Triplicate cores were collected on 3 occasions: August 2010, June 2011, and August 2011. Samples of the tidal creek bed were taken by hand using clear polyvinyl chloride cores (5 cm diameter × 32 cm tall). The cores were inserted directly into the creek bed sediment at low tide to a depth of ~15 cm, while keeping the sediment surface undisturbed. The cores were then stored in coolers to keep them cool and dark. In the same creeks, water was collected at high tide, filtered to 0.2 μm, and stored in 20 l carboys. Both cores and filtered site water were transported back to Boston University for incubation, where they were placed in a water bath in an environmental chamber set to ambient creek water temperature.

Cores were left in the water bath overnight with air gently bubbling the overlying water. In the morning, this water was carefully siphoned out and replaced with the filtered site water. The cores were filled to the top and then fit with gas-tight lids without any air headspace or bubbles. A magnetic stir bar (~30 revo- lutions min−1) on the core tops allowed for enough water movement to prevent stratification within the core but not enough to disturb the sediment surface. Cores were incubated in the dark. Replicate water samples were collected in gas-tight 12 ml Exetainer vials (Labco) for later analysis of dissolved N2 and Ar. These samples were taken at 5 time points until the dissolved oxygen in the cores dropped at least 2 mg l−1 (62.5 μmol O2 l−1, typically 6 to 8 h). Dissolved oxygen concentrations were measured using a luminescent dissolved oxygen meter (Hach) at each time point. Once the desired drop in oxygen was attained, final samples were taken, and the incubation was stopped before the dissolved oxygen level reached 2 mg l−1 (hypoxia). The caps were then removed from the cores, and the overlying water was gently bubbled with air overnight. The next morning, the water was carefully siphoned off and replaced with filtered site water. Again, we capped the cores with a gas-tight lid and no air headspace for the second incubation for nutrient fluxes. Water samples for nutrient analysis were also collected at 5 time points until the oxygen dropped by at least 2 mg l−1. Samples for NO3−, NO2−, NH4+, and dissolved inorganic phosphorus (DIP) analysis were filtered through a 0.7 μm glass fiber filter (Whatman GF/F), collected in 30 ml acid washed, polypropylene bottles that had been leached with Milli-Q water, and stored frozen until analysis. At the end of the incubations we removed sub-samples of sediment from each core (down to 1 cm in 0.5 cm increments) for sediment density, porosity, and chlorophyll a (chl a) analysis. These samples were also frozen prior to analysis.

Water samples were analyzed for dissolved N2 and Ar using the N2/Ar technique on a quadrupole membrane inlet mass spectrometer against an air- equilibrated, deionized water standard at constant in situ temperature (Kana et al. 1994). This method requires no sample preparation and only a small sample size (<10 ml) and allows rapid sampling (20 to 30 samples h−1) with a precision of < 0.03 % for N2 /Ar. This method gives net N2 fluxes only. That is, it is a measurement of total denitrification minus total N fixation, and specific N transformation pathways cannot be distinguished. For this reason, positive N2 fluxes indicate net denitrification, while negative fluxes indicate net N fixation.
Dissolved inorganic nitrogen (DIN) and DIP sam- ples were analyzed colorimetrically on a SEAL Auto- analyzer 3 using standard colorimetric techniques (Strickland & Parsons 1968, Grasshoff et al. 1983). The practical detection limits of the SEAL Autoanalyzer 3 are 0.247, 0.066, 0.013, and 0.005 μM for ammonium, nitrate, phosphate (PO43 –), and nitrite, respectively (K. Czapla pers. comm.). We also routinely compare our laboratory standards to external standards (OSIL environmental instruments and systems) for each nutrient species; this comparison shows that analysis of each species has a precision of < 4.0 %.

We calculated fluxes from the sediment by running a linear regression through each of the 5 incubation sample points plotted against time. Any flux regressions with an r2 > 0.65 were considered significant (Prairie 1996). The slope of the linear regression was then multiplied by the volume of overlying water in the core and divided by the core area to arrive at a flux of either dissolved gas or nutrients in μmol m−2 h−1. Positive flux values indicate a release from the sediment, while negative values signify sediment uptake. (From: Vieillard and Fulweiler (2012) Marine Ecology Progress Series 147: 11-22. DOI:10.3354/meps10013).