Dissolved iron and its speciation in a shallow eutrophic lake and its inflowing rivers

Dissolved iron and its speciation in a shallow eutrophic lake and its inflowing rivers

ARTICLE IN PRESS WAT E R R E S E A R C H 41 (2007) 775 – 784 Available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/watres D...

391KB Sizes 1 Downloads 46 Views

ARTICLE IN PRESS WAT E R R E S E A R C H

41 (2007) 775 – 784

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

Dissolved iron and its speciation in a shallow eutrophic lake and its inflowing rivers Takashi Nagaia,b,, Akio Imaia, Kazuo Matsushigea, Kunihiko Yokoic, Takehiko Fukushimab a

National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba, Ibaraki 305-0053, Japan Environmental Modeling and Creation, Integrative Environmental Sciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, Tennoudai, Tsukuba, Ibaraki 305-8572, Japan c Division of Natural Science, Osaka Kyoiku University, Kashiwara, Osaka 582-8582, Japan b

art i cle info

ab st rac t

Article history:

It has been suggested that iron is a limiting factor on bloom-forming cyanobacteria in lake

Received 25 August 2006

water. Although the availability of iron for phytoplankton depends significantly on its

Accepted 8 October 2006

speciation, little is known about iron speciation in natural lake water. We investigated the

Available online 8 January 2007

horizontal distribution and temporal variation of dissolved iron and its chemical speciation

Keywords:

in Lake Kasumigaura and its two inflowing rivers. Concentrations of dissolved iron and its

Iron

organic ligands, determined by cathodic stripping voltammetry, clearly decreased as lake

Organic ligand

water flowed from the river entry points toward the center of the lake, indicating their

Speciation

riverine origin. The fraction of iron occurring in organic complexes tended to increase with

Voltammetry

the water flow. These results suggest that most of the dissolved iron in river water forms unstable soluble species, such as inorganic iron; thus, these unstable iron species may be scavenged in the mouths of rivers, and stable organic complexes of iron may flow to the center of the lake. Furthermore, most of the dissolved iron (88.2–99.9%) was present as organic complexes, and the inorganic iron level in Lake Kasumigaura (pFe0 value ¼ 7.8–13.6) was similar to that observed in the open ocean. This result suggests that iron is an important factor determining the structure of the phytoplankton community in Lake Kasumigaura. & 2006 Elsevier Ltd. All rights reserved.

1.

Introduction

It is now well documented that iron limits primary production in the open ocean (e.g., Martin et al., 1994). Furthermore, although the concentration of dissolved iron in lake water is greater than that in seawater, it has been shown that iron can also limit phytoplankton growth in natural lake waters (Clasen and Bernhardt, 1974; Lin and Schelske, 1981; Evans and Prepas, 1997; Hyenstrand et al., 2000; Twiss et al., 2000). Lake Kasumigaura is a shallow, unstratified, eutrophic lake in Japan, and it has been suggested that iron may limit the

growth of bloom-forming cyanobacteria in the lake (Yagi et al., 1987; Aizaki and Aoyama, 1995). The limitation is thought to be caused by the low availability of iron in the lake (Imai et al., 1999). Bioavailability of iron for phytoplankton is changed significantly by iron complexation with dissolved organic matter. It is therefore essential to determine the chemical speciation of iron as well as its concentration in order to examine the effect of iron on phytoplankton growth. Inorganic iron (Fe0 , the sum of free hydrated and hydrolyzed ferric iron species) is directly available for phytoplankton uptake,

Corresponding author. Environmental Modeling and Creation, Integrative Environmental Sciences, Graduate School of Life and Environmental Sciences, University of Tsukuba, Tennoudai, Tsukuba, Ibaraki 305-8572, Japan. Tel.: +81 29 853 7189; fax: +81 29 853 4210. E-mail address: [email protected] (T. Nagai). 0043-1354/$ - see front matter & 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2006.10.038

ARTICLE IN PRESS 776

WA T E R R E S E A R C H

whereas organic-iron complexes are kinetically less labile than Fe0 and are thus thought to be unavailable (Hudson and Morel, 1993). Therefore, organic and inorganic iron must be considered separately to understand the effect of iron speciation on phytoplankton growth. However, little is known about the chemical speciation of iron in lake water. In the oceanic water column, most of the dissolved iron is strongly complexed with organic ligands (Gledhill and van den Berg, 1994; Rue and Bruland, 1995; van den Berg, 1995; Boye et al., 2001). Cathodic stripping voltammetry (CSV) was employed in these studies to determine the chemical speciation of iron as well as its concentration. However, it was difficult to study metal complexation by CSV in lake water because it contains a high concentration of organic matter, part of which is surface-active. In our previous study, the problem associated with CSV measurement in lake water (instability caused by interfering organic matter) was overcome by using a high concentration of 1-nitroso-2-naphthol (NN, 50 mM), a well-conditioned voltammetric cell, and diluted samples (Nagai et al., 2004). The aim of our present study was to understand the dynamics of dissolved iron and its speciation in lake water by investigating their horizontal distribution and temporal variations in Lake Kasumigaura. Inaba et al. (1997) suggested that dissolved iron in Lake Kasumigaura was mainly of river origin. Therefore, we extended our study area to the two main rivers flowing into Lake Kasumigaura. We used the CSV method using NN with bromate at pH 8.1 to determine the concentration of dissolved iron and its organic ligands. Some restrictions in our CSV determination are as follows: we defined ‘‘dissolved iron’’ as iron that can pass through a 0.2mm filter; we fixed the pH at 8.1 in the speciation experiment to calculate the conditional stability constant for the complexation of iron(III); and we did not consider redox speciation. Recently, argument was put forward based on theoretical grounds that the CSV method has inherent kinetic limitations and an equilibrium-based interpretation of the CSV measurements of iron (III) speciation is clearly questionable (Town and van Leeuwen, 2005). The validity of the CSV technique was deliberated in the section of discussion.

2.

Materials and methods

2.1.

