Basin-scale hydrological response to leaky wooden dam installation

Leaky wooden dams are commonly incorporated into rivers as part of restoration efforts to increase channel roughness and force geomorphic complexity, slowing the flow in the headwaters and aiming to desynchronise flows to reduce downstream flood risk. These structures are (dis)connectivity agents, working to decrease longitudinal connectivity whilst simultaneously increasing floodplain connectivity and encouraging water storage.

Most numerical modelling of leaky wooden dams at the basin scale does not consider sediment transport at spatial resolutions fine enough to appropriately represent the dams as individual features. Due to the paucity of both spatially- and temporally-distributed sediment transport data, there is also a high level of uncertainty regarding the influence of leaky wooden dams on basin hydrology over time, yet it is important that we consider the geomorphological influence of these structures and how their evolution influences flood hazard, particularly given that extreme storms are becoming increasingly common.

This study implements a heuristic behavioural approach within the landscape evolution model CAESAR-Lisflood to assess the broad influence of leaky wooden dams on a 32 km² prototype catchment with a mixture of first, second and third order streams. A 20-year spatially-distributed modelled rainfall time series capable of representing convective storms (2020–2040 obtained from the 2018 UK Climate Projections) was used to drive the hydrology across a suite of simulations where leaky wooden dam location in the river network was systematically varied.

Installing leaky wooden dams only on first order streams desynchronised flow and reduced downstream flood peaks by up to 50% whilst retaining the greatest volume of water in the catchment when compared to other stream order combinations. Conversely, installing leaky wooden dams on only third order streams increased peak discharge by over 10% for 22% of storm events owing to the presence of fewer structures and therefore reduced opportunity for desynchronisation of peak flows from the various sub-catchments. Most importantly we detail how storm sequencing, and the capacity of the active channel, plays an important role in exacerbating flood risk, with frequent, yet relatively minor, storms increasing peak discharge despite the presence of leaky wooden dams. As such where leaky dam interventions are installed plays a critical role in their efficacy in mitigating flood peaks and should be given more consideration by practitioners.