Physical behavior of the SEEDS iron-fertilized patch by sulphur hexafluoride tracer release

Progress in Oceanography Progress in Oceanography 64 (2005) 111–127 www.elsevier.com/locate/pocean

Physical behavior of the SEEDS iron-fertilized patch by sulphur hexafluoride tracer release Daisuke Tsumune a,*, Jun Nishioka a, Akifumi Shimamoto b, Shigenobu Takeda c, Atsushi Tsuda d a

Central Research Institute of Electric Power Industry, Environmental Science Department, 1646 Abiko, Abiko-shi, Chiba-ken 270-1194, Japan b The General Environmental Technos Co., Ltd., Chuou, Osaka 541-0052, Japan c Department of Aquatic Bioscience, University of Tokyo, Bunkyo, Tokyo 113-8657, Japan d Ocean Research Institute, University of Tokyo, 1-15-1 Minamidai, Nakano, Tokyo 164-8639, Japan Available online 12 April 2005

Abstract The first iron (Fe) – fertilization experiment in the western North Pacific was carried out using SF6 to trace the Fefertilized water mass. A solution in 10,800 liters of seawater of 350 kg of Fe and 0.48 M of SF6 tracer was released into the mixed layer over a 8 · 10 km area. On the first underway transects through the patch after the Fe release, we observed a significant increase of dissolved Fe (ave. 2.89 nM). The fertilized patch was traced for 14 days by on-board SF6 analysis. A Lagrangian frame of reference was maintained by the use of a drogued GPS buoy released at the center of the patch. The patch moved westward at a rate of 6.8 km d
1. Mixed layer depth increased from 8.5 to 15 m during the experiment. Horizontal diffusivity was determined by the change of SF6 concentration in the patch. The horizontal diffusivity increased during the experiment. We evaluate here the fate of Fe in a Fe-fertilized patch using the dilution rate determined from sulphur hexafluoride (SF6) concentration. Dissolved Fe concentrations subsequently decreased rapidly to 0.15 nM on Day 13. However, the dissolved Fe half-life of 43 h was relatively longer than in previous Fe-enrichment studies, and we observed a larger increase of the centric diatom standing stock and corresponding drawdown of macro-nutrients and carbon dioxide than in the previous studies. The most important reason for the larger response was the phytoplankton species in the western North Pacific. In addition, the smaller diffusivity and shallower mixed layer were effective to sustain the higher dissolved Fe concentration compared to previous experiments. This might be one reason for the larger response of diatoms in SEEDS.
2005 Elsevier Ltd. All rights reserved. Keywords: SF6 tracer; Dilution rate; Experimental methods; Patch; Fe behavior

*

Corresponding author. Tel.: +81 4 7182 1181; fax: +81 4 7183 2966. E-mail address: [email protected] (D. Tsumune).

0079-6611/$ - see front matter
2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.pocean.2005.02.018

112

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

1. Introduction Several in situ iron (Fe) fertilization experiments have been performed with the general goal of evaluating whether Fe availability controls phytoplankton production in high nitrate low chlorophyll (HNLC) waters in the equatorial Pacific and Southern Ocean (Boyd et al., 2000; Coale et al., 1996; Martin et al., 1994). The Fe-limitation hypothesis has also been investigated in the western subarctic North Pacific by a mesoscale, single Fe infusion called the Subarctic Pacific Iron Experiment for Ecosystem Dynamics Study (SEEDS) (Tsuda et al., 2003). In all of these in situ Fe-enrichment experiments, sulphur hexafluoride (SF6) concentrations were used as proxies for the enriched Fe concentration without any effects of biochemical processes (Law, Watson, Liddicoat, & Stanton, 1998). SF6 concentration decreases by physical dilution and also air-sea gas exchange. Physical dilution rate can be determined from the SF6 concentration when considered together with the loss by air-sea gas exchange (Liss & Merlivat, 1986; Wanninkhof, 1992). SEEDS showed the largest biological and chemical responses to Fe supply (Tsuda et al., 2003) among all such experiments to date in HNLC areas: in the equatorial Pacific, IronEx I and II (Martin et al., 1994; Coale et al., 1996) and in the Southern Ocean, SOIREE (Boyd et al., 2000) and EisenEx (Gervais, Riebesell, & Gorbunov, 2002). The most likely reason for the greater level of induced bloom was the stronger response of diatom species in the western North Pacific (Tsuda et al., 2003). In addition, SEEDS might have had some advantages in physical conditions in comparison with the previous experiments. Except for IronEx I (Martin et al., 1994), multiple Fe infusions were conducted in the previous Fe enrichment studies: IronEx II (Coale et al., 1996), SOIREE (Boyd et al., 2000) and EisenEX (Gervais et al., 2002). However, Bowie et al. (2001) suggested that multiple infusions brought some difficulties in understanding the quantitative changes in chemical Fe species in the Fe-enriched patch in SOIREE, and it is more difficult to compare experimental results with natural events. Therefore, the single infusion of Fe in the SEEDS study has advantages for understanding the Fe dynamics during the experiment. Progress in our understanding of the biogeochemical cycle of Fe in seawater is one goal of these in situ Fe-fertilization experiments. In this paper, we summarize the methodology and behavior of the SEEDS Fe-enriched patch, and also changes in the concentrations of Fe in the patch during the induced phytoplankton bloom.

