EPA Response to Comments on U.S. Environmental Protection Agency Proposed Rule, "Control of Air Pollution from Motor Vehicles: Tier 3 Motor Vehicle Emission and Fuel Standards," RIN 2060-AQ86

   1) Preamble, Table I-1  -  Please explain the rationale for choosing the NMOG+NOx levels for each year?  Are these the limits and years that CA finalized in their LEV III program?   What cost/benefit analysis did EPA conduct to compare the various compliance options such as:

         a. setting the standards every 3 or 5 years apart 
         b. more gradually reducing the emissions over time as opposed to requiring a 46% reduction in the first year of the program.
            
      EPA Response:  The first year of the program has an immediate, single step down to a fleet-average curve that matches the California LEV III curve, which is important in order to quickly achieve nationwide harmonization.  Most current (Tier 2) vehicles are already performing or capable of performing well below the Tier 2 standards, and manufacturers agree that this first step is preferable to a separate federal phase-in that would not match the LEV III phase-in.   We believe that our proposed early credit program will help facilitate this initial step in stringency.
      The value of the annual progression in subsequent years of the program is that it would provide each manufacturer with a predictable regulatory path many years into the future and, by strategic averaging within their unique product lines, would provide maximum flexibility in how they introduce their chosen technologies.  We did not specifically analyze a phase-in schedule with more abrupt step functions, but such a program would almost certainly tend to favor some manufacturers and not others, depending on fleet mix, vehicle design cycles, and relative technological challenges at different points in the phase-in. 
   2) PM standards  -  Will EPA explain the rationale for making PM standards apply on a vehicle-by-vehicle basis?  Did EPA conduct an analysis to compare the costs/benefits of using a fleet average PM standard versus an individual vehicle standard? 
      EPA Response:  Current vehicles are generally significantly over-performing relative to the Tier 2 PM standards, which were also per-vehicle standards.  Since the intent is to prevent new technologies from "backsliding" on PM emissions, fleet averaging around the standard would counterproductively allow higher-than-current emissions from some vehicles.  We did not analyze a fleet-averaging approach to the PM standards.    
      
         a. Is this also what California used?

      EPA Response:  Yes, California's LEV III program established per-vehicle maximum PM standards.  

         b. Did EPA apply the same individual vehicle standards in Tier 2?
      EPA Response:  Yes, Tier 2 also had per- vehicle maximum PM standards.
         c. What is the increment cost/benefit of obtaining additional PM reductions from vehicles going from Tier 2 to Tier 3?  How many tons of PM emissions reductions would be achieved on an annual basis through the vehicle PM standards compared to the fuel standards?
            
            EPA Response:  As discussed in preamble Section III.B.4, the reductions in direct PM emissions are projected to result solely from the proposed vehicle tailpipe standards.  By 2030, we project a reduction of about 7,500 tons annually.  The fuel sulfur standards would reduce NMOG and NOx emissions, which are precursors to ambient PM, but they would not reduce directly emitted PM.  The modeled changes in ambient PM and the resulting monetized benefits reflect the combined effect of changes in emissions of all these PM precursors (e.g., direct PM, NMOG, and NOx).  The costs associated with the proposed Tier 3 PM standard are related to PM measurement, as described in RIA Chapter 2.3 (see page 2-25).

   3) What's the rational for making early reduction credits expire after 5 years?  This policy would seem to limit the amount of early reductions that manufacturers would try to make and therefore reduce the overall air quality and health benefits of the program?  Did EPA run any sensitivity analyses on different lengths of time for the expiration of early reduction credits such as 10 years, 15 years, or no limit?
      EPA Response:  We are proposing a significant, but limited, lifetime for early Tier 3 credits (5 years).  This proposal is consistent with EPA's past practice of establishing reasonable restrictions on credit life.  EPA believes that a credit life of five years represents an appropriate balance between two factors.  Early credits serve as an incentive to encourage early compliance and to promote an orderly upgrading of emissions control technology across a manufacturer's fleet.  However, there is a potential that large numbers of credits accumulated early in the program could delay progress on necessary technological improvements later in the program.  Therefore, as in past programs, EPA is again proposing reasonable restrictions on the time period over which early credits can be used, in order to limit the effect of early credits on the long-term emissions reductions anticipated to result from the new standards.
      The 5-year lifetime for early credits aligns with the approach used in LEV II and LEV III, as well as the basic credit carry-forward provisions in the EPA/NHTSA GHG/fuel economy rules for heavy-duty and light-duty vehicles (with some exceptions).  (Note that the reference to Tier 2 in the preamble was erroneous and we will correct that.) Consequently, in discussions the vehicle manufacturers discouraged longer credit life, as it would delay achieving their desire for a harmonized 50-state program.
   4) What are the CO and formaldehyde benefits for HDVs?  What are the air quality benefits in terms of bringing into attainment any areas with the CO NAAQS?         
      EPA Response:  The rulemaking analysis does not project HDV benefits separately from the overall Tier 3 program benefits.  As for light-duty vehicles, our reasons for including CO and HCHO standards for HDVs at the proposed levels is to prevent new engine and emission control designs that would result in increases in CO and HCHO emissions compared to current vehicles.   These standards are set at levels that should be easily met through compliance with the NMOG+NOx standards.  There are currently no CO nonattainment areas.  As of September 27, 2010, all CO nonattainment areas were redesignated as maintenance areas.   
RIA Questions
   * What would be the effect of including LD 2017 -- 2025 and HD GHG rules in the baseline for this rule?    
      EPA Response:  Including the 2017-2025 LD GHG and HD GHG standards in the baseline will not substantially modify our results or conclusions.  Our analysis of the Tier 3 proposal is based on the relative differences in emissions impacts, air quality impacts and benefits between future year scenarios with and without the Tier 3 standards.  Had we modeled the 2017-2025 LD GHG and HD GHG rules, they would be included in both the reference and the control scenarios; therefore, the relative differences that characterize our inventory, air quality, and benefits analyses would not change with the inclusion of these rules. Some of the changes to the vehicle fleet expected to result from the 2017-2025 LD GHG standards, such as a greater penetration of downsized engines along with further downsizing, could somewhat lower the cost of complying with Tier 3 standards.  The reduced fuel consumption associated with the 2017-2025 LD GHG standards could result in somewhat lower costs for sulfur control, because less gasoline will need to be desulfurized.   We will be including both the 2017-2025 LD GHG and HD GHG rules in the baseline for the final rule's analysis.

