Effects of plant community and phosphorus loading rate on constructed wetland performance in Florida, USA

We evaluated the effectiveness of constructed wetlands with varying plant communities for phosphorus (P) reduction from the Everglades Agricultural Area runoff in south Florida. Weekly or biweekly water samples from the inflow and outflow regions of I I test cells (2,000 m(2)) were analyzed for various forms of P and other selected water quality variables between January 2002 and August 2004. Test cells located at the north site received water with a high average total P (TP) concentration (72 mu g L-1), while test cells located at the south site received water with a lower average TP concentration (43 mu g L-1). These test cells were dominated by an emergent vascular plant-cattail (Typha latifolia), submerged aquatic vegetation or SAV (Najas guadalupensis, Chara sp., Ceratophyllum demersum, and Hydrilla verticillata), or algal periphyton (mixed with Eleocharis cellulosa and Utricularia spp. in the south site only). Under a constant hydraulic loading rate (9.27 in yr(-1)), these test cells removed P effectively, with removal efficiencies of 56%-65% at the north site and 35%-62% at the south site. The mass removal rate and rate constant at the north site were also higher than at the south site. Soluble reactive P (SRP) and particulate P were the major forms at inflow and were removed effectively by all of the test cells. The removal of dissolved organic P was significant (similar to 60%) in the cattail and periphyton test cells, but no removal was detected in the SAV test cells. At the north site, P removal efficiency of the cattail test cells was slightly higher than that of the SAV test cells. At the south site, periphyton test cells were the best performers. The removal of SRP was positively correlated with the removal of calcium in the majority of the test cells, pointing to the potential importance of co-precipitation of calcium carbonate and SRP. Direct plant uptake, wetland filtering, microbial degradation, and co-precipitation with calcium carbonate were mechanisms thought to be responsible for P removal in these wetlands. Outflow TP concentration, an important measure for restoration performance, increased continuously with the increases in the TP mass loading rate at the north site, but peaked at approximately 30 mu g L-1 when the TP mass loading rate reached 0.5 g m(-2) yr(-1) at the south site.