This work addresses human stressors and their impacts on fish assemblages at pan-European scale by analysing single and multiple stressors and their interactions. Based on an extensive dataset with 3105 fish sampling sites, patterns of stressors, their combination and nature of interactions, i.e. synergistic, antagonistic and additive were investigated. Geographical distribution and patterns of seven human stressor variables, belonging to four stressor groups (hydrological-, morphological-, water quality- and connectivity stressors), were examined, considering both single and multiple stressor combinations. To quantify the stressors' ecological impact, a set of 22 fish metrics for various fish assemblage types (headwaters, medium gradient rivers, lowland rivers and Mediterranean streams) was analysed by comparing their observed and expected response to different stressors, both acting individually and in combination. Overall, investigated fish sampling sites are affected by 15 different stressor combinations, including 4 stressors acting individually and 11 combinations of two or more stressors; up to 4 stressor groups per fish sampling site occur. Stressor-response analysis shows divergent results among different stressor categories, even though a general trend of decreasing ecological integrity with increasing stressor quantity can be observed. Fish metrics based on density of species 'intolerant to water quality degradation' and 'intolerant to oxygen depletion" responded best to single and multiple stressors and their interactions. Interactions of stressors were additive (40%), synergistic (30%) or antagonistic (30%), emphasizing the importance to consider interactions in multi-stressor analyses. While antagonistic effects are only observed in headwaters and medium-gradient rivers, synergistic effects increase from headwaters over medium gradient rivers and Mediterranean streams to large lowland rivers. The knowledge gained in this work provides a basis for advanced investigations in European river basins and helps prioritizing further restoration and management actions.
Woody riparian vegetation (WRV) benefits benthic macroinvertebrates in running waters. However, while some functions are provided by WRV irrespective of surrounding and catchment land use, others are context‐specific. In recent large‐scale studies, effects of WRV on macroinvertebrates were therefore small compared to catchment land use, raising the question about the relevance of WRV for restoration. Model‐based recursive partitioning was used to identify context‐dependent effects of WRV on the macroinvertebrates' ecological status in small (catchment area 10–100 km2) lowland (n = 361) and mountain (n = 748) streams. WRV cover was quantified from orthophotos along the near (500 m) and far (5000 m) upstream river network and used to predict the site's ecological status. Agricultural, urban and woodland cover at the local and catchment scales along with hydromorphology were considered as partitioning variables. In rural agricultural landscapes, the effect of WRV on the ecological status was large, indicating that establishing near‐upstream WRV can improve the ecological status by as much as two of the five classes according to the EU Water Framework Directive. Even in urban landscapes, effects of far‐upstream WRV were large if catchments had a moderate share of agricultural land use in addition. The beneficial effects of WRV were only limited in purely urban catchments or in a multiple stressor context. Synthesis and applications. While woody riparian vegetation (WRV) can even improve the ecological status in urban settings, it is especially relevant for river management in rural agricultural catchments, where developing WRV potentially are effective measures to achieve good ecological status.
Woody riparian buffers (hereafter, ‘woody buffers’) are frequently considered as important to mitigate the effects of stressors on streams and rivers. While several individual studies addressing nutrients, pesticides, water temperature and different biotic components support this conjecture, no study has addressed the effects of woody buffers on riverine biota at country‐wide scales. We used a comprehensive dataset from sampling sites on 1082 catchments in France, comprising samples of benthic invertebrates, along with data on river size, physico‐chemistry, hydromorphology, riparian and catchment land use and woody buffers at sampling sites and upstream. Using partial least square modelling, we delineated the effects of the different environmental variables on two benthic invertebrate metrics, separately for siliceous and calcareous rivers. Overall, models explained 49% (calcareous) and 39% (siliceous) of the variation in benthic invertebrate metrics. Direct effects of woody buffers on benthic invertebrate metrics were marginal, while physico‐chemical conditions and catchment land use explained most of the deviance. Direct and indirect effects of woody buffer together covered up to 6% (upstream scale) plus 2% (local scale) of the explained variability. Synthesis and applications. In this national‐scale study, on 1082 catchments, we investigated the potential of woody buffers to mitigate the effects of catchment‐scale and local‐scale stressors on macroinvertebrate biodiversity. Our results underline that the establishment of woody buffers is not necessarily a sufficient measure to solve the problem of deteriorating riverine macroinvertebrate communities, especially in catchments prone to intense land use. Nevertheless, two main outcomes included that local woody patches are not sufficient and that woody buffers should be established along longer river stretches. Also, accompanying catchment‐scale measures should be promoted to reduce the effects of intense land use and pollution to a level that enables woody buffers to be effective as well.