One of the greatest challenges in concentration trend analysis—carried out with the specific purpose of determining if pollution control efforts have been successful—is accounting for meteorological influences, which can cause a great deal of variability in concentrations and are often large enough to obscure changes in anthropogenic emissions. A variety of statistical approaches (such as regression-based modeling and extreme value approaches) can be used to filter out meteorological influences from a trend; however, the results often vary, depending on the method and the specific concentration metric used.4 For all these reasons, most pollutant concentration trends cited by EPA and other organizations (including the trends noted in this report) should be used with caution (Box 6-4).

Data Availability

EPA issues a yearly National Air Quality and Emission Trends Report (EPA 2000e), which provides a general summary of the nation’s air quality. The summary includes the results of an analysis of air pollutant trends based on data obtained from the monitoring networks described above. EPA also issues a yearly Toxics Release Inventory (TRI) (EPA 2003m), which contains information about the types and amounts of approximately 650 chemicals that are released from certain industries and federal facilities. The reported information includes annual estimates of emissions as well as releases to water and land and quantities of chemicals sent to other facilities for waste management. However, the trends reports and TRI reports are intended for a general audience and do not contain enough information to facilitate more extensive analyses by the scientific and technical community. For such purposes, EPA has developed the Aerometric Information Retrieval System (AIRS), a framework that provides access to monitoring data from different sources. Although EPA has made efforts to update AIRS, problems remain with the system that limit its utility. All monitoring data are not stored in AIRS, and data stored in AIRS have not been “standardized”; for example, the data have not been converted to uniform measuring units or to universal standard time. Most important, the data are not available in real time (as are, for instance, data from the National Weather Service). Although EPA maintains the AIRNow web site where individuals can access information on present and past air quality conditions at localities throughout the United States, the information is qualitative and thus not suitable for quantitative analyses. Many states have begun to make their air quality monitoring data available in real time over the


A trend analysis for CO in Fairbanks and an approach by Reddy (2000) for estimating the probabilities of future CO exceedances in Denver are described in earlier NRC reports (NRC 2002b, 2003b).

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement