EVALUATION OF NITRATE-N FLUXES FROM A TILE-DRAINED WATERSHED

Intensively cropped watersheds of the Midwestern US have been identified as important sources of nutrients within the Mississippi river basin that contribute to hypoxic conditions in the Gulf of Mexico. Midwestern watersheds have been artificially drained to a large extent, to allow settlement and modern agricultural use of the land. Despite the importance of tile drainage and our need to understand nutrient flux dynamics in Midwest watersheds, there are few long-term data sets that summarize nutrient fluxes from tile-drained watersheds in this region. Actual NO3-N loads from small-order watersheds would be particularly helpful, because understanding the distribution and timing of fluxes helps us evaluate the potential to modify them through Best Management Practices (BMPs). In this study, we evaluated discharge and NO3-N fluxes from a central Iowa watershed (5134 ha) occurring from mid-1992 through 2000. The watershed outlet and two tile-drained sub-basins of 493 and 863 ha were included in the analyses. About 85% of the watershed is in row crop (corn/soybean) production, with no significant animal production or land applications of manure.Continuous flow monitoring and change-in-flow-triggered automatic sampling methods were implemented; sampling intervals ranging from 0.25 hour to a maximum of 7 days. Periods with ice in the stream were omitted due to icing effects on the instruments and flow calibrations. Sample numbers (n) were 1182 and 1454 at the tiled sub-basins, and 2276 at the outlet, which represented flow durations between 2098 and 2401 days. The total NO3-N load for the entire record was 168 kg ha−1 from the watershed outlet, and 176 and 229 kg ha−1 from the sub-basins. The outlet had greater total discharge (1831 mm) and smaller flow-weighted mean NO3-N concentration (9.2 mg L−1) than the sub-basins, while the larger sub-basin had greater discharge (1712 vs 1559 mm) and mean NO3-N concentration (13.4 vs 11.4 mg L−1) than the smaller subbasin. Concentrations exceeding 10 mg L−1 were common, occupying 71 to 81% of the flow record at the sub-basins, but only 31% of the record at the outlet. Fluxes of NO3-N exceeding 0.1 kg ha−1 d−1 occupied 23 to 31% of the flow record, but exported 74 to 79% of the total NO3-N load.Relationships were established between water flows and NO3-N fluxes by aggregating data across 7-week periods to minimize autocorrelation effects. Results indicate that NO3-N was generally not diluted by large flows, except during 1993 flooding. The outlet, in fact, showed smaller NO3-N concentrations during low flows. The discharge - NO3-N flux relationships showed log-log slopes near 1.0 for the sub-basins, and 1.2 for the outlet. The difference occurred because, presumably, in-stream processes (assimilation, denitrification), and/or dilution by denitrified baseflow decreased NO3-N concentrations and fluxes between the tiled sub-basins and the outlet, with the largest effect occurring during low flows. Flow contributions from non-cropped areas near the outlet would also have contributed to the smaller concentrations observed there.We estimated denitrification of sub-basin NO3-N fluxes in a hypothetical wetland using published data relating denitrification rates and water temperature. We then assumed that average long-term soil temperatures could be used to estimate the temperatures of tile drainage waters. If both temperature and NO3-N inflow rate (from the tile outlets) could limit denitrification, then about 20 of the NO3-N load would have been denitrified by a wetland constructed to meet USDA-approved criteria under Iowa's Conservation Reserve Enhancement Program. The low fraction largely results from non-dilution of NO3-N by large flows, and poor timing of NO3-N fluxes relative to warm summer months when denitrification rates would be optimal. Results indicate that wetland BMPs would not achieve water quality goals in this tiledrained watershed without improved management of agricultural nitrogen for greater crop-use efficiency.