2. Methods The first Fe-enrichment experiment in the western North Pacific was carried out by R.V. Kaiyo Maru from June 28 to August 6, 2001. The first leg was from June 28 to 9 July, 2001 as a pre-survey to select the experimental site and the second leg was from July 14 to August 6 in 2001 for the Fe-enrichment experiment. The pre-survey around the experimental site was conducted with XBT observations, an underway survey and vertical bottle sampling to determine the biological, chemical and physical heterogeneity of the experimental area. Fe was supplied on July 18–19 over 23 h, and the Fe-enriched water mass (hereafter patch) was traced for 14 days. Tracing of the Fe-enriched patch was conducted with a combination of SF6 and the drogued GPS buoy, similar to those of previous experiments (Law et al., 1998). 2.1. Iron and SF6 tracer release The method of SF6 tracer release in SEEDS was similar to the previous Fe-fertilization experiments (Law et al., 1998). The saturated SF6 solution was prepared in the port of Kushiro after the first leg. Two 2000 L steel tanks were filled with seawater from the experimental area defined during the pre-survey. Therefore, it was not necessary to adjust the salinity of the solutions. SF6 solubility is very low for seawater; therefore, we employed the same method as in a previous experiment to make the SF6 solution effectively

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

113

(Upstill-Goddard, Watson, Liss, & Liddicoat, 1990). Fig. 1 shows a schematic drawing of the tanks for preparing the SF6 solution. An acrylic cylinder was placed on the top of the tanks, making a headspace of about 5 L for effective bubbling. Pure SF6 (99.9999%) gas was injected into the seawater in the tanks directly through air stones at the rate of 150–200 mL min
1. In addition, pure SF6 gas in the headspace was re-circulating for additional bubbling at a rate of 36 L min
1 by an air-tight pump. To avoid contamination of SF6 in the ship, an exhaust line led to the outside. The bubbling of SF6 was continued for 24 h in each tank to achieve a saturated SF6 solution. The SF6 concentration was measured to be 1.2 · 10
4 M L
1 by a headspace sampling method using an Electron Capture Detector (ECD) (Ledwell & Watson, 1991). This value was of the same order as saturated SF6 concentration in seawater, 2.0 · 10
4 M L
1 (Ledwell & Watson, 1991). As a result, a total of 0.48 M of SF6 was dissolved in the 4100 L seawater in the two tanks. During release of the SF6 tracer, a meteorological balloon (The Weather Balloon Mfg. Co., Ltd. Japan) within the airtight steel tank was filled with water to replace the volume of expanding headspace to prevent degassing of SF6. The Fe solution (FeSO4 Æ 7H2O) was prepared at the experimental site just before Fe and SF6 tracer release. Surface seawater was pumped into four 3000 L tanks (total 10,800 L) on deck and acidified to pH  1.8 with concentrated HCl (Wako Chemical, Inc. Japan). Then, 350 (6250 M) kg of Fe as sulphate heptahydrate (1740 kg of FeSO4 Æ 7H2O; Yoneyama Chemical, Inc. Japan) was added to the acidified seawater and mixed by a stirrer. Fig. 2 shows a schematic drawing of the system for release of Fe and SF6 tracer. 4100 L of SF6 saturated solution and 10,800 L of Fe solution were injected at a depth of 5 m in the 10 m mixed layer over an area of 8 · 10 km. The ship moved at 5 knots for 23 h during the Fe and tracer release. The flow rate of Fe solution was 8 L min
1, and that of SF6 solution was 3 L min
1. Therefore, 260 g of Fe and 3.6 · 10
4 M of SF6 were released per minute. Times and quantities comprising the Fe release in the SEEDS experiment are given in Table 1. Expected average concentrations of Fe and SF6 were estimated to be about 7.8 nM and 1200 fM in the patch, respectively, comparable to the maximum measured dissolved Fe concentration of 6.02 nM on Day 1 and those of the IronEx I and the SOIREE (Bowie et al., 2001; Martin et al., 1994). A Lagrangian reference frame was employed for the release and samplings in the Fe-enriched patch to minimize the effect of advection (Stanton, Law, & Watson, 1998). A GPS buoy with WOCE-style drogue attached at about 20 m depth was used as a reference for the Lagrangian frame. A computer display on board updated the relative position of the ship to the GPS buoy every 10 min during the Fe and SF6 tracer

Fig. 1. Schematic drawing of the steel tank used to make the SF6 solution.