Study area and sample collection

Lake Kasumigaura is the second largest lake in Japan, and it is located 50 km northeast of Tokyo. It has a surface area of 171 km2, a mean depth of 4 m, and a maximum depth of 7.3 m. The lake is well known for eutrophication and is so shallow that vertical stratification is easily destroyed by a moderately strong wind. The lake has two large bays, Takahama-iri and Tsuchiura-iri. The Koise and Sakura rivers are the main rivers influent into Takahama-iri and Tsuchiura-iri, respectively. Water tends to flow through the lake from the northwest, where Takahama-iri and Tsuchiura-iri are located, to the southeast, to the effluent Hitachitone River. The apparent retention time is 2.7 months for Takahama-iri, 7.8 months for Tsuchiura-iri, and 8.3 months for the central basin. However, the average time needed for water to flow through the two

41 (2007) 775– 784

bays to the center of the lake may be around 3 months because the wind-driven currents are fairly strong. Lake-water samples were collected at four stations (Fig. 1). Station 1 (St. 1) is in the deepest part of Takahama-iri and is affected directly by the Koise River; St. 2 is at the center of Takahama-iri; St. 3 is located in Tsuchiura-iri; and St. 4 is located in the center of the lake. River water samples were collected in the Koise and Sakura rivers at downstream sites (Fig. 1). Water sampling was performed in April, July, and October 2002 and in January 2003. Surface-water samples for iron analysis were collected directly into 250-ml high-density polyethylene bottles. At the same time, separate samples were collected for analyses of chlorophyll a (Chl a), nutrients, and dissolved organic carbon (DOC). The samples were immediately cooled in an ice cooler and brought back to the laboratory. The samples for iron analysis were then filtered through a 0.2-mm pore-size polycarbonate membrane filter (Nuclepore, Whatman, Brentford, UK). A portion of the filtrates was stored at 3 1C in Teflon vials after acidification to pH 2.5 with HCl for the later determination of total dissolved iron. The remaining filtrate was stored frozen (30 1C) in high-density polyethylene bottles until analysis of iron speciation. Separate samples were filtered through precombusted (450 1C for 4 h) glass-fiber filters (Whatman GF/F, average pore size 0.7 mm). The filters were used for Chl a analysis, and the filtrates were used for nutrients and DOC analyses. The high-density polyethylene bottles were cleaned by soaking in 3 M HCl for 3 days and then rinsing with Milli-Q water (MQ, resistance 18.3 MO; Millipore, Billerica, MA, USA). Teflon vials were cleaned by soaking in 3 M HCl for 3 days, and then soaking in 2 M HNO3 for 3 days, and finally rinsing with MQ.

Fig. 1 – Sampling sites in Lake Kasumigaura and its inflowing rivers; () and (’) represent the sampling points of lake and rivers, respectively.

ARTICLE IN PRESS WAT E R R E S E A R C H

2.2.

Equipment and reagent for iron analysis

The voltammetric system consisted of a Princeton Applied Research (PAR, Oak Ridge, TN, USA) 303A static mercury drop electrode connected to a PAR 394 voltammetric analyzer. The working electrode was a hanging mercury drop (medium size), the reference electrode was Ag/AgCl in 3 M KCl saturated with AgCl, and the counter electrode was a platinum wire. Solutions in the Teflon voltammetric cell were stirred with a Teflon-coated magnetic stirring bar driven by a PAR 305 electric stirring motor. MQ was used for reagent and sample preparation. A 0.02 M stock solution of 1-nitroso-2-naphthol (NN) was prepared in methanol (‘‘infinity-pure’’ grade, Wako, Osaka, Japan). A 1 M stock solution of tris(hydroxymethyl)aminomethane (Tris) was adjusted to pH 8 with HCl (Suprapur grade, MERCK, Darmstadt, Germany). A 0.4 M stock solution of potassium bromate and a 5 M stock solution of NaCl were prepared in MQ. Iron contaminants were removed from the Tris, potassium bromate, and NaCl stock solutions (the potassium bromate and NaCl stock solutions were buffered at pH 8 with 10 mM Tris) by adding 20 mM NN and passing the mixture through a Sep-Pak C18 cartridge (Waters, Milford, MA, USA; precleaned with methanol, HCl, and then MQ). Iron standard solutions were prepared by diluting a 100-ppm Fe standard (Wako) with MQ and acidifying to pH 2.5 with HCl.

2.3.

Determination of total dissolved iron

Samples for the determination of total dissolved iron were UVirradiated prior to analysis to decompose interfering organic compounds. Samples (10 ml) were placed in acid-washed quartz tubes and then UV-irradiated with a 400-W low-pressure Hg lamp for 60 min. UV-irradiated samples were diluted with an appropriate amount of MQ (10  for lake waters and the Koise River in July 2002 and 50  for other river waters), and 10 ml aliquots of the diluted solutions were pipetted into Teflon vials. Ten microliters of a 0.02 M NN solution (final concentration 20 mM) and 100 ml of a 5 M NaCl solution (final concentration 50 mM) were added to the samples. The pH was approximately neutralized using ammonia solution, and 100 ml of a 1 M Tris solution (final concentration 10 mM) was added (final pH 8.1). The solution was deaerated by purging for 4 min with nitrogen gas, and 250 ml of a 0.4 M potassium bromate solution (final concentration 10 mM) was added prior to the voltammetric scan. Deposition onto a fresh mercury drop was carried out for 30 s at 0.15 V while the solution was stirred. The stirrer was stopped, and 10 s later the potential was scanned in differential pulse stripping mode (pulse height 20 mV) from 0.15 to 0.7 V at a scan rate of 20 mV s1. This measurement was repeated with three standard additions of iron to the sample sufficient to double the peak height, and quantification was made by the standard addition method. The detection limit of our method was 1.8 nM (3  the standard deviation of the blank measurements).

2.4.

