Investigating process drivers of Natural Flood Management and its flood risk reduction potential across scales

Before-after monitoring of small-scale restoration activities in blanket peatlands, e.g., revegetation and gully blocking, suggests they can also deliver significant Natural Flood Management (NFM) benefits (reduce and delay floodpeaks). However, we still lack a clear understanding of the underlying processes driving NFM effects; and doubts remain about whether interventions will retain their impact when implemented at scales large enough to reduce flooding in downstream communities. We examine the impact of the two interventions at a range of scales from the 1 hectare micro-catchment scale at which a Before-After-Control-Intervention (BACI) study has been undertaken, to the 40 km² scale (at which flooding begins to affect residential properties). We calibrate the Generalised Multistep Dynamic (GMD) TOPMODEL rainfall-runoff model to different BACI experimental catchments each representing an intervention scenario. Through numerical experimentation with the calibrated parameters, we estimate the impact-magnitude of different process drivers. Our findings confirm the NFM benefits of these restoration-focused interventions at the micro-catchment scale. In both interventions and in our largest storms, floodpeak attenuation is primarily due to roughness reducing the floodwave speed and thus thickening the overland flow (kinematic storage). More conventional, static storage (i.e. interception + ponding + evapotranspiration), becomes important only in smaller storms. Finally, we use the parameter-sets identified by calibrating to the BACI catchments to extend our findings to the 40 km² Glossop catchment. Glossop has experienced several damaging floods in the last 50 years and has received appreciable recent restoration activity. Here we use GMD-TOPMODEL in a second set of modelling experiments to estimate downstream impact of existing interventions and to examine the impact of alternative scenarios of spatially distributed intervention configurations.