The study of dendritic ecological networks (network theory) has provided new insights about the spatial structure of stream ecosystems and the mechanisms of colonization and extinction that regulate stream communities.  However, applications of these concepts for conservation and management are in their nascent stages.  In this synthesis paper, we explore the links between network theory and the practice of stream conservation and management.  First, we review the origins of network theory in streams.  Second, we identify heuristic principles for conservation and management.  Third, we evaluate how such principles could benefit stream conservation and management efforts.

Results/Conclusions

A review of the literature demonstrated that stream networks can influence local populations and communities by regulating dispersal among connected streams.  From this premise, we identified 3 heuristic principles for stream conservation and management: (1) Conservation success should improve as the grain size of management units approaches the grain size of dispersal within stream networks; (2) A mismatch between the spatial structure of sample locations and dispersal networks may bias interpretations of local environmental quality in streams; and (3) Connectivity in stream networks can affect the movement of tissue-based contaminants via organismal dispersal.  We presented 3 case studies to illustrate how these principles may improve stream conservation and management.  First, we evaluated how stream network connectivity affects the distribution of native brook trout (Salvelinus fontinalis) and can inform conservation planning units (principle 1).  Second, we evaluated data from a stream-fish bioassessment program in West Virginia.  We demonstrated that stream sites connected to large rivers obscured fish assemblage relations to local environmental quality, in contrast to stream sites which lacked such riverine connectivity (principle 2).  Third, we evaluated the proposed regulation of selenium from fish tissue concentrations, demonstrating how network-scale dispersal could affect type I and II error probabilities (principle 3).