Rates of denitrification and nitrous oxide (N 2 O) formation, and the sources of N 2 and N 2 O, were examined by the isotope-pairing technique in three U.K. estuaries (Colne, Humber, Conwy), which ranged from extremely nutrified to oligotrophic. Nitrous oxide concentrations were supersaturated throughout the year with annual averages of 603% in the Colne, 158% in the Humber, and 133% in the Conwy, indicating that the estuaries were sources of atmospheric N 2 O. Denitrification and N 2 O formation were from benthic processes, and not from water-column processes. Generally, both denitrification and N 2 O formation decreased down the estuary as nitrate concentrations lowered. The proportion of N 2 and N 2 O derived from nitrate in the overlying water column (D w ) also decreased with nitrate concentration, while that from benthic coupled nitrification-denitrification (D n ) increased. Scaled to the total estuary area, in the hypernutrified Colne, water-column nitrate was the main source of N 2 and N 2 O; in the moderately nutrified Humber, D w and D n contributed approximately equally, while in the oligotrophic Conwy, coupled nitrification-denitrification (D n ) was the main source of N 2 and N 2 O. For the first time, the formation of N 2 O from either the nitrification or denitrification steps of D n was also determined. In the Colne, that from the nitrification step predominated at the top of the estuary but decreased down the estuary, while in the Humber that from denitrification dominated at the top and decreased down the estuary. In the oligotrophic Conwy, there were approximately equal contributions.Bacterial denitrification in estuarine sediments, the reduction of nitrate to dinitrogen (N 2 ) and nitrous oxide (N 2 O), has been identified as a process leading to nitrate removal from estuaries. Therefore the capacity of an estuary to remove anthropogenic inputs of nitrate is important in modifying the load of nitrate to coastal waters (Nedwell and Trimmer 1996;Barnes and Owens 1998;Trimmer et al. 1998;Nedwell et al. 1999; Dong et al. 2000a,b). Rates of denitrification vary considerably in different estuaries, at different sites within an estuary, and at different times of the year. As a result, the extent of nitrate removal by estuarine denitrification is variable from less than 1% to 100% of total nitrate input into an estuary (Nielsen et al. 1995;Ogilvie et al. 1997a;Barnes and Owens 1998;Trimmer et al. 1998; Dong 1 Corresponding author (nedwd@essex.ac.uk).
AcknowledgmentsWe thank John Green for excellent technical support in the field, and we are grateful to J. Allen (University of Hull), B. Emmett (CEH, Bangor), and P. Williams (University of Wales, Bangor) for their kindness in letting us use their laboratories for our field work in the Humber and Conwy estuaries. We thank two anonymous reviewers whose comments substantially improved an early version of this paper.