Hydraulic performance of leaky barriers for water resource management and climate change adaptation using nature-based solutions

Abstract

Flooding is one of the most significant natural hazards affecting river systems and urban catchments worldwide, and its impacts are expected to intensify under future climate change conditions. Nature-Based Solutions (NbS), including engineered log jams and leaky barriers, are increasingly being investigated as sustainable alternatives to conventional hard-engineering flood management. These structures have the potential to attenuate flood peaks, increase upstream water storage, and improve ecological resilience while reducing reliance on traditional grey infrastructure. This research investigated the hydraulic performance of channel-spanning leaky barriers through HEC-RAS 1D steady-state modelling, an energy-slope validation method, and laboratory flume experiments across six barrier configurations at the University of Waikato. Calibration against 46 runs drawn from three published datasets (Muhawenimana et al., 2021; Huang et al., 2022; Schalko, 2018/2019) showed that barrier-induced resistance, rather than bed friction, dominates the hydraulic response of channel-spanning leaky barriers, with calibrated Manning's n values of 0.086–1.426. A backwater rise prediction framework, extended with an effective solid volume fraction correction for barriers with a clear bed gap, achieved a combined Nash–Sutcliffe Efficiency (NSE) of 0.932 across 308 observations from four datasets; the correction proved essential, recovering NSE for the two gapped datasets from strongly negative values to above 0.87. A piecewise predictive equation for Manning's n was then developed from laboratory observations, expressing n as a power-law function of upstream depth and a cumulative array parameter (CA) for unsubmerged flow, and as a power-law decay from a maximum value with increasing submergence once the barrier is overtopped. The unsubmerged equation achieved an NSE of 0.853 and a mean absolute percentage error of 7.49%, using only geometric and hydraulic parameters known prior to a flood event. The research demonstrates that this relationship may be incorporated directly into HEC-RAS 1D through an iterative coupling with the backwater rise model, and sets out a physically consistent pathway, proposed but not yet directly tested, for incorporating the same relationship into two-dimensional and catchment-scale morphodynamic frameworks such as CAESAR-Lisflood. The outcomes of this research contribute to sustainable flood management practice and strengthen the evidence base for Nature-Based Solutions in climate change adaptation.