April 2015 Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2013: Revision to Refinery Emissions Estimate Overview of Method in 2014 Inventory (estimates for 1990-2012) Inventory estimates of refinery emissions included in the petroleum systems source category have been based on emission factors that were derived from studies conducted by EPA and the American Petroleum Institute (API) between 1992 and 1996. The emission factors were multiplied by activity data obtained each year from the U.S. Energy Information Agency's Gross Inputs to Refineries, and Oil and Gas Journal's World-wide Refining Report. The previous Inventory method for the petroleum systems source category included methane (CH4) emissions released to the atmosphere as fugitive emissions, vented emissions, emissions from operational upsets, and emissions from fuel combustion by certain equipment at refineries. Fugitive and vented carbon dioxide (C02) emissions from petroleum refineries were also included. Combustion C02 emissions from fuel use (e.g., in process heaters) are accounted for in the fossil fuels combustion source category, and not taken into account in the petroleum systems source category. Overview of Revision Implemented in 2015 Inventory Petroleum refineries have been reporting to the Petroleum Refineries source category (subpart Y) of the Greenhouse Gas Reporting Program (GHGRP) since 2010, and now four years of data from 2010 to 2013 are available. In the 2015 Inventory, EPA has revised the refineries portion of the petroleum systems source category calculations to use the data available from GHGRP subpart Y. Revised Approach for 2015 Inventory Stakeholder comments to the 2014 Inventory supported use of the GHGRP data on petroleum refineries to update the GHG Inventory estimates. Stakeholders noted significant differences between GHGRP results and the GHG Inventory estimates for this source. Taking into account this feedback, EPA noted in the "Planned Improvements" section of the 2014 Inventory that analysis of GHGRP data for petroleum refineries is a priority for the 2015 Inventory. EPA then solicited feedback from expert reviewers1 and public reviewers on a proposed revised methodology for the 2015 Inventory that is based on GHGRP data. Reviewers generally supported the revised methodology. Petroleum refineries report under subpart Y of the GHGRP. All petroleum refineries, regardless of size, are required to report to GHGRP (i.e., there is no threshold for this source).2 Emissions from petroleum refineries are estimated using methods prescribed in the GHGRP including direct measurement and engineering calculations. Tables 1 and 2 below compare 2012 process-level refinery emissions calculated in the 2014 Inventory to the corresponding 2012 GHGRP emissions (reflecting submissions from facilities as of August 18, 2014) for CH4 and C02, respectively. 1 Every year, the Inventory undergoes an expert review period during which a first draft of the document is sent to a select list of technical experts outside of EPA. The purpose of the Expert Review is to encourage feedback on the methodological and data sources used in the current Inventory, especially for sources which have experienced any changes since the previous Inventory. This memorandum references feedback from 2015 Inventory expert reviewers. 2 While this source category has no threshold and therefore all facilities in this source category were required to report emissions to the GHGRP from 2010-2012, due to GHGRP provisions that allow facilities to discontinue reporting under specified circumstances, it is possible that not all refineries will continue to report after 2012. The GHGRP allows for a facility to discontinue reporting if a facility's annual reports submitted under the rule show that emissions were less than 25,000 metric tons of CO2 equivalent per year for 5 consecutive years, or less than 15,000 metric tons of CO2 equivalent per year for 3 consecutive years. 1 ------- April 2015 Table 1. 