The Taking Stock dataset is developed by looking at the information that is comparable among the national PRTR programs of North America. While Canada, Mexico and the United States have similar pollutant release and transfer registers, there are significant differences among them. Some of the most important differences include the number of chemicals listed and the types of industrial sectors covered.

Because the matched 2004 dataset for the three countries is relatively small (including 56 chemicals and nine industry sectors), Taking Stock also provides analyses based on the larger 2004 matched dataset for Canada and the US. In addition, the report presents trends in pollutant releases and transfers from Canada and the US since 1998 (users can also search Taking Stock Online for Canada-US trends dating back to 1995).

Taking Stock uses the following categories for presenting information on pollutant releases and transfers:

• "on-site releases" describes releases that occur at the facility-i.e., chemicals put into the air or water, injected into underground wells, or put in landfills "inside the fenceline."
• "off-site releases" describes chemicals sent off-site to other locations for disposal, as well as metals sent to treatment, sewage, and energy recovery.
• "total releases on- and off-site" or simply "total releases" is the sum of on- and off-site releases.
• "total releases (adjusted)" is the sum of on- and off-site releases minus those off-site releases that are reported as on-site releases by another facility.
• "transfers to recycling" describes chemicals sent off-site for recycling
• "other transfers for further management" describes chemicals (other than metals) sent for treatment and energy recovery and to sewage plants.
• "transfers for further management" encompasses: (1) chemicals sent for recycling and (2) other transfers for further management, i.e., chemicals (other than metals) sent for treatment and energy recovery and to sewage plants.
• "total reported amounts" describes the sum of all of the above categories: on- and off-site releases, recycling, and other transfers for further management. All releases as reported are included. While not perfect, this is the closest estimate available from the matched North American PRTR data of the total amount of chemicals arising from a facility's activities that need to be managed.

The Taking Stock datasets include an "adjustment analysis" that adjusts the total release numbers for "double-counting." Double-counting can occur when a facility sends chemicals for disposal or metals to treatment, sewage, or energy recovery to another facility that also reports on its releases and transfers. This creates the possibility of the same chemicals being reported twice: once as an off-site release by the first facility, and again as an on-site release by the second facility. It is not necessary to adjust releases when considering total reported amounts, which are an estimate of total amounts generated that require handling or management. Double-counting became more likely with the addition of hazardous waste management/solvent recovery facilities to TRI in 1998.

The categorization used in this report includes, as part of off-site releases, those metals that are sent off-site to disposal, to treatment, for energy recovery, or to sewage. This categorization is needed to make the data comparable. TRI has a special method for classifying transfers of metals in which transfers of metals to sewage, treatment, or energy recovery are considered releases because metals are not destroyed by treatment or burned in energy recovery.

While it may seem confusing at first to those who are accustomed to seeing "releases" refer to on-site activities and "transfers" refer to off-site activities, this categorization has several benefits. It aggregates similar activities; for example, all chemicals that are landfilled are called releases, regardless of where the landfill is located. It preserves the sense of location of releases, either on or off the site of the facility. The approach also recognizes the physical nature of metals and acknowledges that metals sent to disposal, sewage, treatment, and energy recovery are not likely to be destroyed or burned and so may eventually enter the environment.

Matched Datasets

The data are "matched" for a particular span of years, that is, they are based on chemicals and industrial sectors that are common to TRI, NPRI, and RETC for the year(s) in question.

The chemicals and industry sectors included depend on the year(s) chosen for analysis:

• 2003 (and subsequent years) datasets include all matched chemicals and industries,
• 2002 (used to present year-to-year changes since 2002) dataset includes all matched chemicals except carbonyl sulfide, which was added to NPRI for 2002, and all matched industries,
• 1998 (and subsequent years) dataset (used to present year-to-year trends since 1998) does not include chemicals added since 1998, chemicals whose reporting threshold has changed since 1998, or the Petroleum Bulk Terminals industry added to NPRI in 2002, and
• 1995 (and subsequent years) dataset (used to present year-to-year trends since 1995) does not include chemicals added since 1995, chemicals whose reporting threshold has changed since 1995 and industry sectors added to TRI or NPRI reporting requirements since 1995.

Note that some persistent bioaccumulative toxic chemicals (PBTs) are not included in the "matched" data sets. Dioxins and furans, hexachlorobenzene, and polycyclic aromatic compounds (PACs/PAHs) are not included because which facilities and which amounts are required to report are different in the different countries' datasets.