Plant trait characteristics vary with size and eutrophication in European lowland streams

Abstract

Previous studies investigating community-level relationships between plant functional trait characteristics and stream environmental characteristics remain scarce. Here, we used community-weighted means to identify how plant traits link to lowland stream typology and how agricultural intensity in the catchment affects trait composition.We analysed plant trait characteristics in 772 European lowland streams to test the following two hypotheses: (i) trait characteristics differ between plant communities in small and medium-sized streams, reflecting adaptations to different habitat characteristics, and (ii) trait characteristics vary with the intensity of agricultural land use in the stream catchment, mediated either directly by an increase in productive species or indirectly by an increase in species that efficiently intercept and utilize light.We found that the communities in small streams were characterized by a higher abundance of light-demanding species growing from single apical meristems, reproducing by seeds and rooted to the bottom with floating and/or heterophyllous leaves, whereas the community in medium-sized streams was characterized by a higher abundance of productive species growing from multi-apical and basal growth meristems forming large canopies.We also found indications that community trait characteristics were affected by eutrophication. We did not find enhanced abundance of productive species with an increasing proportion of agriculture in the catchments. Instead, we found an increase in the abundance of species growing from apical and multi-apical growth meristems as well as in the abundance of species tolerant of low light availability. The increase in the abundance of species possessing these traits likely reflects different strategies to obtain greater efficiency in light interception and utilization in nutrient-enriched environments. Synthesis and applications. Our findings challenge the general assumption of the EU Water Framework Directive compliant assessment systems that plant community patterns in streams reflect the nutrient preference of the community. Instead, light availability and the ability to improve interception and utilization appeared to be of key importance for community composition in agricultural lowland streams. We therefore suggest moving from existing approaches building on species-specific preference values for nutrients to determine the level of nutrient impairment to trait-based approaches that provide insight into the biological mechanisms underlying the changes. We recommend that existing systems are critically appraised in the context of the findings of this study.

Keywords: Water Framework Directive; assessment; eutrophication; functional traits; growth form; macrophyte; nutrients; plant trait characteristics; river; vegetation.

Figure 1

Conceptual figure depicting two non‐mutually exclusive hypotheses on mechanisms driving changes in plant composition under nutrient‐enriched conditions in lowland streams. The first hypothesis suggests that eutrophication directly promotes productive species forming large standing crops, building on the assumption that nutrients are limiting for plant growth, whereas the second hypothesis suggests that species efficiently capturing or utilizing light are promoted. According to the second hypothesis, species with low light requirement and species capable of concentrating their photosynthetic active biomass in the uppermost waters by possessing single and multi‐apical growth meristems should increase in abundance.

Figure 2

Ordination plots of the coinertia analysis (COIA) between species traits (18 traits) and species abundances (77 plant species). In total, 772 stream sites were included in the analysis, all of which were located in the central Baltic region of Europe. The first plot (a) shows how species cluster groups (1–5) associate with the first two axes of the COIA. Large light grey circles indicate mean abundance based on the ordination score of species within each cluster group (normed row scores from the COIA) where small circles represent the score of individual species. Dark grey squares indicate mean trait‐based ordination score of species within each cluster group (normed row scores from the COIA) where small squares represent the score of individual species. The lines between the circles and the squares indicate mean distances in species abundance‐based scores and trait‐based scores within each cluster group (i.e. the level of association between species distributional patterns and ecological traits). The second plot (b) shows how traits associate with the first two axes of the COIA (a projection of the canonical weights of species traits) and thus indicates which traits are related to which cluster group in (a). See Table 2 for definitions of the trait names.

Figure 3

Mean trait values [community‐weighted means obtained through coinertia analysis (COIA)] in the identified species groups 1–5 for the range of traits used to characterize the plant communities. Error bars indicate standard error. Horizontal line indicates the grand mean for all sites. See Table 2 for definitions of the trait names.

Figure 4

Plots showing mean local abundance (a) and distribution (b) of species belonging to cluster groups 1–5 in small (R‐C1; black points) and medium‐sized (R‐C4; white points) streams. Error bars indicate standard error. Local abundance was calculated as the average abundance for each site where the species was present. Distribution was calculated as the sum of sites where the species was present divided by the total number of sites.

Figure 5

Plots showing mean trait values for the range of traits used to characterize the plant communities in small (black bars) and medium‐sized (white bars) streams. Error bars indicate standard errors of the grand mean for the respective stream types. The traits are ordered according to difference in mean trait values between small and medium‐sized streams, with the largest difference first (NE) and the smallest difference last (Ansule).* = significant differences between mean values (P < 0·05). See Table 2 for definitions of the trait names.