in However, the default equipment populations in TexN for 2012 and 2018 do not account for the decline in agricultural production that has been occurring in the region, which CAPCOG expects to continue from 2012 through Whereas TexN predicts fairly significant growth in agricultural equipment populations over this period, the continued, steady decline the region s agricultural production strongly suggests the need to adjust equipment population projections for these source types in years beyond 2007 for this region. The default equipment populations also do not account for what data is available on agricultural mower populations from earlier Censuses of Agriculture. Based on CAPCOG s review of Pechan s survey, CAPCOG determined that the best available growth surrogate for all six of these equipment types was the regional cattle inventory as reported in the USDA s annual survey. Based on a review of the equipment ratios developed by Pechan from its 2007 survey, all of these equipment types are more closely associated with cattle production or hay production than cotton, wheat, or other crop production. And since hay is often used to feed cattle and often is co-located with cattle production, hay production closely follows cattle production. The wheat, hay, cotton, and other crop production equipment ratios are directly comparable since they have the same type of denominator (acres harvested). It is also possible to make the cattle equipment ratios more directly comparable to the other ratios by multiplying them by an equivalent acreage of hay production. CAPCOG calculated a ratio of cattle per acre of forage harvested based on the statewide total of 13,709,543 head of cattle and calves and the 5,264,287 acres of forage harvested reported in the 2007 Census of Agriculture. The table below shows the Pechan equipment ratios along with the adjusted cattle equipment ratios. Table 89: Pechan Equipment Ratios for Two-Wheel Tractors, Balers, Mowers, Sprayers, Tillers, and Swathers by Farm Type Equipment Cotton (per 1K acres) Cattle (per 2,604 head) Hay (per 1K acres) Wheat (per 1K acres) Other (per 1K acres) Two-Wheel Tractors Hay Balers Mowers Sprayers Tillers Swathers Not only are the equipment ratios for cattle and hay much higher than for the three other farm types, cattle and hay production are also by far the dominant agricultural activities within the region, making up the primary activity at 62% and 11% of the region s farms in the region, respectively. Since the USDA s annual survey no longer reports data on hay production, and given the close association of hay 85

86 production and cattle inventories, CAPCOG is confident that the best growth surrogates available for all six of these equipment types are cattle inventories reported in USDA s annual survey. Table 90: Farms, Forage Acreage, and Cattle and Calves Inventories for Selected Farm Types, Statewide Farm Type (NAICS Code) Farms Forage Acreage Cattle and Calves Surgarcane, Hay, and Other Crop Farming (11193, 11194, 11199) 47,265 1,677, ,269 Cattle Ranching (112111) 124,992 2,796,639 8,396,943 All 247,437 5,264,287 13,709,543 CAPCOG obtained the cattle inventories for each county in the region from January 1, 1995, through January 1, 2013, from the USDA s annual survey. The table below shows the region s total cattle inventory as of January 1 st of each year, representing the previous year s inventory level (January 1, 2013, represents the inventory for 2012). The data show a steady decline in cattle inventories over this time period, including a sharp decline in 2011 due to large-scale sell-offs of herds due to the drought that year. Figure 19: Cattle Inventories in CAPCOG Program Area as of January 1st, , , , , , , ,000 y = -12,215.69x + 25,118, R² = 0.79 Cattle Linear (Cattle) 450, , CAPCOG created a modified NATION.GRW file that sets growth factors equivalent to the cattle inventory equation below, based on the best line fit shown above. ( ) 86

87 CAPCOG then ran TexN for two-wheeled tractors, hay balers, agricultural mowers, sprayers, tillers, and swathers for 2012 and 2018, with the only modification being the NATION.GRW file. CAPCOG then used the AGGREGATE.OUT files for the two runs to calculate emission rates (tons per piece of equipment per ozone season day). CAPCOG then obtained the default 2007 equipment populations to serve as a baseline for projecting 2012 and 2018 equipment populations. Since the 2007 equipment populations are based on Pechan s 2007 survey and production data from the 2007 Census of Agriculture, it should provide the most robust baseline estimate available for that year. CAPCOG then calculated growth factors for 2012 and 2018 from a 2007 baseline by dividing the predicted cattle inventories for those years by the actual cattle inventory for 2007 as reported in the annual survey. ( ) ( ) CAPCOG then multiplied the 2007 populations by the appropriate growth factors in order to obtain 2012 and 2018 equipment populations. One additional adjustment was made for agricultural mower populations in order to account for data from older Census of Agriculture data on this equipment type. From 1974 to 1992, the Census of Agriculture asked respondents to report the number of agricultural mowers they operated on their farms. CAPCOG developed an equipment population adjustment factor based on the ratio of mowers reported in the 1987 and 1992 Census of Agriculture and the number of mowers that Pechan s equipment ratios would yield for statewide populations in 1987 and In order to update the equipment populations, CAPCOG used the equipment ratios for mowers that Pechan developed for TCEQ based on a 2007 survey, along with agricultural production data from the 1987 and 1992 Census of Agriculture. CAPCOG applied the ratios developed by Pechan to the 1987 and 1992 populations in order to produce equipment populations. The table below shows the data that were used in these calculations for the region. Table 91: Pechan Mower Equipment Ratios and Derived Statewide Equipment Populations for 1987 and 1992 for Texas Product Mowers Qty. Units Cotton ,601 Hay ,604 Acres Hvst. Acres Hvst. Ratio (pieces per unit of production) 1987 Production 1992 Production 1987 Mowers 1992 Mowers ,349,755 3,620, ,252,216 3,607,387 4, ,

88 Product Mowers Qty. Units Wheat ,271 Other ,563 Beef Cattle Acres Hvst. Acres Hvst. Ratio (pieces per unit of production) 1987 Production 1992 Production 1987 Mowers 1992 Mowers ,649,104 3,726, ,270,240 5,567,641 2, , ,336 Head ,020,910 13,242,832 12, ,971.1 Since the actual number of agricultural mowers were available for each county in the region for each of these Censuses, it was possible to directly compare the equipment populations derived from Pechan s survey to the actual reported equipment inventories for those years. The table below shows a comparison of the 1987 and 1992 Census of Agriculture s statewide count of mowers and estimates derived from the equipment ratios developed by Pechan, which were applied to 1987 and 1992 production data from the Censuses of Agriculture conducted for those years. Table 92: Comparison of Survey-Derived Statewide Mower Populations and Census of Agriculture Populations Estimate Population Derived from Pechan Survey 12, ,971.1 Census of Agriculture Population 21, ,898.0 Scaling Factor (census/survey) The scaling did not change much between the two years, indicating a robust relationship between the survey-derived results and the actual equipment populations as reported in the Census of Agriculture. CAPCOG averaged the scaling factor and then calculated the 2007 equipment populations for each county using the 2007 Census of Agriculture production data, the Pechan equipment ratios, and the average scaling factor ( ). CAPCOG then applied this adjustment factor to 2012 and 2018 equipment populations projected from the 2007 baseline described above. While the 2007 Census of Agriculture does include data on hay balers, the vast majority of hay balers today operate off of a power-take-off bar (PTO) from a tractor. Only about 3% of hay balers are selfpropelled, based on the number of hay balers in the NONROAD 1998 and 2000 base year equipment populations and the nation-wide number of hay balers reported in the 1997 and 2002 Census of Agriculture; NONROAD has a 1998 base-year population of 17,993 gasoline-powered hay balers and 2000 base-year equipment population of 5,529 diesel-powered hay balers, compared to 773,141 hay balers reported nation-wide in the 1997 Census of Agriculture and 766,457 in the 2002 Census of Agriculture. Pechan s report estimated a state-wide total of 13,188 self-propelled hay balers. This represents 31 percent of the statewide total of 43,206 hay balers reported in the 2007 Census of Agriculture, and 56% of all of the hay balers reported in the base year equipment populations for NONROAD, leading CAPCOG 88

89 to believe that the equipment population estimate is likely high. One reason for this might be the wording of the questions in Pechan s survey. The survey first asked whether the respondent owned or operated balers and then asked how many it operated and what percent were powered by diesel, gasoline, LPG, or CNG. Given that Pechan s estimate of diesel-powered balers in Texas exceeds the nation-wide population estimate in the NONROAD model for 2000, and given the similarity in fuel type distribution to agricultural tractors, CAPCOG believes that many respondents likely reported the engine type of the tractor used to pull the hay baler rather than reporting that the baler had its own, independent engine. Table 93: Comparison of Hay Baler and Agricultural Tractor Fuel Type Distributions in Pechan Study Equipment Diesel Gasoline LPG CNG TOTAL Ag. Tractors (#) 134,608 4,960 1, ,025 Balers (#) 12, ,189 Ag. Tractor (%) % 3.517% 0.999% 0.034% % Balers (%) % 2.684% 1.691% 0.030% % In Pechan s survey, 16% of respondents primarily engaged in raising cattle (128 of 788 responses) and 26% of respondents primarily engaged in producing hay (162 of 622 responses) reported owning or operating hay balers. Pechan s survey instrument did not specifically ask whether the hay balers being reported were self-propelled, or whether they used a power-take-off bar. ERG s 2012 survey for CAPCOG, on the other hand, did specifically ask survey respondents whether their hay balers were PTOdriven or self-propelled, and of the 108 survey respondents (of which 95 reported primarily hay or livestock production) none reported that their hay balers were self-propelled. Since ERG was not able to collect any responses from farm management companies, which would be more likely to own and operate self-propelled equipment, it did not recommend setting the equipment populations for this equipment type to zero. To some extent, a higher equipment population may still help represent the added emissions a tractor would produce with the added load of operating a PTO-driven implement such as a hay baler, so using Pechan s estimates for hay balers are not necessarily as problematic as it might seem given the wide disparity between Pechan s survey results for hay balers ERG s survey results for hay balers. 89

90 Figure 20: 2012 and Wheel Tractor, Baler, Mower, Sprayer, Tiller, and Swather Equipment Populations 2,500 2,000 1,500 1, Wheel Tractors Balers Mowers Sprayers Tillers > 6 HP Swathers 2012 Default 2012 Updated 2018 Default 2018 Updated Section Other Equipment: Forage Harvesters The Census of Agriculture began reporting the number of self-propelled forage harvesters at the county level in These are large pieces of equipment that are used to collect and process grass and row crops for use as animal feed. CAPCOG used the 2007 equipment counts and projected them to 2012 and 2018 using a best-fit line for the region-wide cattle inventory reported in the USDA s annual survey from January 1, 2003 January 1, 2013 (each representing the inventory of the immediately preceding year) and applied a best fit line to obtain an estimate for the annual change in head of cattle over this time period. CAPCOG then projected that forward from January 1, 2013, to January 1, 2019, to represent the number of head of cattle in the region in CAPCOG used 2002 as a baseline since there were also equipment populations available for There relatively smooth line indicates that this is likely a robust, long-term trend, possibly as a result of farms being purchased and converted into housing for the fast-growing population in the region. 90

91 Figure 21: Head of Cattle as of January 1, , , , , , ,000 y = x + 3E+07 R² = , , The following table shows the cattle inventory by year for 2007, which was used as the base year for forage harvester populations, for 2012, and for the projections to 2018 based on a loss of 16,115 cattle per year from 2012 to Table 94: Baseline and Projected Cattle Populations by Year, 2007, 2012, and Year Cattle Ratio to , , , CAPCOG then multiplied the 2007 forage harvester equipment populations in each county by the regionwide ratios of 2012 and 2018 cattle inventories to The equipment populations by year for each county are listed below. Table 95: Forage Harvester Equipment Population by County and Year County Bastrop Blanco Burnet Caldwell Fayette Hays Lee

