Twenty years experience with constructed wetland systems in Denmark – what did we learn? H. Brix, H.-H. Schierup and C.A. Arias Aarhus University, Department of Biological Sciences, Ole Worms Alle ´ , Building 1135, 8000 Aarhus C, Denmark Abstract Full scale constructed wetland systems for wastewater treatment have been in operation in Denmark since 1983, mainly for the treatment of domestic sewage from small villages. The systems are constructed as soil-based horizontal subsurface flow systems but, because of low soil hydraulic conductivity, surface runoff is evident in most of the systems. Two decades of experience show that soil-based systems are generally efficient in removing suspended solids and BOD, but the removal of nitrogen and phosphorus is lower (typically 30–50%) and the systems do not nitrify ammonium. Contrary to earlier claims, the reeds do not increase the hydraulic conductivity of cohesive soils as much as necessary to secure sub-surface flow. Operation needs of soil-based reed beds are low and normally restricted to emptying of the sedimentation tank, cleaning of the distribution system and mowing of the grass around the system. The dead plant material and accumulated litter on the surface of the systems improve performance after the initial years. A significant number of systems have been shut down or extended with other technologies in order to meet new effluent standards, particularly demands for nitrification. New constructed wetland systems are either compact vertical flow systems which provide good nitrification, willow systems with no discharge or restored wetland systems for nitrate removal. If efficient removal of phosphorus is required, this is achieved by chemical precipitation in the sedimentation tank. Keywords Constructed wetland; Phragmites; reed bed; root-zone method; vertical flow; wastewater Introduction Constructed wetland systems for the treatment of wastewater have been in operation in Den- mark since 1983 when the root-zone method was introduced from Germany (Brix, 1987b). Numerous Danish municipalities, who at that time were in the process of establishing improved wastewater treatment for small communities, were attracted by this new low-cost technology, particularly because of the excellent performance claims. For all parameters (BOD, N, P, pathogens, heavy metals, etc.) removal efficiencies much better than 90% were expected using reed bed surface areas of approximately 5 m 2 per person equivalent (Kickuth, 1980, 1982). These performance claims were largely based on data from a system in Othfresen in Germany. However, the data from this system have later been seriously questioned (Brix, 1987a). After a few years it became obvious that the initial treatment performance promises were not fulfilled (Schierup and Brix, 1990), and contrary to earlier belief, the growth of the reed did not increase the hydraulic conductivity of the substrate (Brix and Schierup, 1989; Haberl and Perfler, 1990; Brix, 1998). Therefore, new designs based on gravel and vertical flow including gravel-based horizontal subsurface flow systems, two-stage systems (Brix and Johansen, 1999) and vertical flow systems were developed (Brix and Arias, 2005a, 2005b). In addition, different means of achieving sustained phosphorus removal were tested (Arias et al., 2001; Brix et al., 2001; Ger- vin and Brix, 2001; Arias and Brix, 2005). This paper focuses on the experiences with the soil-based constructed reed beds in Denmark that have now been in use for more than two decades. What is the status of these systems in Denmark today? How well do they perform? And has their treatment performance changed over time? Water Science & Technology Vol 56 No 3 pp 63–68 Q IWA Publishing 2007 63 doi: 10.2166/wst.2007.522