Determination of iron speciation

The conditional stability constants and complexation capacities of the natural iron-complexing ligands in the fresh-

41 (20 07) 77 5 – 784

777

water samples were determined by a competitive ligand equilibration-CSV (CLE-CSV) method (Gledhill and van den Berg, 1994; Nagai et al., 2004). Appropriate sample amounts (10 ml for lake waters and the Koise River in July 2002, and 2 ml for the other river waters) were diluted to 100 ml with MQ. The diluted solutions were mixed with NN (final concentration 50 mM), Tris (final concentration 10 mM), and NaCl (final concentration 50 mM). Appropriate amounts of a 1.79 mM (100 ppb) iron standard solution were pipetted into Teflon vials in nine different concentrations, and 10 ml of the mixture was pipetted into each vial. The added iron, NN, and natural complexing ligands were allowed to equilibrate overnight with gentle shaking. At the same time, the voltammetric cell was conditioned in the remaining 10 ml of the mixture. The iron complexed by the added NN was determined by CSV after purging and addition of potassium bromate (final concentration 10 mM). The voltammetric procedure was the same as that described above. Details for the iron speciation calculations were described previously (Nagai et al., 2004). Ligand concentrations (CL) and conditional stability constants (K0 FeL ¼ [FeL]/[Fe3+][L0 ]) were determined by linear least-squares regression of the data fitted to the following equation (Gledhill and van den Berg, 1994): ½Fe labile=½FeL ¼ ½Fe labile=CL þ ða0Fe þ a0FeNN Þ=ðCL K0FeL Þ,

(1)

where [Fe labile] is the concentration of iron complexed by the added NN as well as inorganic iron, and [FeL] is the concentration of iron complexed by the natural organic ligand, L. Values for a0 Fe and a0 FeNN (the a-coefficients for inorganic complexation of iron and complexation of Fe3+ by NN) of 1012.3 and 1016.7 were used, respectively (Nagai et al., 2004). The concentration of inorganic iron [Fe0 ] originally present in the sample was calculated from a0 Fe, CFe (concentration of total dissolved iron), CL, and K0 FeL on a thermodynamic equilibrium basis. The fraction of iron originally present as organic species was calculated from {(CFe[Fe0 ])/ CFe}  100, and pFe0 is defined as log[Fe0 ].

2.5.

Other sample analyses

Water temperature, pH, and dissolved oxygen (DO) were measured at sampling sites. Electric conductivity (EC) was measured at the laboratory with a TOA CM-20E conductivity meter (DKK-TOA Co., Tokyo, Japan). Chl a concentration was estimated spectrophotometrically after extraction with 100% methanol. DOC was measured with a Shimadzu TOC-5000 total-organic carbon analyzer equipped with a Pt catalyst on quartz wool (Shimadzu Co., Kyoto, Japan). Inorganic nutrients (ammonium, nitrite, nitrate, and phosphate) were measured with an autoanalyzer (TRAACS 800, Bran+Luebbe, Tokyo, Japan).

3.

Results

Water temperature, pH, EC, Chl a, DOC, and nutrients (ammonium, nitrite, nitrate, and phosphate) in Lake Kasumigaura, the Sakura River, and the Koise River are shown (Table 1). Water flowed in two directions: (1) from the Koise

ARTICLE IN PRESS 778

WA T E R R E S E A R C H

River to St. 1, St. 2, and then St. 4; (2) from the Sakura River to St. 3 and then St. 4. Nitrate concentration tended to decrease in the flow direction. Phosphate concentration did not show a clear decreasing trend in the flow direction, indicating more than one origin including river. The distribution of Chl a indicated that primary production in the two bays was higher than that in the center of the lake. DOC concentrations in the lake water were ranged from 2.61 to 5.68 mg C l1 and higher than those in the river waters, which ranged from 1.15 to 2.59 mg C l1. The concentrations of total dissolved iron and its speciation are shown in Table 2. The concentration of total dissolved iron ranged from 35 to 254 nM in Lake Kasumigaura and from 47 to 2910 nM in the two rivers. A low level of dissolved iron concentration (47 nM) in the Koise River in July 2002 was observed, but it is regarded as an abnormal value because of regional heavy rainfall on the day before sampling. The dissolved iron concentration showed a clear decrease in the direction of water flow, except for the Koise River in July 2002. This indicated that the rivers are the main source of iron to the lake water. Lake sediment is another source of iron when the sediment becomes reductive. However, the DO values of the lake water from the surface to the bottom did not fall below 4 mg l–1 (Fig. 2) because of its shallowness. Therefore, the release of iron from the lake sediment seems to be small. With respect to temporal variations of dissolved iron, the concentration in the lake water was relatively high in October

41 (2007) 775– 784

2002 and low in April 2002. On the other hand, there was no specific temporal trend in the river waters. The ligand concentrations and conditional stability constants for the lake-water samples were determined from iron titration and linear plots of [Fe labile]/[FeL] versus [Fe labile] (Fig. 3). The distribution of the ligand concentration was similar to that of the dissolved iron concentration, indicating that natural organic ligands of iron(III) in Lake Kasumigaura were also of river origin. The log values of the conditional stability constants were 25.1–26.2, and there was no particular variation. The lack of variation points out a single source of organic ligands in Lake Kasumigaura. According to calculations on the chemical speciation of iron, most of the dissolved iron was present as organic complexes in Lake Kasumigaura. The percentage of iron occurring as organic species in the river waters was relatively low and tended to increase with the water flow (shown clearly in October 2002). When the ligand concentration was higher than the dissolved iron concentration, nearly all (499.9%) of the dissolved iron was present as organic complexes, owing to the very high conditional stability constants of the natural organic ligands. A large variety in pFe0 values (representing negative logarithm of the inorganic iron level) was found in the lake and river waters (range 6.3–13.6). pFe0 values were relatively low in the river waters and tended to increase with the water flow.

