Flow resistance parameterization for large wood in two‐dimensional depth‐averaged models: Effects of flow condition, blockage, and sediment

Summary

Large wood (LW) significantly alters flow and bed morphology in river systems, necessitating accurate parameterization in two‐dimensional (2D) hydraulic models. This study investigated the effects of LW‐induced flow resistance in 2D depth‐averaged models using a combination of systematic flume experiments and numerical modeling. Sixteen controlled laboratory experiments were conducted, varying flow conditions, LW blockage ratios (full‐span vs half‐span) and sediment presence. Rectangular porous LW structures of constant porosity ( 0.66) were placed in a recirculating flume, and flow data including water surface elevation and velocity were recorded. Drag coefficient () and Manning's roughness were calibrated for each case using Gaussian Process Optimization to minimize model error based on measured water depths and flow partitioning. Results show that LW flow resistance parameterization is more complex than previously assumed. In addition to LW characteristics, it also depends on flow condition, blockage and sediment transport. Sediment presence and the resulting bed topography exert significant influence on flow resistance parameterization, followed by LW blockage configuration. Flow resistance parameters displayed wide variability even with fixed LW porosity, with ranging from 14 to 65 and Manning's from 1.17 to 2.6. These large values are equivalent resistance coefficients that represent the bulk drag of the emergent porous LW lumped over its footprint, not bed‐surface roughness values. While both and Manning's yield equivalent flow resistance in 2D modeling, 's dimensionless nature may offer broader generalizability. Importantly, the study confirms that parameter calibration is essential, especially for cases involving sediment transport and complex bed morphology. Rather than a single empirical formula, the study yields regime‐dependent guidance: where the bed stays relatively flat the resistance can be treated as a function of LW characteristics, whereas calibration is recommended where sediment‐shaped scour produces strongly three‐dimensional flow. These findings provide practical guidance for river modelers and engineers incorporating LW effects into flood‐risk assessment and restoration design.