Executive Summary
   * How are the average fleet standards determined (referred to in Table ES-2)? 
      EPA Response:  Please refer to Section IV.4.a of the preamble for a discussion of the proposed vehicle NMOG+NOx average fleet standards for SFTP. 
Chapter 2
   * What is the source for projected vehicle sales presented in table 2-26 (Page 2-24)? 
      EPA Response:  The sales shown are taken from AEO 2011 and are consistent with the sales projections that support the joint MYs 2017-2025 light-duty GHG/CAFE rules and the MYs 2014-2018 heavy-duty GHG rule.
   * It would be useful to revise or reorganize the discussion of industry learning-by-doing in Section 2.1.3.  As currently drafted, it is difficult to follow.
      EPA Response:  This discussion is essentially a copy/paste of the discussion developed jointly by EPA and NHTSA in support of GHG/CAFE rulemakings and reviewed/approved by OMB in the past.  We prefer not to revise the text given the efforts toward consensus behind it.
Chapter 3
   * Pg 3-2, According to AAM summer fuel surveys, the average octane of finished gasoline has remained constant around 89-90 (R+M)/2 over the past decade (refer to Figure 3-2). The reported octane is higher than expected because AAM takes roughly an equal number of regular and premium gasoline samples as part of their North American Fuel Survey. In reality, the majority of consumers fill up on 87 octane regular-grade gasoline based on price and other factors. According to EIA's Petroleum Marketing Annual, regular-grade gasoline represents over 85 percent of U.S. sales.2 Accordingly, we believe our proposed 87-88.4 R+M/2 specification for test fuel is representative of in-use gasoline. However, the AAM survey does tell us one important thing about octane  -  refiners are doing their best not to give it away. We anticipate that this trend will continue into the future as E15 replaces E10.  
         o Comment:  There is still large percentage of octane given in the mid-continent because of the operating practices of the pipeline system (Megallan) in the region, although they are attempting to changes this starting this year.
            
            EPA Response:  While the AAM survey data suggests refiners have been match blending for octane, we agree that not all have been historically.  However, this analysis is in support of a new certification fuel that will likely be used for many years.  Thus we are trying to reflect fuel that will be used well out into the future.  Magellan is planning on switching over their entire system later this year, and other pipeline systems already have.   Thus, we believe that it is appropriate to assume that match blending for octane will occur and that assuming 87 octane for our certification test fuel is appropriate.  
            
   * Pg 3-17, Shifting to E15 test fuel will result in a T50 range that's about 15 percent lower than today's E0 test fuel and about 10 percent lower than today's in-use E10. The proposed 170-190°F T50 range for E15 test fuel was determined by interpolating between the T50 of current 9-psi E10 market fuel which averages around 195°F (according to AAMB) and the estimated T50 of E20 which spans a narrow range of plus or minus several degrees Fahrenheit centered around165°F (according to our EPAct test program5). We believe the proposed T50 range adequately characterizes future 9-psi E15 market fuel. 
         o Comment:  EPA should seek comment on the applicability of this assumption.  This assumption presumes that another grade of CBOB and RBOB will and can be economically transported and stored in the existing infrastructure for an initially small volume of E15 sales
            
            EPA Response:  The impact of E15 on T50 and distillation here does not assume a special grade of RBOB/CBOB will be produced and distributed for blending with E15.  For the most part, the effect of more ethanol on T50 is independent of match blending vs splash blending and results directly from the fact that ethanol has a relatively low boiling point.  We already seek comment on this very issue in the preamble.
   * Pg 3-18, We expect that refiners will continue to back off on octane production once the transition from E10 to E15 has occurred.
         o Comment:  An full explanation of why this would occur needs to be presented.  This assumes that the transportation, storage and retail infrastructure could cost effectively accommodate multiple octane grades and RVP types where E10 RFG, 9psi, and 10psi Rvp CBOB exist.
            
