Evaluation of Hydrology in an Agricultural Watershed and Nitrogen Removal by Constructed Wetlands

Abstract

There has been a particular concern on excessive nitrogen loading in coastal watersheds because agricultural lands are located adjacent to environmentally sensitive waters. The constructed wetland at a watershed outlet is recognized as an environmentally safe wastewater treatment system, but the results still vary. The objectives of this study were to investigate the effects of in-stream processes in agricultural watersheds of the lower coastal plain, to determine nitrogen transformations in constructed wetlands, and to propose a modeling approach for quantifying nitrogen transport and in agricultural watersheds and constructed wetlands. The first step was a literature review on nitrogen retention in agricultural streams, nitrogen transformation in constructed wetlands, and design criteria of constructed wetlands for efficient nitrogen removal. In the second step, the daily outflow from a 1037 ha agricultural watershed at Open Grounds Farm, Carteret County, NC was evaluated using the DRAINWAT model. As a third step, vegetation succession and nitrogen removal efficiency were monitored, and a sequential nitrogen model in constructed wetlands was evaluated. Mainly due to problems with flow meters and lack of sufficient measured data, hydrology simulation results were not adequate to estimate nitrogen loading. For this reason, Hydrologic modeling for watershed focused on the sensitivity analysis using four input parameters: Manning’s roughness coefficient, maximum depressional storage, calculation methods of potential evapotranspiration (PET), and channel bedslope. PET calculation methods were most sensitive among those four parameters. Two Thornthwaite based methods using different correction factors (Thorn 1 and Thorn 2) were tested. Thorn 2 correction factors obtained from the agricultural watershed were more accurate than Thorn 1 correction factors, which were averaged from two agricultural and one forested watershed. Three hydrophyte species were successfully established in the constructed wetland. Good removal efficiencies of constructed wetland system for all nitrogen species were observed. Simulation results of sequential nitrogen model were not yet reasonable due to the limited number of sample data, a long waterway before the first cell of the constructed wetland, and non-calibrated average rate constant of the model. This model may predict the concentration of nitrogen species successfully with proper calibration with more measured data.