92 County Llano Milam Travis Williamson Total Since forage harvesters does not have a distinct equipment profile in NONROAD or TexN, CAPCOG developed an equipment profile using collected data from used farm equipment listings on tractorhouse.com. Based on 67 self-propelled forage harvesters listed, all of which were diesel-powered, CAPCOG estimated an average horsepower rating of 588 horsepower (+/- 44 at a 95% confidence level), and allocated all of the equipment populations into the HP horsepower bin and updated the average HP rating for the bin to 588. Section Other Equipment: Cotton Pickers and Strippers Cotton pickers and strippers are high-horsepower, self-propelled equipment used to harvest cotton. As a baseline for the cotton picker and stripper equipment population, CAPCOG used the number of cotton pickers and strippers in each county from the 2007 Census of Agriculture. For Bastrop, Burnet, Fayette, Hays, and Lee Counties, only the total number of farms that reported owning a cotton picker or stripper was reported, not the actual number of cotton pickers or strippers. For these counties, CAPCOG first calculated ratio of cotton pickers and strippers to farms reporting cotton pickers and strippers (1.2), and then multiplied the number of farms reporting a cotton picker or stripper by the ratio. In order to project 2012 and 2018 equipment populations from that baseline, CAPCOG used data on cotton acreage harvested from the USDA s annual survey. In order to estimate 2012 populations, CAPCOG multiplied the 2007 populations by the ratio of cotton acreage harvested in 2012 to cotton acreage harvested in 2007 for the whole region from the County Agricultural Production Survey (CAPS). CAPCOG used the 2012 values for 2018, since the data do not seem to indicate any clear pattern in the number of acres of cotton harvested from either or from 2007 to 2012, as the chart below shows. If a trend line were used, the 2018 equipment count would only be 3% lower than the 2012 population. Given the erratic nature of the cotton harvesting data for the region, CAPCOG believes that using the 2012 populations for 2018 is reasonable. 92

93 Figure 22: Acres of Cotton Harvested by Year 70,000 60,000 50,000 40,000 y = x R² = ,000 20,000 10, The table below shows the 2007 equipment populations obtained from the 2007 Census of Agriculture, and the projected 2012 and 2018 equipment populations. Table 96: 2007, 2012, and 2018 Cotton Picker and Stripper Populations by County County Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson TOTAL As was needed for forage harvesters, CAPCOG also needed to develop an equipment profile for cotton pickers and strippers apart from the broader other agricultural equipment category. CAPCOG obtained used cotton picker and stripper listings from tractorhouse.com in order to estimate the average horsepower for this equipment type. The average horsepower rating for the 364 cotton pickers and strippers reviewed was 318 (+/- 13 at a 95% confidence level). CAPCOG allocated all of the equipment populations into the HP horsepower bin and updated the average HP rating for the bin to

94 Section Other Equipment Aside from Forage Harvesters and Cotton Pickers CAPCOG updated the equipment populations for other agricultural equipment in a manner similar to the procedure used for two-wheel tractors, balers, mowers, sprayers, tillers, and swathers. CAPCOG projected the 2007 equipment populations to 2012 and 2018 (except for CNG equipment, which did not have default equipment populations for 2012 and 2018) using the growth factors described in section Table 97: 2012 "Other Agricultural Equipment" Populations County Gasoline LPG Diesel Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson TOTAL Table 98: 2018 "Other Agricultural Equipment" Populations County Gasoline LPG Diesel Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson TOTAL Section Activity Updates CAPCOG updated annual activity levels for all agricultural tractors and irrigation sets, and obtained activity estimates for two specific equipment types categorized under other agricultural equipment: forage harvesters and cotton pickers. For all other equipment types, CAPCOG used default annual 94

95 Annual Activity (hours/year) 2012 and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties activity levels. CAPCOG also retained all of the default weekday/weekend and seasonal activity allocations. Section Agricultural Tractors CAPCOG modeled five different annual activity levels for five different tractor horsepower groupings in order to reflect the 2012 survey results. The chart below shows the default NONROAD and TexN annual activity levels compared to the activity levels CAPCOG used. CAPCOG found statistically significantly different activity levels at the <40 HP, HP, HP, HP, and 100+ HP ranges, and used these activity levels for all fuel types within those ranges. This resulted in much lower activity levels than TexN defaults. The following chart shows a comparison between NONROAD, TexN, and CAPCOG s updated activity levels. Figure 23: Comparison of Spark-Ignition (S.I.) and Compression-Ignition (C.I.) Annual Activity Estimates (hours/year) <40 HP S.I HP S.I HP S.I HP S.I HP S.I. <40 HP C.I HP C.I HP C.I HP C.I HP C.I. Default CAPCOG Updated Section Other Equipment CAPCOG used the annual usage reported in used forage harvester and cotton picker listings on tractorhouse.com and the difference between 2013 and the model year in order to calculate the average annual activity levels for these two types of other agricultural equipment. CAPCOG calculated an average annual use for forage harvesters of 337 hours per year (+/- 47 at a 95% confidence level, N=67) for forage harvesters based on used equipment listings on tractorhouse.com. Average annual 95

96 activity for the 364 cotton pickers and strippers was 263 hours per year (+/- 22 at a 95% confidence level). CAPCOG did not change the activity estimates for the remaining other equipment. The following table These equipment types were exclusively high-horsepower (over 100 HP), diesel equipment, so CAPCOG did not update the TexN default activity levels for any spark-ignition other agricultural equipment or any diesel-powered equipment under 100 horsepower. The chart below shows the default TexN activity levels for spark-ignition (S.I.) other agricultural equipment (gasoline, LPG, and CNG) and compression-ignition (C.I.) other agricultural equipment (diesel), as well as the average annual activity calculated for forage harvesters and cotton pickers and strippers. Figure 24: Activity Levels for "Other Agricultural Equipment" TexN S.I. TexN C.I. Forage Harvesters Cotton Pickers/Strippers Section Irrigation Sets CAPCOG used the 2008 Farm and Ranch Irrigation Survey s data on fuel expenditures for irrigation in order to estimate average annual activity for each engine type. By using energy price information, CAPCOG was able to calculate the total quantities of fuel consumed by the irrigation pumps reported in the 2008 survey. The price for natural gas came from the Energy Information Administration s (EIA s) Natural Gas Prices for industrial consumers in Texas in 2008 (since farming and ranching establishments are classified as industrial establishments for EIA reporting). 98 For diesel, gasoline, and LPG prices, CAPCOG used data from the USDA on prices paid by farmers and ranchers in the southern 98 Energy Information Administration. Texas Natural Gas Prices. Last accessed August 8,

97 plains region in The following table shows the statewide expenditures for irrigation fuel, the price per unit of energy, and the calculated fuel consumption totals for Table 99: Texas Irrigation Fuel Expenses, Prices, and Quantity Consumed, 2008 Fuel Type Expenses ($1,000) Price ($ per unit) Units Fuel Consumed Natural Gas 194, MCF 21,707,478 MCF LPG 3, Gallon 1,540,140 Gallons Diesel 32, Gallon 9,062,710 Gallons Gasoline 1, Gallon 426,514 Gallons The total fuel consumed can then be divided by the number of pumps to get the average fuel consumed per pump per year. That fuel consumption rate can then be divided by the fuel consumption rate for each fuel type when operated at 8,760 hours per year in order to obtain a factor that can be used to calculate activity levels. CAPCOG performed a TexN run for 2008 with the base year NONROAD equipment populations and a 8760 hours/year activity level in order to obtain annual fuel consumption rates. Since TexN reports natural gas consumption in gallons, it is necessary to convert MCF into gallons. CAPCOG used a ratio of gallons per cubic foot. Implicit in these calculations is the assumption that the horsepower distribution for these irrigation sets was the same in 1998 or The updated fuel consumption rates and the 8760 hours/year default fuel consumption rates are presented in the table below. Table 100: Comparison of Updated and 8760 Hour/Year Irrigation Fuel Consumption Rates (gallons per piece per year) Fuel Type Fuel Consumed Pumps Conversion Fuel Consumption Updated Fuel 8,760 hours Activity Ratio Natural Gas 21,707,478 MCF 12,998 7,481 gallons/mcf 12,492,939 gal/pump/year 33,927,212 gal/pump/year LPG 1,540,140 gal. 417 n/a 3,693 gal/pump/year 53,502 gal/pump/year Diesel 9,062,710 gal. 2,428 n/al 3,733 gal/pump/year 20,819 gal/pump/year Gasoline 426,514 gal. 907 n/a 470 gal/pump/year 28,875 gal/pump/year Multiplying 8,760 hours per year by these ratios yields updated activity levels. These are presented in the chart below, along with the default NONROAD and TexN activity levels. 99.National Agricultural Statistics Service. Agricultural Prices 2008 Summary. Prices Paid: Fuels, Regions, and United States, April , page 203. United States Department of Agriculture, August Last accessed August 8,

98 Figure 25: Default TexN and Updated Irrigation Set Activity Levels (hours/year) 3,500 3,000 2,500 2,000 1,500 Default Updated 1, Natural Gas LPG Gasoline Diesel The updated activity levels for the two most prevalent fuel types natural gas and diesel are remarkably similar to the activity levels calculated by Pechan based on its 2007 statewide survey. The activity level for natural gas irrigation sets is only 10% higher than the Pechan estimate, and the activity level for diesel irrigations sets is only 2% higher than the Pechan estimate. However, for the less prevalent fuel types LPG and gasoline the update activity levels were 79% and 95% lower, respectively. Section Growth and Scrappage Updates CAPCOG updated the NATION.GRW file used by TexN to calculate the age distributions for each equipment type in order to reflect state-wide or region-wide equipment population data and projections. CAPCOG also updated the scrappage function for several types of equipment in order to produce age distributions that more closely matched ERG s 2012 survey results. The table below shows the growth factors used by CAPCOG and whether the scrappage curve was updated or not. For comparison, the default growth rates for diesel, gasoline, LPG, and CNG agricultural equipment is also presented. Table 101: Growth Factors Used in Agricultural Equipment Emissions Modeling Equipment Types Ag Tractors <40 HP 1998 Growth Factor 2000 Growth Factor 2005 Growth Factor 2010 Growth Factor 2015 Growth Factor 2025 Growth Factor Scrappage Updated? YES 98

99 Equipment Types 1998 Growth Factor 2000 Growth Factor 2005 Growth Factor 2010 Growth Factor 2015 Growth Factor 2025 Growth Factor Scrappage Updated? Ag Tractors HP YES Ag Tractors 100+ HP YES Combines YES Irrigation Sets YES Cotton Pickers and NO Strippers All Other Equipment NO Default Diesel N/A Default Gasoline N/A Default LPG N/A Default CNG N/A CAPCOG calculated these growth factors as the ratio of various equipment population estimates to a 1996 base year times The assumptions CAPCOG used for these calculations were as follows: Ag Tractors and Combines: o For 1992, 1997, 2002, and 2007, CAPCOG used the region-wide equipment populations as reported in Census of Agriculture ( HP population estimated using the 40+ HP population in 1992 and the percentage of 40+ HP tractors that the 100+ HP tractors made up in the 1997 Census). o For , , and , CAPCOG interpolated the values. o For , CAPCOG extrapolated the exponential growth between 2002 and 2007 (4.37%, 1.48%, 1.44%, and 4.53% annual reductions for <40 HP tractors, HP tractors, 100+ HP tractors, and combines, respectively). Irrigation Sets: o For 1988, 1993, 1998, 2003, and 2008, CAPCOG used statewide total of engine-powered irrigation pumps, o For , , , and , CAPCOG interpolated the statewide populations, o For 2009 and after, CAPCOG used the annual decrease in combustion-powered irrigation equipment (441 per year) using a best-fit line for the 1988, 1993, 1998, 2003, and 2008 equipment populations. Cotton Pickers: o For 1992, 1997, 2002, and 2007, CAPCOG used the region-wide equipment populations as reported in Census of Agriculture. o For , , and , CAPCOG interpolated the values. 99

100 o For , CAPCOG used the ratio of cotton acreage harvested in that year to the acreage harvested in 2007 as reported in the annual USDA survey. o For , CAPCOG assumed the same population as in All Other Equipment (Two-Wheel Tractors, Balers, Mowers, Sprayers, Tillers, Swathers, Forage Harvesters, and Other Equipment aside from Cotton Pickers and Forage Harvesters): o For 1992, 1997, 2002, and 2007, CAPCOG used the statewide tractor population as reported in Census of Agriculture as a surrogate. o For , , and , CAPCOG interpolated the values. o For , CAPCOG extrapolated the linear growth between 2002 and 2007 (resulting in a decrease in the growth factor value of 19 for each year after 2007). CAPCOG updated the scrappage curves for agricultural tractors in order to produce a tractor age distribution that more closely matched the age distribution in ERG s 2012 survey. The 2012 survey results showed a relatively flat distribution that did not match the shape generated by the default scrappage curve, and which could be roughly reproduced using a scrappage function that assumes that farmers and ranchers will keep using a tractor until it has reached its full useful life, or: CAPCOG consulted with Rick Baker from ERG prior to performing this update in order to assess whether it would be an appropriate application of ERG s survey data. Rick s response was, your conclusion that the overall tractor age distribution is drastically skewed relative to the NONROAD default is completely defensible - this can be demonstrated using something like a Chi-square analysis grouping the observed and NONROAD distributions into a few categories - e.g., 0-5 yrs, 5-10 yrs, 10-20, 20+. It could be easily demonstrated that the simplification you propose - essentially a step function with the step at 2 x useful life - quantitatively matches the observed distribution much more closely than the default NONROAD logistics curve. 100 He also stated that your recommendation, while of course a simplification of the real-world situation, is an improvement over the default assumption. 101 The figure below shows a comparison of the 2012 ERG survey age distributions for the horsepower range, which was the most prevalent horsepower range in ERG s survey, compared to the age distribution by CAPCOG using the updated scrappage and growth data, as well as the default TexN age distribution, the age distribution using updated annual activity and a default NATION.GRW file, and the age distribution using updated annual activity and updated growth data in the NATION.GRW file without an updated scrappage curve. As the figure shows, the age distribution used by CAPCOG is the one that matches the survey data the closest. 100 Rick Baker. communication to Andrew Hoekzema, February 14, Rick Baker. communication to Andrew Hoekzema, February 20,