114

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127 SF6 Tanks Add seawater

Add seawater

Fe Tanks

Bladder

stirrer

Bladder

stirrer

Drain cock

Drain cock

flow rate controller

flow rate controller

Flow sensor

controller

Flow sensor

Manual needle Valve

Manual needle Valve

Mixer

P strainer

Drain cock

Drain cock

Magnet pump (7-10 L/min)

Magnet P pump (1.5-3.5L/min)

strainer

Pump unit release Fig. 2. Schematic drawing of the system for release of Fe and SF6. Table 1 Quantities and times of iron release Date of release Start GMT

Finish GMT

18-Jul (Day 0) 9:45

19-Jul (Day 1) 8:30

Amount of FeSO4 Æ 7H2O (kg)

Release grid size (km2)

Depth of surface Mixed layer (m)

Theoretical iron increase (nM)

1740 (Fe: 350)

80

10

7,8

release. During the release of Fe and SF6, the GPS buoy advected toward the west at 0.1 m s
1. The Lagrangian coordinate navigation relative to the buoy successfully covered the planned 8 · 10 km experimental patch (Fig. 3). 2.2. Sampling during SEEDS experiment Continuous sampling of surface waters by propeller surveys (along a track shaped like a propeller, see Fig. 13), was conducted to determine the position and shape of the patch using the ship
s pumping system with an intake 6 m below the surface for measurements of SF6. SF6 concentration was measured using a sparge and cryogenic trap system based on Upstill-Goddard et al. (1990) and Watanabe, Shimamoto, and Ono (2003). Seawater was pumped to the ship
s laboratory continuously and supplied to a glass sparge tower. SF6 was accumulated in two Parapak Q traps and determined by the Electron Capture Detector (ECD). This system can measure the SF6 concentration at intervals of 10 min.

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127 4 8 .6 0

10

4 8 .5 8

8

115

6

4 8 .5 6

2 4 8 .5 2

NS (km)

Latitude (degree)

4 4 8 .5 4

4 8 .5 0

0 -2

4 8 .4 8 -4 4 8 .4 6

-6

4 8 .4 4

-8

4 8 .4 2

- 10 1 6 4 .7 6 1 6 4 .8 0 1 6 4 .8 4 1 6 4 .8 8 1 6 4 .9 2 1 6 4 .9 6 1 6 5 .0 0

Lon gitu d e (d e gre e )

- 10

-8

-6

-4

-2

0

2

4

6

8

10

EW (km)

Fig. 3. Trajectories of ship and buoy in the regular coordinate system and Lagrangian coordinate system referenced to the GPS buoy.

To observe Fe concentration during the propeller survey, clean surface water (1.5–3 m depth) was collected using a towed-fish, metal-free sampling system, which consists of a towed-fish covered with metalfree epoxy-paint, and Teflon tubing (ID 12 mm) covered by PVC. The tubing opened through the center front of the fish and carried sample water from the top of the towed fish to a clean-air tent located in the ship
s main laboratory. The fish was set on the ship
s side. The sample water was moved by an air-driven Teflon pump. The sampling tube was split into two channels in order to provide filtrate, from an 0.22 lm cartridge-type Durapore filter (Opticap, Millipore), and an unfiltered sample. Samples were collected every 5–10 min. Concentrations of chlorophyll a and nutrients in samples also collected from the towed fish were measured (methods are given in Kudo et al., 2005). The ship moved through the patch several times at various angles. The ship trajectory was navigated in the Lagrangian reference frame based on the reference GPS buoy, similarly to the release methods. The speed of the ship during the propeller survey was 10 knots. The propeller surveys were conducted on Day 1, 3, 6, 8, 10 and 12, and also before the in-patch sampling to determine the in-patch sampling site. Discrete vertical sampling was conducted to characterize vertical profiles at in-patch and out-patch stations on Days 2, 4, 7, 9, 11 and 13. Hydrographic data were collected by a CTD, and metal-free seawater samples were also collected from 5, 10, 20, 30, 50 and 70 m depth using modified 10-L Niskin X samplers. The samplers were coated inside with Teflon, the drain cocks were replaced with all-Teflon stop cocks, and acid-cleaned inside at the onshore laboratory. The samplers were suspended by a Kevlar hydro-wire and tripped using Teflon coated messengers. For sub-sampling from Niskin X samplers, acid cleaned 0.22 lm Durapore filters (Millipac 100, Millipore) in a polycarbonate housing were connected to the sampler spigot. Then the filtrate was collected under gravity pressure. Samples collected from the surface mixed layer (5, 10 m) for Fe analysis were immediately size-fractionated by a clean in-line filtration system using a 200 kDa polyethylene hollow-fiber ultrafilter (Nishioka, Takeda, Wong, & Johnson, 2001; Nishioka et al., 2003) in the clean-air tent. In addition, sampling for cross-section profiling analysis through the patch was conducted on Daya 3, 8 and 12. The locations of the cross-section lines were decided based on previous propeller surveys. Seven to eight sampling stations were set on the line through the patch, and discrete vertical sampling was conducted by the same method as at the in-patch and out-patch stations.

