Sediment denitrification in waterways in a rice-paddy-dominated watershed in eastern China

Abstract

Rice-paddy-dominated watersheds in eastern China are intensively cultivated, and lands with two crops receive as much as 550–600 kg ha–1 year–1 of nitrogen (N), mainly through the addition of N-based fertilizers. However, stream N concentrations have been found to be relatively low. Waterways in the watersheds are assumed to be effective “sinks” for N, minimizing its downstream movement. We directly measured net sediment denitrification rates in three types of waterways (ponds, streams/rivers, and a reservoir) and determined the key factors that control net sediment denitrification. Such information is essential for evaluating the impact of the agricultural N cycle on the quality of surface water. The pond–stream–reservoir continuum was sampled every 2 months at nine sites in an agricultural watershed between November 2010 and December 2011. Net sediment N2 fluxes/net sediment denitrification rates were determined by membrane inlet mass spectrometry and the N2/Ar technique. A suite of parameters known to influence denitrification were also measured. Net denitrification rates ranged between 28.2 ± 18.2 and 674.3 ± 314.5 μmol N2–N m–2 h–1 for the streams, 23.7 ± 23.9 and 121.2 ± 38.7 μmol N2–N m–2 h–1 for the ponds, and 41.8 ± 17.7 and 239.3 ± 49.8 μmol N2–N m–2 h–1 for the reservoir. The mean net denitrification rate of the stream sites (173.2 ± 248.4 μmol N2–N m–2 h–1) was significantly higher (p < 0.001) than that of the pond sites (48.3 ± 44.5 μmol N2–N m–2 h–1), and the three types of waterways all had significantly higher (p < 0.01) mean net denitrification rates in summer than in other seasons. Linear regression and linear mixed effect model analysis showed that nitrate (NO3 −–N) concentration in surface water was the primary controlling factor for net sediment denitrification, followed by water temperature. Using monitoring data on NO3 −–N concentrations and temperature of the surface water of waterways and an established linear mixed effect model, total N removed through net sediment denitrification in the pond–stream–reservoir continuum was estimated at 46.8 ± 24.0 t year–1 from July 2007 to June 2009, which was comparable with earlier estimates based on the mass balance method (34.3 ± 12.7 t year–1), and accounted for 83.4 % of the total aquatic N. However, the total aquatic N was only 4.4 % of the total N input to the watershed, and thus most of the surplus N in the watershed was likely to be either denitrified or stored in soil. High doses of N in a rice-paddy-dominated watershed did not lead to high stream N concentrations due to limited input of N into waterways and the high efficiency of waterways in removing N through denitrification.