Argon Analysis in Surface Waters in the

Subscriber access provided by Bethel University

Environmental Measurements Methods

Potential of Nitrogen/Argon Analysis in Surface Waters in the Examination of Areal Nitrogen Deficits Caused by Nitrogen Fixation Oliver Schmale, Mattis Karle, Michael Glockzin, and Bernd Schneider Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b06665 • Publication Date (Web): 22 May 2019 Downloaded from http://pubs.acs.org on May 30, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25

Environmental Science & Technology

1

Potential of Nitrogen/Argon Analysis in Surface Waters in the Examination of

2

Areal Nitrogen Deficits Caused by Nitrogen Fixation

3

Oliver Schmale*, Mattis Karle, Michael Glockzin, and Bernd Schneider

4

*Correspondence to:

5

Oliver Schmale (oliver.schmale@io-warnemuende)

6

Leibniz Institute of Baltic Sea Research (IOW), Seestrasse 15, D-18119 Rostock

7

1

ACS Paragon Plus Environment


Page 2 of 25

8

ABSTRACT

9

In marine systems, the loss of nitrogen caused by denitrification in oxygen-deficient zones is

10

balanced by nitrogen fixation mediated by cyanobacteria, which may form extensive blooms in

11

surface waters. In this study, by determining the concentration ratio of nitrogen (N 2) and argon

12

(Ar) in air equilibrated with surface water we were able to detect changes in the N2 concentration

13

attributable to N2 fixation. For this purpose, surface water was pumped continuously into a spray-

14

type equilibrator while the air in the equilibrator’s headspace was analyzed by mass

15

spectrometry. After laboratory tests and model analysis to evaluate the sensitivity of our N2/Ar

16

approach, feasibility studies were conducted in the central Baltic Sea in the summer of 2015,

17

during the development of a cyanobacterial bloom. Our results showed that N2 deficits

18

accumulated during periods of low wind and increasing surface water temperatures. A

19

comparison of our results with the N2 deficits calculated from changes in the partial pressure of

20

carbon dioxide in surface water indicated a similar trend. By demonstrating the ability of the

21

N2/Ar approach to resolve N2 deficits in surface water caused by N2 fixation, our study

22

contributes to assessments of the N2 fixation efficiency of cyanobacterial blooms.

23

24 25 2


Page 3 of 25


26

INTRODUCTION

27

The concentrations of dissolved gases in seawater are controlled by physical and biological

28

processes. The physical processes include temperature-, salinity-, and wind-driven gas exchange

29

with the atmosphere, bubble injection, and the interaction between different water bodies through

30

mixing. In marine biogeochemistry, the inert character of argon (Ar) is used to separate these

31

physical effects from the biological activities that affect gas concentrations. Because oxygen (O2)

32

and Ar have similar physicochemical properties, dissolved O2/Ar ratios can be employed to

33

estimate net community production and to resolve regional phenomena in the surface water

34

linked to photosynthesis and respiration.1, 2 Biological processes involved in the marine nitrogen

35

cycle and associated with changes in the concentration of molecular nitrogen (N2) include

36

denitrification, ammonium oxidation (anammox reaction), and N2 fixation.3 Analogous to

37

oxygen-related processes, the molecular nitrogen (N2) excess resulting from water column

38

denitrification and the anammox reaction in oxygen-deficient zones was estimated by measuring

39

N2/Ar ratios .4, 5

40

In the ocean, the loss of dissolved inorganic nitrogen (DIN) compounds due to denitrification and

41

anammox reactions is mainly balanced by N2 fixation, i.e., the conversion of molecular into

42

organic nitrogen, in reactions carried out by diazotrophic organisms. Most of the N2 fixation in

43

the central Baltic Sea and Gulf of Finland can be attributed to the cyanobacteria Nodularia

44

spumigena, Aphanizomenon spec. and Dolichospermum spec. 6, which form annually recurring

45

blooms from July to the beginning of August. There is general concern that eutrophication and

46

climate change will favor the proliferation of these harmful blooms.7 Consequently, an

47

understanding of their development is essential for the ecological management of the Baltic Sea

48

and other coastal areas. 3



Page 4 of 25

49

Among the methods used to estimate the contribution of N2 fixation, and therefore the activity of

50

bloom-forming cyanobacteria, to the nitrogen (DIN) budget of the Baltic Sea are: (i) the

51

incubation of discrete water samples with 15N-labeled molecular nitrogen to obtain instantaneous

52

fixation rates 6, 8, (ii) the establishment of basin-wide mass balances for total nitrogen 9, 10, (iii) the

53

deployment of measuring devices by voluntary observing ships (VOS) to obtain high-resolution

54

data on the carbon dioxide partial pressure (pCO2) in surface water, thus enabling determinations

55

of net community production and estimates of the nitrogen demand.11, 12

56

In this study, we demonstrate that the small changes in the N2 concentration arising from N2

57

fixation can be detected by a novel method based on mass spectroscopy of the N2/Ar ratio in air

58

equilibrated with surface water. To this end, a continuous-mode equilibrator system coupled to a

59

mass spectrometer (Equi-MS) was developed and its performance optimized and evaluated in

60

laboratory experiments. Then, using a simple physico-chemical model, we assessed the potential

61

and limitations of the N2/Ar approach for studies of the biogeochemical processes that decrease

62

or increase the N2 concentration. Finally, the feasibility of our N2/Ar approach in the detection of

63

N2 deficits in surface water caused by N2 fixation was tested in a field study in the central Baltic

64

Sea. The results were compared with those of an established approach for the quantification of N2

65

fixation capacities.