            EPA Response:  Yes, in the case of E15, it assumes many hurdles are overcome, and we have already added discussion to the preamble in response to prior OMB comments seeking comment on E10 vs E15 for certification test fuel.  However, this comment is raising a short-term issue while we are looking at fuel changes for the long term.  Over the long term we would expect that the market would in fact match blend more and more for octane.  We believe the text is already clear that this is a longer term view.
Chapter 4
   * Pg 4-4, The average amount of sulfur in crude oil refined in the U.S. is about 10,000 ppm.
         o Comment:  EIA data indicates that the average crude sulfur content run in U.S. refineries is about 14,000 ppm.  See link: http://www.eia.gov/dnav/pet/pet_pnp_crq_dcu_nus_a.htm 
            
            EPA Response:  We agree and will change it to 14,000 ppm.  This was a carryover from prior language, and our cost analysis is already based on 14,000 ppm.  
            
   * Please provide fuller discussion on the permitting analysis (4-31). 
      
      EPA Response:  There is already a cross reference to a lengthy discussion of this in section V.B. of the preamble.
      
   * We suggest changing the title of the section 4.3.3 to distinguish from Chapter 9 and be consistent in using person-hour.  
      
      EPA Response:  We changed the title of section 4.3.3 to "Employment constraint analysis".  We changed work hours to "person-hours" in this section.  

Chapter 5
   * The first part of the task is to attempt to determine how the refinery is meeting the 30 ppm sulfur gasoline limit as a basis for what needs to change.  Major portions of the data needed to accurately determine the current operation of a refinery is not publicly available and even with some information available to EPA from company-confidential EIA data there remains much data that must be estimated or obtained from refiners willing to cooperate.  The quality of the data estimations have a critical impact on the cost results developed.  It is suggested that EPA explicitly seek comment from refiner's where data was not available and assumptions were used.   
      
      EPA Response:  As discussed below, we have added detail to the request for comments in the preamble to address these comments.
      
   *  Pg 5-1, When used to model the cost of nationwide fuel control programs on the entire refining industry, LP models are usually used to model groups of refineries in geographic regions called Petroleum Administration for Defense Districts (PADDs).
         o Comment:  LP models are more widely used by refiner's to help them optimize a particular refinery.  Very few refinery LP models are used to model the industry in aggregate.  The handful of LP models that exist are usually used by or for government entities such as EPA and DOE or NGOs that are reviewing regulation that impact the industry.  The commenter suggests that EPA reword this sentence and the next sentence to say that the aggregate industry LP models could potentially underestimate or overestimate the cost of the program depending on the accuracy of the input data and assumptions.
            
            EPA Response:  We will add a sentence which states:  Based on the quality of input data to these LP models and the assumptions made for complying with a regulatory requirement, LP refinery models may overestimate or underestimate the program costs. 
            
   * Pg 5-6, In 2017, we estimated that approximately half the gasoline would be 10 percent ethanol and the about the other half would be 15 percent. To model the emissions impact of the different ethanol blends, we modeled two reference cases, one with 100 percent E10 and the other with 100 percent E15.
         o Comment:  Suggest that EPA discuss the analysis and assumptions made regarding these estimates.  Did EPA consider the feasibility of the existing infrastructure of accommodating another octane and RVP grade for E15 gasoline blendstock that would be needed in order to implement these policies?  Did EPA estimate the cost of this additional infrastructure?  Please explain.
            
            EPA Response:  We are aware of the issues associated with expanding the use of ethanol beyond E10, including the infrastructure issues.  However, they are of very little relevance to the Tier 3 proposal, as they are in the baseline ("reference case") for the Tier 3 rulemaking.  Their only relationship to the Tier 3 proposal is that in order to model the emission and cost impacts of Tier 3 in the future, we first had to establish a future reference case.  In conducting the analysis for the RFS2 final rulemaking in 2009, we assumed certain volumes of ethanol use in the future through blending of E10 and E85.  For this analysis, we assumed the same ethanol volumes and that E15 use would grow out into the future in lieu of E85.  We can revisit this assumption for the final rulemaking as appropriate.  
            
   * Pg 5-24,We estimate that 66 percent of the volume of NGL and LSR are hydrotreated before being blended into gasoline and have a very low sulfur content of approximately 1 ppm.
         o Comment:  What is the basis for this estimate?  Please explain in the text.
             
            EPA Response:  This is discussed in detail in chapter 5.1.4.2. beginning on page 5-37.
            
   * Pg 5-24, For the Tier 3 control case we assumed that all of the NGLs and LSR were hydrotreated and therefore had an average sulfur content of 1 ppm. 
Comment:  Suggest EPA seek comment on this assumption. 
EPA Response:  See section 5.1.4.2 for more detail.  This is an assumption that only tends to increase our costs.  We added some text to the fuel cost section of preamble section VII .B which requests comments on this and other issues raised by the comments:  "We request comment on all aspects of the methodology described in Chapter 5 of the draft RIA used for estimating the cost of the program using the refinery-by-refinery cost model.  In particular, we request comment on: the methodology for estimating the sulfur content of the FCC naphtha; how FCC naphtha sulfur levels are impacted by both an FCC pretreater and an FCC unit; to what extent desulfurization costs for complying with Tier 3 vary due to the sulfur level entering the existing FCC postreaters installed for Tier 2; the estimated level of hydrotreating of light straight run (LSR) naphtha and natural gas liquids (NGL) occurring today; and the level of hydrotreating of LSR and NGLs expected to occur under this program." 
   * Pg 5-29, Information we received from the vendors also indicated that refiners with both a pretreater and a postreater are producing FCC naphtha that ranges from 200 to 450 ppm before being processed by the postreater.
         o Comment:  Suggest that EPA seek comment from refiner's concerning this assumption. 
            
