Evaluating the effect of seasonal temperature changes on the efficiency of a rhizofiltration system in nitrogen removal from urban runoff.

The study presents an evaluation of nitrogen removal efficiency of a pilot-scale rhizofiltration system in Pretoria, South Africa. The rhizofiltration system was divided into two sections, one side planted with common reeds (Phragmites australis) and the other side was without plants kept as a control. The objective of the study was to evaluate the influence of seasonal temperature on the removal of nitrogen species from the simulated urban runoff using the rhizofiltration system. The final effluent from the filter was collected bimonthly at different sampling points for 10 months after an application time of 5 min and 25 min. Duplicate samples were taken to determine the concentrations of TKN (Total Kjeldahl nitrogen), ammonium, nitrate and chemical oxygen demand (COD) for the raw influent and final effluent from the rhizofiltration system. Temperature and pH were determined on-site. During the monitoring period, there was no significant difference in the inflow concentration of ammonium in colder and warmer months for both planted and control sides. Furthermore, the composition of the feed medium to the rhizofilter was kept the same in both cold and warm season and for both planted and control sides. The removal of ammonium in colder and warmer months was not significant in both systems. At an average temperature increase of 5.2 °C in the warmer months, the ammonium removal efficiency in the planted side increased by 7.5%, while for the control side the removal efficiency increased by 2.4%. The difference in removal was not significant between the averages of effluent ammonium after an application time of 25 min in colder versus warmer months for the planted and control sides of the system. Furthermore, an increased nitrification rate was more evident in the planted than in the control side, which was subsequently denitrified. It was observed that 60.4% of nitrate concentration was potentially removed in the planted side whereas 45.4% was potentially denitrified in the control side. These results suggest positive correlation between nitrate concentration and the potential for denitrification. The nitrate removal efficiency dropped to 32.2% for the planted site and to 26.1% for the control system in colder months. Temperature had an effect on nitrogen removal, since nitrogen removal efficiency decreased in colder months. Complete nitrogen removal could not be achieved under the operating conditions.