EDITOR'S NOTE: This is 1 of 12 papers prepared by participants attending the workshop “Risk Assessment in European River Basins–State of the Art and Future Challenges” held in Liepzig, Germany on 12–14 November 2007. The meeting was organized within the framework of the European Commission's Coordination Action RISKBASE program. The objective of RISKBASE is to review and synthesize the outcome of European Commission FP4–FP6 projects, and other major initiatives, related to integrated risk assessment–based management of the water/sediment/soil environment at the river basin scale.

Stream risk assessment and restoration requires understanding of the controlling factors and the scale at which they act. The role of hydromorphology, along with physicochemistry, was for a long time neglected, and scale issues were barely tackled. In this study, both the role of hydromorphology and the relevance of scale are studied. For this purpose, the macroinvertebrate community of the stream is used as the scale of the target biota. Next, the following research question is dealt with: At which scale, and to what extent, do hydrology and morphology along with physicochemistry explain stream macroinvertebrate distribution? Three data sets were used: The European AQEM study, the Dutch streams study, and an extensive habitat-preference study. Ordination was used to relate the macroinvertebrate species composition to the (hydromorphological) environment for both the European study and the Dutch stream study data. To explore the strength of one or more variables in explaining the macroinvertebrate distribution over the sampling sites, the fraction of the sum of canonical eigenvalues was used as a measure. To determine the preference for a specific habitat type of each macroinvertebrate species in the habitat preference study, the index of representation was calculated. The European study showed that streams within a more limited geographic area tend to carry macroinvertebrates whose distribution is better explained by stream stretch and in-stream variables. However, even within stream type catchment and stream valley, variables almost equally add to the explanation of the macroinvertebrates distribution. The explanatory power of hydrological and physicochemical variables increased toward smaller scales, and morphological variables showed an equal explanatory power over the different scales. In the Dutch streams study, stream level was much better explained in comparison to the habitat level. Geographical, morphological, and physicochemical variables were strong explanatory variables. For both habitat and stream, the stream stretch variables contributed most to the explanation of macroinvertebrate distribution, whereas microhabitat variables were less explanatory. The habitat preference study supported the observation that habitat provided less explanation than stream stretch. Only 15% of the macroinvertebrate species showed a clear habitat preference; none showed an obligatory one. In conclusion, stream macroinvertebrates distribution is best explained by local stream–stretch variables, provided those variables are contained within a catchment and stream valley context. Differences in vulnerability and biotic capacity between macroinvertebrate species determine the assemblage present. Applying this knowledge in water management means that any risk assessment and restoration effort needs a hydromorphological context.