2012 Refinery CH4 Emissions from GHG Inventory and GHGRP Subpart Y GHG Inventory Source 2012 Inventory Emissions (MT CH4) GHGRP Emission Source 2012 GHGRP Emissions (MT CH4) Vented and Fugitive Emissions Fixed Roof Tanks 278 All storage Tanks 1,592 Floating Roof Tanks 9 System Blowdowns 14,378 Uncontrolled Blowdowns 5,136 Asphalt Blowing 624 Asphalt Blowing 292 Fuel Gas System 1,210 Equipment Leaks3 2,704 Process Vents 5,505 CEM Vents 48 Loading Operations 19 Wastewater Treating3 198 Cooling Towers3 248 Sub-Total 16,945 15,743 Combustion Emissions Atmospheric Distillation 387 Vacuum Distillation 180 Thermal Operations 96 Catalytic Cracking/Reforming/ Fluid Coking 1,473 Catalytic Cracking 182 Catalytic Reforming 157 Catalytic Hydrocracking 78 Hydrorefining 34 Hydrotreating 472 Alkylation/Polymerization 96 Aromatics/lsomeration 13 Lube Oil Processing 0 Engines 118 Flares 20 Flares 13,229 Delayed Coking 1,201 Coke Calcining 24 Sub-Total 1,834 15,928 Total 18,780 31,671 a - "Equipment leaks" fugitive emissions in Subpart Y include emissions from valves, pumps, compressors, pressure relief valves, connectors, flanges, and open-ended lines—and are most analogous to the "fuel gas system" source in the current Inventory.3 The current Inventory fugitive emission factors for wastewater treating and cooling towers include other non-component-based losses (e.g., evaporative).4 EPA's used the existing methodology to fully account for emissions from these sources, as shown in Table 3 below. As shown in Table 1, several CH4 emission sources are grouped differently between the two data sets. The 2014 Inventory estimates vented emissions from fixed roof tanks separately from fugitive emissions from floating roof tanks, 3 Subpart Y cites Protocol for Equipment Leak Emissions Estimates (EPA-453/R-95-017, NTIS PB96-175401); see section 2.2.2 of the document. 4 The basis for the current Inventory methodology regarding cooling towers and wastewater treatment (Methane Emissions from the U.S. Petroleum Industry EPA-600/R-99-010, Table 5-11) cites AP-42 Chapter 5.1 for the cooling tower emission factor and Model for Evaluation of Refinery and Synfuels VOC Emission Data Vol. I, EPA-600/7-85-022a for the wastewater treatment emission factor. Per AP-42 Section 5.1.3.5: "Atmospheric emissions from the cooling tower consist of fugitive VOCs and gases stripped from the cooling water as the air and water come into contact." Per Model for Evaluation of Refinery and Synfuels VOC Emission Data: "Primary sources of VOC emissions from wastewater treatment systems are evaporative emissions..." 2 ------- April 2015 and estimates combustion CH4 emissions from process heaters and boilers by individual process unit. The 2014 Inventory also estimates wastewater and cooling tower CH4 emissions, which are not reporting elements under GHGRP subpart Y. The combustion emissions reported under GHGRP subpart Y are limited to combustion reactions that occur within specific process equipment. The fuel combustion emissions from conventional stationary combustion units at refineries including heaters, boilers, engines and turbines are not reported under subpart Y, but are reported under subpart C (and are not included in Table 1). Note, combined vented and fugitive CH4 emissions in Table 1 are similar between the two data sets; and combustion CH4 emissions from flares in particular result in higher combustion emissions in the GHGRP Subpart Y data set compared to the current GHG Inventory. As shown in Table 2, the 2014 Inventory methodology for petroleum systems accounted for asphalt blowing as the only source of C02 emissions from refineries. Table 2 matches the current GHG Inventory sources and therefore presents only the asphalt blowing C02 emissions source. Table 2. 2012 Refinery Carbon Dioxide Emissions from GHG Inventory and GHGRP National Emission Inventory Source 2012 Inventory Emissions (MT C02) Subpart Y Emission Source 2012 Subpart Y Emissions (MT CO2) Asphalt Blowing 13,937 Asphalt Blowing 116,850 In addition to asphalt blowing, the GHGRP subpart Y accounts for six other sources of C02from refineries: CEM data (i.e., process-specific stack measurements), coke calcining, catalytic regeneration, process vents, sulfur recovery, and flaring. EPA sought comment from 2015 Inventory expert and public reviewers on implementing an approach consistent with IPCC guidelines across the 2015 Inventory—adding flaring C02 emissions and non-combustion C02 emissions (i.e., process vents) to the GHG Inventory; reviewers generally supported implementing such a revision. In the 2015 Inventory, the C02 emission factors for flaring and process vents shown in Table 4 below have been applied for all years of the time series. As with CH4, the GHGRP accounts for C02 emissions from