101 and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties Figure 26: Cumulative Age Distribution of HP Tractors by Model Year 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% ERG TexN Default CAPCOG Updated TexN Default with Updated Annual Activity TexN Default with Updated Annual Activity and Growth In order to quantitatively evaluate the difference, CAPCOG performed chi-squared tests for the observed and expected distribution of the tractors surveyed using the age distributions generated by CAPCOG s updated scrappage function and the default scrappage function (with the activity level and population growth estimates updated in both cases). CAPCOG used the number of tractors in each emissions tier as the basis for the analysis since this is the most relevant aspect of the impact of the age distribution on emissions. The table below shows the values used for the chi-squared test and the p- value of the results: Table 102: Chi-Squared Test of Independence for HP Tractor Age Distributions Model Years Tier Observed Expected: Default Scrappage Expected: Updated Scrappage 1996 and Earlier Chi 2 test p-value n/a n/a

102 These results mean that for a 95% confidence level, the observed age distribution was statistically significantly different from the expected age distribution when using the default scrappage function, but not statistically significant from the age distribution when using the updated expected scrappage function. The most likely explanation for these data is that farmers and ranchers are less likely to invest in new equipment in a sector with declining production than one with increasing production, and would tend to continue making repairs to older equipment until it became completely unusable. Given these results, CAPCOG decided to use the updated scrappage function for all tractors, as well as for combines and irrigation equipment. All three equipment types have many years of equipment population data from the Census of Agriculture and experience similar patterns in population decreases between 2002/2003 and 2007/2008 that almost certainly continued through 2012 given trends in regional agricultural production. An additional support for the application of these data to combines is the similarity to the reported age distributions of combines and 100+ HP tractors in the 2007 Census of Agriculture: 10% and 11% of the regional equipment populations were manufactured in the past 5 years, respectively. CAPCOG did not change the scrappage function for forage harvesters, cotton pickers, two-wheel tractors, balers, mowers, sprayers, tillers, swathers, or any other agricultural equipment. CAPCOG s review of the age distribution of the forage harvesters and cotton pickers indicated that these equipment populations followed a significantly different age distribution pattern that was more similar to the default distribution generated by the default NATION.GRW file used by TexN. Since there was no clear evidence that would suggest the other equipment types would exhibit scrappage patterns similar tractors and combines, CAPCOG used the existing scrappage curve for these other equipment types. Section Emissions Modeling and Calculations CAPCOG used TexN to generate emissions rates for all agricultural equipment. Due to the complexity of some of the updates, CAPCOG could not simply create a single 2012 scenario and a single 2018 scenario to incorporate all of the changes to equipment populations, engine horsepower ratings, and activity levels simultaneously. As a result, CAPCOG had to perform numerous TexN runs in order to obtain the emissions data necessary for these updates. This section provides details on the steps CAPCOG took to use TexN and equipment population data to generate emissions totals. Section Agricultural Tractor Emissions Modeling In order to model the emissions for agricultural tractors, CAPCOG set up and ran the following model runs for 2012 and 2018: 100+ Horsepower, Horsepower, Horsepower, Horsepower, <40 Horsepower Diesel, <40 Horsepower Gasoline, and <40 Horsepower LPG. 102

103 For some of the horsepower ranges for gasoline tractors and for all of the LPG tractors, there is no way to directly model the emissions in TexN. As a result, CAPCOG used existing data fields and modified the inputs in order to produce emissions results that would have been produced if the appropriate fields were available. For HP gasoline-powered tractors, CAPCOG adjusted the equipment population for the HP range to increase it to account for the HP tractors. For HP tractors and HP tractors, CAPCOG used the light commercial generator sets (SCC and ) data entry fields in order to model the emissions from LPG and gasolinepowered tractors in those ranges. CAPCOG also used light commercial generators to simulate the HP and HP ranges for LPG tractors. Since generator sets have a higher load factor (68%) than tractors (62% for spark-ignition tractors), CAPCOG adjusted the activity level downward in order to simulate the actual engine activity from these tractor s engines. This produced the following adjusted activity levels: <40 HP: 107 hours/year HP: 133 hours/year HP: 198 hours/year HP: 269 hours/year In order to incorporate the LPG tractors in the HP range into the emissions estimates, CAPCOG increased the equipment populations for the HP range in order to simulate the additional engine capacity from the and HP ranges. This is necessary because TexN does not include any LPG equipment below 25 horsepower. ( ) ( ) ( ) ( ) CAPCOG ran TexN for each scenario using the corresponding updated NATION.GRW file as described in Section CAPCOG then consolidated all of the aggregate output files in order to develop the emissions estimates for each SCC code for each county. Section Combine Emissions Modeling CAPCOG performed default TexN runs for 4-stroke and diesel combines for 2012 and 2018 using the updated NATION.GRW file as described in section CAPCOG then summed the population and emission data for each SCC for each county from the AGGREGATE.OUT file, and then scaled the emissions totals to match the updated equipment populations described in section

104 Section Two-Wheel Tractors, Balers, Mowers, Sprayers, Tillers, and Swathers Emissions Modeling In order to obtain emissions rates for two-wheel tractors, balers, mowers, sprayers, tillers, and swathers, CAPCOG performed default TexN runs with the updated NATION.GRW file described in section for 2012 and 2018 and divided the emissions totals by the equipment populations. CAPCOG then summed the population and emission data for each SCC for each county from the AGGREGATE.OUT file, and multiplied the emissions totals by the ratio of the updated equipment populations to the default equipment populations. Section Other Agricultural Equipment Emissions Modeling CAPCOG modeled the forage harvester, cotton picker and stripper, and all remaining other equipment emissions using separate TexN runs. For forage harvesters and cotton pickers and strippers, CAPCOG used the diesel other agricultural equipment SCC, set the populations for all HP bins to zero other than the HP bin, and then entered each county s equipment population. CAPCOG updated the average HP for the HP bin to reflect the profiles described in sections and CAPCOG then entered the appropriate activity level and ran TexN using the NATION.GRW files described in section For cotton pickers and strippers, CAPCOG encountered a problem with TexN that it was able to resolve, but which will be described here for the purpose of enabling the reproduction of CAPCOG s emissions estimates. CAPCOG s 2012 and 2018 runs, for some reason, produced aggregate output files that included population, activity, and emissions for Blanco and Llano counties, despite CAPCOG entering zero in those fields. CAPCOG notified ERG of this problem and ERG is looking into what programming issue may have caused this problem. CAPCOG ran the same scenario on two different computers to make sure there wasn t an error specific to the files used by one machine, and in both cases, this problem occurred. In addition, for the 2012 scenario (but not for the 2018 scenario), TexN generated increased population totals that were higher than the totals entered into TexN for the scenario, which generated higher emissions totals. In order to address the first issue, CAPCOG simply set the emissions totals for Blanco and Llano County to zero for both years. In order to correct the 2012 scenario s output, CAPCOG simply scaled the emissions based on the ratio of its actual equipment populations and the populations reported in the AGGREGATE.OUT file. For all remaining other agricultural equipment, CAPCOG ran default 2012 and 2018 TexN scenarios using the updated NATION.GRW file described in section 5.1.3, except that for the HP, HP, and HP bins for the diesel-powered other agricultural equipment, the population was set to zero. This was to reflect that most of the emissions from higher-horsepower equipment under this category were likely already been accounted for in the forage harvester and cotton picker/stripper emissions estimates. While CAPCOG did not attempt to net out the equipment populations for this source category, the much lower emissions rates for diesel-powered equipment should enable better represent the emissions for all of the remaining diesel-powered other agricultural equipment. 104

105 Section Irrigation Set Emissions Modeling Due to the small number of irrigation sets for some fuel types in some counties and TexN s rounding to the nearest tenth in the population input fields, CAPCOG decided to generate emission rates in TexN that reflected the updated data and then apply those rates to the small equipment populations outside of TexN. CAPCOG used the updated activity levels in the activity input fields, and used the updated NATION.GRW file described in section prior to running TexN. Section Emissions Totals for Agricultural Equipment The following table shows the updated CO, NO X, and VOC emissions, by county, for 2012 and Table 103: 2012 and 2018 Emissions by County for Typical Ozone Season Weekday (tons per day) County 2012 CO 2012 NO X 2012 VOC 2018 CO 2018 NO X 2018 VOC Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total These estimates represent a significant increase in emissions relative to the default estimates for 2012 and Region-wide, these estimates represent a 19% increase in NO X emissions for 2012 and a 47% increase for The differences are not uniform, however. The figure below shows the difference in NOX emissions estimates by county. For several counties, the overall emissions estimates wound up quite close to the existing estimates, despite the scope of the updates performed. However,Bastrop, Burnet, Fayette, Hays, Lee, and Llano Counties, changed quite substantially. 105

106 Figure 27: Default and Updated Agricultural Equipment NO X Emissions by County (tons per day) Default 2012 Updated 2018 Default 2018 Updated The table below shows the region-wide emissions estimates by equipment type. Table 104: 2012 and 2018 Emissions by Equipment Type for Typical Ozone Season Weekday (tons per day) Equipment Type 2012 CO 2012 NO X 2012 VOC CO NO X 2018 VOC Two-Wheel Tractors Agricultural Tractors Combines Balers Sprayers Mowers Tillers Swathers Forage Harvesters Cotton Harvesters Other Agricultural Equipment Irrigation Sets Total

107 Section Agricultural Equipment Spatial Allocation Improvements Agricultural equipment was allocated based on CropScape raster data for 2012, obtained from the United States Department of Agriculture (USDA) National Agricultural Statistics Service 102. CropScape provides land cover data for the contiguous United States at 30-meter resolution, derived from multispectral satellite images. CAPCOG assigned each equipment type to various land uses, as described in CAPCOG s 2013 spatial allocation improvement report. 103 Equipment types were matched to the land uses most likely to require that type of equipment for agricultural production. There were a few deviations from the methodology described in the 2013 report: A few counties did not have any grid cells reporting the kind of intensive crop farming that a twowheeled tractor or tiller would typically be used for. CAPCOG used all agricultural land as the spatial allocation factor for these situations. Forage harvesters, which are considered other agricultural equipment, were allocated to hay and pasturelands. Cotton pickers and strippers only were allocated to land used for cotton production. Other agricultural equipment aside from forage harvesters and cotton pickers were allocated to all agricultural lands evenly. The table below shows the land use types that CAPCOG assigned to each equipment type were assigned to. The column headings reflect the last four digits of the equipment types SCC code: 5010: 2-wheel tractors, 5015: Agricultural tractors, 5020: Combines, 5025: Balers, 5030: Mowers, 5035: Sprayers, 5040: Tillers, 5045: Swathers, 5055: Other Agricultural Equipment, and 5060: Irrigation Sets. Table 105: Agricultural Equipment Land Use Assignments for Spatial Allocation Category Corn X X X X X Cotton X X X X* X Rice X X X X X Sorghum X X X X X Soybeans X X X X X gmu.edu/cropscape/, accessed 6/3/ CAPCOG. Spatial Allocation Surrogate Updates for Selected Area and Non-Road Sources in the Austin-Round Rock Metropolitan Statistical Area. August _Development_of_Updated_Spatial_Surrogates_for_Selected_Area_and_Non-Road_Sources_Final.pdf 107

