Ecosystem responses to channel restoration decline with stream size in urban river networks.

Abstract

Urban streams are often severely impaired due to channelization, high loads of nutrients and contaminants, and altered land cover in the watershed. Physical restoration of stream channels is widely used to offset the effects of urbanization on streams, with the goal of improving ecosystem structure and function. However, these efforts are rarely guided by strategic analysis of the factors that mediate the responsiveness of stream ecosystems to restoration. Given that ecological gradients from headwater streams to mainstem rivers are ubiquitous, we posited that location within a river network could mediate the benefits of channel restoration. We studied existing stream restorations in Milwaukee, Wisconsin, to determine (1) whether restorations improve ecosystem function (e.g., nutrient uptake, whole-stream metabolism) and (2) how ecosystem responses vary by position in the urban river network. We quantified a suite of physicochemical and biological metrics in six pairs of contiguous restored and concrete channel reaches, spanning gradients in baseflow discharge (19-196L/s) and river network position (i.e., headwater to mainstem). Hydrology differed dramatically between the restored and concrete reaches; water velocity was reduced 2- to 13-fold while water residence time was 50-5,000% greater in adjacent restored reaches. Restored reaches had shorter nutrient uptake lengths for ammonium, nitrate, and phosphate, as well as higher whole-stream metabolism. Furthermore, the majority of reaches were autotrophic (i.e., gross primary production>ecosystem respiration), which is not common in stream ecosystems. The difference in ecosystem functioning between restored and unrestored reaches was generally largest in headwaters and declined to equivalence in mainstem restorations. Our results suggest that headwater sites offer higher return on investment compared to larger downstream channels, where ecosystem responsiveness is low. If this pattern proves to be general, the scaling of ecosystem responses with river size could be integrated into planning guidelines for urban stream restorations to enhance the societal and ecological benefits of these expensive interventions.