66

67

MATERIAL AND METHODS

68

Fundamentals and limitations of the N2/Ar method

4


Page 5 of 25


69

The N2/Ar method was applied to estimate N2 fixation by measuring the surface water

70

concentration of N2. As noted above, this same method is used to estimate the N2 excess caused

71

by denitrification (e.g. 4). Given the similar physico-chemical properties of N2 and Ar, the N2/Ar

72

concentration ratios in seawater correspond within narrow limits to saturation with atmospheric

73

N2 and Ar (denoted as (N2/Ar)sat), at the respective temperature and salinity as long as no N2

74

transformations occur

75

concentration will be reflected in a shift of the N2/Ar ratio and can be determined by multiplying

76

the difference in the N2/Ar ratio by the Ar concentration 14, as shown in Eq. (1):

77

∆𝑁2 = ((𝐴𝑟2 ) − (𝐴𝑟2 )

78

where the element symbols represent the concentrations of those elements.

79

However, the use of (N2/Ar)sat as a non-biotic reference for the actual measured N2/Ar ratio is an

80

approximation and may be associated with considerable uncertainties if the N2 concentration is

81

expected to change by only a few percent, as in the case of N2 fixation. One reason for these

82

problems is due to the seasonality of the surface water temperature in combination with the

83

different temperature coefficients of the solubility constants for N2 and Ar. The solubility of N2

84

and Ar decreases by 1.99% and 2.16% per Kelvin, respectively.15 Consequently, (N2/Ar)sat values

85

run through a minimum/maximum corresponding to the seasonality of the sea surface

86

temperature (SST, Fig. 1). However, the seasonality of the real N2/Ar ratio lags behind that of

87

(N2/Ar)sat because gas exchange and re-equilibration with atmospheric N2 and Ar do not occur

88

spontaneously but are delayed with respect to the temperature change. Furthermore, Ar gas

89

exchange is faster than N2 gas exchange because Ar has a higher diffusivity (lower Schmidt

90

number).

𝑁

𝑁

𝑠𝑎𝑡

13

. Hence, any biogeochemically induced changes in the N2 (ΔN2)

(1)

) 𝐴𝑟

5



Page 6 of 25

91

To estimate possible deviations between the N2/Ar ratios of the “real” abiotic concentrations and

92

those of (N2/Ar)sat, N2/Ar ratios were simulated by a simple model that takes into account the

93

different scenarios in the Baltic Sea. A sinus-shaped SST seasonality (0–20°C) of the water

94

column with a defined mixed layer depth (zmix) was considered. The initial concentrations of N2

95

and Ar at the start of the calculations were set to saturation values. SST was then changed at time

96

sat steps of 1 day and the difference between the “new” ct+1 and the previous ct was used to calculate

97

the exchange with the atmosphere (F) of either N2 or Ar for each time step, as shown in Eqs. (2)

98

and (3):

99

𝐹 = 𝑘660 (660)

𝑆𝑐

−0.5

𝑠𝑎𝑡 ) (𝑐𝑡 − 𝑐𝑡+1

(2)

100

with:

101

𝑘660 = 0.24 u2

102

where k660 is the gas exchange transfer velocity at a Schmidt number of 660 [cm h-1]; Sc is the

103

Schmidt number which is reciprocally proportional to the diffusivity of the considered gas in

104

seawater; and u is the wind speed [m s-1].16,17

105

Dividing the flux (F) by the mixed layer depth yields the changes in N2 and Ar concentrations in

106

units of mmol m-3 from t to t+1.

107

For the determination of csat we first calculated the solubility constant as a function of

108

temperature and salinity15 which was then multiplied with the atmospheric partial pressures of the

109

respective gases. For the calculation of the latter, we used standard atmospheric pressure as well

110

as the N2 and Ar mole fractions in dry air and assumed water vapor saturation at the sea-air

111

interface.

(3)

6


Page 7 of 25


112

N2/Ar seasonality was simulated for a scenario in which conditions favor mid-summer production

113

fueled by N2 fixation: low wind speed (u = 4 m s−1), a shallow surface mixed layer (zmix = 5 m),

114

and a temperature maximum by the end of July of 20°C.

115

Three seasonal cycles were simulated for a spin-up of the model calculations. Deviations from a

116

steady state were already negligible after the second cycle. The seasonality of (N2/Ar)sat (Fig. 1,

117

blue curve) showed minimum and maximum values of 36.9 and 38.1, respectively, coinciding

118

with the minimum and maximum surface water temperature. In contrast, both the amplitude and

119

the phase of the simulated “real” N2/Ar ((N2/Ar)real; Fig. 1, red curve) were slightly different. The

120

difference (N2/Ar)real – (N2/Ar)sat corresponds to a bias in the calculated ΔN2 (δ(ΔN2), Fig. 1,

121

black curve). The asymmetric annual cycle of δ(ΔN2) was due to the superposition of the two

122

effects: different temperature coefficients of the solubility constants and different transfer

123

velocities (Schmidt numbers) for N2 and Ar.

124 125

Figure 1. Simulation of the N2/Ar ratios for a sinus-shaped seasonality of temperature:

126

Spontaneous equilibration by gas exchange (red line) and delayed equilibration due to slow gas

127

exchange (green line). The difference between the N2/Ar ratios at equilibrium (N2/Ar, saturation)

128

and under simulated real conditions (N2/Ar, real) corresponds to a bias of the calculated ΔN2

129

(δ(ΔN2), black line) for the performed scenario (Tmax=20°C, zmix =5 m, u = 4 m s-1). The blue dots 7



Page 8 of 25

130

in the third period of the simulations indicate the start (15.6.) and termination (15.8.) of the time

131

period when the bloom of cyanobacteria (i.e., nitrogen fixation) are most common in the Baltic

132

Sea.

