Application of Innovative Methodologies to Advance Understanding of River Restoration

Abstract

Globally, rivers and associated floodplain habitats are highly degraded by anthropogenic stressors. This reduces their ecological complexity and the functions and some of the services that they provide. For about three decades, river restoration has attempted to revert this degradation, but it is still a young discipline that lacks robust frameworks for evaluating whether interventions achieve their intended ecological outcomes. My dissertation shows how the use of innovative methodologies in the fields of automated content analysis (ACA) and functional ecology can advance river restoration as a mature scientific discipline by 1) allowing researchers to synthesize relevant trends across thousands of publications and 2) contributing to the development of a generalizable framework for how restoration influences ecosystem recovery. Specifically, I validated the use of ACA methods to synthesize relevant trends in the river restoration literature and used a trait-based framework to quantify the response of riparian forests to land-use change and various restoration strategies, using case studies from the Garonne (SW France) and Ebro (NE Spain) river basins. Overall, unsupervised and semi-supervised ACA performed better than supervised classification at synthesizing relevant trends within the river restoration literature. However, issues with model validity meant that not all relevant concepts, such as specific restoration strategies or indicators used, were captured by these methods. Thus, demonstrating that while ACA methods can synthesize relevant trends, they are not a complete replacement for traditional literature review methods. Plant functional responses to riparian restoration were primarily driven by reproductive strategies and species environmental preferences, and these results were consistent for restored riparian forests sites across river catchments in France and Spain. Specifically, hydrologically reconnecting riparian habitats (e.g., through channel widening) was necessary to recover similar functional composition and diversity as mid-seral riparian forests. On the other hand, the abandonment of forestry practices (Populus x euramericana plantations) within the floodplain resulted in an incomplete but overall positive functional recovery, driven by an increase in plant functional strategies characteristic of mature riparian forests. My dissertation therefore provides examples of how applying such innovative methodologies can contribute to the growth of river restoration as a scientific discipline.