Table 1 – Temperature, pH, EC (electric conductivity), Chl a, and nutrients in Lake Kasumigaura, the Sakura River, and the Koise River

April 2002

July 2002

October 2002

January 2003

Koise Sakura St. 1 St. 2 St. 3 St. 4 Koise Sakura St. 1 St. 2 St. 3 St. 4 Koise Sakura St. 1 St. 2 St. 3 St. 4 Koise Sakura St. 1 St. 2 St. 3 St. 4

Temp. (1C)

pH

EC (mS cm 1)

Chl a (mg l1)

DOC (mg C l1)

PO4 (mg P l1)

NH4 (mg N l1)

NO2 (mg N l1)

NO3 (mg N l1)

19.1 18.7 14.8 14.1 14.8 14.0 23.4 25.3 27.6 26.1 27.2 24.8 18.0 19.1 21.5 21.7 21.7 21.3 4.8 4.8 3.7 4.0 4.1 3.9

7.8 7.7 8.6 8.5 8.9 8.4 7.1 7.2 8.1 7.6 7.9 7.6 6.5 6.8 7.3 7.6 7.2 7.3 7.7 7.8 7.8 8.0 7.7 7.6

220 284 241 276 292 281 186 226 276 336 357 376 157 380 221 331 313 385 166 284 263 298 296 305

6 53 156 119 120 55 3 10 114 68 53 34 6 11 64 39 13 26 2 6 31 39 21 32

2.27 2.49 3.41 3.48 3.26 3.24 1.94 2.48 4.05 3.75 3.72 3.55 2.59 2.05 2.61 3.80 3.16 3.72 1.15 1.65 3.73 3.58 3.43 5.68

25 4 4 2 3 1 28 20 14 3 1 11 26 8 4 19 38 51 14 4 4 3 11 10

146 40 28 21 3 1 90 104 61 31 18 55 111 65 19 8 43 11 288 383 66 5 42 3

55 46 27 10 9 ND 21 29 9 5 11 21 39 27 36 52 14 2 43 57 21 11 11 5

2660 1330 730 173 255 ND 2420 1180 104 7 146 37 2610 1680 2470 684 1300 564 2700 1950 1530 817 1080 554

ND, not detected (o1 mg N l1).

ARTICLE IN PRESS 779

41 (20 07) 77 5 – 784

WAT E R R E S E A R C H

Table 2 – Dissolved iron and its speciation in Lake Kasumigaura, the Sakura River, and the Koise River, including dissolved iron concentration (D-Fe), ligand concentration (CL), conditional stability constant (K0 FeL), percentage of iron occurring as organic species (% Organic), and inorganic iron (Fe0 ; pFe0 ¼ log[Fe0 ]) log K0 FeL

[CL] (nM)

D-Fe (nM) Avg

SD

Avg

SD

Avg

SD

% Organic

pFe0

April 2002

Koise Sakura St. 1 St. 2 St. 3 St. 4

2910 1840 74 47 75 35

205 282 2 4 7 5

2650 2050 95 97 127 72

258 326 21 6 34 19

25.7 25.9 26.2 25.6 25.1 25.7

0.1 0.4 0.2 0.3 0.3 0.1

90.9 499.9 499.9 499.9 499.9 499.9

6.6 12.6 13.4 13.3 12.7 13.4

July 2002

Koise Sakura St. 1 St. 2 St. 3 St. 4

47 911 127 72 61 44

2 47 9 5 9 5

68 1310 151 109 96 68

7 172 18 20 18 21

25.9 25.7 25.8 25.5 26.1 25.8

0.1 0.3 0.4 0.6 0.6 0.8

499.9 499.9 499.9 499.9 499.9 499.9

13.3 13.0 12.7 12.9 13.6 13.3

October 2002

Koise Sakura St. 1 St. 2 St. 3 St. 4

1320 1950 254 65 116 69

230 262 7 9 18 6

764 1720 252 114 105 96

59 410 26 24 35 12

25.4 26.2 25.5 25.7 25.4 26.2

0.4 0.6 0.3 0.4 0.3 0.5

57.7 88.4 99.3 499.9 90.8 499.9

6.3 6.6 8.8 13.3 8.0 13.5

January 2003

Koise Sakura St. 1 St. 2 St. 3 St. 4

743 686 133 47 67 48

80 20 11 6 7 4

605 719 117 74 89 80

61 138 11 5 10 20

26.1 25.7 25.5 25.9 25.7 25.9

0.4 0.3 0.3 0.3 0.3 0.4

81.5 499.9 88.2 499.9 499.9 499.9

6.9 12.0 7.8 13.3 13.0 13.4

Average (Avg) and standard deviation (SD) are shown (n ¼ 3).

4.

Discussion

4.1.

Validity of the CSV technique

Town and van Leeuwen (2005) raised questions about the validity of the CLE–CSV techniques on theoretical ground. They suggested that the large stability constants determined by overnight equilibration are likely to result from nonequilibrium condition and therefore be overestimated. Responding to their suggestion, van den Berg (2006) investigated the reaction kinetics to establish how long was required for the reaction to attain equilibrium. He found that the CSV peak height for iron using 2,3-dihydroxynaphthalene (DHN) as competing ligand increased gradually, reaching a plateau after 150–200 min. Hence, in terms of this reaction kinetics, overnight equilibration in the CSV technique has no kinetic problem. Furthermore, the complex stability constant, log K0 Fe0 L ¼ 11.54, determined using DHN, was identical to that found previously (11.6 after correction of log K0 Fe0 L ¼ 21.6 to the Fe0 -based constants using log a0 Fe ¼ 10) by CSV in competition against NN. The similar values of the conditional stability constant obtained using DHN and NN suggest that equilibrium among iron, natural ligand (L), and NN should be established by overnight equilibration. Therefore, there is an apparent disagreement between the much slower kinetics