108 Category Sunflower X X X X X Peanuts X X X X X Barley X X X X X Spring Wheat X X X X X Winter Wheat X X X X X Dbl Crop WinWht /Soybeans X X X X X Rye X X X X X Oats X X X X X Millet X X X X X Alfalfa X X X X X X** X Other Crops X X X X X X Misc Vegs & Fruits X X X X X X Onions X X X X X X Peas X X X X X X Other Tree Crops X X X X X X Pasture/Hay X X X X X X** X *Cotton pickers **Forage harvesters CAPCOG developed a technique to allocate county-level emissions estimates for agricultural equipment to photochemical grid cells based on the relative share of acreage for matching land use types. Below is an example of the land use patterns for Travis County in 2012 that were used for this spatial allocation process. 108

109 Figure 28: CropScape Land Cover Data for Travis County, 2012 The information below describes the steps undertaken to obtain and use the data to create the spatial surrogates. First, when downloading the CropScape data, Specify Projection was checked, and Degrees Lat-Long, WGS 84 datum was selected. Geospatial Modeling Environment was used to obtain grid cell-level land cover statistics. Geospatial Modeling Environment (GME) is a free stand-alone toolset available online 104. The process for using GME for this project is outlined later in this section. CropScape data were downloaded for the study area and brought into ArcGIS along with shapefiles representing counties in the CAPCOG area and a TCEQ-provided 4-km grid shapefile. All shapefiles were re-projected to WGS 84 to match the CropScape raster using the Project tool in ArcGIS. The grid was then clipped to each individual county, creating individual grids for each county. Individual CropScape rasters were extracted for each county using the Extract by Mask tool in ArcGIS. These steps were necessary because emissions inventories are aggregated at county level, so ratios for each grid cell

110 must represent the proportion of emissions of the county total; in other words, each county had to be processed individually. Geospatial Modeling Environment was used to obtain counts of CropScape gridcodes for each 4-km grid cell using the Isectpolyrst script. The Isectpolyrst script intersects a raster with a polygon layer, giving counts of each cell value that fall within the features of a polygon layer, automatically appended to the attribute table of the polygon layer, with fields for each value. The script was set up with the following settings: In: Clipped 4-km county grid shapefile Raster: Clipped county CropScape raster Prefix: User must choose a name for the count fields in the attribute table, this can be any word up to six characters Thematic = TRUE (identifies raster data as categorical) Proportion: FALSE Metrics: N/A. Leave blank Allowpartialoverlap: FALSE Medquant: FALSE The figure below shows the output of the script in the shapefile attribute table. Counts are on the far right (newv1, etc.). New is the prefix that is user-defined in the script, followed by a V and the Z-value of the raster being counted (in this case, CropScape gridcode). Figure 29: Example of Attribute Table with Gridcode Counts From Isectpolyrst Script. (Only three columns of gridcodes are shown) The attribute tables for each county were then exported to Excel, and ratios were calculated for each county by using standard Excel formulas (totaling the count of each gridcode present in the county and dividing the count for each cell by the total count for that county). Once land use (gridcode) ratios were calculated, ratios of agricultural equipment emissions sources for each cell were calculated. This was done by adding the gridcode counts for each land use type that corresponded to an equipment SCC code for each cell, totaling the cell counts, and dividing the individual cell values by the total to allocate the portion the county total for that SCC code for each cell. Below is an example of the gridded NOX emissions from agricultural tractors for

111 Figure 30: PAVE Plot of Agricultural Tractor NO X Emissions, 2012 Section 5.2 Construction and Mining Equipment CAPCOG updated the construction and mining equipment emissions estimates for two diesel construction equipment (DCE) subsectors modeled in TexN the Mine and Quarry Operations (MQ) subsector and the Heavy Highway Construction (HH) subsector and updated the spatial allocation surrogates for both of those subsectors and the Landfill Operations (LF) subsector. For the MQ subsector, CAPCOG developed emissions estimates for each mine and quarry in the program area for both 2012 and For the HH subsector, CAPCOG developed emissions estimates for each highway construction project underway in the summer of 2012 for Bastrop, Caldwell, Hays, Travis, and Williamson Counties. Finally, for the LF subsector, CAPCOG allocated the default emissions estimates for Travis and Williamson Counties (the only two counties in the region with landfill operations emissions) to the landfills in each county based on reported disposal rates for 2012 and projected capacity and disposal rates for All other construction and mining equipment emissions estimates and spatial allocations were based on default TexN model runs and spatial allocation surrogates (urban area). The table below summarizes the updates CAPCOG made for each DCE subsector. 111

112 Table 106: Summary of DCE Subsector Emissions Estimates and Spatial Allocation Updates DCE Subsector Heavy- Highway Construction Landfill Operations Mining and Quarry Operations All Other DCE DCE Sector # , 11-22, 24 MSA Counties 2012 Other Counties 2012 Spatial Allocation 2012 MSA Counties 2018 Other Counties 2018 Spatial Allocation 2018 The table and figure below summarize the region-wide default and updated emissions estimates of ozone season day (OSD) emissions for all construction and mining equipment. Table 107: Default and Updated Region-Wide Construction and Mining Equipment Emissions (tons per ozone season day) Inventory CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Default Updated Difference

113 Ozone Season Weekday Emissions (tons) and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties Figure 31: Default and Updated Region-Wide Construction and Mining Equipment Emissions Default 2012 Updated 2012 Default 2018 Updated CO NOX VOC Section Mining and Quarry Operations DCE Subsector CAPCOG developed emissions estimates for each mine and quarry in the region using research conducted by the Alamo Area Council of Governments (AACOG) for CAPCOG in AACOG reviewed 2011 aerial photography of each of the active mine and quarry sites listed in the Mine Safety and Health Administration s (MSHA s) Mine Data Retrieval System (MDRS) extended mine search. 106 The data obtained from these sites were used in order to obtain equipment populations for each quarry and mine where aerial photography was available. CAPCOG obtained updated MSHA data on pit labor-hours for each mine and quarry for 2012 and 2011 and increased the 2011 equipment populations by the ratio of 2012 pit-labor-hours to 2011 pit-labor-hours where data was available for For any locations where 2011 photography was not available, CAPCOG applied an equipment ratio developed by AACOG based on the number of annual pit labor-hours as reported by MSHA. AACOG used the total number of equipment pieces observed by the annual pit-hours for all locations in both the AACOG and CAPCOG areas where aerial photography was available. CAPCOG multiplied these ratios by the annual pit labor-hours for each site in operation in 2012 where aerial photography was not available. 105 Alamo Area Council of Governments. Mining and Quarrying Equipment Emission Inventory for CAPCOG counties and Milam County. January 16, U.S. Department of Labor, Mine Safety and Health Administration. Extended Mine Search Page. Mine Data Retrieval System. Note: all mines were identified by querying the system by identifying active mines by county except for the coal mine in Lee County, which did not appear in searches by county. This mine was identified by querying the database by commodity (coal). 113

114 Table 108: Mine and Quarry Diesel Construction Equipment Population Ratios (Pieces per 10,000 Annual Labor-Hours) Equipment Type SCC AACOG CAPCOG + Milam Combined Rollers Scrapers Bore/Drill Rigs Excavators Cranes Graders Rock Trucks Water Trucks Rock Crushers N/A N/A N/A Loaders Backhoes Bulldozers Tractors Other Construction Equipment The following table shows the total number of mines and quarries in each county, and the total number of pit labor-hours for all mines and quarries in For a full listing of the mines and quarries and in the region that were active in 2012, please see Appendix D. Table 109: Active Mines and Quarries and Pit Labor-Hours, 2012 County Active Locations Pit Labor-Hours Bastrop 4 62,895 Blanco Burnet ,610 Caldwell 0 0 Fayette 9 132,015 Hays 4 115,595 Lee 1 591,590 Llano 4 21,952 Milam 2 29,883 Travis 6 176,077 Williamson 24 1,074,471 Total 72 2,480,940 In order to project growth to 2018, CAPCOG used the 2012 data and prior production data CAPCOG had obtained for 2006, 2008, and Since this time period straddles the economic recession that started at the end of 2008, the use of a 2006, 2008, or 2011 baseline might not provide the most accurate representation of the production potential for a given county. In every county except for Blanco County, production in 2012 was higher than in 2011, indicating that there is likely to be continued growth corresponding with overall economic growth in the region since the recession. In some about half of the counties, though, the 2012 production totals are less than the totals in 2006 or 2008, indicating that the 114

115 mining and quarrying activity has not fully recovered from the recession. CAPCOG also did not feel comfortable using the change from 2011 to 2012 to project to 2018, since mines and quarries have physical constraints to them and in some counties, there was very high growth from 2011 to % in Williamson County, and 55% in Llano County that would not be expected to continue at the same rate. Thus, CAPCOG used that the highest value of labor-hours from the four years for which CAPCOG obtained data provided the best estimate of future activity for a given county in this sector. The table below shows these growth factor data for each county. Table 110: Mine and Quarry Growth Factors by County County Max Year Growth Factor Bastrop 65,028 67,397 57,013 62, Blanco 0 0 1, Burnet 348, , , , Caldwell n/a Fayette 177, , , , Hays 109, , , , Lee 292, , , , Llano 18,311 19,848 14,122 21, Milam 208, ,968 28,454 29, Travis 370, , , , Williamson 1,324,572 1,124, ,532 1,074, TOTAL 2,622,349 2,767,474 2,194,877 2,480,940 n/a For Rock Crushers, AACOG obtained permit data from the TCEQ for equipment populations and average horsepower ratings. CAPCOG did not apply any growth factors for this category since they require a permit. The following tables show the estimated equipment population totals for each county. 115

116 Table 111: 2012 Mine and Quarry Equipment Populations by County Equipment Bastrop Blanco Burnet Fayette Hays Lee Llano Milam Travis Williamson Total Roller Scraper Bore/Drill Rigs Excavators Cranes Graders Rock trucks Water Truck Rock Proc. Eq Loaders Backhoes Dozers Tractors Other Const. Eq

117 Table 112: 2018 Mine and Quarry Equipment Populations by County County Bastrop Blanco Burnet Fayette Hays Lee Llano Milam Travis Williamson Total Roller Scraper Bore/Drill Rigs Excavators Cranes Graders Rock trucks Water Truck Rock Proc. Eq Loaders Backhoes Dozers Tractors Other Const. Eq

118 Annual Hours of Use 2012 and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties The average horsepower ratings and activity for quarries for most equipment types are based on AACOG s 2011 survey (described in their 2013 report for CAPCOG). The horsepower and activity data for Tractors and Other Construction Equipment are TexN defaults. The horsepower ratings and activity levels for the one coal mine in the region come from ERG s 2009 Update of Diesel Construction Equipment Emission Estimates for the State of Texas Phase I and II. 107 The figures below show the TexN defaults and updated estimates for annual activity and average horsepower rating for each equipment type included in CAPCOG s 2012 and 2018 inventories. For equipment types marked with an asterisk in the figures below, the equipment was observed by AACOG at mines and quarries in the region, but were not included in the mine and quarry operations profile, meaning that the model would assume that these equipment types are not present at mines or quarries. As a result, there is no default annual activity for these equipment types for the mine and quarry operations DCE subsector. There was also no default for average annual hours of use for rock crushers and tractors, despite their inclusion in the mine and quarry operations DCE subsector profile. AACOG segmented the Off-Highway Trucks equipment type ( ) into two types of trucks rock trucks and water trucks each of which have different activity and horsepower profiles. Figure 32: Default and Updated Annual Hours of Use for Mine and Quarry Operations DCE TexN Default Updated-Coal Mine Updated-Quarries 107 Eastern Research Group, Inc. Update of Diesel Construction Equipment Emission Estimates for the State of Texas Phase I and II. Table 3-1. Coal Mining Equipment Survey Results. Prepared for the Texas Commission on Environmental Quality. July 31, DCE_EI_Update.pdf. Last accessed August 16,

