Modelling wet weather sediment removal by stormwater constructed wetlands: Insights from a laboratory study

Abstract

Summary Constructed wetlands are now commonly used to control polluted urban stormwater discharges. A laboratory study was conducted to investigate the treatment of solids in these systems. Three mesocosm stormwater wetlands (vegetated with a well-established canopy of different densities) and one mesocosm non-vegetated pond were used, all sized to achieve particle fall number ( N f , a ratio between the times of the particle travel in horizontal and vertical directions) and Particle Shear Velocity Reynolds Number, Re ∗ , which are reflective of full-scale systems. The mesocosm vegetated systems had also similar turbulent Reynolds Numbers ( Re T ) to those funds in full-scale systems. Ten groups of steady-state experiments were carried out, all with different hydraulic loadings and sediment inflow concentrations (also maintained within the ranges found in real systems during wet weather). Samples were taken along the mesocosms and analysed for Total Suspended Solids concentrations (TSS) and Particle Size Distribution (PSD). It was found that both Re ∗ and Re T do not significantly influence the trapping of sediments, and therefore the particle re-suspension induced by water flow is not important for sedimentation in constructed stormwater wetlands. Vegetation density was found not to be an important factor, while particle diameter, and flow characteristics (e.g., flow rate and velocity) do influence trapping efficiency of particles. It was concluded that sediment trapping correlates strongly with particle fall number, N f , and therefore can be explained by this single non-dimensional number. A simple non-linear two-parameter regression model is proposed for prediction of particle trapping efficiency in constructed stormwater wetlands. However, further work is needed before the method can be used in practice. The aim of the ongoing work is to test whether the proposed model could be used across a number of real stormwater constructed wetlands without any further calibration. The data collected from a number of stormwater treatment systems in Melbourne, Australia, will be used in this study.