emergent plants such as cattails, whereas floating plants like water hyacinth are most effective in subtropical and tropical wetlands (Reddy and DeBush, 1987).
In contrast, diversion of streams or runoff into unmanaged natural wetlands appears to provide only limited long-term nutrient removal. Although such wetlands may assimilate nutrient inputs during the growing season, a large outflow of nutrients released from dead vegetation the following spring may offset the nutrients stored the previous growing season. In addition, large losses of nutrients from wetlands during high-flow, intensive rain events or through channelization tend to counterbalance the net storage of nutrients during longer periods of low or moderate flow rates (Richardson, 1988).
In a few cases, streams flowing into lakes have been treated by adding phosphorus-precipitating chemicals (iron, aluminum), but because volumes of water that need to be treated generally are large (compared with municipal wastewater), this usually is not a cost-effective approach. Iron is preferred for in-stream treatments because it has fewer toxicity problems than does aluminum, but binding of phosphorus to iron requires continuously aerobic conditions. Success in lowering phosphorus concentrations has been reported when relatively small flows can be treated. An example is the addition of ferric sulfate to water pumped into Foxcote Reservoir (England) to remove dissolved phosphorus. Although internal phosphorus loading in the reservoir has reduced the treatment's effectiveness, the length of time that Foxcote Reservoir cannot be used as a potable water supply during summer months has decreased (Young et al., 1988).
Wahnbach Reservoir, an important municipal water supply for Bonn, Germany, is protected from nutrient, silt, and organic matter loading from its main tributary by a prereservoir detention basin and phosphorus elimination plant (Bernhardt, 1980; Clasen, 1989). Water from the detention basin is treated with iron to remove phosphorus and then filtered through an ion exchanger and a series of activated carbon and sand filters. The plant removes 95 to 99 percent of phosphorus, coliform bacteria, algae, and turbidity; 77 percent of the water's biochemical oxygen demand; and 58 percent of the dissolved organic carbon. Grossly enriched river water is converted into nearly drinkable water before it enters the reservoir. Costs of this project have not been published.
Best Management Practices Numerous best management practices (BMPs) (Table 4.8) have been developed to decrease losses of soil,