Error assessment of nitrogen and oxygen isotope ratios of nitrate as determined via the bacterial denitrification method

Currently, bacterial denitrification is becoming the accepted method for δ¹⁵N‐ and δ¹⁸O‐NO determination. However, proper correction methods with international references (USGS32, USGS34 and USGS35) are needed. As a consequence, it is important to realize that the corrected isotope values are derived from a combination of several other measurements with associated uncertainties. Therefore, it is necessary to consider the propagated uncertainty on the final isotope value. This study demonstrates how to correctly estimate the uncertainty on corrected δ¹⁵N‐ and δ¹⁸O‐NO values using a first‐order Taylor series approximation. The bacterial denitrification method errors from 33 batches of 561 surface water samples varied from 0.2 to 2.1‰ for δ¹⁵N‐NO and from 0.7 to 2.3‰ for δ¹⁸O‐NO, which is slightly wider than the machine error, which varied from 0.2 to 0.6‰ for δ¹⁵N‐N₂O and from 0.4 to 1.0‰ for δ¹⁸O‐N₂O. The overall uncertainties, which are composed of the machine error and the method error, for the 33 batches ranged from 0.3 to 2.2‰ for δ¹⁵N‐NO and from 0.8 to 2.5‰ for δ¹⁸O‐NO. In addition, the mean corrected δ¹⁵N and δ¹⁸O values of 132 KNO₃‐IWS (internal working standard) measurements were computed as 8.4 ± 1.0‰ and 25.1 ± 2.0‰, which is a slight underestimation for δ¹⁵N and overestimation for δ¹⁸O compared with the accepted values (δ¹⁵N = 9.9 ± 0.3‰ and δ¹⁸O = 24.0 ± 0.3‰). The overall uncertainty of the bacterial denitrification method allows the use of this method for source identification of NO. Copyright © 2010 John Wiley & Sons, Ltd.