Publication Detail
A Cursory Analysis of EMFAC7G: Reconciling Observed and Predicted Emissions
|
UCD-ITS-RR-94-08 Research Report |
Suggested Citation:
Washington, Simon P. (1994) A Cursory Analysis of EMFAC7G: Reconciling Observed and Predicted Emissions. Institute of Transportation Studies, University of California, Davis, Research Report UCD-ITS-RR-94-08
California Air Resources Board staff have proposed modifications to the current EMFAC7F mobile source emissions model. The proposed model EMFAC7G, while still in the preliminary stages of development, suggests increases in the prediction of critical NOx, CO, and HC emissions inventories by about 8%, 80%, and 45% respectively. These increases result from new approaches to modeling cold and hot starts, trip length and number of trips by vehicle age, 'off-cycle' emissions, high emitting vehicles, the dispersion of emissions from cold and hot starts, and other parameters.
Published results from 'top-down' emissions research, such as remote sensing, tunnel, and ambient air quality monitoring studies, indicate that the current emission models may under-predict motor vehicle emissions. The proposed modeling changes will increase model-predicted emissions and will help to reconcile the noted differences between model-predicted and field-measured pollutant concentrations. The proposed changes to EMFAC reflect the CARB staff's commitment to both 'bottom up' and 'top down' assessments of emissions inventory estimation accuracy, and represent a significant effort and improvement over the current version. However, we still must question whether the bottom-up changes to EMFAC will finally reconcile the differences between 'top down' emissions approaches.
This cursory analysis suggests that the currently proposed (and released) EMFAC7G improvements are still likely to underestimate NMOG emissions from the motor vehicle fleet by roughly 25% to 75%, while CO emissions will be either over-estimated by about 40% or under-estimated by about 30%. In effect, EMFAC7G will likely still under-predict NMOG emissions, but may do a reasonable job predicting CO emissions.
Among the reasons for this remaining under-estimation are the poor link between vehicle activity models and air quality models, EMFAC methodological errors, the presence of inadequately modeled high-emitting vehicles and high-emitting vehicle activity, and uncertainty involved with remote sensing, tunnel, and ambient air quality monitoring studies. Given the nature of these uncertainties (most contribute to under-prediction of emissions), and the conservative approach used to estimate the discrepancies above, the real-world difference between observed and predicted emissions is likely to be even greater than is presented in this paper.
The results presented here are preliminary, and are based upon analysis of the Preview EMFAC7G model, which may turn out to be significantly different than the final EMFAC7G in terms of methodological changes incorporated. The results, therefore, are contingent upon the proposed EMFAC7G model being adopted without significant change, and upon the accuracy of the CARB staff spreadsheet analyses of EMFAC7G model impact on current vehicle fleet emission estimates.
Published results from 'top-down' emissions research, such as remote sensing, tunnel, and ambient air quality monitoring studies, indicate that the current emission models may under-predict motor vehicle emissions. The proposed modeling changes will increase model-predicted emissions and will help to reconcile the noted differences between model-predicted and field-measured pollutant concentrations. The proposed changes to EMFAC reflect the CARB staff's commitment to both 'bottom up' and 'top down' assessments of emissions inventory estimation accuracy, and represent a significant effort and improvement over the current version. However, we still must question whether the bottom-up changes to EMFAC will finally reconcile the differences between 'top down' emissions approaches.
This cursory analysis suggests that the currently proposed (and released) EMFAC7G improvements are still likely to underestimate NMOG emissions from the motor vehicle fleet by roughly 25% to 75%, while CO emissions will be either over-estimated by about 40% or under-estimated by about 30%. In effect, EMFAC7G will likely still under-predict NMOG emissions, but may do a reasonable job predicting CO emissions.
Among the reasons for this remaining under-estimation are the poor link between vehicle activity models and air quality models, EMFAC methodological errors, the presence of inadequately modeled high-emitting vehicles and high-emitting vehicle activity, and uncertainty involved with remote sensing, tunnel, and ambient air quality monitoring studies. Given the nature of these uncertainties (most contribute to under-prediction of emissions), and the conservative approach used to estimate the discrepancies above, the real-world difference between observed and predicted emissions is likely to be even greater than is presented in this paper.
The results presented here are preliminary, and are based upon analysis of the Preview EMFAC7G model, which may turn out to be significantly different than the final EMFAC7G in terms of methodological changes incorporated. The results, therefore, are contingent upon the proposed EMFAC7G model being adopted without significant change, and upon the accuracy of the CARB staff spreadsheet analyses of EMFAC7G model impact on current vehicle fleet emission estimates.
Prepared for the Union of Concerned Scientists.