conventional stationary fuel combustion units (heaters, boilers, engines, and turbines) under subpart C. Including C02 emissions from fuel combustion reported under GHGRP subpart C in the refineries portion of the petroleum systems source category of the Inventory would result in double-counting with the fossil fuels combustion source category in the Inventory. EPA used the 2010 through 2013 GHGRP subpart Y emissions for the sources identified on the right side of Tables 1 and 2 to replace current petroleum systems source-level estimates for the 2015 Inventory. Inventory reviewers generally agreed with this approach. For the two sources currently included in the Inventory but not reported by the GHGRP, cooling towers and wastewater treating, the previous Inventory method was used to estimate emissions from these sources. Inventory reviewers generally agreed with this approach, but raised concern as to whether counting emissions from these two sources over the entire time series is an accurate approach. 3 ------- April 2015 Revised Approach for Calculating 1990-2009 Estimates Since GHGRP data only cover recent years of the Inventory time series, EPA developed an extrapolation approach to develop consistent emissions estimates back to 1990. EPA used throughput-based emission factors developed from recent GHGRP data in conjunction with publicly available throughput data from Department of Energy/Energy Information Administration (e.g., refinery feed data) to effectively scale GHGRP emissions to reflect activity in earlier years. Reviewers commented that in addition to considering refinery throughput dynamics over time, changes in refinery complexity over time ideally would be taken into account when scaling emissions for earlier years; however, a methodology to do so was not proposed. Table 3. Comparison of Previous and Revised Refinery CH4 Emission Factors Previous Inventory CH4 Emission Factor Draft Inventory Emission Factor (scf CHVmbbl refinery feed) Fixed RoofTanks 20.6 scf/mbbl heavy crude input All storage Tanks 10.71 Floating RoofTanks 587 scf/tank System Blowdowns 137 scf/mbbl total crude input Uncontrolled Blowdowns 51.03 Asphalt Blowing 2555 scf/mbbl asphalt production Asphalt Blowing 2.78 Fuel Gas System 439,000 scf/refinery Equipment Leaks 26.05 Process Vents 43.14 CEM Vents 0.33 Loading Operations 0.24 Wastewater Treating 1.88 scf/mbbl total crude input Wastewater Treating 1.88 Cooling Towers 2.39 scf/mbbl total crude input Cooling Towers 2.39 Atmospheric Distillation 3.61 scf/mbbl total crude input Vacuum Distillation 3.61 scf/mbbl unit throughput Thermal Operations 6.01 scf/mbbl unit throughput Catalytic Cracking/ Reforming/ Fluid Coking 13.97 Catalytic Cracking 5.17 scf/mbbl unit throughput Catalytic Reforming 7.22 scf/mbbl unit throughput Catalytic Hydrocracking 7.22 scf/mbbl unit throughput Hydrorefining 2.17 scf/mbbl unit throughput Hydrotreating 6.50 scf/mbbl unit throughput Alkylation/Polymerization 12.6 scf/mbbl unit throughput Aromatics/lsomeration 1.80 scf/mbbl unit throughput Lube Oil Processing 0.0 scf/mbbl unit throughput Engines 0.006 scf/hp-hr Flares 0.189 scf/mbbl total crude input Flares 123.98 Delayed Coking 16.63 Coke Calcining 0.27 Table 4. Comparison of Previous and Revised Refinery C02 Emission Factors Previous Inventory CO2 Emission Factor Draft Inventory Emission Factor (scf COz/mbbl refinery feed) Asphalt Blowing 20,736 scfC02/mbbl asphalt production Asphalt Blowing 367 - - Process Vents 19 - - Flaring 839 4 ------- April 2015 Results The impact of revising the refinery emission factors to use those listed on the right side of Tables 3 and 4 is presented in Table 5 below. The emissions of CH4 and C02 are compared separately. The impact of the new CH4 factors is an increase of CH4 emissions for years across the time series that ranges from 40 to 80 percent. The impact of the new C02 factors is an increase in C02 emissions for years across the time series by two orders of magnitude. Table 5. Impact of Revised Emission Factors-2014 and 2015 Inventory Estimates for Petroleum Refineries 1990 1995 2000 2005 2010 2013 2014 Inventory (mtch4) 17,622 18,055 19,277 19,342 18,586 N/A 2015 Inventory (MTCH4) 27,493 28,650 30,976 31,206 26,807 33,964 2014 Inventory (MTCCh) 17,981 18,686 21,080 20,478 15,265 N/A 2015 Inventory (MTCCh) 4,070,108 4,241,363 4,585,706 4,619,709 3,835,619 5,540,353 5 ------- |