133

Positive δ(ΔN2), which cause an underestimate of ΔN2, were the largest during spring when the

134

temperature increase was the strongest. For the period from mid-June to mid-August when

135

cyanobacteria blooms are common in the central Baltic, δ(ΔN2) ranged between 0.2 and 1.1 µmol

136

L-1. The simulations presented in Fig. 1 for a whole year was based on a mid-summer scenario

137

regarding zmix and u. Nonetheless, the results for a typical winter scenario, i.e., higher wind speed

138

and deeper mixing, showed a similar pattern for δ(ΔN2), but with somewhat larger amplitudes.

139

Our exemplary and simplified simulations of the influence of seasonal surface water temperature

140

variations on the N2/Ar ratios provided a sensitivity test of the N2/Ar method. The results are not

141

intended for practical use; rather, they indicate the possibility of errors in the determination of N 2

142

concentration changes caused by biogeochemical processes such as N2 fixation and

143

denitrification. In reality, the wind speed and mixed layer depth are subject to short-term

144

fluctuations, and the seasonality of the surface water temperature is not a continuous function of

145

time, but frequently pulse-like. To account for such conditions, repeated measurements on the

146

basis of a few days may provide a realistic N2/Ar reference ratio for the start of the N2 fixation by

147

cyanobacteria. This could be facilitated by the deployment of an automated equi-MS system on a

148

VOS line (see section “Comparison with the pCO2 approach”).

149

Our simulations did not account for the impact of air bubble injection on the determination of

150

ΔN2. Previous studies have demonstrated that the injection of air bubbles may generate a

151

significant surface water N2 and Ar disequilibrium that, depending on the bubble size, may also

152

affect the N2/Ar ratio (e.g. 18, 19). Small bubbles (bubble diameter < 150 m) may collapse by the 8


Page 9 of 25


153

complete dissolution of the entrapped air, thereby transferring the atmospheric N2/Ar signature

154

into the surface water. This process would favor the enrichment of the less-soluble gas species

155

and thus result in an increase of ΔN2 that counteracts the N2 loss by N2 fixation. The effect of the

156

injection of large air bubbles on the N2/Ar ratio is far more complex since these bubbles rise back

157

to the sea surface and the entrapped air is incompletely dissolved in the seawater. Modeling the

158

effect of bubbles on the N2/Ar ratio of seawater must therefore include several parameters (e.g.,

159

bubble size spectrum, penetration depth, rising speed, transfer velocity, surfactants), which was

160

beyond the scope of this study. However, to assess the possible effect of bubbles on our field

161

measurements and on N2 fixation studies in the Baltic Sea in general, the meteorological and

162

hydrographic conditions that allow for the massive occurrence of cyanobacteria in the Baltic Sea

163

must be taken into account. Studies have shown that cyanobacterial blooms occurring during

164

June–August are episodic events triggered by the low-wind conditions (about < 4 m/s), which

165

typically last for 1–3 weeks.20, 21 The shallow thermal stratification (5–10 m) that develops in the

166

calm waters increases the exposure of the cyanobacteria to solar radiation

167

conditions, the generation of bubbles can be reasonably excluded and a possible effect on N2

168

fixation studies neglected. Furthermore, the development of extreme temperature gradients at the

169

thermocline (up to 10°C over a depth interval of only 10 m) stabilizes the stratification, such that

170

mixing (entrainment) with deeper water masses that may have been subjected to denitrification is

171

widely prevented.

22

. Under these

172 173

Measurement Procedure

174

The concentrations of N2 and Ar in surface water were measured by coupling an equilibrator

175

system with a mass spectrometer (Equi-MS system), as shown in Fig. 2. 9



Page 10 of 25

176 177

Figure 2. Diagram of the Equi-MS system.

178

The approach is based on the equilibrium of atmospheric gases between the water and gas phases

179

within the headspace of the spray-type equilibrator (E1, total cell volume 2.2 L, water volume

180

500 mL). Seawater enters the equilibrator at the top, through two nozzles, at a flow rate of 5.5 L

181

min-1 and leaves the chamber through a siphon at the bottom. Temperature is measured using a

182

temperature probe (PT100, precision of 0.1°C) inserted into the cell. To achieve equilibrium

183

between the headspace and the seawater, gas (air) is circulating in a closed loop (gas volume in

184

the tubing 0.5 L) through the equilibrator at a flow rate of 600 ml min-1 (membrane pump, P1,

185

KNF Neuberger). A Peltier cooling trap (PK1, Gröger & Obst) reduces the water content in the

186

gas stream and the subsequent water guard protects the mass spectrometer from water leakage.

187

From this pre-dried gas stream a constant flow of 50 ml min-1 is directed into the sampling loop

188

of the mass spectrometer (adjusted with valve V1). For gas analyses, a three-way valve (V2)

189

opens and an external membrane pump (P2) transports the gas along the inlet of the mass

190

spectrometer (flow rate 15 ml min-1). The dead volume between the three-way valve and

191

membrane pump is flushed for 1 min before the gas analyses are conducted for an additional

192

minute, leading to a total removal of 30 ml of gas from the equilibrator gas cycle (< 2% of the

193

total gas volume in the Equi-MS system). The removal of the gas sample is balanced by pre-

194

equilibrated air from the headspace of an auxiliary bubble-type equilibrator (E2, total cell volume

10


Page 11 of 25


23

195

0.35 L, water volume 0.23 L,

). The auxiliary equilibrator is open to the atmosphere and the

196

headspace gas in the chamber is circulated by an external pump (P3) through the water column.

