Seawater samples were collected aboard the R/V S.A. Agulhas II (from CTD Niskin bottles and the underway intake) and immediately frozen at \u221220\u00b0C until analysis. Hydrographic data for each profile station were acquired from a conductivity-temperature depth (CTD) sensor mounted on the Niskin bottle rosette.
\n
An aliquot of each seawater sample was treated with sulfamic acid to remove nitrite (Granger & Sigman, 2009). Both nitrate+nitrite and nitrate-only samples were measured for concentration by chemiluminescence (Braman & Hendrix, 1989) and isotope composition using the denitrifier method (Sigman et al., 2001). Briefly, sample nitrate or nitrate+nitrite is converted to N\u2082O gas by denitrifying bacteria and measured for N isotope composition by GC-IRMS with an on-line N\u2082O extraction and purification system (Sigman et al., 2001; Casciotti et al., 2002; Weigand et al., 2016). The pooled standard deviations (1\u03c3) of replicate denitrifier measurements was 0.09\u2030 for nitrate+nitrite \u03b4\u00b9\u2075N and 0.06\u2030 for nitrate-only \u03b4\u00b9\u2075N.
\n
Hydrographic profiles were acquired by a Sea-Bird conductivity-temperature depth sensor during each cast. The isotopic composition of N\u2082O was measured at Princeton University by GC-IRMS using a Thermo MAT 253 mass spectrometer with a purpose-built on-line N\u2082O extraction and purification system.
\n
See methods section of Smart et al. (2020) for more detail.
Isotope ratio measurements were referenced against automated injections from an N\u2082O from a gas cylinder, and then calibrated to N\u2082 in air using two international reference materials IAEA-N3 and USGS-34 that were included in every run. An in-house N\u2082O standard was also measured along with each batch of samples to monitor mass spectrometry.