116

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

All unfiltered and filtrate samples were buffered at pH 3.2 with 10 M formic acid – 2.4 M ammonium formate buffer solution in the laminar-flow, clean-air hood in the clean-air tent. Concentrations of Fe (III) in the buffered samples were determined with an automatic Fe (III) flow injection analytical system (Kimoto Electric, Ltd.) using chelating resin for concentration and chemiluminescence detection (Obata, Karatani, & Nakayama, 1993; Obata, Karatani, Matsui, & Nakayama, 1997). The determined Fe is the chemically labile species, which strongly reacts with 8-hydroxyquinoline resin at pH 3.2. In this paper, observed Fe concentrations are defined as: Total (unfiltered), labile particulate (>0.22 lm), dissolved (<0.22 lm), colloidal (200 kDa–0.22 lm) and soluble (<200 kDa) (Nishioka et al., 2001). Niskin bottle samplings were carried out to measure vertical profiles of the SF6 concentration, however, these bottle samples could not be determined due to SF6 contamination. Therefore, we mainly focus on the horizontal behavior of the patch in this paper.

3. Results and discussion 3.1. Physical conditions The center of the western subarctic gyre was located at 49–50N, 165E, and its southern boundary at 47N, 165E from the pre-survey XBT observations. An experimental area was chosen at 48.5N, 165E. Chlorophyll a, fCO2 and Fv/Fm (quantum efficiencies of algal photosystem II) were approximately constant in this area during the pre-survey (Tsuda et al., 2003). The mixed layer depth was defined as the depth where the difference of potential temperature to that at the sea surface was over 1 C. On the release date (July 18, defined as Day 0), the mixed layer depth was 8.5 m. From Day 0 to Day 4, the mixed layer depth increased from 8.5 to 10 m and the sea surface temperature also increased from 7.5 to 9 C. The mixed layer depth then increased to 15 m and sea surface temperature also increased to 9.5 C by Day 7. After Day 7, the mixed layer depth and sea surface temperature were relatively constant (Fig. 4). Increase of sea surface temperature corresponded to the increase of air temperature. Temporal changes of mixed layer depth and sea surface temperature outside the patch were similar to those inside the patch; the mixed layer depth changed from 8 to 15 m during experiment (Fig. 4). These mixing depths were shallower than those of previous experiments: about 30 m at the IronEx I site (Stanton et al., 1998) and 65 m at the SOIREE site (Boyd et al., 2000). From Day 0 to Day 6, the daily mean wind speed increased from 5 to 10 m s
1. Temporal changes of the mixed layer depth (Fig. 4) corresponded to the changes of wind speed.

4. Patch behavior After the Fe release, the GPS buoy which had been deployed at the center of the Fe patch was advected toward the west at about 7.9 · 10
2 m s
1 (6.8 km d
1) (Fig. 5). It moved away from the center of the patch during the experiment due to wind stress (Stanton et al., 1998), therefore, it was re-deployed after the propeller sampling of Days 3, 6, 8, 10 and 12 (Fig. 5). Sea surface height (SSH) on Day 9 (July 27th 2001) by TOPEX/ERS-2 is shown in Fig. 6 with the trajectory of the center position of the patch from July 18 to 30. The Fe solution and the GPS buoy were released in a mesoscale eddy, and the movement of the patch was anticlockwise along a contour of SSH, which suggest SSH imagery by satellite is useful to predict the movement of a Fe-enriched water mass. The SF6 concentrations decreased progressively due to ocean mixing and air-sea gas exchange, and Fe concentrations also decreased due to oceanic mixing and biochemical processes. The first propeller survey throughout the patch after the Fe release (within 4.5–11.5 h from completing the release) showed a

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

117

Wind speed (m s-1)

20

15

10

5

Temperature (degree C)

0 12

Air Sea surface

11 10 9 8 7 6

Mixed layer depth (m)

20

In patch Out patch

18 16 14 12 10 8 -2

0

2

4

6

8

10

12

14

Days from fertilization

Fig. 4. Temporal changes of the wind speed, air and sea surface temperature, and mixed layer depth for 14 days.

significant increase of dissolved Fe. Mean concentration in the patch was 2.89 nM, and maximum concentration was 6.02 nM in the first propeller survey on Day 1. Dissolved Fe concentrations subsequently decreased, and the loss rate gradually decreased (Fig. 7). The dissolved Fe half-life of 43–69 h was relatively longer than in previous Fe-enrichment studies (28–40 h at the IronEx I, Gordon, Johnson, & Coale, 1998 and 14–34 h at the SOIREE, Bowie et al., 2001).