197

The mole fractions of N2, Ar and H2O are determined using a commercially available quadrupole

198

mass spectrometer (QMS, GAM 200, InProcess Instruments). The gas flow from the gas inlet

199

into the vacuum system is reduced to approximately 1 cm3STP min-1 by a 3.2-m-long capillary

200

(inner diameter of 50 m). A high-vacuum of ~6×10-6 mbar is maintained by a mini-diaphragm

201

roughing pump (KNF Neuberger) and a HiPace 80 turbo pump (Pfeiffer Vacuum). The gas

202

species are ionized by electron impact ionization (ionization energy of 70 eV). The produced ions

203

are sorted by a quadrupole mass analyzer and detected by an electron-multiplier. N2, Ar, and H2O

204

are quantified from the signal intensities at mass-to-charge ratios (m/z) of 29, 40, and 18

205

respectively. In contrast to the detection of N2 at m/z 28, the use of m/z 29 shows less mass

206

interference with other gas species and avoids the large difference of signal intensity between m/z

207

28 and m/z 40. Within a measurement cycle of 6 s, each gas species is measured for 1 s, with the

208

remaining time used to define the baseline of the detector signal. This cycle is repeated ten times

209

to derive a solid database leading to a total run time for the gas analyses of 1 min. Before each

210

sequence of gas analyses, the mass spectrometer is calibrated against atmospheric air. An external

211

pump (P4) delivers the gas sample, which is dried by passing through a phosphorus pentoxide

212

(P4O10, Sicapent® with indicator, Merck) and silica gel (SiO2, silica gel with indicator, Roth)

213

column and directed by a multiposition valve (V3, VICI Valco Instruments) and the three-way

214

valve to the gas inlet of the mass spectrometer. The Quadstar 32-bit software (InProcess

215

Instruments) is used to operate the mass spectrometer and analyze the data.

216

The partial pressure of a gas i in the headspace of the equilibrator cell (piequi) is calculated from

217

the mole fraction of the gas component in dry air (xidry), the headspace pressure (ptotal) measured 11



Page 12 of 25

218

outside the equilibrator cell, and the partial pressure of water (pH2O), which is assumed to be at

219

saturation level for the given temperature and salinity.24 The relationship is expressed by Eq. (4).

220

𝑝𝑖𝑒𝑞𝑢𝑖 = 𝑥𝑖

221

The concentration of the dissolved gas ci (mol kg-1) is then calculated according to Henry’s law

222

using the Bunsen coefficient  and the molar gas volume (Vmi), as shown in Eq. (5).

223

𝑐𝑖 =

224

Since the concentration changes of N2 in surface water attributable to N2 fixation are expected to

225

be small, an evaluation of the changes in the concentrations of N2 and Ar requires a high degree

226

of precision and accuracy. We therefore conducted a series of laboratory tests in which the Equi-

227

MS system was coupled to a circulation thermostat (Huber CC-K15 with Pilot ONE, Peter Huber

228

Kältemaschinenbau) filled with 15 L of tap water. To achieve stable equilibrium conditions

229

between the tap water and laboratory air, the water temperature in the thermostat was fixed at 20

230

±0.02°C and laboratory air was permanently introduced into the water reservoir through glass frit.

231

From this reservoir, water was continuously circulated through the equilibrator cell by an external

232

pump with a flow rate of 5.5 L min-1. To determine N2 concentration changes the Equi-MS

233

measurement cycles were conducted in the following order: (i) calibration of the MS with outside

234

air, as described above; (ii) measurement of the mole fractions of N2, Ar, and H2O in laboratory

235

air; and (iii) measurement of these gases in the headspace of the equilibrator cell. The cycle was

236

repeated every 60 min and a record of seven consecutive cycles was used to define the standard

237

deviation (SD) and detection limits for N2 and Ar gas measurements in the equilibrator cell

238

(Table 1). The values of these indicators were then applied to define the SDs of the N2/Ar ratios

239

and N2. In addition, the measured concentrations of N2 and Ar in tap water (Cmeas) were

𝑑𝑟𝑦

(4)

∙ (𝑝𝑡𝑜𝑡𝑎𝑙 − 𝑝𝐻2𝑂 )

𝛽𝑖 ∙𝑝𝑖

(5)

𝑉𝑚𝑖

12


Page 13 of 25


240

compared with the saturation concentrations of these gases (csat, calculated after 15), as shown in

241

Table 1.

242

Table 1. Performance indicators of the Equi-MS system obtained by repeated measurements of

243

the gases in ambient air and in the headspace of the equilibrator cell (absolute standard deviation,

244

aSD). The detection limit is defined as twice the aSD. Gas

aSD

Detection

aSD

limit

29

(N2)

Detection

aSD

limit

N2/Ar

(ppm)

(ppm)

(mol L-1)

(mol L-1)

182

364

0.13

0.26

(mol L-1)

0.05 40

Ar

12

24

0.02

aSD N2

cmeas

csat

tap water

tap water

(mol L-1)

(mol L-1)

540.1

549.1

13.9

14.1

0.7

0.04

245 246

Another important feature of the Equi-MS system is the equilibration time (τ), which is the time

247

required for the establishment of a “new” equilibrium in response to changing water masses.

248

Using our system onboard a moving ship, τ together with the ship’s speed determines the

249

temporal resolution of the measurements and may be interpreted as the “footprint” of the

250

measured change in the N2/Ar ratio. To calculate τ, we performed “step” experiments, in which

251

the Equi-MS system was successively connected to two different water sources. The system was

252

first supplied with tap water previously equilibrated with ambient air in a temperature-controlled

253

bath (T = 20 ± 0.02°C). After a stable N2/Ar signal was obtained from the MS, the water supply

254

was switched to the original tap water, which was not at equilibrium with the ambient air at T =

255

14°C and had a lower equilibrium N2/Ar ratio, (N2/Ar)eq. Due to gas exchange, a progressive

256

decrease in the measured headspace N2/Ar ratio was registered, with (N2/Ar)eq reached after

13



Page 14 of 25

257

about one hour (Fig. 3, inset). The temporal development of (N2/Ar)t in the headspace was

