Abstract
Four constructed freshwater riparian marshes in northeastern Illinois, USA, each experimentally subjected to either high or low loadings of turbid river water over a 3‐yr period, were investigated for phosphorus retention and fate. Average phosphorus concentrations decreased by 64‐92% in low‐flow wetlands (to 11‐40 mg P/L) and by 53‐90% (to 12‐57 mg P/L) in high‐flow wetlands. Intensive sampling at 4‐h intervals during spring and summer of the final year showed phosphorus mass decreases of 81‐74% for low‐ and high‐flow wetlands, respectively. During this frequent sampling, inflow averaged 176 mg P/L, while outflows were 34 mg P/L for the low‐flow wetland and 45 mg P/L for the high‐flow wetland. Spatial patterns of isopleths showed clear gradients in high‐flow wetlands and near‐homogeneous concentrations in low‐flow wetlands. Annual budgets illustrated that most inflowing phosphorus was retained through sedimentation with some capacity for phosphorus retention by macrophytes and a lesser amount by microbial (periphyton and planktonic) communities. A simple Vollenweider‐type model predicted well the retention of phosphorus in these wetlands for low‐ and high‐flow wetlands with 3 yr of data, i.e., the model was calibrated and verified with six independent data sets. A mean deviation of only 9.3% between measured and simulated retention was found when the retention coefficient was k = 6.51 wk‐1. Adding a factor for seasonal temperature fluctuation did not improve the fit, whereas using a model where the retention is a function of flow provided the best overall fit to the data (mean deviation of 6.3% between data and predictions). These constructed wetlands retained about the same amount of phosphorus per unit area (0.5‐3 g P°m‐2°yr‐1) as have several other natural and constructed wetlands receiving similar concentrations of phosphorus.