California Gasoline Predictive Models, and CARBOB Model Development

This page last reviewed February 8, 2018

CaRFG3 Predictive Model Development Process
The California Air Resources Board is currently in the process of developing a new CaRFG3 Predictive Model. This page contains information about fuel/vehicle emission studies, vehicle emissions inventory, permeation, MIR factors, and other related information to the development of the CaRFG3 Predictive Model.

2007 Draft Predictive Model

ARB Staff has released a preliminary draft of the Predictive Model.

  • 2007 Draft Predictive Model v3.0 (March 26, 2007) (XLS-311k) 
  • 2006 Draft Predictive Model v2.0 (December 14, 2006) (XLS-305k)
  • 2006 Draft Predictive Model v1.0 (September 14, 2006) (XLS-306k)

Predictive Model Database

Five recently released fuel/vehicle emission studies have been added to the current California Phase 3 Reformulated Gasoline Predictive Model database:

  1. ExxonMobil Research and Engineering Company: LEV/ULEV Gasoline Oxygenate Study (1999).
  2. Toyota Motor Co.: Effects of Ethanol on Emissions of Gasoline LDVs (2000).
  3. Alliance of Automobile Manufacturers (AAM) and Association of International Automobile Manufacturers (AIAM): Sulfur Oxygen Vehicle Emissions Test Program (2001).
  4. CRC Project No. E-60: The Effect of Fuel Sulfur on NH3 and Other Emissions from 2000-2001 Model Year Vehicles (2003).
  5. CRC Project No. E-67: Effects of Ethanol and Volatility Parameters on Exhaust Emissions (2006).
  • Updated California Reformulated Gasoline Predictive Model Database (May 4, 2006).

  • The ARB staff would like to receive input regarding the viability of adding vehicle emissions test data from the Mexican Petroleum Institute into the CaRFG3 Predictive Model database. These data will be used to update the Predictive Model. (XLS-25k)

Vehicle Emissions Inventory

ARB staff is updating the methodology used to develop temperature and relative humidity profiles. These new profiles will be developed to represent meteorological conditions for the California 8-hour ozone standard. These profiles will not be part of vehicle emissions inventory model (EMFAC), but will be used to estimate vehicle emissions by technology group to update the Predictive model.


The need for a study of the permeation effects of ethanol became apparent when in late 1999 California banned the use of MTBE in gasoline. With this ban, which became effective starting in year 2004, ethanol became the only oxygenate approved for use in California gasoline. California must quantify the permeation effects of ethanol because California's statutes require that any increase in fuel emissions be off-set with a similar reduction from other sources. Therefore, the Board directed staff to report on the potential for hydrocarbon emissions increases associated with increased motor vehicle fuel system permeability resulting from the blending of ethanol into California gasoline.

From late 2002 to June 2004, the study was conducted by Coordinating Research Council (CRC) sponsored by the California Air Resources Board. The results of this study would assist the stockholders and the State of California to evaluate the emission benefits as well as the CaRFG 3 Gasoline Program. In January 2005, a second phase of the permeation study was initiated to test higher ethanol content fuels.

  • Final Report: Fuel Permeation from Automotive Systems: E0, E6, E10, E20 and E85 - CRC Project No. E-65-3 (PDF-657k)
  • Interim Report: Fuel Permeation From Automotive Systems: E0, E6, E10 and E85 - CRC Project No. E-65-3 (PDF-462k)
  • Final Report: Fuel Permeation from Automotive Systems - CRC Project No. E-65 (PDF-874k)

Copies of both reports and the supporting data are also available from the Coordinating Research Council.

MIR Factors

ARB staff is releasing updates to specific reactivity values complied using the updated MIR values from the list adopted by ARB in 2003. The draft 2006 MIR list was presented for consideration by the reactivity subgroup and there was consensus on its use in the 2006 draft Predictive Model. This new 2006 draft MIR list is presented below. The 1998 and 2003 MIR lists are also provided below for reference.

Using the 2006 draft MIR list, staff has calculated average reactivity values for evaporative and exhaust emissions for speciated data from in-use testing of California compliant fuels at the ARB El Monte laboratory. The data sets include:

  • Speciated evaporative emissions (diurnal and hot soak) data sets from in-use testing at the ARB El Monte laboratory (updated December 13, 2006) (ZIP-5.3MB) 
  • Speciated exhaust emissions from in-use testing at the ARB El Monte laboratory (updated December 13, 2006) (ZIP-1.4MB) 
  • Permeation data sets from the CRC E-65 study are available from the permeation section of this webpage

Average Reactivities



Diurnal/Resting Loss

Hot Soak

Running Loss


Average MIR (Draft)






In March 2006, ARB staff released a draft modeling analysis for assessing the potential air quality impacts of increased permeation VOC emissions relative to reduced CO emissions for California vehicles using ethanol gasoline. This document describes the problem to be addressed, previous related work, and the air quality modeling results. It is not intended to estimate the air quality impact of ethanol vs. MTBE gasoline.
  • The Ozone Impact of Permeation VOC Relative to Carbon Monoxide (PDF-70k)

Public Comments

  • Renewable Energy Action Project (REAP) Comments on the Exxon Dataset Proposed for Inclusion in the 2006 CARB Predictive Model Update (PDF-49k)
  • WSPA Comments on the Mexican Dataset Proposed for Inclusion in the 2006 CARB Predictive Model Update (PDF-25k)
Current CaRFG3 Predictive Model
Procedures for Using the California Model for California Reformulated Gasoline Blendstocks for Oxygenate Blending (CARBOB) (PDF - 68k) - Adopted April 25, 2001
Motor Vehicle Emission Testing Option
- California Procedures for Evaluating Alternative Specifications for Gasoline Using Vehicle Emissions Testing  (PDF - 84k) - Amended April 25, 2001