2.1. Study area
The Gulf of Riga, situated in the southeastern part of the Baltic Sea, is an ecologically unique subbasin of the Baltic Sea due to the significant impact of the river runoff. This brackish water inlet is characterized by its shallow depth, fluctuating salinity levels, and low water residence time influenced by both marine and freshwater inputs. In the context of this study, a location at a depth of 44 meters in the central region of the Gulf of Riga (Fig. 1; map made with ArcMap 10.6) was selected to minimize the direct effects of river discharge or coastal processes.
The subbasin of the Gulf of Riga is relatively eutrophied and its phytoplankton seasonal succession closely mirrors the general pattern observed in temperate coastal waters (Jurgensone et al. 2011; Labucis et al. 2017; Purina et al. 2018). This succession features a spring bloom of diatoms, followed by dinoflagellates and ciliates in later stages. During the summer, the dominant phytoplankton species shifts to cyanobacteria, notably A. flosaquae, often cohabiting with chlorophytes and cryptophytes. The summer bloom is followed by a second peak of diatoms in the autumn period, as noted in previous studies.
2.2. Sampling procedure
The surface water (0.1 m) was collected throughout the period of summer bloom from July to August of 2022 with the Latvian Institute of Aquatic Ecology Vessel “Ronis2” and used to estimate N2-fixation rates under different cyanobacteria biomasses and proportions of N2 fixing species.
To reflect on the continuously changing abiotic conditions of the Gulf and the responses of diazotrophic cyanobacteria to them, specifically DIN:DIP ratio and cyanobacteria biomass, additional water volume for DIP enrichment experiments was acquired during three distinct environmental conditions (Fig. 2). For Experiment No. 1, seawater for incubations was obtained under conditions where the DIN:DIP ratio exceeded 16, making it unfavorable for the growth of diazotrophic cyanobacteria. For Experiment No. 2, seawater was collected when the DIN:DIP ratio was favorable, although the cyanobacterial biomass was low. For Experiment No. 3 the collection of seawater was done when the DIN:DIP ratio was favorable and the phytoplankton community was dominated by cyanobacteria.
The water samples were collected using a Van Dorn water sampler. Upon sampling, the sub-samples for phytoplankton analysis were fixed with acid Lugol’s solution and transported to the laboratory for further analysis based on HELCOM Monitoring Guidelines (HELCOM 2021). The surface water for nutrient analysis and incubation experiments was transported to the laboratory for further treatment in a coolbox.
2.3. DIP enrichment mesocosm and N2-fixation rate incubation experiments
For DIP indoor aquatic mesocosm incubation experiments 25 L of surface water was collected and brought to the laboratory. Three parallel DIP enrichment treatment incubations (7 L each) were set up in glass tanks: control with no treatment (C) and enrichment with DIP, by adding 5 mL of 10 µmol phosphate stock solution at the start (0 h) of the incubation (DIP5) and at the start and after 24 hours of the incubation (DIP10). Incubation experiments were carried out in a thermochamber (IL3-25A, Jeiotech Lab Companion) equipped with LED lamps (Commel 406 − 202, 8W, 6500K) set to 550 µmol photons m− 2 s− 1. Subsamples of treated seawater were collected 3 times during the 48 h incubation for the determination of cyanobacteria biomass and heterocyst abundance (250 mL), as well as for DIN and DIP analysis (1 L) for all runs.
For the N2-fixation rate determination, the remainder of each sample from indoor aquatic mesocosm incubation experiments was transferred to 500 mL Duran glass bottles. Thereafter, the bottles were sealed with bromobutyl rubber stoppers and 50 mL of the seawater was replaced with 50 mL of MiliQwater enriched with 15N2 through an injection. The enriched MiliQ water used to spike incubation bottles was pre-prepared based on adjusted methodological recommendations by Klawonn et al. (2015). For each sampling occasion, 900 mL of degassed MiliQ water (sonicated for 1 h, 60°C) was transferred to a Tedlar gas sampling bag (push lock valve, 1 L) and spiked with 9 mL of 15N2 gas (Sigma-Aldrich, 98 atom% 15N). The filled gas sampling bag was left for 24 h at 8°C before being used to dissolve the 15N2 gas bubble in the MiliQ water. After enriching the seawater, the bottles were placed back in the thermochamber under 550 µmol photons m− 2 s− 1.
The samples were then incubated for 12 h. After incubation, three replicates (50–100 mL each) from each bottle and of the bulk volume used to fill incubation bottles were filtered on a pre-combusted (500°C for 2 h) 13-mm diameter GF/F filter for isotopic signature analysis with an elemental analyzer (EuroEA-3024, EuroVector S.p.A, Italy) coupled with a continuous flow stable isotope ratio mass spectrometer (Nu-HORIZON, Nu Instruments Ltd, UK). The isotope ratio mass spectrometry analysis (IRMS) was performed in the Laboratory of Analytical Chemistry, University of Latvia. Isotope ratios were reported relative to atmospheric nitrogen for δ15N as parts per thousand (‰).
