Emissions Data. Unlike air quality data, emissions estimates are available for longer periods of time, going back to 1900 for SO2, NOx, and VOC; 1940 for PM-10 and CO; and 1970 for lead. These estimates, however—put together with great care by EPA—are not based on direct measurements as are air quality data. They are calculated by taking into consideration a number of factors that vary or evolve from year to year. Those factors include quantity and quality of fuels consumed by each end use category (i.e., residential, commercial, industrial, and utility), age structure of the vehicles in each category of motor vehicles (e.g., cars, light and heavy trucks), production from the various industrial, manufacturing, and other pollution‐ generating processes, control technologies and practices employed by the various emissions sources, and growth rates for the various economic and emissions sectors. To compile a long-term emissions inventory requires not only engineering knowledge of the processes and equipment contributing to emissions but also historical knowledge of the evolution of those processes and their controls and of operation and maintenance practices.
For any specific pollutant, different methodologies are used to derive estimates for different time periods. The methodologies used to estimate emissions for 1900-39 are different from those used for 1940-84 and for 1985 to the present. Every year EPA updates the latest set of emissions, and sometimes fine-tunes some of the earlier numbers as well. Although these estimates are currently the best available, because of the many assumptions that have to be made in developing emissions for each source category, uncertainties inevitably creep in. Accordingly, quantitative changes in emissions from one period to another should be viewed cautiously. Nevertheless, those estimates provide a good indication of the general trends in emissions.