258

expressed by a simple exponential function:

259

(N2/Ar)eq – (N2/Ar)t = [(N2/Ar)eq – (N2/Ar)t=0] ∙ exp – t/

260

Eq. (6) implies that the gas exchange of N2 and Ar occurs with identical equilibration times. This

261

is not exactly the case because the exchange rate of Ar is slightly faster than that of N2 (lower

262

Schmidt number); nonetheless, the effect on the calculation of τ for the ratio N2/Ar is slight. For

263

the determination of τ, we plotted ln[(N2/Ar)eq-(N2/Ar)t] vs. time (t, Fig. 3). According to Eq. (6),

264

the reciprocal slope of the respective regression line gives the e-fold equilibration time (τ). A

265

value of τ = 12 min was obtained, which means that it takes 12 min until 63 % of an initial

266

(arbitrary) disequilibrium is compensated by gas exchange.

(6)

267 268

Figure 3. The N2/Ar ratio as a function of time (inset) and the logarithmic presentation of the

269

difference between N2/Ar at any time t and at equilibrium as a function of time. The reciprocal

270

slope of the regression line corresponds to the e-fold equilibration time ().

271 272

Field measurements

14


Page 15 of 25


273

To demonstrate that the Equi-MS system is capable of detecting N2 loss due to N2 fixation, we

274

participated in Meteor cruise M117, conducted in order to study a cyanobacterial bloom in the

275

Baltic Sea.25 Surface water was permanently pumped from 3 m below the sea surface to the Equi-

276

MS system at a flow rate of 5.5 L min−1. Sea surface temperature, salinity and meteorological

277

data (e.g., wind speed) were recorded using a thermosalinograph (SBE 21 SeaCAT, Sea-Bird

278

Scientific) and the ship’s weather station.

279

The experimental set-up was the same as that used in our laboratory-based measurements. N2/Ar

280

ratios in the equilibrator’s headspace were measured at hourly intervals. At an e-fold

281

equilibration time of about 12 min, this time difference guaranteed that the N2/Ar ratio in the

282

headspace was at equilibrium with that of the gasses dissolved in seawater. The equilibration time

283

τ also determines the spatial resolution (footprint) of the N2 concentration measurements. A value

284

of three times the equilibration time (95% equilibrium) and a cruising speed of 10 knots yielded a

285

spatial resolution of about 5 nautical miles, which is sufficient for large-scale records of N2

286

fixation. However, in highly dynamic surface water regimes, the presence of strong temperature

287

variations may complicate the choice of an appropriate temperature for the calculation of the

288

reference N2/Ar ratio. This problem can be mitigated by reducing the equilibration time, which is

289

easily done by reducing the volume of the equilibrator’s headspace/tubing, a value directly

290

proportional to τ. 26

291

Another problem related to the circumstances of our cruise were measurements made while the

292

ship remained at a station to conduct other research activities. Due to the deployment of samplers

293

and various instruments as well as the frequent re-positioning of the ship, the thermal

294

stratification of the upper surface water was disrupted. This may have caused erroneous

295

measurements because cyanobacterial blooms develop preferably in the shallow surface layer that 15



Page 16 of 25

296

forms under the calm conditions typical of mid-summer. To eliminate potential artifacts from our

297

dataset, we rejected data obtained at a ship’s speed of < 3 knots. Nevertheless, if sampling-related

298

perturbations are avoided, then N2/Ar data acquisition at a ship speed < 3 knots is possible and is

299

suggested for investigations in areas subject to complex oceanographic conditions (e.g., filaments

300

in upwelling regions).

301 302

RESULTS AND DISCUSSION

303

Surface water N2 deficits detected using the N2/Ar approach

304

Our field investigations focused on a time slot of the research cruise that was characterized by a

305

general decrease in wind speed and an increase in SST, which occurred between July 27 and

306

August 12 (Fig. 4a). The cruise can be subdivided into three main parts in terms of its

307

oceanographic/meteorological features and the impact of those features on the surface water N2

308

pattern.

309 16


Page 17 of 25


310

Figure 4. (a) Color-coded sea surface temperature (SST), wind speed, and (b) N2 values

311

measured along a ship track in the central Baltic Sea in summer 2015. In (b) data points in the

312

eastern Gotland Sea that belong to Part 3 are framed by black lines to better distinguish these data

313

from the overlapping data points belonging to Part 1. The black line in (b) indicates the VOS

314

Finnmaid track used to calculate N2 values from surface water measurements of the CO2 partial

315

pressure.

316

Part 1. Between July 27 and August 4, relatively strong wind speeds of up to 13 m s -1 were

317

detected and N2 values of around zero reflected atmospheric equilibrium (Fig. 4b). However, an

318

oversaturation of N2 in the surface water at some measuring points was indicated by the positive

319

N2 values. We assume that the increase in N2 values observed in Part 1 may have been due to

320

wind-induced breaking waves and the concomitant entrainment of atmospheric bubbles and their

321

dissolution in the surface water.13,

322

atmospheric gas composition (N2/Ar = 83), which was quite different from the surface water ratio

323

of N2/Ar in equilibrium with the atmosphere.

324

Part 2. Around August 3, satellite-based SST measurements showed relatively cold temperatures

325

along the eastern coast of Sweden (Fig. 5a). These could be attributed to the upwelling of cold

326

deep-water masses and they were well-reflected in the surface water temperatures recorded

327

during the research cruise between the southern tip of Gotland and east of Öland (drop in the SST

328

from 15.8°C to 11.4°C, Fig. 4a). The temperature pattern correlated with a positive shift in N2

329

(up to 5.3 mol kg-1 near Öland, Fig. 4b) and suggested that denitrification in suboxic and anoxic

330

zones near the sediment28 contributed to the N2 content in the surface water, due to an upward

331

transport of bottom water to the sea surface. The upwelling of deep cold water and the subsequent

27

These air bubbles carried the N2/Ar signature of the

17



Page 18 of 25

332

warming at the sea surface would further support the positive shift in N2 due to a mechanism

333

comparable with the warming of surface water during spring (Fig. 1).