suggested by Town and van Leeuwen and the reported experimental data. This discrepancy suggests that more complex reaction mechanisms exist in real natural water samples. In their theory, the reaction of metal complexation proceed by Eigen mechanism (dissociative mechanism), and the reaction rate depends on the rate constant for water substitution, kw. However, if the reaction of complexation proceeds via associative mechanism, an outer-sphere complex is not produced, and therefore their theory cannot be applied when estimating the reaction rate. The water substitution and ligand exchange rates for [Fe(H2O)6]3+ were reported to be consistent with the associative mechanism (Grant and Jordan, 1981; Swaddle and Merbach, 1981). When the associative mechanism works, kw does not limit the reaction rate of complexation, and therefore the theory by Town and van Leeuwen is clearly inappropriate in describing the reaction. Van den Berg (2006) also suggested that it is more realistic to consider the iron complexation as a reaction between Fe0 (sum of free ion and hydrolyzed ferric species, predo3+ 0 with minantly Fe(OH)03 and Fe(OH) 4 ) and L rather than as Fe 0 L . The water substitution reaction for hydrolyzed iron, [Fe(H2O)5OH]2+, has been reported to proceed by dissociative mechanism (Grant and Jordan, 1981; Swaddle and Merbach, 1981). However, even in this case, the theory of Town and van

ARTICLE IN PRESS 780

41 (2007) 775– 784

WA T E R R E S E A R C H

DO (mg l-1) 6

8

10 12

6

8

10 12

6

8

10 12

6

8

10 12

0

Depth (m)

2

4 St. 1 St. 2 St. 3 St. 4

6 April 2002

July 2002

October 2002

January 2003

Fig. 2 – Vertical distribution of dissolved oxygen (DO) in Lake Kasumigaura. The bottom depths at St. 1, 2, 3, and 4 are 2.9, 4.3, 3.4, and 6.3 m, respectively.

5 A-1

40

[labile Fe] / [FeL]

Peak current (nA)

50

30 20 10 0

A-2

4 3 2 1 0

0

10 20 Fe concentration (nM)

30

0

10 20 Labile Fe (nM)

30

50

150

B-1

200

[labile Fe] / [FeL]

Peak current (nA)

5

100

0

B-2

4 3 2 1 0

0

50

100

150

Fe concentration (nM)

0

100

Labile Fe (nM)

Fig. 3 – Typical titration curves and linear plots: (A1) titration curves in St. 4, April 2002; (A2) linear plots in St. 4, April 2002; (B1) titration curves in Koise River, April 2002; (B2) linear plots in Koise River, April 2002. For measurement, samples were diluted 10 times for St. 4, April 2002, and 50 times for Koise River, April 2002.

Leeuwen should not simply be applied when considering the iron complexation as a reaction between Fe0 and L0 . The water substitution rate for [Fe(H2O)6]3+ greatly increase through complexation with organic ligands. For example, the water substitution rate for [FeEDTA(H2O)] is 7  107 s1, which is five orders of magnitude higher than that for [Fe(H2O)6]3+ (Schneppensieper et al., 2001). Furthermore, the ligand

substitution rate for aqua-hydroxo complex, [Fe(dapsox) (H2O)OH] is 500 times higher than that for diaqua complex, [Fe(dapsox)(H2O)2]+ (Ivanovic-Burmazovic et al., 2006). The theory by Town and van Leeuwen includes nothing about the above-mentioned mechanism, which increase the reaction rate. When the water substitution rate for iron is 109 s1, ligand exchange reactions in CSV analysis will be in

ARTICLE IN PRESS

4.2.

Dissolved iron

The horizontal distribution of dissolved iron and the vertical distribution of DO at each sampling site suggest that dissolved iron in Lake Kasumigaura is mostly of river origin. Dissolved iron concentrations decreased suddenly from the rivers to the entrance of the lake, and the proportion of the iron occurring as organic species increased with water flow (Fig. 4). These results suggest that most of the dissolved iron in river water is of unstable soluble forms and thus only stable organic complexes of iron would remain in the water and reach the center of the lake. Inorganic iron(III) is known to quickly form hydrolyzed polymer species, so the polymerized iron would precipitate and be scavenged when it reached the lake. The temporal variation of dissolved iron was very large in the river waters, whereas it became small as the water flowed from the rivers to the center of the lake (see error bars in Fig. 4). This also suggests that dissolved iron is present in unstable forms in the river waters and relatively stable forms in the lake water. Not only inorganic iron but also organic complexes of iron are likely to be scavenged with the water flow. Organic ligand concentrations also decreased suddenly from the rivers to the lake (Table 2). Dissolved iron (defined as o0.2 mm in this study) includes filterable colloidal particles. Tanizaki et al.

100

2500

1500

90

1000 85

500 0

% Organic Fe

95

2000

80 Koise

St. 1

St. 2

St. 4 100

2500

95

2000 1500

90

1000 85

500 0

% Organic Fe

Fe concentration (nM)

equilibrium within several hours, even though the K0 FeL of natural organic ligands is 1025 M1. No iron(II) is thought to exist in equilibrated oxic water. Although iron(II) seems not to exist in the appearance, cycle of reduction-oxidation is actually repeated. When iron(III) is reduced to iron(II), stability constants of iron complexation with natural organic ligands will greatly decrease. The value of kw for Fe(OH)2+ and Fe(OH)+2 do not differ markedly from the kw value for Fe2+ (Town and van Leeuwen, 2005) and this suggest that the dissociation rate greatly increase by iron reduction because of much lower stability constant of iron(II) complexes with ligands. Chemical reactions occurring in natural water samples are not only simple ligand exchange between added ligands and natural ligands. There are various organic ligands, inorganic ligands, and several metals in natural water samples. Therefore various ligand exchange reactions, metal exchange reactions, and double exchange reactions will occur in the process of equilibration. The several exchange reactions would increase the overall reaction rate. The reaction of metal complexation is catalyzed by low concentrations (108–107 M) of other metal ions such as Hg, Cd, Zn, and Pb (Tabata and Tanaka, 1980). The reaction is also catalyzed by amino acids such as tryptophan, phenylalanine, and tyrosine (Tabata and Tanaka, 1988). The catalytic effect of amino acids is thought to be caused by hydrophobic interaction of metalamino acids complexes. Consequently, the rate of ligand exchange in natural water samples would be greater than that thought by Town and van Leeuwen. Their theory that cannot explain the experimental results is rather inappropriate. Therefore, it can be reasonable to state that our CSV technique is unlikely to face the kinetic problem during the course of experiment.

