Evaluation of the use of landscape classifications for the prediction of freshwater biota: synthesis and recommendations

This paper summarizes and synthesizes the collective results that emerged from the series of papers published in this issue of J-NABS, and places these results in the context of previously published literature describing variation in aquatic biota at landscape spatial scales. Classifications based on landscape spatial scales are used or are being evaluated for use in several countries for aquatic bioassessment programs. Evaluation of the strength of classification of different approaches should provide insight for refinement of existing bioassessment programs and expedite the development of new programs. The papers in this series specifically addressed the degree to which descriptions and classification of landscape features allow us to account for, and thus predict, variation in the composition of biota among individual sites. In general, we found that although landscape classifications accounted for more biotic variation than would be expected by chance, the amount of variation related to landscape features was not large. Thus, large-scale regionalizations, if used alone to specify expected biotic conditions, will likely have limited use in aquatic bioassesments, where it is critical to specify expected conditions as accurately and precisely as possible. Landscape classifications can play an important additional role, however, by providing an initial stratification of site locations to ensure that different landscape features are adequately represented in a sampling program. In general, we believe a tiered classification based on both reach-level and larger-scale landscape features is needed to accurately predict the composition of freshwater fauna. One potential approach entails use of landscape classifications as a means of refining or augmenting classifications based on local habitat features, which appear to account for substantially more biotic variation than larger-scale environmental features. These results have significant implications for how assessment and monitoring programs at local, state/province, and national levels should be designed.