Developing Biological Indicators: Lessons Learned from Mid-Atlantic Streams - Brochure

As part of the U.S. Environmental Protection Agency’s (EPA) Environmental Monitoring and Assessment Program (EMAP), a survey of water chemistry, land use, riparian condition, and channel morphology was conducted to understand how human influence alters fish, invertebrate and periphyton assemblages. During 1993-1996, 296 sites were sampled for fish, 583 for invertebrates and 317 for periphyton. A primary goal of the Mid-Atlantic Integrated Assessment (MAIA) study was to define biological indicators for each assemblage that could be used to assess stream condition at the regional level. During the course of the project, researchers working independently derived different approaches to data analysis and reported different results regarding the relationship between human influence and biological change. To build consensus among the scientists involved, EPA sponsored a series of workshops to create a consistent approach for testing and selecting biological indicators for fish, invertebrates and periphyton. This document presents some of the issues from those workshops that were most challenging to resolve. The probabilistic sampling approach was the most efficient method for obtaining an unbiased estimate of regional condition and is a cornerstone of EMAP. The coarse level of resolution has made this design somewhat controversial for application at the state level, but alternative sampling designs fail to yield data that can be applied across the region. Although much attention was given to database design, data management issues often drove the agenda. Major issues not anticipated during the design phase arose in the course of analyzing the data. Final results were delayed, for example, by lengthy discussions of how to count invertebrate taxa with incomplete phylogeny and how to combine periphyton data from soft algae and diatom counts that were derived from different laboratory methods. One of the greatest challenges for the MAIA project was selecting the best measure of human disturbance from among the hundreds of potential variables. Through discussion and consensus, a set of variables was selected that summarized human influence at multiple spatial scales and included measures related to specific types of disturbance and measures that integrated across the landscape. During the process of testing biological attributes for their association with disturbance, concerns repeatedly arose regarding whether the results would be robust in new contexts. As a consequence, safeguards were introduced at many steps of the analysis to avoid circular reasoning. Similar statistical tests and criteria were used to select from among the list of candidate metrics for fish, invertebrates and periphyton. All three assemblages showed strong and consistent associations with general measures of disturbance, but showed different sensitivities to individual stressors. Diatoms were more sensitive to water chemistry and invertebrates were more sensitive to riparian condition. Multimetric indexes for each assemblage showed a similar ability to detect both differences in sites and change through time (trend). The indexes for each assemblage represent statistically reliable and biological meaningful monitoring tools for assessing and reporting the biological condition of wadeable streams in the Mid-Atlantic.