334 335

Figure 5. Color-coded SSTs from (a) August 3 and (b) August 7 (data from AVHRR Pathfinder

336

Version 5.2 obtained from the US National Oceanographic Data Center and the Group for High-

337

Resolution Sea Surface Temperature and reanalyzed by the Integrated Climate Data Center). 29 (c)

338

True-color image from August 7 indicating the accumulation of cyanobacteria at the sea surface

339

(data from MODIS Aqua, NASA Goddard Space Flight Center, Ocean Ecology Laboratory,

340

Ocean Biology Processing Group).

341

Part 3. In the following part of the cruise (August 5–12), the continuous decline of the wind

342

speed and the associated breakdown of the upwelling cells were reflected in the increase in SST

343

(Figs. 4a and 5b). In addition, the N2 values declined (Fig. 4b) and correlated negatively with

344

the SST (Fig. 6). This clearly indicated persistent N2 fixation by cyanobacteria. Schneider et al.12

345

showed that temperature increases correlate strongly with N2 fixation activity and concluded that

346

the exposure of cyanobacteria to solar radiation is the most important factor regulating the

347

intensity of N2 fixation in the Baltic Sea. This conclusion was supported by the true-color satellite

348

images obtained for this study, which indicated that the calm weather conditions during the cruise

349

reduced the turbulence in the mixed layer and favored the accumulation of cyanobacteria at the 18


Page 19 of 25


350

sea surface (Fig. 5c). The images also showed that cyanobacteria were not homogenously spread

351

over the sea surface but had accumulated in patches.

352 353

Figure 6. Relationship between SST and N2 for the time span July 30 to August 13.

354 355

Comparison with the pCO2 approach

356

A confirmation of the surface water N2 deficits detected using the N2/Ar method was obtained

357

from automated measurements of the surface water pCO2 made from the VOS Finnmaid.12

358

Between July 30 and August 12, the ship crossed the eastern Gotland Sea and the entrance to the

359

Gulf of Finland seven times. This period of time, corresponding to Part III of our cruise, was

360

characterized by high N2 values in the surface water. From the high-resolution time series data

361

for the mean pCO2 (Surface Ocean CO2 Atlas, SOCAT v4 data, http://www.socat.info/), we used

362

the data from the northern Gotland Sea (Fig. 4b) to characterize the temporal development of N2

363

fixation in the northern part of the central Baltic Sea. However, since our N2/Ar measurements

364

were performed at different times and at different places, a direct comparison of the two datasets

365

is only conditionally possible.

19



Page 20 of 25

366

At the beginning of August, the total CO2 concentration (CT), derived from the pCO2

367

measurements

368

investigations (Fig. 7, inset). Since the DIN pool (Swedish National Monitoring Program,

369

https://sharkweb.smhi.se) was entirely exhausted, the drop in CT (~40 µmol L-1) was attributed to

370

biomass production fueled by N2 fixation. Dividing the organic carbon equivalent by the Redfield

371

C/N ratio (6.625

372

(blue dots/line) as ΔN2 together with the range of variability that corresponded to the spatial

373

variability of CT within the considered area in the northern Gotland Sea.

30

, started to decrease and reached a minimum by August 13, at the end of our

31

) yielded the amount of “fixed” nitrogen.12 The latter is presented in Fig. 6

374 375

Figure 7. ΔN2 values for the individual measurements made between July 30 and August 13

376

obtained from N2/Ar measurements (red dots) and as inferred from continuous pCO2 records

377

obtained from a cargo ship (blue dots/line). Open blue circles indicate the range of the spatial

378

variability. The shadowed green area in the inset displays the mean total CO2 decrease (obtained

379

from the pCO2 data) during the period of N2 fixation in the northern Gotland Sea (CT = blue line,

380

SST = red line).

381

The uncertainties and problems associated with the choice of the reference N2/Ar ratio for the

382

determination of ΔN2 were discussed above. Here we applied the (N2/Ar)sat value from the

383

beginning of N2 fixation activity, on August 4, for the following fixation period (Part III). The 20


Page 21 of 25


384

constant (N2/Ar)sat implied that the gas exchange of N2 and Ar was negligible, an assumption

385

justified by the strong reduction in gas exchange related to the low wind speed. The general trend

386

in the mean CO2-derived ΔN2 was widely reproduced by the temporal change in ΔN2 determined

387

from our N2/Ar method (Fig. 7, red dots). However, this method resulted in a large scatter of the

388

ΔN2 values, perhaps due to the patchiness of the cyanobacterial distribution (Fig. 5) and thus of

389

their N2 fixation activity. In this regard, it must be taken into account that the N 2/Ar

390

measurements were made while the ship was cruising at a speed of about 10 knots. Thus,

391

measurements performed in the course of a single day may refer to water masses separated from

392

each other by 100 miles or even more. Hence, the variability of ΔN2 reflects not only small scale

393

patchiness but may also be caused by regional gradients of the N2 fixation activity.

