Functional and taxonomic plant diversity for riverbank protection works: bioengineering techniques close to natural banks and beyond hard engineering.

ABSTRACT

Erosion control is a major issue in the Prealps region since piedmont is subject to both intense flood hazards and anthropic pressure. Riverbank protections may have major impacts on local ecosystem functioning and ecological corridor continuity. This study aimed to estimate the effects of the types of riverbank protection technique (from pure riprap to pure bioengineering) on the taxonomic and ecological composition of plant communities in comparison with unmanaged riverbanks as the referential system. Thirty-eight embankments were sampled in the foothills of the French and Swiss Alps. Four distinct riverbank techniques were analyzed and natural young willow stands were chosen as the referential system. At each site, vegetation was sampled along three transects from the waterline to the top of the riverbank. Plant communities were characterized using biological group composition (growth forms and life history, life strategies and distribution in space and time) and functional diversity indices (MFAD, FDc and wFDc). We identified 177 distinct plant species on 38 sites. Higher species richness levels were observed on bioengineered banks (from an average of 12 species recorded on ripraps to 27 species recorded on bioengineered banks) strongly dominated by Salicaceae species, especially for fascine and cribwall banks. Functional analyses of plant communities highlighted significant differences among bank types (p-value 0.001) for all selected biological groups. Competitive - ruderal strategy, rooting shoots, stems or leaves that lie down or break off, and unisexual - dioecious, as well as pioneer plants and low shrubs (<4 m tall) distinguished bioengineered bank types. Functional diversity indices confirmed these differences among bank types (MFAD p-value 0.002; FDc p-value 0.003; wFDc p-value 0.005). Riprap always showed the lowest levels on functional diversity indices, fascine and cribwall banks were at the medium level and finally mixed and natural banks the highest level. These results confirm the low ecological potential of purely hard engineering techniques and highlight the similarity of bioengineered techniques and unmanaged riverbanks.