Modelling of stormwater treatment in biofilters using MIKE+ : Possibilities and limitations

Abstract: As research has expanded on the environmental impact of stormwater on receiving ecosystems more focus is now being put on the quality issues of stormwater. Biofilters are one of many nature-based solutions that have been developed for that purpose and are also the subject of this study. In order to plan and implement biofilters, predictive models can be useful tools to forecast their performance on a given site. The aim of this study was to investigate the possibilities and limitations of modelling the treatment of stormwater in a biofilter using MIKE+. This was done by first trying to model the hydrological conditions of a biofilter from a study site in Sundsvall and thereafter coupling a water treatment model created in ECO Lab.  The results showed that there were some notable differences between the parameters affecting the hydrological flow in reality and what is currently possible to model in MIKE+. It was seen that the Soakaway node used to model biofilters needs to be complexified in order to properly model the hydrological conditions of biofilters. The main improvements required are that the flow attenuation should occur within the Soakaway node rather than before or after and that it should be based on soil properties. The retention volume also needs to be integrated in the node and a varying exfiltration rate is believed to be required to fit the varying nature of evapotranspiration.  The hydrological model was seen to have a great impact on the water treatment model and some limitations with the program and the used model were identified. Regarding ECO Lab, the program is currently unable to consider interevent processes when the biofilter is empty. These processes can however be of great importance for some contaminants. The program also assumes a time dependency whereas it has been seen that a short retention time may be sufficient to achieve good reduction efficiency. The model used is a highly lumped conceptual model with few parameters so further research aiming at the creation of correction factors for the main affecting parameters is believed to be required in order to avoid design specific calibration. Calibration should also occur over a longer time period in order to consider the variability of stormwater.