Scaling of colonization processes in streams: Parallels and lessons from marine hard substrata

Understanding colonization processes specifically requires a comprehensive understanding of the frequency and scales over which dispersal occurs, which in turn determines the degree of linkage within and between populations and possibly results in metapopulation structure. What is our level of such knowledge for stream systems! We examine colonization processes in streams using a framework produced by analogous literature on sessile, marine species, which are comparatively well-known and may provide insights for stream biota. For simplicity, we separate dispersal into two, somewhat artificial scales: microscale (between habitat patches - cm to m) and mesoscale (between groups of patches - tens to thousands of metres), and consider dispersal links between them. Traditional views on dispersal developed similarly in both systems, with a strong initial focus on mesoscale dispersal followed by an awareness that some species do not disperse usually over these scales and that there are a wide range of dispersal profiles. In both habitats, there are few data on actual dispersal distances, and longevities are sometimes used to infer them instead. Organisms can be transported by water currents (marine larvae, stream larvae) or wind (adult insects), the directions and strengths of which are relatively predictable at mesoscales. However, behavioural choices of organisms during dispersal can change their potential dispersal distances and directions markedly. Additionally, predictability of transport processes at microscales is very poor. As a consequence, simple, lone biological measures (like longevity) or simple, lone physical measures (like discharge) are useless for predicting dispersal frequencies and scales in most cases. Mortality during dispersal is also extremely important but there are few data; this represents a major information gap in both sets of literature. Finally, if organisms end dispersal by searching for and being able to respond to specific cues, then we may be able to predict colonization by looking at the distribution of such cues; different cues are likely though to vary greatly in space and time. There is potential for dispersal 'structure' to develop in rivers of different hydrology, and for sets of correlated life-history characters to result in dispersal at particular scales, but there are very few stream studies that bear on these issues unambiguously. Progress in understanding the scales of dispersal, in both habitats, will require a lot more studies designed formally to test clear hypotheses about the scales of dispersal, rather than continuing a status quo of generating essentially anecdotal information.