             EPA Response:  A request for comment was added to the preamble which is contained in the response above to the second comment on page 5-24.
            
   * Pg 5-29, After we have calculated the sulfur level of the FCC feed we must then take into consideration the impact the FCC unit itself has on the sulfur level of the FCC naphtha. We reviewed several literature sources7,8 and found that the FCC naphtha sulfur level can be accurately determined by dividing the FCC feed sulfur level by 20 for refineries with an FCC feed pretreating unit. For refineries without an FCC feed pretreater, the FCC naphtha sulfur levels can be calculated by dividing the desulfurized FCC feed sulfur level by 10. 
   * Comment:  Suggest EPA seek comment on this methodology to determine the level of variance there may be relative to the 1/20th estimate. 
      
      EPA Response:  A request for comment was added to the preamble which is contained in the response above to the second comment on page 5-24. 
      
   * Pg 5-32, The costs for the FCC postreater revamps submitted by one of the vendors, however showed that for low (0  -  400 ppm) and...
         o Comment:  This is a key economic assumption.  How many refineries fall into this category? 
            
            EPA Response:  See response below for comments on Table 5-26.
            
   * Pg 5-32, According to this vendor these cases also had a 4 to 5 year catalyst cycle length, equivalent to the higher capital cost cases even though a second stage reactor was not required.
         o Comment:  EPA should seek comment on this assumption.  Axen's NPRA-12-08 paper indicates that adequate cycle length may not be achieved via revamp of pretreaters. 
            
            EPA Response:  
            We did not assume any revamping of pre-treaters, so this comment appears to be irrelevant.  
   * Pg 5-32, In our refinery-by-refinery model, however, we had multiple refineries with FCC feed sulfur levels in the 0  -  400 and >1,200 ppm categories that use this vendor's postreating technology. In order to apply this vendor's cost estimate to cases of low (0  -  400 ppm) and high (>1,200 ppm) sulfur feed categories we adjusted this vendors 400  -  1,200 ppm postreater revamp cost based on the cost differentials between the three FCC naphtha sulfur levels in the other vendors' revamp estimates. 
   * Comment:  Suggest EPA seek comment on this cost adjustment approach and other alternatives. 
      
      EPA Response:  A request for comment was added to the preamble which is contained in the response above to the second comment on page 5-24. 
      
   * Pg 5-34, Table 5-26 - It would be helpful if there was a table that translated this table (5-26) to cents per gallon. 
      EPA Response:  These costs cannot be expressed as cents per gallon because they are not proportional to gallons.  In addition, they are inputs into coming up with costs per gallon for a 30,000 BPSD FCC postreater,, not results.   Instead we added a footnote in the fuel cost section (section 5.2) of the draft RIA which says:  "Approximately 55% of this desulfurization cost is comprised of the variable cost, and the majority of that is due to the cost of recovering the octane lost when the hydrotreater unit saturates the octane-rich olefins contained in the FCC naphtha.  Another 30% of the cost is due to the capital cost amortized over the volume of gasoline.  Finally, about 15% of the cost is due to the fixed operating cost which includes the maintenance of the new equipment and taxes."  
      
   * Also it would be helpful if EPA provided the number of refineries that fell into each revamp category.  
      
      EPA Response:  We have added a footnote to the table providing how many refineries we modeled that fall into the 3 different categories.   The footnote states that of the refineries that are expected to revamp their FCC naphtha hydrotreater for the no ABT case, 36 have FCC naphtha sulfur levels in the 0  -  400 ppm range, 20 have FCC naphtha sulfur levels in the 400  -  1200 range, and 13 refineries have FCC naphtha sulfur levels greater than 1200 ppm.
      
   * How did EPA arrive at the cost estimates for varying fuel scenarios (e.g., no ABT, AB but not T, ABT)?  What does this imply about the price of the sulfur credits?
      
      EPA Response:  Section 5.1.5.2 describes the ABT analysis.  For no ABT we assumed each refinery needed to meet the 10 ppm average standard (averaging only within the refinery itself across its own batches of fuel).  For our primary cost analysis we assumed only "trading" within the same company to be conservative.  Since we anticipate some trading between companies, but cannot assess the degree to which this will occur, we also show a sensitivity analysis showing full trading nationwide.  In all cases the analysis assumes that the lowest cost per gallon refineries make the investment.   Thus, we essentially created a list of refineries, including those with gasoline sulfur levels which are already low and those needing to lower their gasoline sulfur.  The list started from the lowest-cost refinery increasing to the highest-cost refinery and we included the costs for reducing their gasoline sulfur down to both 10 and 5 ppm.  We simply went down the list from the lowest cost refinery to the highest cost refinery until the cumulative sulfur reduction over the refineries on the list achieved the average of 10 ppm in the gasoline pool.   When we modeled the costs for the inter-company trading ABT case, the list of refineries was constrained to only those owned by one company. 
       