781

41 (20 07) 77 5 – 784

Fe concentration (nM)

WAT E R R E S E A R C H

80 Sakura

St.3

St. 4

Fig. 4 – Average concentration of the dissolved iron (bar) and fraction of iron originally present as organic species (J, % organic Fe) at each sampling site associated with the water flows of two directions (Koise river-St. 1-St. 2-St. 4; and Sakura river-St. 3-St. 4). Error bars show the standard deviations of dissolved iron concentration (n ¼ 4).

(1992) reported that colloidal iron represented the majority of the dissolved iron pool in river waters. Therefore, colloidal organic particles containing iron, such as particles originating in soil, could be precipitated with the water flow. Wen et al. (1999) reported that the fractions of high molecular weight colloidal iron (o0.45 mm and 410 kDa) decreased as the water flowed in Galveston Bay (river to estuary). Moreover, Wu et al. (2001) showed that iron and its organic ligands were present in both ‘‘truly dissolved’’ (o0.02 mm) and colloidal (0.02–0.4 mm) size ranges in seawater. These studies suggest the possibility that even chemically stable organic complexes of iron could precipitate and be scavenged if the iron forms colloidal particles.

4.3.

Iron-complexing organic ligands

The log values of the conditional stability constants for the organic ligands (K0 FeL) in Lake Kasumigaura and the two rivers were 25.1–26.2, and were greater than any reported K0 FeL values for seawater: 18.8–19.7 in the North Atlantic (Gledhill and van den Berg, 1994), 19.4–22.5 in the western Mediterranean (van den Berg, 1995), 23.1 in the North Pacific (Rue and Bruland, 1995), 22.7 in the equatorial Pacific (Rue and Bruland, 1997), 20.7–21.7 in the northern North Sea (Gledhill et al., 1998), 21.0–23.0 in the Southern Ocean (Boye et al., 2001), 22.9–24.5 in the Peconic Estuary of the Atlantic Ocean

ARTICLE IN PRESS 782

WA T E R R E S E A R C H

(Gobler et al., 2002), and 20.5–21.4 in the northeast Atlantic Ocean (Boye et al., 2003). The substantially strong complexing ligands in lake water were detected because we applied a high detection window (1015.7–1017.7, calculated with a NN concentration of 50 mM) on CSV measurements. Weaker ligands might be detected if we applied a lower detection window using lower NN concentrations. Although seawater has been successfully analyzed by using NN concentrations of 2–13.6 mM, our attempt to analyze lake water using such NN concentrations was unsuccessful. Rue and Bruland (1995) suggested that two ligands (strong and weak ligand) are resolved from a single set of titration data by using the Scatchard transformation method for the determination of CL and K0 FeL. However, only one linear region was evident in the Scatchard plots of our titration data, and the values of CL and K0 FeL obtained from the Langmuir and Scatchard plots were not significantly different (data not shown). Suppose that undetectable weaker ligands was present, then the pFe0 values would be underestimated, especially for the samples having iron concentrations greater than ligand concentrations. We know very little about the characteristics of ironcomplexing organic ligands. The chemical composition of the organic ligands probably varies significantly depending on where the ligand comes from: a terrestrial source, such as humic substances; or a microbiological source, such as siderophores. In the open ocean, several studies suggest a microbiological source for organic ligands (Rue and Bruland, 1997; Boye et al., 2001; Nakabayashi et al., 2002). The horizontal concentration gradient of organic ligands from the two rivers to Lake Kasumigaura suggests that they are mainly from a terrestrial source. We also suggest that the organic ligands in Lake Kasumigaura are fulvic acid, on the basis of a comparison of the conditional stability constants between our data (K0 FeL ¼ 1025.1–1026.2 M1) and those previously reported: Rue and Bruland (1997) and Boye et al. (2001) suggested that strong iron ligands (K0 FeL ¼ 1022–1023 M1) in seawater might be siderophores; Imai et al. (1999) calculated a conditional stability constant of 41025 M1 for Fe3+ complexed with fulvic acids isolated from the water of Lake Kasumigaura. The production of organic ligands by phytoplankton is well known (e.g., Boye and van den Berg, 2000). However, the absence of a substantial relationship between the ligand concentration and chlorophyll a indicates a small contribution by biological sources to the organic ligands in Lake Kasumigaura.

4.4.

Iron speciation and bioavailability

We found that the inorganic iron level in Lake Kasumigaura (pFe0 valure ¼ 7.8–13.6) was very low. Especially, pFe0 values at the center of Lake Kasumigaura (13.3–13.6) were similar to those in the open ocean despite greater concentrations of dissolved iron. The reported pFe0 values determined by CSV in the open ocean are as follows: 10.6–12.0 in the North Atlantic (Gledhill and van den Berg, 1994); 9.3–13.9 in the northern North Sea (Gledhill et al., 1998); 11.5–13.2 in the Southern Ocean (Boye et al., 2001); 13.1–14.0 in the North Pacific (Rue and Bruland, 1995); and 14.0 in the equatorial Pacific (Rue and Bruland, 1997). Both iron and ligand concentrations in the

