Single water samples for N2O analyses were collected from acid-cleaned, Teflon-coated X-Niskin sampling bottles (OceanTest Equipment) of 8-liter capacity deployed on an epoxy-coated rosette (Sea-Bird Electronics) as described in Noble et al. (2012). Immediately after rosette recovery, water samples were collected by twice overfilling 160 ml glass serum bottles (Wheaton prod. no. 223748) from the bottom up using tygon tubing attached to each niskin bottle. A small headspace was introduced into each sample by pipetting 1 ml of sample off the top. The sample was then poisoned with 100 ul of saturated HgCl2 solution and sealed with gray butyl septa (MicroLiter Analytics prod. no. 20-0025) and aluminum crimps. Poisoned samples were stored for several months in the dark at room temperature before analysis.

Isotopic analyses of N2O were made using a Finnigan DELTA PLUS XP IRMS calibrated for isotopomer-specific measurements (see Frame & Casciotti (2010) for details) and referenced against injections of a calibrated N2O reference tank (d15Nbulk = 0.31‰, SP = -1.2‰, d18O = 40.85‰ as calibrated by S. Toyoda, Toykyo Tech) with quality control checks against analyses of tropospheric air samples. Bottles were purged with He and N2O was trapped on-line with a custom-built purge and trap system (McIlvin & Casciotti, 2010). The molar amount of N2O was determined from each bottle using a constant linear relationship of m/z 44 peak area to molar quantity of N2O which is 2.15 Vs per nanomole of N2O, then converted to a concentration by dividing the molar amount by the volume of sample extracted (154 ml). The ratios of sample N2O to reference gas N2O values of m/z 31/30, 45/44, and 46/44 were corrected for a small linear dependence of these values on the sample peak area (see McIlvin and Casciotti 2010), with all ratios corrected to a peak area of 20 Vs. For a 160 ml water sample containing 2.2 nmole of N2O (typical of the lowest concentrations in this dataset), the precision of the d18O = ±0.25‰, d15Nbulk = ±0.09‰, and SP = ±0.33‰. Additional details about measurement precision are included in McIlvin & Casciotti (2010). Triplicate samples of tropospheric N2O from Woods Hole, MA had _d15Na = 15.0 ± 0.1‰, _d15Nb = −1.9 ± 0.1‰, d_18O = 44.4 ± 0.2‰, _d15Nbulk = 6.5 ± 0.1‰, SP = 16.9 ± 0.1‰, and m/z 44 peak area = 15.6 ± 0.2 Vs. All cruise data are included here except two measurements with background interference.