Sampling and analytical procedures:

The concentration of total nitrogen (TN) for samples was determined by persulfate oxidation (Solorzano and Sharp, 1980) with adaptations (Knapp et al., 2005), and the resulting NO3- was measured by chemiluminescence (Braman and Hendrix, 1989). DON concentration was determined by subtracting the concentrations of PNsusp, NH4+, and NO3-+NO2- (reported in Berthelot et al., 2015, Biogeosciences) from the measured TN concentration of each sample with a propagated error of +/- 0.5 μM. The d15N of the resulting NO3- was measured using the denitrifier method (Casciotti et al., 2002; McIlvin and Casciotti, 2011; Sigman et al., 2001).

The d15N of PNsink was measured using a Thermo Scientific Flash 2000 Elemental Analyzer coupled with a Delta Plus Thermo Scientific mass spectrometer.

The average standard deviation for individual DON concentration measurements was +/- 0.3 µM. DON concentration was determined by subtracting the concentrations of PNsusp, NH4+, and NO3-+NO2- (reported in Berthelot et al., 2015, Biogeosciences) from the measured TN concentration of each sample with a propagated error of +/- 0.5 μM.

The d15N of TN was determined via persulfate oxidation of TN to NO3- (Knapp et al., 2005) and subsequent analysis of NO3- d15N by the denitrifier method, with a propagated error for DON d15N calculated using a Monte Carlo method (Press et al., 1992) of +/- 0.6‰. Samples were calibrated with IAEA N3 and USGS 34 NO3- d15N isotopic reference materials as described in McIlvin and Casciotti, 2011.

The d15N of PNsink was measured using a Thermo Scientific Flash 2000 Elemental Analyzer coupled
with a Delta Plus Thermo Scientific mass spectrometer. The average standard deviation for standards was +/- 0.06‰.