Natwora, Kaela Elizabeth; Sheik, Cody Springer: Nitrogen fixation may provide a significant yet under quantified source of Nitrogen in the Great Lakes [dataset]. PANGAEA, https://doi.pangaea.de/10.1594/PANGAEA.920001 (dataset in review)

Nitrogen fixation (NFix) is an important yet understudied microbial process in aquatic ecosystems and especially in the Laurentian Great Lakes (LGL). Early work suggested the contribution of NFix in the LGL is minimal to the nitrogen budget. However, recent work has shown during bloom events, NFix can help alleviate nitrogen limitations. Thus, we sought to revisit NFix in the LGL and comprehensively sample from near and offshore stations and with depth to understand the spatial variability of NFix.

Nitrogen fixation rates were quantified using an adapted acetylene reduction assay (Stewart et al. 1967; doi:doi:10.1073/pnas.58.5.2071). Water samples were collected from discrete depths (Table 1) using a CTD-rosette water sampler equipped with 12 - 8L Niskin bottles and outfitted with a Sea-Bird multi-parameter profiler (conductivity, temperature and depth). Two liters of water were collected in triplicate from each depth and concentrated onto a 0.22µm pore size MF- Millipore Membrane filter using a vacuum manifold at low pressures to minimize cell breakage. Given the vast difference in lake depths between lakes and even within lakes, at each sampling station we took three depths, a surface (5m) sample (constant at all stations), mid- water column (based on max station depth) and near-bottom water (5m from bottom) (SI Table 1). The mid-water collection did not necessarily coincide with the thermocline, if it was even present. In preliminary testing in Lake Superior, we determined that approximately 2L of water was necessary to capture enough biomass to quantify rates in oligotrophic waters. In areas of high biomass, such as Lake Erie, water was filtered until the filter was clogged, which were all below 1L. Filtrate volume was noted, and final nitrogen fixation values were corrected based on volume filtered. Following filtration, filters were transferred to 50mL serum vials, submerged in 25mL of filtrate (lake water) from the same depth it originated from, and capped with a Teflon stopper. Acetylene gas was generated shipboard prior to sampling by combing 1g calcium carbide (CaC2) and 100ml of deionized water in a side arm flask attached to a 1L Tedlar gas sampling bag (EnviroSupply & Service). Samples were spiked with 1mL of acetylene gas and incubated at near in situ conditions (temperature and light) in the lab for 24 hours. For deep samples, incubations were done in the dark. After 24 hours, the incubations were terminated using 5mL of trichloroacetic acid (TCA) and stored in the dark at 4°C until measurements could be made in the lab. Acetylene gas concentration (peak area integration) was measured using an Agilent 6890 Plus GC System equipped with a flame deionized detector (FID) and a GS-Carbon Plot column 100/120 mesh (Agilent 113-3122). The GC-FID parameters are as follows: the carrier gas He2 was set at a flow rate of 30cm/s, the oven temperature was held isothermally at 125°C for 3 minutes, with an a split injection of 1:20 at 250°C. Sample injection volumes were 100 µL taken from the headspace of the vial. The retention times for acetylene and ethylene were observed at 1.8 and 1.9 minutes, respectively. For each station and at each depth, the following series of blanks and controls were used: 1) kill standard, 2) acetylene + DI water 3) acetylene + filtrate. Peak area integrations were calculated at retention times 1.9, and a correction ratio of 1:4 of N2 fixed to ethylene formed were factored in to determine molar N2 fixation rates (nmol/L/day) (Peterson and Burris 1976; doi:10.1016/0003-2697(76)90187-1).