Gasoline Properties Over Time

This webpage provides the public with data on gasoline fuel properties and how they have changed over time due to both EPA standards and shifts in market dynamics. The results are compiled from data provided to EPA by refiners, gasoline blenders and importers to verify compliance with our gasoline fuel quality standards.

The analysis and presentation of the gasoline property data is provided by the following two reports:

• Fuel Trends Report: Gasoline 1995 – 2005 (PDF)(268 pp, 35.9 MB, EPA420-R08-002, January 2008)
• Fuel Trends Report: Gasoline 2006 – 2016 (PDF)(123 pp, 2.15 MB, EPA-420-R-17-005, October 2017)

Important notes regarding the data include:

• The data shows how gasoline properties have changed due to the implementation of EPA fuel quality standards over time as well as changes in the marketplace. See: What the Data Shows
• The gasoline properties after 2005 were adjusted to account for the downstream blending of ethanol to better indicate the gasoline properties at retail.  See: How the Data was Adjusted and Reported
• Summaries of the gasoline property data for the years 1995 through 2016 are summarized under the tabs:
  • Conventional Gasoline;
  • Reformulated Gasoline; and,
  • All US Gasoline (Except California)
• Although these figures are calculated from data drawn from compliance reports submitted by refiners, the figures here do not represent actual compliance information used to determine whether any specific regulatory party has satisfied their statutory and regulatory requirements.

If you have questions or request information, please contact the appropriate support or help line found on the Support and Help page.

About the Data

• This Web page provides the public with estimated average retail gasoline fuel properties and how they have been trending over time due to
  • EPA standards
  • market shifts
• The data presented here  is collected  by EPA from
  • refiners
  • gasoline blenders
  • importers
• The data are reported to EPA for every batch of gasoline produced at the refinery, blended at terminals or imported into to the U.S., and sold in the U.S.
  • does not apply to exported gasoline
  • does not include the gasoline sold in California
• This data set represents the most accurate and comprehensive set of data available for gasoline quality in the U.S.
  • The 1995 through 2005 data was already posted based on a previous report¹ and is reposted along with new gasoline property data for the years 2006 to 2015
• The batch data are reported as measured at the refinery/importer, not at the retail station, so we have adjusted the results to approximate actual retail fuel properties, as described under the tab “How the data was adjusted and reported,” to account for the downstream blending of ethanol
• Because most of the data is reported from refineries and the gasoline market served by each refinery is uncertain and variable, we cannot use this data to make conclusions about the gasoline properties in any specific retail market at any specific point in time

Other Data

• Other data exists which also characterizes gasoline properties
  • Reformulated Gasoline (RFG) Survey is conducted every year and reported to EPA
    • This is a large retail survey that collects data on a range of fuel properties for thousands of fuel samples collected each year from retail stations in RFG areas across the country
    • However, it is limited to only gasoline samples collected from RFG areas, not conventional gasoline areas
  • Retail surveys by the Association of Automobile Manufacturers (AAM) and the previous fuel surveys by TRW
    • These are spot surveys sampling a small fraction of all the gasoline sold at retail for a couple points in time per year

What the Data Shows

• The data shows how gasoline properties have changed due to the implementation of EPA fuel quality standards over time as well as changes in the marketplace
  • RFG Program
    • Took effect in 1995 and 1998
    • Relative to the 1990 baseline gasoline properties (provided in the charts) the RFG program caused a reduction in RVP, sulfur, benzene, aromatics, T50 and T90, while causing an increase in the use of oxygenates
  • Tier 2 sulfur standards and the beginning of Tier 3 sulfur standards
    • Tier 2 Phased-in over the period from 2004 to 2006, although the averaging, banking and trading provisions provided an incentive for refiners to begin reducing their gasoline sulfur prior to 2004, and some small refiner exemptions continued until 2011
    • Tier 2 caused average gasoline sulfur to decrease from about 260 parts per million (ppm) down to about 30 ppm
    • Tier 3 took effect January 1, 2017, although to take advantage of early credits, refiners had already been reducing sulfur levels in 2014 and 2015
    • Tier 3 has already caused the average gasoline sulfur level to drop below 30 ppm in 2014 and 2015, and it is expected to continue to steadily decline to 10 ppm by 2020
    • Associated effects of gasoline desulfurization
      • A reduction in olefins
      • An increase in RVP
        • Under the RFG program, lower sulfur gasoline allows refiners to increase RVP while maintaining the same overall environmental performance
        • This, along with the reduction in the number of low-RVP and RFG areas, has caused summertime RVP levels to increase slightly over time
  • Mobile Source Air Toxics program (MSAT2)
    • MSAT2 benzene reduction provisions began in 2011, although refiners began to reduce their gasoline benzene levels in 2007 to generate early credits
    • Caused the average benzene levels in conventional gasoline to decrease from about 1.15 volume percent to about 0.60 volume percent
    • The MSAT2 program caused conventional gasoline benzene levels to decrease to nearly the same benzene levels as reformulated gasoline
  • Renewable Fuel Standard (RFS) program
    • Began in 2006 and has resulted in the blending of much more ethanol into the gasoline pool, such that nearly all gasoline contained 10 volume percent ethanol by 2013
    • Due to the use of ethanol, the weight percent of oxygen has increased substantially in gasoline
    • Ethanol, with its high octane value has also allowed a significant reduction in the aromatic content of the gasoline
    • Other direct effects of blending in ethanol are described under the tab “How the data was adjusted and reported”
• Initially, the RFG program resulted in RFG having very different fuel properties compared to conventional gasoline - the implementation of the subsequent fuel programs has caused conventional gasoline and RFG to be very similar in fuel properties
  • The primary difference now between RFG and conventional gasoline is that, during the summer, RFG has much lower RVP than most of conventional gasoline
• The marketplace has also caused changes in fuel quality
  • As is evident by API gravity and E300 values, gasoline has been becoming “lighter” over time as heavier hydrocarbons previously blended into the gasoline pool by refineries have been shifted into the distillate pool to meet the rising demand for these products
  • The increased blending of ethanol also played a role in this change in API gravity and E300