R software v.3.6.1 (R Core Team 2019) was used for data visualization and analysis of the data collected during sampling and after the mesocosm incubation experiments. To assess statistically significant differences between treatments in each experiment, a t-test was conducted. An overview of the relationship between biological variables associated with diazotrophic activity was provided based on Spearman's rank correlation with the coefficient of significance set at α = 0.05.
2.4. Phytoplankton community composition and nutrient analysis
To determine the composition of the phytoplankton community, 10 mL subsamples were examined using an inverted microscope. Individual cells were counted following the HELCOM Monitoring Guidelines (HELCOM 2021). Subsamples were allowed to settle in a sedimentation chamber for 12 hours and were then counted at magnifications of 200x and 400x, as per Uthermöl (1958). The total count in all subsamples exceeded 500. The wet weight of the phytoplankton biomass was expressed in mg m− 3 and computed by following Olenina et al. (2006). Phytoplankton organisms were identified to the most specific taxonomic rank, and their scientific names and classifications adhered to the accepted binomial nomenclature of the World Register of Marine Species (version 2021). The specific growth rate of cyanobacteria was calculated using the results obtained from three data points collected during the 48-hour incubation, based on the equation:
$$\mu =\frac{\text{ln}\left({N}_{2}\right)-\text{l}\text{n}\left({N}_{1}\right)}{{t}_{2}-{t}_{1}}$$
where µ is the specific growth rate, and N1 and N2 are the biomass of A. flosaquae or N. spumigena at the onset of the experiment (t1) and subsampling time (t2), respectively.
The DIN as sum of ammonia, nitrite and nitrate with a limit of detection (LOD) of 0.1 µmol L− 1 and DIP with a LOD of 0.01 µmol L− 1 were determined according to Grasshoff and Ehrhardt (1999). DIP and ammonium were determined using the molybdenum blue and indophenol blue methods, respectively. Nitrite and nitrate were measured after reduction to nitrite in a copper-coated cadmium column, followed by the nitrite reaction with an azo dye.
2.4. Calculation of N2 fixation rate
The 15N atom% in the enriched samples was expressed as the sum of added and naturally abundant 15N following the same calculation steps described in detail by Liepina-Leimane et al. (2022). First, the total of dissolved N2 in the sample was determined by employing the seawater temperature, salinity and N2 solubility coefficients (A0, A1, A2, A3, B0, B1 and B2) provided by Hamme and Emerson (2004):
lnC = A0 + A1TS + A2T2S + A3T3S + SB0 + B1TS + B2T2S
Here C represents the N2 concentration at equilibrium with the atmosphere, while T is the temperature (°C) and S is the salinity of the collected sample. Subsequently, the nitrogen isotope ratio in the atmosphere (15 N/14 N = 0.366 atom%) was used to determine the 15N content in the sample. Finally, the N2 solubility formula was used to calculate 15N in the 50 mL of enriched water, presumed to exclusively contain 15N2 due to degassing before enrichment. It is essential to acknowledge that a theoretical estimation of 15N2 gas dissolution may lead to some underestimation of rates (White et al. 2020).
Following incubation, in order to utilize a mass balance approach and determine N2-fixation rate as described by Montoya et al. (1996), the δ15N values were then converted to absolute abundance ratio A (15N atom%) of the 15N enrichment in particulate N:
$$A=100*\frac{({10}^{-3}{\partial }^{15}N+1{\left){(}^{15}N{/}^{14}N\right)}_{atmosphere}}{1+({10}^{-3}{\partial }^{15}N+1{\left){(}^{15}N{/}^{14}N\right)}_{atmosphere}}$$
The remaining incubated sample in the Duran bottles was used to filter a second batch of aliquot subsamples on pre-combusted (at 500°C for 2 hours) 24 mm diameter GF/F filters for N% analysis (Elementar Vario El III). As a last step, the bulk seawater was filtered on a nitrocellulose membrane (Millipore, 45 mm diameter, 0.45 µm pore size) to determine suspended particulate matter (SPM). N2-fixation rate (NFR, nmol N L⁻¹ h− 1) was then estimated using the following equation (Montoya et al. 1996):
$$NFR=\frac{{A}_{PN}-{A}_{PN\left(0\right)}}{{A}_{N2}-{A}_{PN\left(0\right)}}*\frac{PN}{\varDelta t}$$
with the variables defined as follows: APN(0) is the 15N enrichment of particulate N before the incubation, APN is the 15N enrichment of particulate N after the incubation, AN2 represents the sum of naturally abundant and added 15N, and PN represents particulate nitrogen concentration at the end of the incubation period calculated based on the corresponding samples N% and SPM expressed as nmol N L⁻¹.