41 (2007) 775– 784

lake waters are higher than those in oceanic waters. As a result, inorganic iron levels did not differ significantly among lake water and oceanic water regardless of dissolved iron concentration. We now consider the effect of iron speciation on the phytoplankton community in Lake Kasumigaura. In eutrophic Lake Kasumigaura, cyanobacteria, such as Microcystis spp. or Planktothrix spp., were the dominant species of phytoplankton for quite a long time. However, they were hardly observed in 2001 and 2002, and diatoms such as Cyclotella spp. have become dominant (CGER, 2004). Total nitrogen (TN, 540–1383 mg l1) and total phosphorus (TP, 77–197 mg l1) in the center of Lake Kasumigaura at 2001 to 2002 (CGER, 2004) are suitable levels for cyanobacteria dominance. Therefore, this change in dominant species may have something to do with iron requirements and its availability for each algal species. Iron availability to phytoplankton depends on the inorganic iron (Fe0 ) concentration (Hudson and Morel, 1993). Therefore, iron limitations on phytoplankton growth may occur even if the dissolved iron concentration is quantitatively adequate. According to Imai et al. (1999), M. aeruginosa can grow at a pFe0 of more than 12.9 but not at a pFe0 of 13.9 in the defined culture media. Therefore, the pFe0 values at the center of the lake (13.3–13.5) were close to the boundary beyond which M. aeruginosa is not able to grow. Which species wins when iron competition occurs between cyanobacteria and diatoms? According to Brand (1991), cyanobacteria have a relatively higher cellular iron requirement than do other phytoplankton. However, the relationship between iron and phytoplankton growth is highly complicated and is not understood completely. Recently, the availability of organically complexed iron at the low concentration of inorganic iron has been discussed, but the availability remains still unknown. Hutchins et al. (1999) suggested that cyanobacteria and diatoms possess different uptake strategies for organically complexed iron, and therefore, iron competition between cyanobacteria and diatoms may depend on the chemical nature of the iron complexes. Further studies are definitely required to obtain the solid evidence of a close relationship between iron availability and algal species succession in Lake Kasumigaura, namely, (i) direct field observations confirming an iron limitation on the growth of cyanobacteria; (ii) investigation of the details of phytoplankton iron requirements and the availability of several iron species to each species of phytoplankton; and (iii) characterization of iron-complexing organic ligands.

5.

Conclusions

This is the first study to demonstrate horizontal distribution and temporal variation of dissolved iron, including its chemical speciation, in a lake. From our results, we suggest that dissolved iron in Lake Kasumigaura is mainly of river origin and that most of the dissolved iron in river water forms unstable soluble species. Thus, these unstable iron species may be precipitated at the mouth of river. Our data also suggest that organic ligands in Lake Kasumigaura may be

ARTICLE IN PRESS WAT E R R E S E A R C H

terrestrial fulvic acids. The pFe0 values (inorganic iron level) in Lake Kasumigaura were similar to those in the open ocean, even though the concentration of dissolved iron in Lake Kasumigaura was much higher. It is therefore possible that iron speciation and iron availability are important factors in determining phytoplankton species succession and its diversity in Lake Kasumigaura.

Acknowledgments This work was supported by JSPS Research Fellowships for Young Scientists to T.N., by Grant-in-Aid for JSPS Research Fellows (No. 17  7257, 2005), and by Grant-in-Aid for Scientific Research (No. 17310013) from the Japan Society for the Promotion of Science. Sampling was supported by the GEMS/Water Trend Monitoring Project at Lake Kasumigaura. We thank the members of the project for their cooperation. R E F E R E N C E S

Aizaki, M., Aoyama, K., 1995. Changes of phytoplankton composition on the cyanobacterial blooms forming experiment using large outdoor continuous flow-through ponds. In: Fukushima, T., Aizaki, M.,(Eds.), Aoko (Water-blooms of Blue-Green Algae); Measurement, Occurrence, and Factors on its Growth. National Institute for Environmental Studies F-72-’95/NIES. National Institute for Environmental Studies, Ibaraki, Japan, pp. 105–114 (In Japanese). Boye, M., van den Berg, C.M.G., 2000. Iron availability and the release of iron complexing ligands by Emiliania Huxleyi. Mar. Chem. 70, 277–287. Boye, M., van den Berg, C.M.G., de Jong, J.T.M., Leach, H., Croot, P., de Baar, H.J.W., 2001. Organic complexation of iron in the Southern Ocean. Deep-Sea Res. I 48, 1477–1497. Boye, M., Aldrich, A.P., van den Berg, C.M.G., de Jong, J.T.M., Veldhuisb, M., de Baar, H.J.W., 2003. Horizontal gradient of the chemical speciation of iron in surface waters of the northeast Atlantic Ocean. Mar. Chem. 80, 129–143. Brand, L.E., 1991. Minimum iron requirements of marine phytoplankton and the implications for the biogeochemical control of new production. Limnol. Oceanogr. 36 (8), 1756–1771. CGER 2004. Lake Kasumigaura Database: /http://www-cger2.nies.go.jp/moni-e/gems/database/kasumi/index.htmlS. National Institute for Environmental Studies, Center for Global Environmental Research, Ibaraki, Japan. Clasen, J., Bernhardt, H., 1974. The use of algal assays for determining the effect of iron and phosphorus compounds on the growth of various algal species. Water Res. 8, 31–44. Evans, J.C., Prepas, E.E., 1997. Relative importance of iron and molybdenum in restricting phytoplankton biomass in high phosphorus saline lakes. Limnol. Oceanogr. 42, 461–472. Gledhill, M., van den Berg, C.M.G., 1994. Determination of complexation of iron(III) with natural organic complexing ligand in seawater using cathodic stripping voltammetry. Mar. Chem. 47, 41–54. Gledhill, M., van den Berg, C.M.G., Nolting, R.F., Timmermans, K.R., 1998. Variability in the speciation of iron in the northern North Sea. Mar. Chem. 59, 283–300. Gobler, C.J., Donat, J.R., Consolvo III, J.A., San˜udo-Wilhelmy, S.A., 2002. Physicochemical speciation of iron during coastal algal blooms. Mar. Chem. 77, 71–89.

