Summary

An increasing number of river sections have been restored in the past few decades but only a small number of these projects have been monitored. The few monitoring studies mainly investigated single organism groups, reported contrasting results, and rarely did investigate the influence of catchment, river or project characteristics. In this study, we compiled a harmonized dataset on the effects of hydromorphological river restoration measures on biota based on a standardized monitoring design to minimize scatter due to methodological differences. A broad range of response variables was recorded to draw conclusions on the effect of restoration on biota in general, including habitat composition in the river and its floodplain, three aquatic and two floodplain-inhabiting organism groups, as well as food web composition and aquatic land interactions as reflected by stable isotopes. Additional data on factors potentially constraining or enhancing the effect of restoration were compiled to identify conditions which favour restoration success. The main focus was dedicated to investigate the effect of restoration extent (as indicated by restored section length and restoration intensity).

Ten pairs of one large and a similar but small restoration project were investigated to address the role of restoration extent for river restoration effects. The restoration effect was quantified by comparing each of the 20 restored river sections to a nearby non-restored, i.e. still degraded section. The large restoration projects were representing good-practice examples in different European regions either targeting medium-sized lowland rivers or medium-sized mountain rivers. Many of the mountain rivers investigated were restored by removing bed and bank fixation, flattening river banks, and partly widening the cross-section (referred to as widening in the following). In the lowland rivers, remeandering and reconnecting oxbows were the most prominent measures besides increasing groundwater levels for restoring wetlands. Moreover, instream measures like large wood and boulder placement have been applied.

We found a significant effect on the number of ground beetle species and on richness and diversity of macrophytes, a moderate effect on fish, and a low effect on macroinvertebrates and floodplain vegetation. This isconsistent with the findings of other studies on single organism groups, except for floodplain vegetation, which usually benefits from restoration but restoration effects were constrained by agricultural land use in our study. Since the effect of restoration was generally higher on terrestrial and semi-aquatic organism groups, we recommend that they are considered in the monitoring and assessment of river restoration projects.

In general, the effect of restoration on community structure, traits, and functional indicators was more pronounced compared to the effects on species number and diversity. These changes in community structure indicate specific functional changes caused by river restoration and can be used to increase our understanding how restoration measures affect aquatic ecosystems, investigate causal relationships, and identify sustainable, (cost-) effective restoration measures. Therefore, we recommend that future restoration projects and monitoring studies should focus more on functional aspects (e.g. species traits, community structure) to investigate how river restoration affects river hydromorphology and biota, which would offer a great opportunity to make fundamental advances in restoration ecology and management.

The factors potentially constraining or enhancing the effect of restoration were partly correlated, which made it difficult to infer causal relationships (e.g. most old projects were located in gravel-bed rivers where mainly widening was the main restoration measure applied, and catchment land use was less intensive). Nevertheless, it was possible to draw some first conclusions on the conditions favouring restoration success:

It has been widely stated that large-scale pressures like water quality and fine sediment loads might constrain the effect of restoration. However, in this study, catchment land use did only affect restoration success for floodplain vegetation, and restoration effect might have been rather constrained by the limited species pool available for re-colonization and dispersal since the organism groups which did benefit most from restoration also have relatively high dispersal abilities (ground beetles, macrophytes). This topic clearly merits further investigation since a limited re-colonization potential would need a completely different restoration strategy compared to reach-scale habitat improvements.

Restoration extent (length of restored section, restoration intensity) was not the main factor determining restoration effects. Most probably, the restoration projects investigated were simply too small to benefit from possible positive effects of restoration extent, which is also supported by other recent studies. Furthermore, project age (time between implementation of the measures and monitoring) only had a positive effect on the aquatic habitat conditions but not on any of the organism groups investigated, possibly due to the young age of most projects investigated. In contrast, project age was identified as one of the most important variables affecting restoration success in the REFORM deliverable D 4.2, stressing the need to further investigate the effect of restoration over time in future studies.

Widening was applied in 11 of the projects investigated and had a significantly larger effect on hydromorphology and several organism groups (e.g. ground beetles, macrophytes) compared to other measures (among others instream measures), which is consistent with the findings of the REFORM deliverable D 4.2, and the widely endorsed assumption that restoring geomorphological processes has a higher effect compared to other measures. Since widening includes a set of measures, it was not possible to investigate the contribution of single measures. Since the positive effect on ground beetles was mainly due to the creation of open pioneer habitats covered by sparse woody vegetation, flattening river banks might already suffice but this has to be further investigated. Moreover, these results do not question the use of instream measures since transferability is limited due to the relatively low number of instream projects investigated in this study, and results of several other studies showing that instream measures generally have a positive effect on different aquatic organism groups.

The results indicated that future restoration projects should aim at increasing and monitoring habitat diversity at spatial scales which are ecologically relevant for the targeted organism groups. Although we found enhanced macro- and mesohabitats, which often is visually appealing, the measures often failed at increasing microhabitat diversity, which in turn was correlated with the effect of restoration on macroinvertebrates. Furthermore, it is not necessarily most important to increase the mere number of habitat types (e.g. habitat diversity) but to restore specific habitats which are of special importance. For ground beetles, the positive effect of widening was mainly due to the strong relationship between ground beetle richness and a specific habitat type: the open pioneer stage covered by sparse woody vegetation, but not to the mere number of habitat types.

The following more specific conclusions can be drawn for the single organism groups and river hydromorphology:

Overall, restoration increased habitat diversity through changes in channel morphology. The dominance of the main channel was significantly reduced, while other channel features such as islands, banks and bars became more frequent. The effect of restoration on hydromorphology was not higher in larger restoration projects compared to smaller projects. The effect of restoration was high for macro- and mesohabitat divsersity but low for microscale substrate composition. Key indicators for identifying restoration success should include parameters at larger spatial scales such as channel adjustments. There is a need to develop terrestrial parameters to assess the lateral dimension of restoration.

In line with other restoration studies no effects of restoration on macroinvertebrates were detected. However, macroinvertebrate richness and diversity was correlated with microhabitat diversity. While restoration projects like widening are visually appealing and increase macro- and mesohabitat diversity, they apparently rarely increase microhabitat diversity relevant for macroinvertebrates and species diversity.

Fish respond in a consistent way to hydromorphological restoration measures by an increase of rheophilic and a decrease of eurytopic fish. The restoration effect increases with habitat quality and length of restored river sections. Future restoration should focus on more dynamic, self-sustaining habitat improvements extending over several kilometres.

Restoration had an overall positive effect on richness and diversity of specific macrophytes (so-called helophytes, emergent plants rooting under water or in wetted soils) but not onemergent and submerged aquatic plants(hydrophytes). Restoration effects were especially high in widening projects located in mountain rivers.

An increase in total ground beetles species richness and richness of habitat specialists could be achieved primarily by creating pioneer patches, for example by river widening, which result in more open banks. Suitable restoration measures should aim on a strong lateral connection between the river and its floodplain. Further research should focus on determining optimal conditions of such pioneer habitats.

Responses of floodplain vegetation were related to changes in trait composition, while general effects on diversity were limited (small restoration projects) or absent (large restoration projects). Few general responses to restoration could be detected because species and trait composition and plant diversity varied substantially between the European regions.

For stable isotopes results supported our hypotheses that trophic length (indicated by Δ15N) as well as diversity of assimilated food sources (indicated by Δ13C) increase with restoration. Δ13C was significantly larger in large restoration projects compared to the corresponding degraded sections, suggesting that macroinvertebrates were feeding from more diverse sources. The results underlined the necessity to limit comparisons to sections within a region, as large-scale differences possibly masked the effects of restoration.