   * Please discuss why EPA assumed that refineries that did not produce 30 ppm fuels to produce 30 ppm fuels?  This seems inconsistent with other assumptions invoked in characterizing the baseline for the fuels section (e.g., CA mandates 15 ppm fuel with 20 ppm cap but the average sulfur level is 10 ppm). 
      
      EPA Response:  This is an analysis for future Tier 3 compliance, so our baseline is Tier 2 at 30 ppm.  Because of the long phase-in flexibility provided for Tier 2 and the fact that our refinery-specific fuel quality data was from 2009, there were a few refineries that were in compliance with Tier 2, but for which we did not yet have actual sulfur data.  For these, we merely assumed they met the average of all the other refineries complying with Tier 2; 30 ppm.
      
   * Should compliance cushion be considered in the analysis (e.g., 5 ppm compliance cushion seen in CA fuel)?
      
      EPA Response:  No. The California standard is a sulfur cap, not an average.  The current Tier 2 standard is an average of 30 ppm and the in-use average is 30 ppm, so there is no over-compliance.  Furthermore, the California in-use sulfur level is not driven by their sulfur cap, which went into effect after their average sulfur level had already fallen to 10 ppm.  California fuel sulfur levels are driven by the emissions performance required of their fuel using their "Predictive Model."  The Predictive Model forced sulfur levels down to about 10 ppm on average as the refiners added 10 percent ethanol.  
      
   * Do different fuel scenarios have different implications for fuel volume or did EPA assume that identical fuel volumes would be produced under the varying scenarios? 
      
      EPA Response:  All scenarios modeled assume the same fuel volumes.  
      
   * EPA discarded information from one of the vendors in assessing the cost to revamp FCC postreaters.  Please discuss the implications of including and excluding this information.  How much will the cost estimates vary when the excluded vendor information is included?
      
      EPA Response:  The vendor is also a refiner and simply provided us with information for controlling sulfur from pre-Tier 2  uncontrolled sulfur levels.  They declined to provide us with the information needed to make it applicable for the Tier 3 proposal.  Thus, it was not appropriate to use it for assessing Tier 3 costs and our peer reviewers concurred.  We have not analyzed the cost implications were we to try to use this vendor's data, nor is it clear how we would modify it to be appropriate for Tier 3 purposes.  Were we to be able to utilize this vendor's information, it would only apply to that subset of the refiners that use that technology.   

Chapter 7
   * Please provide the technical document on the MOVES updates.  Does this document include information on the survey of "in-use vehicles?"

   EPA Response:  This question refers to the statement:  "In producing emission inventory estimates, MOVES combines emission rates with activity patterns derived from surveys of in-use vehicles.  These emissions do not necessarily correlate directly with the test procedures used for compliance; for example, in-use activity shows that more miles are driven per start event than assumed on the FTP."    The docket memo does not explain all MOVES2010 inputs, only those that were changed for this analysis.  Information about the in-use activity data in MOVES is available in  "MOVES2010 Highway Vehicle Population and Activity Data" EPA-420-R-10-026 http://www.epa.gov/otaq/models/moves/420r10026.pdf
            
   * If 80% of the vehicles and 90% of VMT have turned over by 2030, please discuss the need for conducting the analysis to 2050 (7-1).  Why not stop at 2030 especially knowing that EPA is very likely to issue another standard between now and then?
      
      EPA Response:  It is standard practice for us to estimate impacts far into the future.  For example, the RIA for the Tier 2 rule (which regulated vehicles starting in model year 2004) estimated emission impacts until 2030 (a 26-year window) and the recent LDGHG rule also estimated emissions up to 2050.  Because the vehicle fleet turns over slowly (more slowly now than in the past when vehicles were not as durable), estimates into the far future are necessary to capture the full impact of the rules.  The new Tier 3 standards phase in through model year 2025, so while in 2030, 80% of the vehicles are subject to Tier 3 in some way, many are still not subject to the fully phased-in NMOG+NOx standard, and the emission inventory does not yet reflect the full benefits of Tier 3.
            
   * Please discuss the need to include CA in the analysis if the CA fuel is assumed to be 10 ppm in the baseline.  Are there differences in the vehicle standards being proposed compared to the existing CA standards?
      
      EPA Response:  EPA included CA in the analysis because at the time the analysis was conducted, the LEV III program had not been finalized and EPA had not granted California a waiver of preemption under Clean Air Act section 209.
      
   * How do the results of the analysis change when the mix of vehicles explicitly takes into account the LD GHG rule? (Presumably the emission profile would not change but the aggregate level of emissions should have dropped due to the LD GHG rule.)
      
      EPA Response:  We do not expect the future vehicle fleet that is compliant with 2017-2025 LD GHG rule to have different non-GHG emissions than assumed in the Tier 3 rule.  Fuel efficient technologies will not change the fleet's non-GHG emissions or the effectiveness of Tier 3 controls, and improved fuel efficiency does not have a direct impact on criteria pollutants.  Furthermore, non-GHG emissions are a function of the emission control systems, not the amount of fuel burned.  Tier 2 and Tier 3 standards are fleet average standards, so increased penetration of electric vehicles would not change the vehicle fleet's non-GHG emissions.
      