41 (20 07) 77 5 – 784

783

Grant, M., Jordan, R.B., 1981. Kinetics of solvent water exchange on iron(III). Inorg. Chem. 20, 55–60. Hudson, R.J.M., Morel, F.M.M., 1993. Trace metal transport by marine microorganisms: implications of metal coordination kinetics. Deep-Sea Res. I 40, 129–150. Hutchins, D.A., Witter, A.E., Butler, A., Luther III, G.W., 1999. Competition among marine phytoplankton for different chelated iron species. Nature 400, 858–861. Hyenstrand, P., Rydin, E., Gunnerhed, M., 2000. Response of pelagic cyanobacteria to iron additions—enclosure experiment from Lake Erken. J. Plankton Res. 22, 1113–1126. Imai, A., Fukushima, T., Matsushige, K., 1999. Effect of iron limitation and aquatic humic substances on the growth of Microcystis aeruginosa. Can. J. Fish. Aquat. Sci. 56 (10), 1929–1937. Inaba, K., Sekine, T., Tomioka, N., Yagi, O., 1997. Seasonal and longitudinal changes in copper and iron in surface water of shallow eutrophic lake Kasumigaura, Japan. Water Res. 31, 280–286. Ivanovic-Burmazovic, I., Hamza, M.S.A., van Eldik, R., 2006. Solvent tuning of the substitution behavior of a sevencoordinate iron(III) complex. Inorg. Chem. 45, 1575–1584. Lin, C.K., Schelske, C.L., 1981. Seasonal variation of potential nutrient limitation to chlorophyll production in southern Lake Huron. Can. J. Fish. Aquat. Sci. 38, 1–9. Martin, J.H., Coale, K.H., Johnson, K.S., Fitzwater, S.E., Gordon, R.M., Tanner, S.J., Hunter, C.N., Elrod, V.A., Nowicki, J.L., Coley, T.L., Barber, R.T., Lindley, S., Watson, A.J., Vanscoy, K., Law, C.S., Liddicoat, M.I., Ling, R., Stanton, T., Stockel, J., Collins, C., Anderson, A., Bidigare, R., Ondrusek, M., Latasa, M., Millero, F.J., Lee, K., Yao, W., Zhang, J.Z., Friedrich, G., Sakamoto, C., Chavez, F., Buck, K., Kolber, Z., Greene, R., Falkowski, P., Chisholm, S.W., Hoge, F., Swift, R., Yungel, J., Turner, S., Nightingale, P., Hatton, A., Liss, P., Tindale, N.W., 1994. Testing the iron hypothesis in ecosystems of the equatorial Pacific Ocean. Nature 371, 123–129. Nagai, T., Imai, A., Matsushige, K., Yokoi, K., Fukushima, T., 2004. Voltammetric determination of dissolved iron and its speciation in freshwater. Limnology 5 (2), 87–94. Nakabayashi, S., Kuma, K., Sasaoka, K., Saitoh, S., Mochizuki, M., Shiga, N., Kusakabe, M., 2002. Variation in iron(III) solubility and iron concentration in the northwestern North Pacific Ocean. Limnol. Oceanogr. 47 (3), 885–892. Rue, E.L., Bruland, K.W., 1995. Complexation of iron(III) by natural organic ligands in the central north Pacific as determined by a new competitive ligand equilibration/adsorptive cathodic stripping voltammetric method. Mar. Chem. 50, 117–138. Rue, E.L., Bruland, K.W., 1997. The role of organic complexation on ambient iron chemistry in the equatorial Pacific Ocean and response of a mesoscale iron addition experiment. Limnol. Oceanogr. 42 (5), 901–910. Schneppensieper, T., Seibig, S., Zahl, A., Tregloan, P., van Eldik, R., 2001. Influence of chelate effects on the water-exchange mechanism of polyaminecarboxylate complexes of iron(III). Inorg. Chem. 40, 3670–3676. Swaddle, T.W., Merbach, A.E., 1981. High-pressure oxygen-17 Fourier transform nuclear magnetic resonance spectroscopy. Mechanism of water exchange on iron(III) in acidic aqueous solution. Inorg. Chem. 20, 4212–4216. Tabata, M., Tanaka, M., 1980. A kinetic method for the determination of a nanogram amount of mercury(II) by its catalytic effect on the complex formation reaction of manganese(II) with a, b, w, d-tetraphenylporphinesulfonate. Anal. Lett. 13, 427–438. Tabata, M., Tanaka, M., 1988. Importance of hydrophobic interaction in metalloporphyrin formation. Inorg. Chem. 27, 203–205.

ARTICLE IN PRESS 784

WA T E R R E S E A R C H

Tanizaki, Y., Shimokawa, T., Nakamura, M., 1992. Physicochemical speciation of trace elements in river waters by size fractionation. Environ. Sci. Technol. 26, 1433–1444. Town, R.M., van Leeuwen, H.P., 2005. Measuring marine iron(III) complexes by CLE-AdSV. Environ. Chem. 2, 90–93. Twiss, M.R., Auclair, J.C., Charlton, M.N., 2000. An investigation into iron-stimulated phytoplankton productivity in epipelagic Lake Erie during thermal stratification using trace metal clean techniques. Can. J. Fish. Aquat. Sci. 57, 86–95. van den Berg, C.M.G., 1995. Evidence for organic complexation of iron in seawater. Mar. Chem. 50, 139–157.

41 (2007) 775– 784

van den Berg, C.M.G., 2006. Chemical speciation of iron in seawater by cathodic stripping voltammetry with dihydroxynaphthalene. Anal. Chem. 78, 156–163. Wen, L., Santschi, P., Gill, G., Paternostro, C., 1999. Estuarine trace metal distributions in Galveston Bay: importance of colloidal forms in the speciation of the dissolved phase. Mar. Chem. 63, 185–212. Wu, J., Boyle, E., Sunda, W., Wen, L., 2001. Soluble and colloidal iron in the oligotrophic North Atlantic and North Pacific. Science 293, 847–849. Yagi, O., Hagiwara, T., Takamura, Y., Sudo, R., 1987. Limiting nutrients of algal growth in Lake Kasumigaura. Jpn J. Water Pollut. Res. 10, 115–122 (In Japanese).