Performance of grass filters used for stormwater treatment - a field and modelling study

Grass swales and filter strips are among the simplest and most cost-effective form of stormwater control measures. Despite this, relatively few controlled studies have been undertaken to assess their performance, and very few models are available to predict their efficiency. This paper focuses on field performance of grass filters in removal of total suspended solids (TSS) from overland flow, and verification of TRAVA, a model of sediment behaviour in grass filters (published earlier it) Deletic, A., 2001. Modelling of water and sediment transport over grassed areas. Journal of Hydrology 248, 168-182; Deletic, A., 2005. Sediment behaviour in runoff over grassed Surfaces. Journal of Hydrology 301, 108-122). Limited nutrient removal data were collected, at one of the field sites. The paper presents results from controlled field tests undertaken oil a grass filter strip in Aberdeen, Scotland, and a grass swale in Brisbane, Australia. In both studies TSS concentrations were recorded along the grass, for artificial inflow of water and sediment of different flow rates and sediment concentrations. In the Aberdeen study the emphasis was on performance in relation to different sediment particle sizes, while for the Brisbane swale. treatment performance for total nitrogen (TN), and total phosphorus (TP) were also recorded. Both experimental programmes had an unsteady input of pollutants. The Aberdeen strip reduced sediment inflow concentrations by 61-86%, whilst the Brisbane swale removed an average of 69, 46, and 56% of the total loads of TSS, TP and TN, respectively. TRAVA was used to model both the strip and swale. For the Aberdeen strip, the differences between the measured and modelled sediment loading rates were within +/- 25% of measured. Modelled concentrations in outflow from the Brisbane swale were within +/- 17% of measured. The predicted mss of total sediment removed during experiments was within +/- 50% of measured for the Aberdeen strip and +/- 11% for the Brisbane swale. TRAVA is therefore shown to be a reliable tool for prediction of the performance of grass filters used for stormater treatment. (c) 2005 Elsevier B.V. All rights reserved.