394 395

Assessment

396

The results of laboratory tests and a pilot study in the Baltic Sea showed that our N2/Ar approach

397

enables the detection of relatively small decreases in N2 concentrations, attributable to N2

398

fixation. Observations of enhanced N2 concentrations could be explained by upwelling events and

399

bubble entrapment, conditions at which the occurrence of cyanobacteria is unlikely. We consider

400

our first research campaign as a feasibility study for “underway” measurements of N2 fixation on

401

a research vessel. This approach will facilitate large-scale, direct detections of N2 fixation in the

402

Baltic Sea, which is believed to be of the same order of magnitude as the sum of riverine input

403

and the atmospheric deposition of DIN.32,

404

account for the regional variability and episodic character of N2 fixation in the Baltic Sea cannot

405

be reasonably made from sample-based studies such as

33

Reliable measurements based on estimates that

15

N incubations and calculations of the

21



Page 22 of 25

406

total N budget.32 We are aware of the uncertainties and limitations of our N2/Ar approach, which

407

are mainly associated with the choice of an appropriate reference N2/Ar ratio.

408

The development of an automated version of our MS-Equi system and its deployment on a VOS,

409

such as for the above-mentioned pCO2 measurements, will provide further information on the

410

temporal and spatial distribution of N2 fixation. An automated approach may also circumvent

411

some of the problems discussed above, because the generation of N2/Ar ratios as a function of

412

time includes the determination of an appropriate reference N2/Ar ratio. In addition, the

413

applicability of the N2/Ar approach should be validated through tests in different oceanographic

414

settings with similar biogeochemical characteristics. Like the Baltic Sea, other coastal regions are

415

progressively affected by warming, eutrophication and algal blooms. We are confident that, also

416

in those regions, the N2/Ar approach will support the ecological management of these systems.

417

418

AUTHOR INFORMATION

419

Corresponding Author

420

Oliver Schmale (oliver.schmale@io-warnemuende)

421

Author Contributions

422

All of the authors contributed to the writing of the manuscript and gave their approval to the final

423

version of the manuscript.

424

Funding Sources

425

This present work was supported by the Federal Ministry of Education and Research (BMBF)

426

through MOBALAB: Mobiles Analyselabor "Veränderungen der Ostsee". 22


Page 23 of 25


427 428

ACKNOWLEDGMENTS

429

We thank Bernd Sadkowiak for the construction and set up of the equilibrator line, and Stefan

430

Otto for supporting the mass spectrometry work. We also thank Lars Umlauf, Herbert Siegel and

431

Monika Gerth for processing the satellite data, and the captain and crew of the RV Meteor M117

432

for technical support. We acknowledge the use of Rapid Response imagery from the Land,

433

Atmosphere Near real-time Capability for EOS (LANCE) system, operated by NASA's Earth

434

Science Data and Information System (ESDIS) and funded by NASA Headquarters. We thank

435

GHRSST and the US National Oceanographic Data Center for providing sea surface temperature

436

data, obtained in part with support by a grant from the NOAA Climate Data Record (CDR)

437

program for satellites. The work described herein was supported by the Federal Ministry of

438

Education and Research (BMBF) through MOBALAB: Mobiles Analyselabor "Veränderungen

439

der Ostsee".

440 441

REFERENCES

442 443 444 445 446 447 448 449 450 451 452 453 454 455 456

1. Kaiser, J.; Reuer, M. K.; Barnett, B.; Bender, M. L., Marine productivity estimates from continuous O2/Ar ratio measurements by membrane inlet mass spectrometry. Geophys. Res. Lett. 2005, 32, (19), L19605. 2. Tortell, P. D.; Bittig, H. C.; Körtzinger, A.; Jones, E. M.; Hoppema, M., Biological and physical controls on N2, O2, and CO2 distributions in contrasting Southern Ocean surface waters. Glob. Biogeochem. Cycles 2015, 29, (7), 994-1013. 3. Capone, D. G.; Bronk, D. A.; Mulholland, M. R.; Carpenter, E. J., Nitrogen in the marine environment. Elsevier: Amsterdam, 2008; p 1757. 4. Codispoti, L. A.; Brandes, J. A.; Christensen, J. P.; Devol, A. H.; Naqvi, S. W. A.; Paerl, H. W.; Yoshinari, T., The oceanic fixed nitrogen and nitrous oxide budgets : Moving targets as we enter the anthropocene? Scientia Marina 2001, 65, (2), 85-105. 5. Devol, A. H.; Uhlenhopp, A. G.; Naqvi, S. W. A.; Brandes, J. A.; Jayakumar, D. A.; Naik, H.; Gaurin, S.; Codispoti, L. A.; Yoshinari, T., Denitrification rates and excess nitrogen gas concentrations in the Arabian Sea oxygen deficient zone. Deep Sea Research Part I: Oceanographic Research Papers 2006, 53, (9), 1533-1547. 23