How the Data was Adjusted

• Need for adjusting the data
  • RFG is reported as blended with ethanol, while conventional gasoline (CG) typically is not
  • Adjustments were necessary to the CG pool so that the gasoline fuel properties represent CG sold at retail
  • Adjustments were made to the CG data from year 2006 and later, not to the CG data 2005 and earlier, because there was little ethanol use in CG prior to 2006
    • The increase in ethanol use in 2006 is associated with the sudden end to the use of methyl tertiary butyl ether (MTBE)
  • What gasoline was adjusted
    • Batches reported to EPA as conventional blendstock for oxygenate blending (CBOB) were adjusted assuming that denatured ethanol was blended into CBOB at 10 volume percent
    • The volume of ethanol blended into the rest of the CG pool is estimated
      • The volume of ethanol blended into the RFG and CBOB pools was totaled and compared to the total volume of ethanol blended into the gasoline pool as reported by the Energy Information Administration (EIA) in each year
      • The difference was assumed to be blended into the CG pool
    • The volume of gasoline reported was increased to include the volume of ethanol blended downstream of the refinery
• How the adjustments were made to account for the addition of ethanol
  • Sulfur, olefins, aromatics and benzene were adjusted assuming that they were reduced by simple dilution.  These properties were lowered by multiplying the gasoline parameter by 0.902 – adjusting for only the ethanol portion of the denatured ethanol
    • Denaturant was assumed to be gasoline or gasoline like
  • Other adjustments accounted for the non-ideal blending behavior of ethanol
    • RVP was adjusted using the equation:  RVP(E10) = RVP(E0) + 6.2371 * RVP(E0)^-0.794
    • E200 was adjusted using the equation:  E200(E10) = 0.6988 * E200(E0) + 23.182
    • E300 was adjusted using the equation:  E300(E10) = 0.8681* E300(E0) + 12.874
    • API Gravity was adjusted using the equation:  API gravity(E10) = 0.8251* API gravity(E0) +9.5272
    • The adjustments for RVP, E200, E300 and API gravity were made based on a statistical analysis of the data derived by the American Petroleum Institute’s (API) ethanol blending study ²

How the Data was Reported

• For the results from the batch data presented here, each property value is a volume-weighted average based on the volumes reported for each batch
• Weight percent oxygen reporting changed in 2006
  • Prior to 2006, the weight-percent oxygen average values were calculated only when a weight-percent oxygen value greater than zero was reported.  All weight-percent oxygenate values (Ethanol, MTBE, TAME, etc.) were calculated as a group based on the presence of a reported value for oxygen that was greater than zero
  • Starting in 2006, weight-percent oxygen is reported as an average over the entire gasoline pool for each gasoline type/season subcategory
• Reporting of T50 and T90 data
  • T50 and T90 are not reported starting in 2006 due to incomplete reporting and lack of knowledge of which batches to adjust for ethanol content.  E200 and E300 data is complete
• Errors in the data
  • As in any dataset of this magnitude, there are errors in reporting, even after a significant quality control effort by EPA to find and fix perceived errors or omissions
  • If a property value was found to be questionable when assembling this data, we omitted the data from the analysis to avoid potentially biasing the data set
  • The volume of omitted data represented  a very negligible fraction of the entire data set

References:
1. Determination of the Potential Property Ranges of Mid-Level Ethanol Blends, American Petroleum Institute; April 23, 2010.