5. Dilution rate The maximum concentration of SF6 in the patch decreased exponentially with time (Fig. 8). Decrease of SF6 concentration was caused by horizontal and vertical diffusion and air-sea gas exchange. The gas

118

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127 4 8 .8 0

Laitutde (degree)

4 8 .7 5 4 8 .7 0

Day6

4 8 .6 5 4 8 .6 0

Day3 Release

4 8 .5 5

Day12

4 8 .5 0

Day10

Day1

Day8

4 8 .4 5 4 8 .4 0 1 6 3 .6

1 6 3 .8

1 6 4 .0

1 6 4 .2

1 6 4 .4

1 6 4 .6

1 6 4 .8

1 6 5 .0

1 6 5 .2

Longitude (degree)

Fig. 5. Trajectories of the GPS buoy in the regular coordinate system as black dots. Red dots show the trajectories during the release and the propeller surveys. Blue arrows show the re-deployment of the GPS buoy.

Fig. 6. Sea surface height (SSH) (Day 9, July 27, 2001) by TOPEX/ERS-2 (CCAR of Colorado Univ., http://e450.colorado.edu/ realtime/welcome/) and trajectory of the GPS buoy (red squares).

transfer speed of SF6 from ocean to atmosphere depends on the wind speed (Liss & Merlivat, 1986). Wanninkhof (1992) summarized a Schmidt Number for SF6 to calculate the gas transfer velocity k. The flux of SF6 from ocean to atmosphere (F) was calculated by the following equation

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

Fe conc. ( nM)

10 (a)

9

Fe conc. (nM)

8 7 6 5

4

(b)

119

dissolved Fe (IN underway) dissolved Fe (OUT)

3 2 1 0 0

4

2

4

6

8

10

12

14

Days from fertilization

3 2 1 0

1

2

3

4

5

6

7

8

9

10

11

12

13

Days from fertilization Fig. 7. Dissolved Fe (<0.22 lm) concentrations from surface underway transect across the Fe-enriched patch during the experiment. (a) Each datum is plotted vs time from fertilization. Bar shows the ideal concentration when Fe was diluted inside the release area in the mixed layer. (b) Average value in each day of the propeller survey observation vs time from fertilization.

SF6 concentration (moles L-1)

1E-12

1E-13

1E-14 0

1

2

3

4

5

6

7

8

9

10

11

12

13

Days from fertilization

Fig. 8. Maximum concentrations of SF6 in patch observed values. Dots are observed concentrations. Triangles are the concentration without the loss of air-sea gas exchange.

F ¼ kðC w
aC a Þ;

ð1Þ

where Cw is the SF6 concentration in the surface water, a is the Ostwald solubility coefficient, and Ca is the SF6 concentration in the air. In this calculation, maximum concentration at the center of patch was used. Dilution rate before accounting for the effect of the loss by air-sea gas exchange changed on Day 6 or 7. In the first half (until Day 6 or 7), the dilution rate was smaller than in the latter half, which accompanied a wind speed increase during the experimental period (Fig. 4). The vertical and horizontal components of diffusivity cannot be separated in this experiment because we did not measure vertical profiles of SF6 concentration.

120

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

To determine the dilution rate, SF6 concentrations were summarized according to the radius from the center of patch. Patch configuration is idealized as circular to simplify this analysis (Fig. 9). Increases of radius correspond to spatial reduction of SF6 concentration, and SF6 concentration decreased temporally. In general, the maximum radius at which SF6 concentration can be detected is expected to increase over time, however, maximum radius did not change significantly during SEEDS (Fig. 9). This might have been caused by the lack of observations far from the center. Patch radius based on the second moment of the patch distribution, W 2r , was calculated with equation (1) (Martin, Richards, Law, & Liddicoat, 2001)
2 M2 M1 2
; ð2Þ Wr ¼ M0 M0 where M0 ¼

X

Cðri Þ;

M1 ¼

i

X

Cðri Þri ;

M2 ¼

i

X

Cðri Þr2i ;

i

and C(ri) and ri are the SF6 concentration and radius from the center of patch, respectively, for the ith datum. Diffusion coefficients were calculated from the slope of W 2r versus time. In the first half and latter half, apparent diffusion coefficients were estimated to be 9.0 and 5.0 · 101 m2 s
1, respectively (Fig. 10). These values included both horizontal and vertical dilution as mentioned above in this experiment. In IronEx I, horizontal diffusivity was measured to be 2.0–6.0 · 102 m2 s
1 in the equatorial Pacific Ocean (Stanton et al., 1998). Vertical diffusivity was not measured in SEEDS. Vertical diffusivity is considerably smaller than horizontal diffusivity. Therefore, it is considered that horizontal diffusivity in SEEDS was smaller than in IronEx I throughout the observed period. Contours of SF6 concentration are described by a natural neighbor interpolation method (NCAR Command Language, NCL) to confirm how well the SF6 patch can be traced quantitatively. These contours help us to understand the behavior of the patch visually. Contours of SF6 concentration on Days 3, 6, 8, 10 and 12 show the decrease of maximum concentration at the center of the patch (Fig. 11). In general,

D ay3 -1

SF6 concentration (fmol L )

-1

SF6 concentration (fmol L )

100

10

1

D ay6

100

10

1

0

2

4

6

8

10

12

14

16

18

20

0

2

4

1

)

-1

-1

10

10

1

0

2

4

6

8

10

12

14

Radius from patch center (km)

16

18

20

8

10

12

14

16

18

20

D ay12

D ay10

100

6

Radius from patch center (km)

SF6 concentration (fmol L

D ay8

100

SF6 concentration (fmol L )

SF6 concentration (fmol L

-1

)

Radius from patch center (km)

100

10

1

0

2

4

6

8

10

12

14

Radius from patch center (km)

16

18

20

0

2

4

6

8

10

12

14

Radius from patch center (km)

Fig. 9. Observed SF6 concentration vs. radius from the center of the patch.

