Guidelines for Including Data

Once an indicator was chosen and relevant sources of data were identified, the first screen for inclusion was scientific credibility. Again, data were not used simply because they were the “best available” but, based on the professional judgment of the members of each work group, they had to meet the highest standards of the appropriate discipline. The judgments of the work groups were then extensively peerreviewed.

The second criterion for including a data source was that it provide information on a substantial majority of the resource or issue in question. The practical result is that we relied on data sources that covered a majority of states or a significant fraction of coastline. For some indicators, complete coverage is available (such as is provided by remote sensing data). For others, regionally or nationally representative samples are used (such as is provided by monitoring programs employing statistical sampling techniques).

The first draft/prototype of this report, released in 1999, included many examples of data for small areas of the country, as illustrations of the types of results we had hoped to include. Feedback from readers led us to conclude that while such examples are interesting, they obscure the fact that data are not available for a significant fraction of the desired indicators.

Third, to be included in this report, data sources must be from ongoing programs—that is, there must be a reasonable chance that the measurements will be repeated at regular intervals in the future. Although all monitoring and reporting programs are subject to changes in funding and priorities, established programs are clearly different from one-time studies conducted by individual researchers or groups. One-time efforts are extremely valuable because they often break new ground scientifically, and they may provide baselines against which data collected later may be compared, but they do not necessarily form a solid foundation for periodic reporting.

Ideally, data sources used in this report have time trends, but the lack of trends was not a criterion for eliminating data. Where possible, we have attempted to use data from 1950 to the present, with longer historical perspectives included as needed to provide an ecological context for current reporting. These longer-term perspectives include reporting on conditions before European settlement, in the early 20th century, or for other relevant time periods. Many data sources, particularly those based on remote sensing, cover shorter time periods but will illuminate longer-term trends as time goes on and measurements are repeated.

Note as well that there were a number of cases where we were not able to determine which of several possible indicators for an important ecosystem attribute was best, and thus we could not judge whether adequate data are available. We have highlighted these cases, and we hope to work closely with the relevant scientific communities to narrow the range of possible indicators.