119 Average Horsepower Rating 2012 and 2018 Emissions Inventory Updates for the CAPCOG Region and Milam Counties Figure 33: Default and Updated Average Horsepower Ratings for Mine and Quarry Operations DCE TexN Default Updated-Coal Mine Updated-Quarries CAPCOG used TexN to generate emissions rates using the activity and horsepower ratings described above. Since TexN will run separate NONROAD model runs for each DCE subsector that a selected equipment type is found in, it was necessary to set the population and activity levels for all subsectors other than the ones of interest to zero. As mentioned above, several of the equipment types in this inventory were not included in the mine and quarry operations DCE subsector in TexN. For these, CAPCOG selected the first subsector listed, and set the remaining sectors populations and activity levels to zero: Rollers (used City and County Road Construction DCE), Bore/Drill Rigs (used Boring and Drilling Equipment DCE), Cranes (used Cranes DCE), and Other Construction Equipment (used Other Construction Equipment DCE). In order to obtain emissions rates, CAPCOG set the equipment populations for each equipment type for each county to 100 and updated the activity levels and average horsepower ratings to those in the figures above. The coal profiles were applied to the equipment at the Three Oaks coal mine in Lee County, while the quarry profiles were applied to the equipment at all of the other mine/quarry sites in the region. Since AACOG has different profiles for rock trucks and water trucks under SCC Off-Road Trucks, CAPCOG had to perform a separate model run to account for the water trucks. Prior to running TexN, CAPCOG updated the SEASON.DAT file to distribute construction and mining equipment activity uniformly across all months. Using AGGREGATE.OUT files from the TexN runs, CAPCOG calculated the 2012 and 2018 ozone season weekday emissions rates for CO, NO X, and VOC for each equipment type (tons per piece of equipment per ozone season weekday). CAPCOG then multiplied 119

120 these rates by the equipment populations at each mine and quarry in order to obtain emissions estimates for each location. Since the MSHA provides quarterly totals for labor-hours, it was possible to calculate seasonal adjustment factors for each site in order to more accurately reflect summertime usage in CAPCOG calculated the summertime allocations for each mine and quarry using the following equation: ( ( )) ( ( )) Where: Q2 = pit hours for April June, Q3 = pit hours for July September, and Annual = annual pit hours. CAPCOG then calculated the site-specific 2012 summertime activity adjustment factors for each site by dividing the summertime allocation by 25%, and applied these adjustment factors the emissions estimates for each site. In analyzing the patterns of activity at large for the region, there didn t appear to be any systematic seasonal variation in activity. As a result, CAPCOG decided not to adjust the 2018 emissions estimates, which were based on uniform allocation of activity to summer months. The following tables show the default and updated emissions summary for each county, along with the difference in emissions. Location-specific data is available in the Appendix D and in accompanying spreadsheets. Table 113: Default Mine and Quarry Operations DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total

121 Table 114: Updated Mine and Quarry Operations DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Table 115: Difference Between Updated and Default Mine and Quarry Operations DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total As seen above, these updates result in about 10% less NO X emissions in 2012 and about 10% more NO X emissions in 2018 than default TexN runs would produce for the region, although county-specific changes are much more significant. CAPCOG provided AACOG with each mine and quarry s ozone season weekday emissions for 2012 and 2018, along with the latitude and longitude so that AACOG could grid the emissions at the actual mine and quarry locations. Under default spatial allocation factors, these emissions would have been simply allocated to the urban area of each county. In some cases, the emissions from a single mine or quarry are equivalent to what a small point source would emit in a typical ozone season day. Therefore, accurately modeling the location of emissions from this subsector is important. Since mines and quarries typically are located away from urbanized areas, this change in spatial allocation is a significant improvement in the representation of these emissions in the modeling. The figure below shows a PAVE plot of the gridded NO X emissions for the mine and quarry operations for

122 Figure 34: PAVE Plot for Mine and Quarry Operations DCE Subsector NO X Emissions, 2012 Section Heavy-Highway Construction DCE Subsector CAPCOG updated the emissions for the heavy highway construction DCE subsector for Bastrop, Caldwell, Hays, Travis, and Williamson Counties for 2012 using updated, project-specific emissions estimates produced by ERG. 108 In 2012, CAPCOG contracted with ERG to develop equipment use profiles for several types of heavy highway construction projects based on daily work reports (DWR) and highway construction project details obtained from the Texas Department of Transportation (TxDOT), such as lane-miles of construction, description of work performed, and project cost. 109 The table below summarizes these equipment profiles, which are presented as number of piece-days of equipment used for every $1 million in expenses or lane-mile constructed. 108 Eastern Research Group, Inc. Heavy Highway Emission Inventory Update Prepared for the Capital Area Council of Governments. May 31, Eastern Research Group, Inc. Heavy-Highway Emission Inventory Update. Prepared for the Capital Area Council of Governments. April 9,

123 Table 116: Heavy Highway Construction Equipment Profiles (Piece-Days per $M or Lane-Mile of Construction) Equipment Type Bridgework ($M) Turn Lanes (lanemiles) Repair/Resurface (lane-miles) New / Rebuild (lane-mile) Misc. ($M) Crawler Tractor/Dozer Surfacing Equipment 300/600 HP Surfacing Equipment 600/750 HP Excavator Grader Paver Paving Equipment Roller , Scraper Surfacing Equipment Tractor/Loader/Backhoe Wheeled Loader In a follow-up project, CAPCOG contracted with ERG to apply these profiles to the highway construction projects underway in Bastrop, Caldwell, Hays, Travis, and Williamson Counties during the summer of ERG performed 25 project-specific TexN runs for the top highway construction projects underway during this period, and then performed an additional TexN model run with all of the other projects aggregated together for each county, then allocated those emissions to each project based on horsepower-hours of activity. 110 This produced project-specific emissions estimates that could be gridded at the actual construction sites. CAPCOG did not update the 2012 heavy highway construction emissions estimates for Blanco, Burnet, Fayette, Lee, Llano, and Milam Counties, or for any county in The table below summarizes the default and updated emissions estimates for this DCE subsector for each county for 2012 and Eastern Research Group, Inc. Heavy-Highway Emission Inventory Update Prepared for the Capital Area Council of Governments. May 31,

124 Table 117: Default Heavy Highway Construction DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) County CO 2012 NO X 2012 VOC 2012 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Table 118: Updated Heavy Highway Construction DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) County CO 2012 NO X 2012 VOC 2012 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Table 119: Difference Heavy Highway Construction DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) County CO 2012 NO X 2012 VOC 2012 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total

125 CAPCOG used the actual construction project locations in order to spatially allocate the 2012 emissions ERG provided the latitude and longitude coordinates for the start and endpoints for the top 25 projects, and CAPCOG looked up the latitude and longitude data for the remaining projects. CAPCOG then took all of the projects and assigned them to each grid cell in the photochemical modeling grid using the cell s x and y coordinates. Each project was allocated to photochemical grid cells based on the fraction of road segment under construction within a given grid cell. For a handful of projects where the project location was vague (i.e., various county roads ) CAPCOG used a GIS staffer to create spatial surrogates to allocate activity to all county roads equally. The figure below shows a PAVE plot of the NO X emissions from heavy highway construction for By far the largest of the project underway during this time period was the US-290 construction underway in eastern Travis County. Figure 35: PAVE Plot for Heavy Highway Construction DCE Subsector NO X Emissions, 2012 Section Landfill Operations DCE Subsector While CAPCOG did not modify the emissions estimates for LF subsector, it did develop data to spatially allocate the emissions to the actual landfills in operation in 2012 and that are projected to be in operation in 2018 in order to better represent these emissions for modeling purposes. The table below shows the default county-wide emissions from this subsector for 2012 and CAPCOG obtained these estimates by selecting all of the applicable SCC codes for this subsector, setting the equipment 125

126 populations and activity levels for subsectors other than landfill operations to zero, and running the model. 111 The following table shows the emissions totals for each county for 2012 and 2018 from these runs. Table 120: Landfill Equipment DCE Summer Weekday Emissions, 2012 and 2018 (tons per day) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total In order to allocate the county-wide emissions for Travis and Williamson Counties (the only two counties with active landfills in the region), CAPCOG used TCEQ s FY 2011 data on each landfills disposal rates and remaining capacities. 112 For 2012, the emissions were allocated to each of the five landfills in operation in the region based on the number of tons disposed in Since one landfill (BFI Sunset Farms Landfill in Travis County) is only projected to have capacity through 2015, its disposal rate (tons per year) was allocated to the other three landfills in Travis County for CAPCOG looked up the grid cells containing these landfills in order to obtain the X and Y coordinates. The following table shows these data. Table 121: Landfill Equipment DCE Spatial Allocation for 2012 and 2018 Landfill County 2011 Tons 2012 County Fraction Closure Date 2018 County Fraction Austin Community Recycling and Disposal Travis 255, Facility BFI Sunset Farms Landfill Travis 612, Please see section for details on problems that were encountered when CAPCOG attempted to run the TexN model for specific DCE subsectors in order to obtain the default emissions estimates and how CAPCOG resolved these problems. 112 Texas Commission on Environmental Quality. Municipal Solid Waste in Texas: A Year in Review; FY 2011 Data Summary and Analysis. AS-187/12. Revised November Last accessed August 21, Page

127 Landfill County 2011 Tons 2012 County Fraction Closure Date 2018 County Fraction IESI Travis County C&D Landfill Travis 102, Texas Disposal Systems Landfill Travis 575, Williamson County Recycling and Disposal Facility Williamson 249, The following figure shows a PAVE plot of the 2012 NO X emissions from the landfill operations DCE subsector. Figure 36: PAVE Plot for Landfill Operations DCE Subsector NO X Emissions, 2012 Section Remaining Construction Equipment Subsectors Since the default method for modeling construction and mining equipment emissions involves allocating the total construction and mining equipment emissions for each county to the county s urbanized area, it was necessary to remove the default emissions for the mine and quarry operations, heavy highway construction, and landfill operations subsectors from each county s total construction equipment emissions. In order to do this, CAPCOG ran default two TexN runs each (one for 2012, one for 2018) for the three subsectors listed above. CAPCOG selected the applicable equipment and set all of the equipment populations and activity levels other than the relevant subsectors to zero. Then, CAPCOG subtracted the default subsector emissions totals from the corresponding aggregated totals for all 127

128 subsectors for each SCC code in each county in order to obtain the county-wide emissions estimates for the remaining 21 DCE subsectors. Table 122: Remaining Construction Subsector Summer Weekday Emissions, 2012 and 2018 (tons per day) County CO 2012 NO X 2012 VOC 2012 CO 2018 NO X 2018 VOC 2018 Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total One minor source of error in this estimation process is that once the TexN graphical user interface (GUI) is opened and equipment population inputs are provided, the program rounds the value to the nearest tenth of a piece of equipment. As a result, the DCE-specific default runs provide a less precise estimate of the equipment populations than simply running a default TexN file without manipulating the equipment populations would provide. This should not produce any systematic bias or error in the estimates, especially since the equipment populations are assigned to individual horsepower bins rather than distributed across a range of bins, reducing the possibility of systematic under-estimation of emissions when there are fractions of equipment populations in a population input field. Without programming changes to TexN or iterative updates to the TexN database, there doesn t appear any readily available solution to this problem. The next section describes other problems that CAPCOG encountered when modeling default emissions for the three specific DCE subsectors that CAPCOG provided modeling updates for in this project and how CAPCOG resolved these issues. Section Notes on Problems with TexN for Construction and Mining Equipment In the process of producing the default emissions estimates for the mine and quarry operations, heavy highway construction, landfill operations, and remaining DCE subsectors, CAPCOG encountered problems with TexN that it was eventually able to resolve, as well be described in this section. The documentation of these problems is necessary in order to enable reproduction of the default emissions estimates described above. CAPCOG set up default runs of TexN version 1.6 for each of the three DCE subsectors CAPCOG as follows: 128

129 1. In the Period and Controls Tab: o Set the Analysis Year and Max Tech Year to 2012 or 2018, as appropriate, o Set the Met Year to Typical Year, o Set the Period to OSD, o Set the Summation Type to Typical Day and Weekday, o Checked all of the post-processing adjustment check boxes, and o Selected all rules enabled; 2. In the Regions tab, selected Bastrop, Blanco, Burnet, Caldwell, Fayette, Hays, Lee, Llano, Milam, Travis, and Williamson Counties; 3. In the Sources tab, selected the equipment types associated with the relevant DCE subsector; 4. In the Population tab: o For each county, clicked on select SCC and clicked select all DCE, and o Set the equipment populations for all subsectors other than the one being analyzed to zero; 5. In the Activity tab, o For each county, clicked on select SCC and clicked select all DCE, and o Set the activity levels for all subsectors other than the one being analyzed to zero; 6. Leave the Fuels and Climates and Retrofit Specifications tabs unchanged; 7. Under Scenario, selected Save and Close; and 8. Re-Opening the scenario and selecting run models. On the first tab, CAPCOG checked the preserve output files option in order to audit the results after the model runs finished. For the default aggregated construction and mining equipment totals, CAPCOG extracted the emissions estimates from a default TexN run for all equipment types for 2012 and Once CAPCOG produced the AGGREGATE.OUT files for each run, it set up a spreadsheet to subtract out the emissions from mine and quarry operations, heavy highway construction, and landfill operations, from the aggregated emissions estimates for all construction and mining equipment emissions in each county. Upon completing these calculations, it became apparent that in the remaining equipment, there were negative values for emissions estimates for some SCCs in some counties. This prompted CAPCOG to review the NR.BMX files for individual DCE subsector runs for each county in order to determine whether they produced results that reflected the inputs CAPCOG had put into the TexN GUI. CAPCOG created a macro that summed the emissions estimates for each county from each subsector in order to determine which, if any, errant emissions estimates had been produced. Figure 37: Non-Zero NR.BMX files Produced from Default 2012 Heavy Highway Construction DCE (#9) Subsector Run DCE * 129