   * How sensitive are the results to the various assumptions about fuel use (e.g., the use of E15, changes in RVP, octane loss)? 
      EPA Response:  See below
      
   * Pg 7-8, While E15 has only limited commercial availability currently, EPA believes it may compete favorably with E85 in the marketplace and could become a major gasoline blend in the future to meet the mandated RFS2 volumes. 
         o Comment:  Suggest EPA provide additional discussion concerning the analysis that the agency has conducted and assumptions made in order to support this statement.  
      EPA Response:  See below
      
   * Pg 7-9, We assumed that E15 utilization for 2001 and later model year vehicles would ramp up to 50 percent by 2017, increasing to almost 100 percent utilization by 2030.
         o Comment:  Does this assume that consumers will refuel with E15 even though their owners manuals (almost all pre-2013 MY vehicles) only provide warranties for the vehicle fuel up to E10 blends not E15?  EPA should seek comment on this assumption.
      EPA Response:  See below
      
   * Pg 7-9, As such, we assumed that E15 utilization in nonroad equipment would ramp up from zero percent in 2017 to almost 100 percent by 2030.
         o Comment:  Suggest EPA seek comment on this assumption. 
      EPA Response:  See below
      
   * Pg 7/12, Due to the additional actions needed to market E15 during the summer in CG areas, it makes sense that more E15 would be used in RFG areas, at least in the near term. Accordingly, we assumed that in 2017, 75 percent of RFG would be E15 and the balance of E15 would be used in CG (resulting in 25 percent E15 utilization and 75 percent E10 utilization in those areas).
         o Comment:  Suggest that EPA detail their analysis and assumptions made concerning how such a large penetration of E15 would occur in metropolitan areas by 2017.  There are a number of significant obstacles to E15 achieving market acceptance.  First, existing owners of MY 2001 and newer vehicles may not be receptive to E15 because manufacturer warranties are not provided for its use.  The transportation and distribution refueling infrastructure will have difficulty accommodating E15 gasoline blendstocks, octane, and RVP grades in addition to E10 octane and RVP gasoline blendstock grades.  Retail outlet owners have expressed concern about the compatibility of their equipment with E15 gasoline.  Additionally, if they do not invest in new refueling infrastructure retailers may be reluctant to forgo sales of a higher margin mid-grade and premium grade E10 gasoline for E15 with a potentially lower sales volume and margin associated with a product such as E15.
            
            EPA Response:  There are a number of comments here related to our E15 assumptions.  We are aware of the issues associated with expanding the use of ethanol beyond E10, including the infrastructure issues, vehicle warranty issues, etc.  However, they are of very little relevance to the Tier 3 proposal, as they are in the baseline or reference case for the Tier 3 rulemaking.  Their only relationship to the Tier 3 proposal is that in order to model the impacts of Tier 3 in the future, we first had to establish a future reference case.  In conducting the analysis for the RFS2 final rulemaking in 2009, we assumed certain volumes of ethanol use in the future through blending of E10 and E85.  For this analysis, we assumed E15 use would grow out into the future.  Since this only affects the reference case, it has little impact on the emission reductions resulting from Tier 3.  We can revisit this assumption for the final rulemaking as appropriate.  
            
   * Please explain the error associated with the 10 ppm CA fuel baseline (7-15)?
            
      EPA Response:  An inconsistency in our approaches for estimating California fuel sulfur levels led us to model an increase in sulfur in California.  More specifically, for the reference case, we used MOVES default fuels for California.  These were developed based on a mix of survey data and other sources, which resulted in a range of sulfur levels, some lower and some higher than 10 ppm.  However, the control case assumed 10 ppm sulfur throughout California.  As a result, although there should have been no change in California fuel properties due to Tier 3, some areas had small modeled increases or decreases.  
            
   * What types of vehicles are in the Tier 2 Bin 2 (7-20)?
      
      EPA Response:  See provided Excel spreadsheet:  iuvp_Bin2_2010203.xls 
      
   * Please explain the cause(s) of the "sharp drop in emissions at the outset of Tier 3 phase in." (7-21, Figure 7-4).
      
      EPA Response:  See the response to preamble question #1.  The drop in emissions  from  MY 2016 to MY 2017 represents the change in emission standards that we are modeling.  Emissions decline from the average emissions in the last year when all the fleet meets Tier 2 standards (160mg/mi in 2016) to the average emissions in the year in which Tier 3 emission standards begin to phase-in (86 mg/mi in 2017). (See Preamble Table IV 1 Proposed LDV, LDT, and MDPV Fleet Average NMOG+NOX FTP Standards (mg/mi) and the paragraph that precedes this table.)
      
   * Please provide the Kansas City Light duty vehicle emission study (7-32).

      EPA Response:  The Kansas City report is available on the web at  http://www.epa.gov/otaq/emission-factors-research/420r08009.pdf  It is not attached here because it is quite large. (28 M)
      
   * Please explain why EPA is projecting decreased lube oil consumption as a result of Tier 3 standards.
      
      EPA Response:  The tighter PM  standard and extended full useful life proposed for Tier 3 will require very low oil consumption.  Oil burners simply should not be able to pass the standard.

