Are spatial distribution and aggregation of wetlands reliable indicators of stream flow mitigation?

Abstract

Increasing legal requirements by local and regional governments to protect hydrosystems and maintain their hydrological services require nowadays the identification of wetlands of particular interest for conservation. Individual wetlands, however, have a limited effect on these services, especially on stream flow at the watershed scale in general. On the other hand, the compound effect of all wetlands does not allow for the identification of wetlands with substantial effects. Building on previous hydrological modelling development and seeking for a compromise between assessing the impact of individual wetlands and all wetlands, we propose the creation of wetland networks (i.e., wetlandscapes) based on: (i) geographic location, (ii) typology, (iii) wetland area, and (iv) contributing area. Thus, we tested the relevance of different ways of building wetland networks using the semi-distributed hydrological model HYDROTEL with a case study, that of the St. Charles River watershed, Quebec, Canada. We defined 48 wetland networks and assessed their capacity to attenuate peak flows and support low flows using 47 years of meteorological data, at both the sub-watershed scale and the watershed scale. Our results clearly demonstrate differences in the networks effect, confirming the relevance of the proposed framework. Although large wetlands with large contributing areas tend to have the strongest effect on high and low flows, we illustrate that small wetlands with large contributing areas can have a significant effect on extreme stream flows. This study also suggests that riparian wetlands might amplify peak flows as well as low flows. The application of the proposed modelling framework to various climates and geological contexts has the potential to be useful to support the design of wetland conservation programs and to meet some legal requirements based on the protection of wetland networks.