Stream-scale flow experiment reveals large influence of understory growth on vegetation roughness

Vegetation is a key source of flow resistance in natural channels and floodplains. It is therefore important to accurately model the flow resistance to inform decision makers and managers. However, it is challenging to predict the resistance of real vegetation, because vegetation models are based on relatively small-scale lab experiments with mostly artificial vegetation. Experimental tests of real vegetation under field conditions are scarce. The purpose of this study is to measure the flow resistance of a submerged willow patch, where small herbaceous vegetation was allowed to grow in between the willow stems to simulate field conditions. Detailed flow velocity measurements were performed during an full scale experiment of flow around a submerged patch of willows. The parameter values of the willow vegetation model, as well as the friction coefficients of the vegetated banks and unvegetated channel bed, were computed simultaneously using Bayesian inference using a 2D hydrodynamic model. Results show that the presence of understory growth greatly affects flow patterns and the value of the effective vegetation density parameter. Measured flow velocities in the patch with understory growth were very low, and the patch has relatively high deflection. After removal of this undergrowth, flow velocities in the patch increased and deflection of the vegetation canopy decreased. We show that estimating vegetation density using an often-used rigid cylinder estimator based on vegetation sampling, underestimated the effective value by more than an order of magnitude. We argue that proposed extensions to existing vegetation models, which can take into account understory growth and reconfiguration, could be tested under field conditions using the approach followed in this paper.