130 DCE Figure 38: Non-Zero NR.BMX files Produced from Default 2012 Mine and Quarry Operations DCE (#23) Subsector Run DCE * 24 Figure 39: Non-Zero NR.BMX files Produced from Default 2012 Landfill Operations DCE (#10) Subsector Run DCE

131 DCE * CAPCOG reviewed all of the TexN inputs to confirm that it had not inadvertently missed setting the populations or activity levels for the other subsectors to zero in some of the counties. Once it confirmed that the inputs would not have caused the modeling error, CAPCOG then ran a MYSQL script to force an update to the activity levels for all subsectors other than the one being modeled to zero in the database. CAPCOG also reviewed the tables in the database to determine whether there was some misclassified equipment that could account for the problem. These investigations did not yield any results. Upon examination of some of the files for each of these county/dce runs that incorrectly produced nonzero emissions output files, CAPCOG was able to determine that TexN appeared to be producing activity inputs for certain equipment types in certain DCE subsectors in certain counties, even though the TexN inputs or those database tables had no activity listed for them. For example, when CAPCOG examined the files for the landfill operations DCE subsector, it found that there were activity inputs listed for offroad trucks ( ) in DCE subsector 17 (scrap and recycling operations) for Blanco, Hays, Milam, Travis, and Williamson Counties, despite off-road trucks not being included in the that subsector s equipment profile. Eventually, CAPCOG was able to come up with a way to manually resolve the problem outside of the model and the TexN model. CAPCOG decided to simply copy and paste a zero-value NR.BMX file into all of the county/dce subsector run folders other than the subsector being analyzed, and then re-ran the TexN postprocessor to obtain corrected emissions estimates in the new AGGREGATE.OUT file that overwrote the existing file. 131

132 CAPCOG contacted Rick Baker and Diane Pruesse at ERG to notify them of the problems encountered and to verify that CAPCOG s work-around to the problem would produce accurate emissions estimates. 113 They indicated that the solution identified by CAPCOG should produce corrected AGGREGATE.OUT files since the post-processor relies on the NR.BMX files. CAPCOG ed them detailed information on the TexN runs and outputs to see if they could re-produce the same problems and to see if they had any fixes for the future. However, the solution identified by CAPCOG provided corrected emissions estimates for those DCE subsectors, CAPCOG used it to generate the default emissions estimates for the mine and quarry operations, heavy highway construction, and landfill equipment emissions estimates and then subtracted these corrected emissions estimates from the aggregated construction and mining equipment emissions estimates for each county in order to obtain the remaining construction equipment emissions estimates. As a final check, CAPCOG wanted to determine whether the aggregated emissions estimates from a default run of TexN for all equipment types would suffer from the same problems. CAPCOG reviewed the ACT.DAT, TX.POP, and NR.BMX files for DCE subsector 17 (scrap and recycling), and found that activity inputs for off-road trucks (687 hours per year) were included in the activity files for each county, and population inputs for off-road trucks were included for Blanco, Hays, Milam, Travis, and Williamson Counties. CAPCOG again reviewed the MYSQL database tables to determine whether off-road trucks were included in that DCE subsector, and could not find anything indicating this. Below is a screenshot of the SCCDCEXWALK table showing that DCE subsector 17 does not appear as an applicable subsector for SCC code Figure 40: Screenshot of MySQL TexN Database Table "SCCDCEXWALK" 113 Andrew Hoekzema. Phone communication with Rick Baker and Dianne Pruesse of Eastern Research Group, Inc. July 5,

133 This investigation indicates that this error recurred in the default TexN run, meaning that the default emissions estimates suffer from this error, and therefore the remaining construction equipment emissions estimates will also suffer from this error. CAPCOG did not have the time to further pursue or rectify this issue prior to submission of emissions inventory data to AACOG, and since it appears to be a fundamental programming problem in TexN, is well beyond the scope of this emissions inventory project to correct anyhow. Therefore, CAPCOG did not perform a full audit of all of the other files to determine to what extent this or other, similar errors may have also occurred for other DCE subsectors in the counties CAPCOG modeled. The remaining construction and mining emissions estimates appear to suffer from the error identified above, but the error likely only produces a relatively small increase in emissions relative to what they should be. For the purposes of this report, CAPCOG is simply documenting the work-around it was able to come up with in order to correct the emissions estimates for the three subsectors it made improvements to for this project. As stated above, the work that would be required to fix the emissions estimates for all of the remaining construction and mining equipment is beyond the scope of this project. Section 5.3 Industrial Equipment CAPCOG updated at least one emissions inventory input parameter for each type of non-road industrial equipment in order to obtain updated 2012 and 2018 emissions estimates for all 11 counties. The table below summarizes CAPCOG s updates. For the activity updates, CAPCOG s updates were limited only to specific fuel types where it had obtained new data. Table 123: Industrial Equipment Updates SCC SCC Description Population 22xx Aerial Lifts 22xx Forklifts 22xx Sweepers/Scrubbers 22xx Other General Industrial Equipment 22xx Other Material Handling Equipment 22xx Refrigeration Units 22xx Terminal Tractors Fuel Distribution HP Distribution Activity The updates to the equipment population estimates and engine parameters for aerial lifts, forklifts, and sweepers/scrubbers are based on local, up-to-date equipment sales data obtained by ENVIRON for CAPCOG in early The updates to the activity estimates for a number of diesel-powered industrial equipment types are based on the same ENVIRON project. The updates to the equipment populations of terminal tractors, engine parameters for refrigeration units, and activity data for LPG 114 Lindhjem, Chris; ENVIRON International Corporation. Memorandum: Industrial Equipment (Forklift) Category. January 18,

134 forklifts were based on a 2005 report ERG prepared for TCEQ. 115 These updates are described in detail in this section. The figure below shows the default and updated ozone season weekday emissions of CO, NOX, and VOC for all industrial equipment across the 11-county region. Figure 41: Default and Updated Industrial Equipment Ozone Season Weekday Emissions for 2012 and 2018 (tons per day) CO NOX VOC Default 2012 Updated 2012 Default 2018 Updated Section Population and Engine Profile Updates for Aerial Lifts, Forklifts, and Sweepers/Scrubbers CAPCOG updated the 2012 aerial lift, forklift, and sweepers/scrubbers equipment populations using the study ENVIRON conducted for CAPCOG in ENVIRON obtained in-use population estimates for 2012 for the Austin-Round Rock MSA (Bastrop, Caldwell, Hays, Travis, and Williamson Counties) based on sales figures. The figure below shows the default and updated equipment populations for all fuel types for the MSA. 115 Wells, Sam; Eastern Research Group, Inc. Data Collection, Sampling, and Emissions Inventory Preparation Plan for Selected Commercial and Industrial Equipment: Phase II. August 31, Last Accessed July 25,

135 Figure 42: Default and Updated Aerial Lift, Forklift, and Sweeper Equipment Populations for the Austin-Round Rock MSA, ,500 3,000 2,952 2,500 2,303 2,000 1,500 TexN Default Updated 1, Aerial Lifts Forklifts Sweepers CAPCOG ratios of updated equipment populations to the default equipment populations for the MSA to the other six counties in the program area (Blanco, Burnet, Fayette, Lee, Llano, and Milam Counties): A 15% increase in the aerial lift equipment population, A 12% decrease in the forklift equipment population, and A 24% increase in the sweeper/scrubber equipment population. CAPCOG allocated the updated 2012 aerial lift equipment populations for each county to each engine type using data provided by ENVIRON in its 2013 report. The figure below shows the default and updated allocations. 135

136 Figure 43: Default and Updated Engine Type Distribution for Aerial Lifts, % 67% Diesel CNG 22% 11% Default 19% 33% Updated LPG 4-Stroke Gas 2-Stroke Gas For forklifts and sweepers/scrubbers, the 2012 populations were allocated to each engine type based on the default distribution in TexN for The engine type distributions for 2012 for each equipment type are listed in the table below. Table 124: Aerial Lifts, Forklifts, and Sweepers/Scrubbers Engine Type Distributions Equipment Type Basis 2-Stroke Gasoline 4-Stroke Gasoline LPG CNG Diesel Aerial Lifts Updated 0.00% 32.89% 18.68% 0.00% 38.8% Forklifts Default 0.00% 1.09% 81.85% 8.44% 8.63% Sweepers/Scrubbers Default 1.78% 11.39% 13.48% 0.04% 73.31% The table below shows a comparison of the default horsepower distributions to the data provided by ENVIRON for each engine type. Although ENVIRON provided updated horsepower data for 4-stroke and LPG aerial lifts, several of the horsepower categories for these fuel types are not included in TexN. As a result, CAPCOG allocated the equipment populations for these fuel types to each horsepower bin based on the default horsepower distributions for 4-stroke and LPG aerial lifts. CAPCOG did update the horsepower distribution for diesel aerial lifts based on ENVIRON s data. Table 125: Default and ENVIRON Aerial Lift HP Bin Distributions by Engine Type HP Range 4-Stroke: Default 4-Stroke: ENVIRON LPG: Default 136 LPG: ENVIRON Diesel: Default Diesel: ENVIRON % 6.4% 0.0% 0.0% 0.0% 0.0% % 0.0% 0.0% 0.0% 0.2% 0.0% % 0.0% 0.0% 0.0% 1.3% 3.8% % 0.0% 0.0% 0.0% 18.3% 2.2% % 23.2% 32.0% 5.6% 22.8% 10.3%

137 HP Range 4-Stroke: Default 4-Stroke: ENVIRON LPG: Default LPG: ENVIRON Diesel: Default Diesel: ENVIRON % 0.0% 0.0% 0.0% 5.4% 19.0% % 54.4% 66.6% 56.3% 38.3% 50.0% % 16.0% 0.0% 38.0% 13.6% 14.7% % 0.0% 1.5% 0.0% 0.2% 0.0% Finally, based on ERG s 2005 report, CAPCOG updated the median life of LPG forklifts at full load to 3,000 hours. Section Transportation Refrigeration Unit Population and Engine Profile Updates CAPCOG updated the Transportation Refrigeration Unit (TRU) equipment populations for each county in the program area using updated fuel type and horsepower distributions based on ERG s 2005 report for TCEQ. ERG s study indicated that based on its research, all TRUs are diesel. In order to reflect this change, CAPCOG obtained default equipment populations for 2012 and 2018 from the TexN model and summed the populations across fuel types in order to obtain the total equipment population for each county. CAPCOG then allocated all of the total equipment population for each county to the diesel fuel type and then allocated to populations to each horsepower bin using ERG s updated horsepower distribution. Table 126: ERG-Updated 2001 Statewide Transportation Refrigeration Unit Population by HP Range HP Min HP Max Default Updated Updated HP Distribution , , , % , ,569 1, % Total Total 15,591 15,645 n/a Section Terminal Tractors Population Updates The NONROAD User s Guide defines terminal tractors as Single driver (typically no passenger seat) offroad trucks used primarily for moving highway trailers around paved areas such as at container ports and other intermodal facilities (rarely used for moving aircraft around airports); also called yard spotters, hostlers, and hustlers. ERG s 2005 report based the estimates for this equipment type exclusively on equipment reported at intermodal facilities. Based on the absence of such facilities within the program area, CAPCOG assumed that the terminal tractor population for each county was zero. 137

