A reduction pond for denitrification of agricultural drainage.

摘要

High-rate denitrification technologies are rarely used in agricultural settings due to high cost and lack of trained operators. Treatment wetlands can be more affordable, but where land is limited, other technologies are needed. The reduction pond process explored in this study may be useful in such circumstances. It retains most of the simplicity of wetlands while increasing process control and land-use efficiency. In contrast to the well-known oxidation pond, a reduction pond reduces oxidized pollutants such as nitrate or other oxyanions. The reducing environment is created by the addition of an electron-rich substrate, molasses in the case of the present study.; The Reduction Pond used in this 4-year study was 3-m deep with a volume of 760 m3. A plastic cover prevented most re-aeration. The provision of molasses at 1.5--3 times the stoichiometric requirement provided all of the electrons and nutrients needed to achieve complete denitrification of brackish, subsurface agricultural drainage containing 50--70 mg/L NO 3--N. The maximum annual nitrate reduction rate was 4.2 mg NO3--N reduced/L/day, while increased molasses dosing is expected to increase this rate to 10 mg/L/day or more. The pond was not stirred, and bacterial cells and precipitated minerals settled to form a sediment. The high hardness (0.1--9 g/L as CaCO3) of the drainage is thought to have promoted efficient cell sedimentation. The upper 15--30 cm of this sediment was flocculated, highly porous, and apparently an active zone of denitrification. Feedback of electron-donor compounds from the sediment was responsible for two-thirds of the total nitrate reduction. The settling and subsequent decomposition of the bacterial cells led to low effluent suspended solids concentrations, eliminated routine sludge handling, and decreased the molasses dosing requirement.; Sufficient water quality data were available to both calibrate and test mathematical models of nitrate reduction. The descriptive and predictive abilities of empirical and mechanistic models were compared. Predictions by the mechanistic model were 2--4 times more accurate than those of the empirical models. With soluble reactants and a cover, the Reduction Pond environment was simple compared to that found in open sewage treatment ponds. The capabilities of the best reduction pond model may represent the upper limit of capabilities to be expected from models of more complex pond environments.