16

18

20

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127 2 .0 x 1 0 1 .8 x 1 0

1 .4 x 1 0

7

7

1

7

2

-1

K = 5 .0 x 1 0 (m s )

7

2

Wr ( m )

1 .6 x 1 0

121

7

2

1 .2 x 1 0 1 .0 x 1 0 8 .0 x 1 0 6 .0 x 1 0 4 .0 x 1 0

7

0

2

5

6

-1

K = 9 .5 x 1 0 (m s )

6

6

6

0

1

2

3

4

7

8

9

10

11

12

13

Days from fertilization Fig. 10. The widths based on the second moment of the patch distribution W 2r versus radius from the center of the patch.

the patch area is expected to increase temporally by horizontal dilution, but the area of the patch in this experiment did not show such a trend, and especially, the patch area was smallest on Day 12 (Fig. 12(a)). Moreover, the total amount of SF6 in the patch calculated from the area of the patch and the mixed layer depth decreased with time, and was always smaller than the total amount of SF6 expected to remain after accounting for air-sea gas exchange (Fig. 12(b)). In these estimations, the patch was defined by concentrations over 2 fM, which might cause considerable underestimation of the patch area during the later period of the observation. Threshold values lower than 2 fM were not effective in estimating a more correct area of the patch, because of the deficiency of observed points in the background areas. Therefore, we focused on the behavior of SF6 and Fe at the center of patch. 5.1. Behavior of released iron Prior to the release of Fe, dissolved Fe concentration (<0.22 lm filtrate) had a nutrient-type distribution characterized by depletion in the surface mixed layer (0.1–0.09 nM), increasing in deep water at the target point (Station PF, 48.5N, 165E) (Fig. 13). Dissolved Fe concentrations from the underway survey also showed very low values in the surface mixed layer around the target point (0.05 ± 0.02 nM n = 28), and these data demonstrate that the dissolved Fe level around the experimental site was low enough to limit phytoplankton growth. Changes in Fe concentration integrated over 0–10 m at the center of the patch, with proportions of each size-fractionated pool, are shown in Fig. 14 and Table 2. Iron was removed from the surface mixed layer by sinking and converted to undetectable form such as intracellular Fe. On Day 9 after Fe release (peak of the phytoplankton bloom), 67% of total Fe compared to the total Fe on Day 2 remained in the surface mixed layer (0–10 m) in the patch. On the other hand, dissolved Fe concentration subsequently decreased during phytoplankton growth in the patch and Fe in the colloidal fraction decreased the most (Figs. 7, 14, and 15) (Nishioka et al., 2003). This disappearance of dissolved Fe resulted mainly from transformation of colloidal Fe to larger particles (Nishioka et al., 2003) and from biological uptake. However, the half-life of dissolved Fe concentration in the SEEDS experiments was longer than in previous experiments, as described above,

122

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

Fig. 11. Contours of SF6 concentration in the Lagrangian coordinate system. Dots show the observational points.

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

123

Patch arae (km 2)

500

400

300

200

100

Release 0 0

2

6

8

10

12

Days from fertilization

(a) 0 .5

Total amount of SF 6( moles)

4

Release

0 .4

0 .3

0 .2

0 .1

0 .0 0

(b)

2

4

6

8

10

12

Days from fertilization

Fig. 12. Temporal change of (a) patch area and (b) total amount of SF6 estimated from figure 11 on the Day 3, 6, 8, 10 and 12.

and the existence of dissolved Fe, mainly as colloidal Fe, in the first half of the experimental period would be advantageous for phytoplankton growth. The smaller diffusivity and shallower mixed layer depth in the first half made the half-life of dissolved Fe longer. Wells (2003) reported that biological uptake of soluble Fe was insufficient to support the observed bloom development in the IronEx II experiment. A similar disproportion was observed in the SEEDS experiment. Calculating from the measured net carbon fixation of 485 mM C m
2 (Day 2–Day 9) and the smallest reported Fe quota for an oceanic diatom, 5 lm Fe M
1 C (Sunda & Huntsman, 1995), the minimum amount of Fe needed to support the observed diatom bloom in SEEDS was 2.4 lM m
2 (Day 2–Day 9). As Wells (2003) pointed out, this value probably underestimates the actual Fe demand, because grazing pressure is not taken into account and large diatoms, dominant phytoplankton in the SEEDS study, have higher cellular Fe quotas (Timmermans et al., 2001). This estimate suggests that uptake of soluble Fe was insufficient to support the diatom growth observed in SEEDS, and indicates that part of another Fe fraction, such as colloidal Fe, became bio-available Fe during the phytoplankton growth. In the patch, 57% of total Fe (an estimate based the level of total Fe on Day 2 defined as 100%) disappeared from the surface mixed layer during the experiment (from Day 2 to Day 13), by dilution, scavenging