457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503

Page 24 of 25

6. Wasmund, W.; Voss, M.; Lochte, K., Evidence of nitrogen fixation by non-heterocystous cyanobacteria in the Baltic Sea and re-calculation of a budget of nitrogen fixation. Mar. Ecol.Prog. Ser. 2001, 214, 1-14. 7. O’Neil, J. M.; Davis, T. W.; Burford, M. A.; Gobler, C. J., The rise of harmful cyanobacteria blooms: The potential roles of eutrophication and climate change. Harmful Algae 2012, 14, 313-334. 8. Montoya, J. P.; Voss, M.; Kahler, P.; Capone, D. G., A Simple, High-Precision, HighSensitivity Tracer Assay for N(inf2) Fixation. Appl. Environ. Microbiol. 1996, 62, (3), 986-993. 9. Eggert, A.; Schneider, B., A nitrogen source in spring in the surface mixed-layer of the Baltic Sea: Evidence from total nitrogen and total phosphorus data. Journal of Marine Systems 2015, 148, 39-47. 10. Larsson, U.; Hajdu, S.; Walve, J.; Elmgren, R., Baltic Sea nitrogen fixation estimated from the summer increase in upper mixed layer total nitrogen. Limnol. Oceanogr. 2001, 46, (4), 811-820. 11. Schneider, B.; Kaitala, S.; Raateoja, M.; Sadkowiak, B., A nitrogen fixation estimate for the Baltic Sea based on continuous pCO2 measurements on a cargo ship and total nitrogen data. Continental Shelf Research 2009, 29, (11), 1535-1540. 12. Schneider, B.; Gustafsson, E.; Sadkowiak, B., Control of the mid-summer net community production and nitrogen fixation in the central Baltic Sea: An approach based on pCO2 measurements on a cargo ship. Journal of Marine Systems 2014, 136, 1-9. 13. Hamme, R. C.; Emerson, S. R., Mechanisms controlling the global oceanic distribution of the inert gases argon, nitrogen and neon. Geophys. Res. Lett. 2002, 29, (23), 35-1-35-4. 14. Benson, B. B.; Parker, P. D. M., Nitrogen/argon and nitrogen isotope ratios in aerobic sea water. Deep Sea Research (1953) 1961, 7, (4), 237-253. 15. Hamme, R. C.; Emerson, S. R., The solubility of neon, nitrogen and argon in distilled water and seawater. Deep Sea Research Part I: Oceanographic Research Papers 2004, 51, (11), 1517-1528. 16. Wanninkhof, R., Relationship between wind speed and gas exchange over the ocean. J. Geophys. Res. 1992, 97, (C5), 7373-7382. 17. Wanninkhof, R.; Asher, W.; Ho, D. T.; Sweeney, C.; McGillis, W. R., Advances in quantifying air-sea gas exchange and environmental forcing. Annual Reviews Marine Science 2009, 1, 213-244. 18. Nicholson, D. P.; Emerson, S. R.; Khatiwala, S.; Hamme, R. C., An inverse approach to estimate bubble-mediated air-sea gas flux from inert gas measurements. Proceedings on the 6th International Symposium on Gas Transfer at Water Surfaces 2011, 223-237. 19. Tortell, P. D.; Bittig, H. C.; Körtzinger, A.; Jones, E. M.; Hoppema ,M., Biological and physical controls on N2, O2, and CO2 distributions in contrasting Southern Ocean surface waters. Global Biogeochem. Cycles 2015, 29, 994–1013. 20. Wasmund, N., Occurrence of cyanobacterial blooms in the Baltic Sea in relation to environmental conditions. Int. Revue ges. Hydrobiol. 1997, 82, 169-184. 21. Lips, I. and Lips, U., Abiotic factors influencing cyanobacterial bloom development in the Gulf of Finland (Baltic Sea). Hydrobiologia 2008, 614, 133-140. 22. Kahru, M.; Savchuk, O. P.; Elmgren, R., Satellite measurements of cyanobacterial bloom frequency in the Baltic Sea: interannual and spatial variability. Mar. Ecol. Prog. Ser. 2007, 343, 15-23. 23. Gülzow, W.; Rehder, G.; Schneider, B.; Schneider von Deimling, J.; Sadkowiak, B., A new method for continuous measurement of methane and carbon dioxide in surface waters using 24


Page 25 of 25

504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533


off-axis integrated cavity output spectroscopy (ICOS): an example from the Baltic Sea. Limnology and Oceanography-Methods 2011, 9, 176-184. 24. Ambrose, D.; Lawrenson, I. J., The vapour pressure of water. The Journal of Chemical Thermodynamics 1972, 4, (5), 755-761. 25. Wurl, O.; Nausch, G.; Nausch, M.; van Pinxteren, M.; Kuss, J.; Loick-Wilde, N., Biochemical processes in upwelling zones of the Baltic Sea - Cruise No. M117 - July 23 - August 17, 2015 - Hamburg (Germany) - Rostock (Germany); Bremen, 2016; p 47. 26. Schneider, B.; Sadkowiak, B.; Wachholz, F., A new method for continuous measurements of O2 in surface water in combination with pCO2 measurements: Implications for gas phase equilibration. Mar. Chem. 2007, 103, 163-171. 27. Schudlich, R.; Emerson, S., Gas supersaturation in the surface ocean: The roles of heat flux, gas exchange, and bubbles. Deep Sea Research Part II: Topical Studies in Oceanography 1996, 43, (2), 569-589. 28. Asmala, E.; Carstensen, J.; Conley, D. J.; Slomp, C. P.; Stadmark, J.; Voss, M., Efficiency of the coastal filter: Nitrogen and phosphorus removal in the Baltic Sea. Limnol. Oceanogr. 2017, 62, (S1), S222-S238. 29. Casey, K. S.; Brandon, T. B.; Cornillon, P.; Evans, R., The Past, Present, and Future of the AVHRR Pathfinder SST Program. In Oceanography from Space: Revisited, Barale, V.; Gower, J. F. R.; Alberotanza, L., Eds. Springer Netherlands: Dordrecht, 2010; pp 273-287. 30. Schneider, B.; Müller, J. D., Biogeochemical transformations in the Baltic Sea Obervations through carbon dioxide glasses. Springer Oceanography 2018, ISSN 2365-7677. 31. Redfield, A. C.; Ketchum, B. H.; Richards, F. A., The influence of organisms on the composition of sea-water. In The sea: ideas and observations on progress in the study of the seas, Hill, M. N., Ed. Interscience: New York, 1963; pp 26-77. 32. Wasmund, N.; Nausch, G.; Schneider, B.; Nagel, K.; Voss, M., Comparison of nitrogen fixation rates determined with different methods: a study in the Baltic Proper. Mar. Ecol. Prog. Ser. 2005, 297, 23-31. 33. Rolff, C.; Almesjö, L.; Elmgren, R., Nitrogen fixation and abundance of the diazotrophic cyanobacterium Aphanizomenon sp. in the Baltic Proper. Mar. Ecol. Progr. Ser. 2007, 332, 107118.

25


Argon Analysis in Surface Waters in the

Recommend Documents