   * Please explain "prevalence of high evaporative emission vehicles in the Colorado field study." (7-34)  Does Tier 3 mandate an increase in the durability of evaporative/fuel systems?
      
      EPA Response:  The intent of the high evaporative field studies done in Colorado was to find out how often "high" evaporative emissions were occurring in the fleet, or the "prevalence".  The studies looked at causes for the high emissions and found vapor leaks were a common source of these excess emissions. This issue had not been addressed since the 1990's and not on the latest enhanced evaporative emissions technology which now dominates the fleet of aging vehicles.  See Section IV.C.5.a in the Preamble.
      
      Yes, Tier 3 does propose an increase in evaporative control system durability.  This is accomplished through two provisions:  The first is by applying the proposed extended useful life requirements to the "zero-evap" standards for most vehicles.  This by definition will necessitate better materials and durability of evaporative system performance.  The second is the proposed leak test and leak standard of 0.02" in Section IV.C .5b of the Preamble.  To support this we are proposing the adoption of CARB's enhanced OBD and 0.020" evap leak detection (see Section IV.C.5.d of the preamble). 
      
   * What is the extent of reversible sulfur loading in Tier 2 vehicles (7-38)?
      
      EPA Response:  We cannot quantify how much sulfur is on the catalysts.  Our test programs can only measure the impacts of the sulfur loading on emissions.   Based on test results from the EPA Sulfur Research Program and on conversations with the Manufacturers of Emission Controls Association (MECA), whatever the current sulfur loading is, it is completely reversible when vehicles experience sufficiently high speed/load driving conditions, and the level of reversible sulfur loading in the in-use fleet is lower on lower sulfur fuel.   
      
   * Please explain the 700 short ton PM increase and the conservative assumptions. (7-44).  
      
      EPA Response:  The sentence immediately after that statement refers the reader to section 7.1.3.2.2 which contains a very detailed explanation of why we believe that the 700 short ton PM increase would not occur.  
      
   * How representative are the "representing counties" in assessing VMT? (7-48)
      
      EPA Response:  As explained in the RIA, the representing counties are chosen so they can be used to compute g/mi factors that will be representative across the group of counties. To assure this, the counties are grouped based on fuel parameters, emission standards, I/M programs and altitude.  However, representative counties are NOT meant to represent VMT.  VMT is estimated for every Continental U.S. county.  The SMOKE model calculates emissions by multiplying the county-specific VMT by  the county-group specific g/mi emission rates produced in the MOVES run.

   * What is the magnitude of the difference between adjusted and unadjusted VMTs? (7-52)?
      
      EPA Response:  These adjustments are designed to address increases in VMT for heavy duty trucks associated with transporting ethanol, and are applied to both the Tier 3 reference and control cases. The adjustments affect a limited number of U.S. counties along projected ethanol transport routes.  Nationwide impacts on VMT are small  -  less than 0.03% of total diesel truck VMT in 2017 and less than 0.05% in 2030. 
   * We suggest updating the discussion on PM NAAQS. 

      EPA Response:  We've updated the PM NAAQS discussion in the version of the draft RIA dated 2/28/13.

   * We suggest deleting "As mentioned in Section 7.2.4.3.1...to attain the NO2 NAAQS." (7-93) 
      
      EPA Response:  We deleted  the first part of that text, see below.
      
      As mentioned in Section 7.2.4.3.1, nonattainment designations for the NO2 NAAQS will
      potentially be occurring in 2016 or 2017. The emission changes from this proposal would go
      into effect during the period when any NO2 nonattainment areas would be working to attain the
      NO2 NAAQS.  

Chapter 8
   * Where is the projected annual consumption referred to in Section 8.1 (Page 8-2)? 

      EPA Response:  RIA Chapter 8.1 was updated to include the annual fuel consumption (see Table 8-2).
      
   * What is assumed rate of income growth between 2000 and 2030 (Table 8-14)?  Assumed to grow linearly? 
      
      EPA Response:  Willingness-to-pay-based benefits are adjusted by multiplying the unadjusted benefits by the appropriate adjustment factor to account for income growth over time using projections provided by Standard and Poor's.  Due to a lack of reliable projections of income growth past 2024, we assume constant WTP from 2024 through 2030. This results in an underestimate of benefits occurring between 2024 and 2030.  For growth between 2000 and 2024, this factor is 1.23 for long-term mortality, 1.27 for chronic health impacts, and 1.08 for minor health impacts.  Note that similar adjustments do not exist for cost-of-illness-based unit values.  For these, we apply the same unit value regardless of the future year of analysis.
      
   * 8.1.4:  Will EPA provide the need for including this section in the RIA? 
            
      EPA Response:  As we mentioned in our response to interagency comments on 2/28/13 (EPA response part 2 to EO_12866_Tier 3_Summary Comments_2_13_2013.docx), we included this analysis to provide substantial and transparent analysis of the impacts of the proposed vehicle and sulfur standards.  Tables III-2 and III-3 from the preamble show that in 2017, the impact of fuel sulfur standards on the existing fleet of Tier 2 and pre-Tier 2 vehicles is quite significant in terms of NOx and VOC reductions.  The proposal would therefore meaningfully reduce ozone concentrations as early as 2017.
      