138 Section Growth Factors For the equipment types that CAPCOG provided updated 2012 populations, CAPCOG projected the populations to 2018 based on the default ratio of 2018 equipment populations to 2012 equipment populations. Section Annual Activity Updates CAPCOG made a number of updates to the annual activity estimates for industrial equipment. CAPCOG updated the LPG forklift annual activity based on ERG s 2005 report, and all diesel-powered industrial equipment other than TRUs based on ENVIRON s 2013 report for CAPCOG. The table below shows the activity estimates for each SCC and fuel type. The updated values are indicated with an asterisk. Table 127: Industrial Equipment Annual Activity by SCC and Fuel Type SCC 2-Stroke 4-Stroke LPG CNG Diesel 22xx * 22xx ,800 1,800 1,270* 1, * 22xx * 22xx ,744* 22xx ,955* 22xx ,341 22xx ,558* Section Updated Temporal Allocation Profile for LPG Forklifts The only update CAPCOG made to the temporal profiles for industrial equipment in the TexN model was to change the LPG forklift weekday/weekend allocation to reflect the survey data collected by ERG in its 2005 report. Table 128: Industrial Equipment Weekdays and Weekend Day Activity Allocations in Season.dat file SCC SCC Description Weekday Weekend Day Updated 22xx Aerial Lifts No Forklifts (4-Stroke) No Forklifts (LPG) Yes Forklifts (CNG) No Forklifts (Diesel) No 22xx Sweepers/Scrubbers No 22xx Other General Industrial Equipment No 22xx Other Material Handling Equipment No 22xx Refrigeration No 22xx Terminal Tractors No CAPCOG further adjusted the weekend emissions to reflect higher usage on Saturdays than Sundays (other than for refrigeration units) using the default ratio of Saturday to Sunday activity in TCEQ s modeling files (7:4). 138

139 Section 5.4 Residential Lawn and Garden Equipment CAPCOG updated the 2012 and 2018 residential lawn and garden equipment emissions estimates using updated data on the number of single-family houses in each county from the U.S. Census Bureau s American Community Survey (ACS) and equipment ratios described in a previous CAPCOG report, as well as updated annual activity, seasonal distribution of activity, and weekend/weekday distribution of activity. Figure 44: Default and Updated Residential Lawn and Garden Equipment Ozone Season Weekday Emissions for 2012 and 2018 (tons per day) CO NOX VOC Default 2012 Updated 2012 Default 2018 Updated Section Equipment Population Updates CAPCOG updated the equipment populations for 2012 and 2018 based on the most recent estimate of single-family detached houses in each county in the region from the ACS, population growth projections from the University of Texas at San Antonio (UTSA), and equipment ratios (number of pieces per household) described in CAPCOG s report on 2006 estimates of emissions for residential lawn and garden equipment. 116 The equipment ratios are based on survey work performed by AACOG in its region and Sonoma Technology in the Baltimore area. Section Equipment Ratios The equipment ratios CAPCOG used to develop equipment populations for each county are divided into two categories: (1) Travis County, and (2) Other Counties. AACOG s study split the equipment ratios into 116 Capital Area Council of Governments. Residential Lawn and Garden Equipment Emissions Inventory for the Austin-Round Rock Metropolitan Statistical Area and Burnet County, July CAPCOG_Residential_Lawn_and_Garden_Equipment_Inventory_Final.pdf. Last accessed July 26,

140 those found in the core urban county in its region (Bexar County) and those in the other counties. CAPCOG s review of Sonoma s responses by county indicated a similar pattern for the Baltimore area. Since AACOG s survey did not specifically ask what kind of rear riding mower respondents owned, CAPCOG used the ratio between rear riding mowers, front riding mowers, and garden tractors in Sonoma s study to disaggregate the riding mowers equipment type. For more on CAPCOG s analysis of these studies, please see CAPCOG s 2013 report on the updated 2006 emissions estimates for this category. The table below shows the composite assumptions of equipment ratios CAPCOG used for the region. Table 129: Residential Lawn and Garden Equipment Ratios (pieces/single-family detached unit household) Equipment Type SCC Travis County Other Counties Lawn Mower 22xx Tiller 22xx Chainsaw 22xx Edger, Trimmer, Etc. 22xx Leaf Blowers 22xx Rear Riding Mowers 22xx Front Riding Mowers 22xx Garden Tractors 22xx Note that CAPCOG did not update the Other equipment category and added front-riding mowers to the residential category of mowers due to their presence among residential equipment reported in Sonoma s Baltimore study. Section Estimated Single Family Detached Housing Units in 2012 and 2018 In order to estimate equipment populations using the equipment ratios described above, CAPCOG applied those ratios to the estimated number of single-family detached housing units in each county for each analysis year. In order to estimate this, CAPCOG used the most recent estimate of single-family detached units (SFDUs) reported for each county from the Census Bureau s ACS, Table DP04. The table below shows the number of households, SFDUs, % of households that are SFDUs, and the Census s estimated population in each county for the most recent year for which data was available, and the ACS sampling time frame used. Table 130: Base Year Housing and Population Data County Year Households SFDUs SFDU % Population ACS Time Frame Bastrop ,427 19,171 65% 74, year Blanco ,403 4,041 75% 10, year Burnet ,931 14,587 70% 42, year Caldwell ,806 8,142 59% 38, year Fayette ,935 10,803 78% 24, year Hays ,833 39,188 63% 163, year Lee ,461 4,858 65% 16, year Llano ,115 9,872 70% 19, year 140

141 County Year Households SFDUs SFDU % Population ACS Time Frame Milam ,304 8,667 77% 24, year Travis , ,342 52% 1,062, year Williamson , ,797 73% 442, year CAPCOG used the base year population and the Census s 2012 population estimates 117 to project the number of SFDUs in 2012, and used the ratio of 2018 to 2012 data from UTSA s projections (using the migration 1.0 scenario) to estimate the number of SFDUs in There are some differences in the estimation methodology for these two data sources, which result in differences in the 2012 population estimates. In particular, it appears that while the Census 2012 population accounts for Bastrop County s reduced population due to the 2011 wildfire, UTSA s does not. Implicitly, this method assumes that the housing mix in the base year remains fixes through Table 131: Estimated Growth of SFDUs by County from Base to 2012 and 2018 County Base Year Census Base Pop. Census 2012 Pop to Base Ratio UTSA 2012 Pop. UTSA 2018 Pop to 2012 Factor Bastrop ,347 74, ,904 95, Blanco ,315 10, ,000 12, Burnet ,802 43, ,635 50, Caldwell ,084 38, ,129 46, Fayette ,520 24, ,265 27, Hays , , , , Lee ,565 16, ,129 18, Llano ,175 19, ,695 20, Milam ,703 24, ,124 26, Travis ,061,203 1,095, ,071,103 1,221, Williamson , , , , U.S. Census Bureau. Table PEPSR6H: Annual Estimates of the Resident Population by Sex, Race, and Hispanic Origin for the US, States, and Counties: April 1, 2010 to July 1, 2012, 2012 Population Estimates. 118 University of Texas at San Antonio, Texas State Data Center. Sex and Race/Ethnicity Total Population by Migration Scenario for in 1-year increments

142 Applying these growth ratios to the base level of SFDUs produces the following estimates of SFDUs for each county in 2012 and Table 132: Single Family Detached Units by County in 2012 and 2018 Year Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson ,278 4,174 14,807 8,281 10,880 40,443 4,869 9,826 8, , , ,320 4,749 16,888 9,696 11,861 54,326 5,298 10,436 8, , ,613 CAPCOG used the default engine type and horsepower distributions in for 2012 and 2018 to allocate the updated county-level equipment populations to the appropriate TexN equipment inputs. The table below summarizes these allocations, which were identical in both years. Table 133: Engine Type and Horsepower Allocation of Residential Lawn and Garden Equipment, 2012 and 2018 Equipment Type 2-Stroke 0-1 HP 2-Stroke 1-3 HP 2-Stroke 3-6 HP 4-Stroke 1-3 HP 4-Stroke 3-6 HP 4-Stroke 6-11 HP 4-Stroke HP 4-Stroke HP 4-Stroke HP Res. Chainsaw Res. Edger, Trimmer, Etc Com. Front Riding Mowers Res. Garden Tractors Res. Lawn Mower Res. Leaf Blowers Res. Rear Riding Mowers Res. Snowblowers Res. Tillers

143 Section Annual Activity Updates CAPCOG updated the annual activity estimates for all residential equipment types (including the front riding mowers included in this inventory) except for other residential lawn and garden equipment in order to incorporate the estimates it developed in its report on 2006 Residential Lawn and Garden Emissions. These estimates reflect survey data completed by AACOG and provide separate estimates for Travis County and other counties in the region. The figure below shows the default and updated activity estimates that were used for 2012 and 2018 emissions estimates. Figure 45: Default and Updated Lawn and Garden Equipment Annual Activity (hours per year) Tillers Chainsaws Edger, Trimmer, Etc. Leaf Blowers Snow Blowers Lawn Mowers Rear Riding Mower Garden Tractors Default Updated-Travis County Updated-Other Counties Section Monthly Allocation of Residential Lawn and Garden Activity CAPCOG also updated the monthly allocation of activity to reflect the local growing season (March to November) and variations in precipitation throughout the growing season, as described in the report on 2006 residential lawn and garden equipment. CAPCOG updated the SEASON.DAT file to reflect these changes. CAPCOG did not update the monthly allocation of chainsaws, which are evenly distributed to each month of the year. This change reduces the annual activity allocated to summer months by 8.6%. 143

144 Figure 46: Default and Updated Monthly Allocation of Residential Lawn and Garden Equipment Activity (except Chainsaws) 18.00% 16.00% 14.00% 12.00% 10.00% 8.00% 6.00% Default Updated 4.00% 2.00% 0.00% Section Weekday/Weekend Allocation of Residential Lawn and Garden Activity CAPCOG also updated the SEASON.DAT file to reflect the results of Sonoma s Baltimore study 119, which showed significantly more weekend activity (65%) for residential lawn and garden usage than the NONROAD/TexN default of 55.5%. This resulted in each weekday s allocation of weekly activity being reduced by 37% and each weekend day s allocation being increased by 46%. 119 Reid, Stephen B., Erin K. Pollard, and Dana Coe Sullivan, Sonoma Technology. Improvements to Lawn and Garden Equipment Emission Estimates for Baltimore, Maryland Emissions Inventory Conference. 144

145 Figure 47: Default and Updated Weekday/Weekend Allocation of Residential Lawn and Garden Activity 35% 32.50% 30% 25% 22.20% 20% 15% 10% 5% 11.10% 7.00% Default Updated 0% Weekday Weekend Day Section 5.5 Aviation: Austin-Bergstrom International Airport CAPCOG updated the emissions for all aviation, auxiliary power unit (APU), and ground support equipment (GSE) source categories at the Austin-Bergstrom International Airport (ABIA) based on a study conducted by ERG for TCEQ in ERG used the Federal Aviation Administration s (FAA s) Emissions Dispersion Modeling System (EDMS) to model the emissions from the activity data submitted by ABIA for 2008, and the inventory was forecast to 2011, 2014, 2017, and 2020 based on the FAA s Terminal Area Forecast (TAF). Since the activity data for ABIA were directly reported, this inventory provided the most representative estimates for ABIA that was available at the time of this project for these sources. The table below shows the reported annual number of landings and take-offs (LTOs) at ABIA by activity type for 2008, 2011, 2014, 2017, and Table 134: Landings and Take-Offs at Austin-Bergstrom International Airport by Activity Type Activity Type Air Taxi, Piston 13,554 8,157 8,707 9,294 9,922 Air Taxi, Turbine 4,610 2,796 2,984 3,184 3,398 APU 51,178 46,319 48,746 51,300 53,989 Commercial 64,275 58,900 61,943 65,144 68,509 General Aviation, Piston 8,642 6,286 6,536 6,795 7,066 General Aviation, Turbine 73,745 46,925 49,683 52,613 55,727 GSE - CNG 1,417 1,113 1,173 1,237 1, Eastern Research Group, Inc. Development of Statewide Annual Emissions Inventory and Activity Data for Airports. July 15, ftp://amdaftp.tceq.texas.gov/pub/offroad_ei/airports/tex/erg_statewide_airport_ei_report_july_2011.pdf Last accessed July 26,