124

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

Fe conc. (nM) 0

0.5

1

1.5

2

0 100

Depth (m)

200 300 400 500 600 700 800 900 Fig. 13. Vertical distributions of dissolved and total Fe (unfiltered) concentration at the experiment site before Fe release. Open points show dissolved Fe concentration. Solid points show total Fe concentration.

Fig. 14. Changes in the 0–10 m integrated total Fe concentration at the center of the patch with the proportion of each size fractionated pools. We defined that 100% as the sum of the Fe amount in every size fraction in the surface mixed layer on Day 2.

and/or conversion into undetectable forms (Fig. 16). However, a significant amount of Fe remained in the Fe-enriched patch as the labile particle fraction (labile particle fraction accounted for 88% of total Fe on Day 13) at the end of experiment (Fig. 14). This Fe in the labile particle fraction in the Fe-enriched patch is

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

125

Table 2 Integrated Fe concentration in each size-fraction at the center of iron enrichment patch (Integrated 0–10 m) Days after Fe release

Soluble Fe (lmol m2)

Colloidal Fe (lmol m
2)

Particulate Fe (lmol m
2)

Total Fe

2 9 13

3.1 0.9 1.2

13.2 2.4 0.4

15.3 18.0 11.8

31.6 21.3 13.4

0

0.0

1.0

Dissolved Fe conc. (nM) 2.0 3.0

4.0

10

1.0

Total Fe conc. (nM) 2.0 3.0

4.0

10

20 Depth (m)

0

0.0

(a)

20

30

30

40

40

50

50

60

60

70

70

80

80

D0-I D2-I D4-I D7-I D9-I D11-I D13-I

(b)

Fig. 15. Vertical Fe distributions at the in-patch station during the SEEDS experiment. (a) Dissolved Fe and (b) total Fe.

0 20

Depth (m)

40

1.2 1.05

Total Fe (Unfiltered)

60

0.9 0.75

0

0.6

20

0.45 0.3

40

Dissolved Fe (< 0.22 µm)

0.15 0

60 -8

–6 –4 –2 0 2 4 6 Distance from the center buoy (km)

8

Fig. 16. Cross-sections of dissolved Fe and total Fe (nM) concentrations across the Fe-enriched patch on Day 12.

126

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

evident in the total Fe vertical section data on Day 12 (Fig. 16). After Day 8, the Si:N consumption ratio increased (Kudo et al., 2005) and the Fv/Fm ratio decreased (Tsuda et al., 2003) in the Fe-enriched patch. These physiological changes indicate that physiological Fe stress occurred for diatoms after Day 8. Therefore, the conversion of dissolved Fe to particulate form will ultimately reduce the bio-availability of Fe newly introduced into the photic zone. 6. Summary We successfully traced an Fe-enriched patch for 14 days by using SF6 tracer and a GPS buoy during the SEEDS experiment. A fast and massive response of diatoms to the Fe supply was observed compared to previous experiments in other HNLC areas. In addition to the response of the phytoplankton species in the western North Pacific, we have shown that smaller diffusivity was effective to sustain the elevated Fe concentration in comparison to previous experiments, and apparently allowed the growth of a very dense phytoplankton bloom. Our results quantify the fate of Fe in a patch fertilized with Fe only once, using the dilution rate determined by SF6 concentration. Acknowledgments We thank Y.W. Watanabe for his useful comments on the manuscript and suggestions for the SF6 analysis. We also thank H. Saito, C.S. Law, A.J. Watson and H.J.W. de Baar for their useful suggestions for the water-mass tracing. We are also grateful to the captain, crew, and scientists aboard the R.V. Kaiyo-Maru in SEEDS experiment. This work was supported by Global Environmental Research Fund from the Ministry of Environment, the Fisheries Agency and CRIEPI research funding.