   * 8.2: this section is inconsistent with the guidance in A-4.  With the current framework, this type of analysis would be more useful if used to look at provision by provision effects.  We suggest using this "version" of cost effectiveness analysis to compare the fuel provision vs. vehicle exhaust provisions vs. vehicle evaporative provisions (and move the discussion to Chapter 7). 
      
      EPA Response:  As we noted in our prior response, we analyzed the combined fuel and vehicle program, in keeping with the fundamental principles of the Tier 3 program: that vehicles and fuels should be analyzed as an integrated system, as has been successfully done in the mobile source emission control program for over a decade; and there should be a harmonized national vehicle program that establishes a comprehensive approach to regulating motor vehicles (both greenhouse gases and non-greenhouse gases), as requested by the Presidential Memo of May 2010.

      As described in detail in the draft proposal and in the scientific literature, there is a first-order relationship between fuel sulfur content and the efficiency of modern three-way catalytic converters.  The Tier 3 proposal addresses the stringency of exhaust gas standards and fuel sulfur level as a system because that is what the scientific literature tells us -- the two are strongly linked.
            
      As such, we determined the program's overall costs as the sum of the vehicle exhaust and evaporative costs, and the fuel costs and we determined the program's benefits as the sum of the impacts on the pre-Tier 3 fleet and Tier 3 fleet.  We cannot separate out the benefits attributable to the vehicle program from the fuel program.
      
      In response to the OMB comment, we will rename RIA Chapter 8 to "Comparison of Program Costs to Program Emission Reductions and Air Quality Benefits" to clearly reflect in the title that it contains an evaluation of the total program.

Chapter 9
   * Cost effect on auto manufacturers: since the agency is not presenting an estimate of employment effects, we suggest deleting Table 9-2 and modifying the discussion appropriately.

      EPA Response:  EPA is using the same partial quantification approach on vehicle impacts that was used in the model year 2017-25 light-duty greenhouse gas rule.  Partial quantification provides insight into the potential magnitude of employment impacts.  In addition, it demonstrates how employment impacts depend on a number of factors, such as the connection between regulatory costs and employment.
   
   * Refinery employment impacts: what is the source for the 1000 engineering jobs and 6000 construction jobs?

      EPA Response:  As stated in the text in Chapter 9.3.2, these employment estimates come from Preamble Section V.A.2.c; they are also found in DRIA Chapter 4.3.3.  They are engineering estimates for the design and construction people-hours associated with refinery projects needed to produce fuels that will meet the proposed standards.

Chapter 10
   * What are the sources of sulfur after the refinery gate?  What are the associated costs of reducing after the refinery gate sulfur? 
      
      EPA Response:  The sources are contamination in the distribution system from higher sulfur levels in other products they come into contact with, and the addition of sulfur containing additives.  Given that diesel fuel is already now lower sulfur, and heating oil is expected to become lower sulfur, we anticipate that any increase in the sulfur content of gasoline from contamination during distribution will be small.  Sulfur containing additives are typically used at low treatment rates, hence their contribution to the average downstream gasoline sulfur content will also be small.   Transmix processors are a special class of a refiner that separates the mixture of gasoline and distillate fuels that is a by-product of pipeline distribution back into gasoline and diesel fuel through simple distillation.   Some high-sulfur distillate inevitably ends up in transmix gasoline product.   The gasoline produced by transmix processors is ~0.2% of the total gasoline consumption.   Hence, the contribution of average gasoline sulfur content is small.  The proposed 15 ppm differential between the refinery sulfur cap and the downstream sulfur cap is necessary to accommodate the infrequent incidences when all of the factors that can increase gasoline sulfur content (contamination during distribution, additives, transmix gasoline product) are coincident and at their maximum.   The costs of reducing these downstream sources of sulfur would not be commensurate with the small potential benefit. 
      
   * What are the reasons for restricting the use of credits to comply with Tier 2?  In section III. B. of the preamble, it states that Tier 2 vehicle program would go on indefinitely but the limitation on the use of credits should be clarified.  
      
      EPA Response:  It is not clear what credits are being referenced.  Discussion of the life of vehicle credits is contained in the response above to preamble comment 3.  Discussion of the life of fuel sulfur credits is contained in section V.E.1.d. of the preamble.  There we state that five years is consistent with the proposed recordkeeping and reporting requirements as well as the current Tier 2 and MSAT2 standard credit life provisions.  Section III.B. is a discussion of our modeling assumptions.  In the reference case, Tier 2 is assumed to remain in place indefinitely.     

   * Have any of the flexibilities applicable to small businesses been considered for the large manufacturers? Implications for costs and benefits?
      
      EPA Response:  The proposal includes a 3-year delay in the fuel standard for small refiners and all small refineries (including those owned by large companies) less than 75,000 bbls/day.  Several other provisions developed and discussed during the SBREFA process that will help small businesses will also provide flexibility to large businesses, including the provisions for additive manufacturers, the downstream sulfur caps, and the hardship provisions.   Likewise for the vehicle standards, we are proposing a delay in compliance for small volume manufacturers in addition to small businesses.
      
   * What are the costs associated with testing the recordkeeping?

      EPA Response:  We do not understand the question.

