Monitoring Stormwater Quality through a Series of Natural and Constructed Treatment Devices: A Case Study from Brisbane, Sub-Tropical Australia

Stormwater best management practices include the installation of a variety of water quality improvement devices to treat urban stormwater runoff. Several of these devices occur within a 1km stretch of an upper urbanised creek catchment. The main channelised stream flows into a sediment basin, then into a constructed wetland, followed by a second deeper wetland. The latter also receives stormwater through a piped drainage system after it passes through a below ground GPT. A minor tributary receives stormwater through a piped drainage system, which also passes through a GPT before entering a natural densely vegetated riparian wetland. The water from both tributaries converges and flows downstream through a natural channel and series of lagoons. The aims of the research project were to determine spatial and temporal changes in pollutant concentration during wet and dry weather and; to assess the effectiveness of the "treatment train" in removing nutrients and suspended solids. Water quality parameters monitored included total and volatile suspended solids and nutrient speciation. TSS in stormwater runoff entering the sediment basin within 12 h of storm event was highly variable and TVS was about 25%. There was little reduction in TSS along the treatment train. However, within 24 hours of a storm event TSS in stormwater had decreased to background concentrations, but was still higher in the wetlands. Higher TSS in the wetlands also occurred in dry weather suggesting resuspension (caused by ducks) of both sediment and organic particulates (TVS was about 50%). TSS was consistently reduced in the downstream natural stream channel. During storm events NH₄ increased in Wetland 2, but NO₃ and PO₄ were reduced. Highest concentrations of NH₄, NO_x, PO₄ in both wet and dry weather were from piped outlets (in particular those from the below ground GPT's). In dry weather, NO₃ concentrations were reduced in the wetlands and natural stream channel. Our study concluded that during storm events the treatment train has little effect on the removal of TSS due to the high flow and short detention times, but some reduction in nutrients did occur. Nutrients in piped drainage were always higher than in the open stream channel suggesting periphyton in the channelised streambed were probably effective in removing dissolved nutrients. The 600 m length of remnant natural channel, lagoons and associated vegetation was effective for water quality improvement in post storm and dry weather. This serves as a reminder that the retention of natural stream channels and buffers of riparian vegetation are an integral component of Best Management Practice for urban stormwater treatment in many catchments.