Recruitment of riparian plants after restoration of geomorphic complexity in northern Sweden

Question: Restoration of channelized streams increases geomorphic complexity but it remains unclear how this interacts with processes that drive future vegetation changes (dispersal, germination and establishment). This study asks if increased geomorphic complexity increases recruitment conditions of sown seeds or affects post-dispersal natural seedling densities. Location: Vindel River catchment, northern Sweden. Methods: We selected seven study streams with paired reaches that differed in the degree to which geomorphic complexity was restored. Basic reaches used simple restoration methods while enhanced reaches additionally added large boulders and woody debris. We sowed seeds of six species at ten locations in each reach in 2014 and counted the number of seedlings after 8wk and the number of naturally occurring seedlings in a plot adjacent to the sowing locations in 2013 and 2014. Using factor analysis based on 34 complexity measurements, overall geomorphic complexity was quantified for eight of the 14 reaches. Results: Total numbers of sown (2014) and natural seedlings (2013 and 2014) summed per reach did not differ between restoration types when tested pair-wise. Enhanced restoration did not always significantly increase geomorphic complexity, which differed considerably between the streams. More complex reaches were steeper, had larger size sediment and more nutrient-poor soils. Total recruitment of sown species significantly decreased with increasing complexity. Numbers of natural seedlings differed considerably from 2013 to 2014, but were not related to complexity. In 2014, a potential parent plant of the same species occurred within the same plot for 71.8% of the natural seedlings that could be identified. Conclusions: The recruitment of sown seeds was affected by overall geomorphic complexity rather than by the enhanced restoration. The absence of a correlation between geomorphic complexity and natural seedlings could indicate that natural seedling dynamics are not solely determined by recruitment conditions, but also by dispersal