A constructed wetland design, consisting of 16 repeating cells was proposed for Henley Brook (Perth, Western Australia) to optimise the removal of FRP from urban stormwater. Three replicate experimental ponds (15×5 m), were constructed to represent at a 1:1 scale a single cell from this design. Three 5 m zones of each pond were sampled: shallow (0.3 m) vegetated (Schoenoplectus validus) inflow and outflow zones and a deeper (1 m), V-shaped central zone. In 1998/99, inflows and outflow waters were intensively sampled and analysed for FRP and Total P. In addition, all major pools of P (plants, sediment) within the ponds, and important P removal processes (benthic flux, uptake by biofilm and S. validus) were quantified.

A removal efficiency of 5% (1998) and 10% (1999) was obtained for FRP. Initial uptake was mainly in plant biomass, although the sediment became an increasingly important sink. Benthic flux experiments showed that anoxia did not cause release of P from sediments, indicating that most of the P was bound as apatite rather than associated with Fe or Mn. The highly coloured waters were believed responsible for the very low biofilm biomass recorded (<1 g.m-2). We have demonstrated that constructed wetlands can be effective for removing FRP immediately after construction, although their long-term removal capacity needs further research.