References Boyd, P. W., Watson, A. J., Law, C. S., Abraham, E. R., Trull, T., Murdoch, R., et al. (2000). A mesoscale phytoplankton bloom in the polar Southern Ocean stimulated by iron fertilization. Nature, 407, 695–702. Bowie, A. R., Maldonado, M. T., Frew, R. D., Croot, P. L., Achterberg, E. P., Mantoura, R. F. C., et al. (2001). The fate of added iron during a mesoscale fertilization experiment in the Southern Ocean. Deep Sea Research II, 48, 2703–2743. Coale, K. H., Johnson, K. S., Fitzwater, S. E., Gordon, R. M., Tanner, S., Chavez, F. P., et al. (1996). A massive phytoplankton bloom induced by an ecosystem-scale iron fertilization experiment in the equatorial Pacific Ocean. Nature, 383, 495–501. Gervais, F., Riebesell, U., & Gorbunov, M. Y. (2002). Changes in primary productivity and chlorophyll a in response to iron fertilization in the Southern Polar Frontal Zone. Limnology and Oceanography, 47, 1324–1335. Gordon, R. M., Johnson, K. S., & Coale, K. H. (1998). The behavior of iron and other trace elements during IronEx I and PlumEx experiments in the equatorial Pacific. Deep Sea Research II, 45, 995–1041. Kudo, I., Noiri, Y., Imai, K., Nojiri, Y., Nishioka, J., & Tsuda, A. (2005). Primary production and nitrogenous nutrients assimilation dynamics during the subarctic pacific iron experiment for ecosystem dynamics study (SEEDS). Progress in Oceanography, doi: 10.1016/j.pocean.2005.02.009. Law, C. S., Watson, A. J., Liddicoat, M. I., & Stanton, T. (1998). Sulphur hexafluoride as a tracer of biogeochemical and physical processes in an open-ocean iron fertilization experiment. Deep Sea Research II, 45, 977–994. Ledwell, J. R., & Watson, A. J. (1991). The Santa Monica Basin tracer experiment: a study of diapycnal and isopycnal mixing. Journal of Geophysical Research, 96, 8695–8718. Liss, P. S., & Merlivat, L. (1986). Air-sea gas exchange rate: introduction and synthesis. In P. Buat-Menard (Ed.), The role of air-sea exchange in geochemical cycling (pp. 113–129). Hingham, MA: D. Reidel. Martin, J. H., Coale, K. H., Johnson, K. S., Fitzwater, S. E., Gordon, R. M., Tanner, S. J., et al. (1994). Testing the iron hypothesis in ecosystems of equatorial Pacific Ocean. Nature, 371, 123–129. Martin, A. P., Richards, K. J., Law, C. S., & Liddicoat, M. (2001). Horizontal dispersion within an anticyclonic mesoscale eddy. Deep Sea Research II, 48, 739–755.

D. Tsumune et al. / Progress in Oceanography 64 (2005) 111–127

127

Nishioka, J., Takeda, S., Kudo, I., Tsumune, D., Yoshimura, T., Kuma, K., et al. (2003). Size-fractionated iron distributions and ironlimitation processes in the subarctic NW Pacific. Geophysical Research Letters, 30(14), 1730. Nishioka, J., Takeda, S., Wong, C. S., & Johnson, W. K. (2001). Size-fractionated iron concentrations in the northeast Pacific Ocean: distribution of soluble and small colloidal iron. Marine Chemistry, 74, 157–179. Obata, H., Karatani, H., Matsui, M., & Nakayama, E. (1997). Fundamental studies for chemical speciation of iron in seawater with an improved analytical method. Marine Chemistry, 56, 97–106. Obata, H., Karatani, H., & Nakayama, E. (1993). Automated determination of iron in seawater by chelating resin concentration and chemiluminescence detection. Analytical Chemistry, 65, 1524–1528. Stanton, T. P., Law, C. S., & Watson, A. J. (1998). Physical evolution of the IronEx-I open ocean tracer patch. Deep-Sea Research II, 45, 947–975. Sunda, W. G., & Huntsman, S. A. (1995). Iron uptake and growth limitation in oceanic and coastal phytoplankton. Marine Chemistry, 50, 189–206. Timmermans, K. R., Gerringa, L. J. A., de Baar, H. J. W., van der Wagt, B., Veldhuis, M. J. W., de Jong, J. T. M., et al. (2001). Growth rate of large and small Southern Ocean diatoms in relation to availability of iron in natural seawater. Limnology and Oceanography, 46, 260–266. Tsuda, A., Takeda, S., Saito, H., Nishioka, J., Nojiri, Y., Kudo, I., et al. (2003). A mesoscale iron enrichment in the western subarctic Pacific induces large centric diatom bloom. Science, 300, 958–961. Upstill-Goddard, R. C., Watson, A. J., Liss, P. S., & Liddicoat, M. I. (1990). Gas transfer velocities in lakes measured with SF6. Tellus B, 42, 364–377. Wanninkhof, R. (1992). Relationship between wind speed and gas exchange over the ocean. Journal of Geophysical Research, 97, 7373–7382. Watanabe, W. Y., Shimamoto, A., & Ono, T. (2003). Comparison of time-dependent tracer ages in the western North Pacific: Oceanic background levels of SF6, CFC-11, CFC-12, and CFC-113. Journal of Oceanography, 59, 719–729. Wells, L. M. (2003). The level of iron enrichment required to initiate diatom blooms in HNLC waters. Marine Chemistry, 82, 101–114.