146 Activity Type GSE - Diesel 86,733 68,096 71,791 75,692 79,813 GSE - Gasoline 18,242 14,322 15,099 15,920 16,786 GSE - LPG 1,792 1,407 1,483 1,564 1,649 Military CAPCOG interpolated the emissions estimates for 2011 and 2014 in order to estimate the 2012 annual emissions and interpolated the emissions estimates for 2017 and 2020 in order to estimate the 2018 annual emissions. CAPCOG took the difference between 2014 and 2011 activity and emissions and between 2020 and 2017 activity and emissions and divided them by three in order to yield the annual change in emissions between these years, and then added those estimated changes to the 2011 and 2017 estimates, respectively, in order to produce the 2012 and 2018 annual estimates. In order to produce ozone season day activity and emissions, CAPCOG divided annual emissions by 365 days. The results are displayed below. Table 135: 2012 Daily ABIA Activity (LTO) and Emissions (tons per day) Description SCC Code LTO CO NO X VOC Air Taxi, Piston Air Taxi, Turbine APU Commercial General Aviation, Piston General Aviation, Turbine GSE - CNG GSE - Diesel GSE - Gasoline GSE - LPG Military TOTAL n/a n/a Table 136: 2018 Daily ABIA Activity (LTO) and Emissions (tons per day) Description SCC Code LTO CO NO X VOC Air Taxi, Piston Air Taxi, Turbine APU Commercial General Aviation, Piston General Aviation, Turbine GSE - CNG GSE - Diesel GSE - Gasoline GSE - LPG

147 Description SCC Code LTO CO NO X VOC Military TOTAL n/a n/a Based on a discussion with a staffer at ABIA, it is likely that there are additional temporal profile details that could be used to improve upon this profile. However, given the time constraints of this project, and given the fact that TCEQ uses uniform distribution across all days for temporal profiles for aviation, CAPCOG believes that this improvement provides a sufficient representation of the relative change in activity and emissions relative to the current modeling files. 147

148 Appendix A: Electronic Files Submitted 1. Task 3.1-Area Source Commercial Fuel Combustion.xlsx 2. Task 3.1-Area Source Industrial Fuel Combustion.xlsx 3. Task 3.1-Area Source Oil and Gas Production Equipment.xlsx 4. Task 3.1-Gridding 2006 ENVIRON.pdf 5. Task 3.1-Gridding 2006 Files.zip 6. Task 3.1-Nonroad Agricultural Equipment 2WT Balers Mowers Tillers Swathers.xlsx 7. Task 3.1-Nonroad Agricultural Equipment Ag Growth Factors.xlsx 8. Task 3.1-Nonroad Agricultural Equipment Cattle Production Nation.grw file.grw 9. Task 3.1-Nonroad Agricultural Equipment Combine Nation.GRW file.grw 10. Task 3.1-Nonroad Agricultural Equipment Combines.xlsx 11. Task 3.1-Nonroad Agricultural Equipment Irrigation Nation.GRW 12. Task 3.1-Nonroad Agricultural Equipment Irrigation.xlsx 13. Task 3.1-Nonroad Agricultural Equipment Other-Cotton Harvesters Nation.GRW file.grw 14. Task 3.1-Nonroad Agricultural Equipment Other-Cotton Harvesters Profile.xlsx 15. Task 3.1-Nonroad Agricultural Equipment Other-Cotton Harvesters.xlsx 16. Task 3.1-Nonroad Agricultural Equipment Other-Forage Harvesters Profile.xlsx 17. Task 3.1-Nonroad Agricultural Equipment Other-Forage Harvesters.xlsx 18. Task 3.1-Nonroad Agricultural Equipment Spatial Allocation.xlsx 19. Task 3.1-Nonroad Agricultural Equipment Spatial Allocation.xlsx 20. Task 3.1-Nonroad Agricultural Equipment Tractor HP Nation.GRW file.grw 21. Task 3.1-Nonroad Agricultural Equipment Tractor 100+ HP Nation.GRW file.grw 22. Task 3.1-Nonroad Agricultural Equipment Tractor LT 40 HP Nation.GRW file.grw 23. Task 3.1-Nonroad Agricultural Equipment Tractors 2012.xlsx 24. Task 3.1-Nonroad Agricultural Equipment Tractors 2018.xlsx 25. Task 3.1-Nonroad Agricultural Equipment Weekday Emissions.xlsx 26. Task 3.1-Nonroad Aviation ABIA Emissions.xlsx 27. Task 3.1-Nonroad Construction 2012 Minus MQ HH and LF.xlsx 28. Task 3.1-Nonroad Construction 2018 Minus MQ HH and LF.xlsx 29. Task 3.1-Nonroad Construction Heavy Highway Report.pdf 30. Task 3.1-Nonroad Construction Heavy Highway 2006 Export Summary.xlsx 31. Task 3.1-Nonroad Construction Heavy Highway 2006 NIF.zip 32. Task 3.1-Nonroad Construction Heavy Highway 2008 Export Summary.xlsx 33. Task 3.1-Nonroad Construction Heavy Highway 2008 NIF.zip 34. Task 3.1-Nonroad Construction Heavy Highway 2012 Report.pdf 35. Task 3.1-Nonroad Construction Heavy Highway 2012.xlsx 36. Task 3.1-Nonroad Construction Mine and Quarry 2012 and 2018.xlsx 37. Task 3.1-Nonroad Construction Mine and Quarry AACOG 2006 and 2008 TexN Outputs.zip 38. Task 3.1-Nonroad Construction Mine and Quarry AACOG pdf 39. Task 3.1-Nonroad Construction Mine and Quarry AACOG 2006 EPS3 Inputs.zip 40. Task 3.1-Nonroad Construction Mine and Quarry AACOG 2006 EPS3 Outputs.zip 41. Task 3.1-Nonroad Construction Mine and Quarry AACOG 2006 OSD.xlsx 42. Task 3.1-Nonroad Construction Mine and Quarry AACOG 2008 NIF.zip 43. Task 3.1-Nonroad Construction Mine and Quarry AACOG 2008 OSD.xlsx 148

149 44. Task 3.1-Nonroad Construction Mine and Quarry AACOG Emissions 2006.xlsx 45. Task 3.1-Nonroad Construction Mine and Quarry AACOG Emissions 2008.xlsx 46. Task 3.1-Nonroad Industrial Equipment Emissions Total.xlsx 47. Task 3.1-Nonroad Industrial Equipment.xlsx 48. Task 3.1-Nonroad Industrical Equipment Forklift Sales by SIC-ENVIRON.xlsx 49. Task 3.1-Nonroad Industrial Equipment NONROAD Defaults-ENVIRON.xlsx 50. Task 3.1-Nonroad Industrial Equipment Population Data-ENVIRON.xlsx 51. Task 3.1-Nonroad Industrial Equipment Report-ENVIRON.pdf 52. Task 3.1-Nonroad Residential Lawn and Garden Equipment 2012.xlsx 53. Task 3.1-Nonroad Residential Lawn and Garden Equipment 2018.xlsx 54. Task 3.1-Nonroad Residential Lawn and Garden Equipment Activity.xlsx 55. Task 3.1-On-Road Idling.xlsx 56. Task 3.1-On-Road MSA Link-Based 2012 Weekday.xlsx 57. Task 3.1-On-Road MSA Link-Based 2018 Weekday.xlsx 58. Task 3.1-Point Sources.xlsx 149

150 Appendix B: Oil and Gas Active Well Counts and Production Data The following data on regular producing oil and gas wells, gas production, casinghead production, oil production, and condensate production for each county from was used to estimate growth factors for pumpjack, compressor engines, and heater/treater emissions for each county. 150

151 Table 137: Oil and Gas Well Counts and Production Data by County, 2006 Data Point Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Oil Wells (Feb.) , , ,902 Oil Wells (Sept.) , ,965 Gas Wells (Feb.) Gas Wells (Sept.) TOTAL Wells (Feb.) , , ,314 TOTAL Wells (Sept.) , ,379 Gas Well Gas (MCF) 186, ,739 11,814, ,011, , ,110 15,052,687 Casinghead (MCF) 159, ,705 10,081, ,065, , ,169,415 Total Gas (MCF) 346, ,444 21,896, ,077, , ,110 38,222,102 Oil (BBL) 90, ,720 1,377, ,376, ,704 1,773 8,415 4,227,749 Condensate (BBL) 15, , , ,683 Table 138: OIl and Gas Well Counts and Production Data by County, 2007 Data Point Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Oil Wells (Feb.) , ,244 Oil Wells (Sept.) , ,472 Gas Wells (Feb.) Gas Wells (Sept.) TOTAL Wells (Feb.) , ,662 TOTAL Wells (Sept.) , ,894 Gas Well Gas (MCF) 172, ,950 10,736, ,450, , ,386 13,391,199 Casinghead (MCF) 137, ,988 8,980, ,715, , ,656,909 Total Gas (MCF) 310, ,938 19,717, ,166, , ,386 33,048,108 Oil (BBL) 181, ,089 1,262, ,136, ,406 1,881 8,186 3,890,321 Condensate (BBL) 16, , , ,

152 Table 139: Oil and Gas Well Counts and Production Data by County, 2008 Data Point Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Oil Wells (Feb.) , ,465 Oil Wells (Sept.) , , ,891 Gas Wells (Feb.) Gas Wells (Sept.) TOTAL Wells (Feb.) , , ,879 TOTAL Wells (Sept.) , , ,307 Gas Well Gas (MCF) 158, ,206 10,226, ,236, , ,646,904 Casinghead (MCF) 209, ,202 7,815, ,720, , ,518,438 Total Gas (MCF) 367, ,408 18,042, ,957, , ,165,342 Oil (BBL) 555, ,159 1,086, ,124, ,094 1,909 14,985 4,125,664 Condensate (BBL) 18, , , ,478 Table 140: OIl and Gas Well Counts and Production Data by County, 2009 Data Point Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Oil Wells (Feb.) , ,741 Oil Wells (Sept.) , ,907 Gas Wells (Feb.) Gas Wells (Sept.) TOTAL Wells (Feb.) ,158 TOTAL Wells (Sept.) ,321 Gas Well Gas (MCF) 139, ,614 9,693, ,913, , ,830,928 Casinghead (MCF) 262, ,031 7,014, ,159, , ,185,080 Total Gas (MCF) 401, ,645 16,708, ,073, , ,016,008 Oil (BBL) 303, , , ,114, ,158 1,814 10,216 3,705,822 Condensate (BBL) 17, , , ,

153 Table 141: Oil and Gas Well Counts and Production Data by County, 2010 Data Point Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Oil Wells (Feb.) , ,906 Oil Wells (Sept.) , , ,998 Gas Wells (Feb.) Gas Wells (Sept.) TOTAL Wells (Feb.) , ,322 TOTAL Wells (Sept.) , , ,408 Gas Well Gas (MCF) 130, ,158 9,300, ,726, , ,248,363 Casinghead (MCF) 154, ,868 8,425, ,235, , ,498,316 Total Gas (MCF) 284, ,026 17,725, ,962, , ,746,679 Oil (BBL) 147, ,153,013 1,105, ,061, ,742 2,061 9,648 3,879,088 Condensate (BBL) 4, , , ,173 Table 142: Oil and Gas Well Counts and Production Data by County, 2011 Data Point Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Oil Wells (Feb.) , ,798 Oil Wells (Sept.) , , ,991 Gas Wells (Feb.) Gas Wells (Sept.) TOTAL Wells (Feb.) ,212 TOTAL Wells (Sept.) ,405 Gas Well Gas (MCF) 120, ,008 8,390, ,531, , ,118,886 Casinghead (MCF) 113, ,840 7,164, ,865, , ,805,407 Total Gas (MCF) 233, ,848 15,555, ,396, , ,924,293 Oil (BBL) 111, ,526,292 1,212, ,027, ,594 4,177 9,442 4,346,450 Condensate (BBL) 3, , , ,

154 Table 143: Oil and Gas Well Counts and Production Data by County, 2012 Data Point Bastrop Blanco Burnet Caldwell Fayette Hays Lee Llano Milam Travis Williamson Total Oil Wells (Feb.) , , ,873 Oil Wells (Sept.) , , ,097 Gas Wells (Feb.) Gas Wells (Sept.) TOTAL Wells (Feb.) , , ,278 TOTAL Wells (Sept.) , , ,497 Gas Well Gas (MCF) 116, ,093 8,908, ,337, , ,473,053 Casinghead (MCF) 99, ,124 6,875, ,823, , ,511,390 Total Gas (MCF) 216, ,217 15,784, ,160, , ,984,443 Oil (BBL) 92, ,755,607 1,246, ,144, ,982 3,403 7,662 4,913,755 Condensate (BBL) 3, , , , ,

155 Appendix C: CAMPO Contract for EPS3 File Preparation for On-Road Link-Based Emissions Inventories 155

156 156