Winkler's method overestimates dissolved oxygen in natural waters: Hydrogen peroxide interference and its implications

Winkler's method overestimates dissolved oxygen in natural waters: Hydrogen peroxide interference and its implications

Marine Chemistry 122 (2010) 83–90 Contents lists available at ScienceDirect Marine Chemistry j o u r n a l h o m e p a g e : w w w. e l s ev i e r. ...

340KB Sizes 0 Downloads 18 Views

Marine Chemistry 122 (2010) 83–90

Contents lists available at ScienceDirect

Marine Chemistry j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / m a r c h e m

Winkler's method overestimates dissolved oxygen in natural waters: Hydrogen peroxide interference and its implications George T.F. Wong ⁎,1, Yao-Chu Wu, Kuo-Yuan Li Research Center for Environmental Changes, Academia Sinica, 128 Sec. 2 Academia Rd., Nankang, Taipei, Taiwan, ROC

a r t i c l e

i n f o

Article history: Received 16 December 2009 Received in revised form 27 May 2010 Accepted 22 July 2010 Available online 4 August 2010 Keywords: Oxygen Hydrogen peroxide Winkler method Community respiration Primary production

a b s t r a c t In the Winkler's reaction scheme used for the determination of dissolved oxygen in natural waters, hydrogen peroxide reacts with iodide to form tri-iodide ion and leads to an over-estimation in the concentration of oxygen. In pure solutions, the reaction is slow but it is accelerated in the presence of Mn+ 2 and reaches completion when Mn(OH)2 is first formed under basic condition before the acidification of the solution, as in this reaction scheme. Each mole of hydrogen peroxide appears as 0.5 mol of apparent oxygen. This interference further exacerbates the recently realized over-estimation by the presence of naturally occurring iodate. Thus, the use of solubility equations of oxygen that are based on manometrically determined equilibrium concentrations, which are free from these chemical interferences, is highly recommended. The presence of hydrogen peroxide may have led to an over-estimation in the global efflux of oxygen from the ocean to the atmosphere by 1% and an over-estimation in the biological contribution to the % saturation anomaly of oxygen by 0.03%. At the present level of sophistication at which oxygen data are being used in data manipulation, these over-estimations should be taken into account and corrected for. Assuming that there is negligible change in their concentrations after sample collection, the distributions of hydrogen peroxide and iodate in the oceans are sufficiently well known that schemes for such corrections may be devised. However, they will degrade significantly the presently claimed precision in the determination of oxygen. Since the concentration of hydrogen peroxide is known to change upon storage, these variations will affect the estimation of community respiration rates and primary production by measuring the changes in the concentration of oxygen in incubated samples. This error may be especially prominent in the estimation of the former. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Dissolved oxygen data from the ocean have been used at an increasing level of sophistication in recent years for tackling some of the critical issues of the day. For example, the deviations of the concentrations of oxygen from their saturation values have been used for estimating net primary production in the ocean (Craig and Hayward, 1987; Platt et al., 1989; Emerson et al., 2001). Inventories of dissolved oxygen in the major ocean basins have been used for constraining the global carbon cycle (Keeling and Garcia, 2002; Plattner et al., 2002). The changes in the concentration of dissolved oxygen in water samples incubated under different defined experimental conditions have been used for the determination of the rates of autotrophic activities and community respiration (Carignan et al., 1998; Chen et al., 2003) and

⁎ Corresponding author. Tel.: +886 2 26539885x856; fax: +886 2 27833584. E-mail address: [email protected] (G.T.F. Wong). 1 Also at: Institute of Hydrology and Ocean Sciences, National Central University, Jungli 320, Taiwan, ROC. 0304-4203/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.marchem.2010.07.006

these measurements have profound implications on our understanding of the metabolic balance in the oceans (del Giorgio et al., 1997; Duarte and Agusti, 1998; Gattuso et al., 1999; Williams and Bowers, 1999; Duarte et al., 1999; del Giorgio and Duarte, 2002; Karl et al., 2003). The validity of all these studies depends critically on the claimed precision and accuracy, at around ±0.1 to ±0.3% (Pai et al., 1993; Emerson et al., 1995; Labasque et al., 2004), in the determination of dissolved oxygen in natural waters. Given an average concentration of dissolved oxygen in surface seawater of 250 μmol kg−1, these precision and accuracy would be equivalent to about ±0.3 to ±0.8 μmol kg−1. In fact, a precision as high as ±0.016%, which is equivalent to about ±0.04 μmol kg−1, has been claimed by some investigators (Williams et al., 2004). At these high levels of claimed accuracy and precision, factors that may affect the results at the level of 10−2 μmol kg−1 may become significant. As a result, many heretofore neglected or overlooked interferences in the determination of dissolved oxygen in natural waters have now become relevant and require a re-examination. The concentration of dissolved oxygen in natural waters, including seawater, has been determined almost exclusively by using methods that are based on the Winkler's reaction scheme. Even when other methods are used, such as electrochemical methods, they are calibrated against

84

G.T.F. Wong et al. / Marine Chemistry 122 (2010) 83–90

these methods (Emerson et al., 2002; Kuss et al., 2006). The chemical reactions involved in the Winkler's scheme are: þ2



Mn þ 2OH →MnðOHÞ2

ð1Þ

4MnðOHÞ2 þO2 þ2H2 O→4MnðOHÞ3

ð2Þ

þ



þ2

4MnðOHÞ3 þ12H þ 4I →2I2 þ 4Mn þ12H2 O





I2 þ I ↔I3

ð3Þ

ð4Þ

The molecular iodine and tri-iodide ion formed are then quantified either by a titration with thiosulfate (Carpenter, 1965a,b; Strickland and Parsons, 1972) or by spectrophotometry (Pai et al., 1993; Labasque et al., 2004; Reinthaler et al., 2006). In this reaction scheme, any oxidizing chemical species in natural water that may react with iodide under acidic conditions to form molecular iodine will interfere with the determination of dissolved oxygen. For example, the oxidation of the excess iodide by oxygen in air under acidic condition has been long recognized and extensive efforts have been made to minimize its effect (Carpenter, 1965a). The interference of nitrite has also been well recognized and dealt with by the freshwater and wastewater community (Clesceri et al., 1998) but it has been basically dismissed by the marine investigators (Strickland and Parsons, 1972; Knap et al., 1997; Grasshoff et al., 1999), except in low oxygen water (Broenkow and Cline, 1966), probably as a result of its low concentrations in most of the oceans. Wong and Li (2009) recently pointed out that iodate, which undergoes the well known Dushman reaction in the presence of excess iodide and acid (Wong and Brewer, 1974), is a long overlooked naturally occurring interfering species in seawater. Its presence can lead to an over-estimation in the concentration of dissolved oxygen in seawater of up to 0.7 μmol kg−1, and such an over-estimation may have significant oceanographic implications. However, since iodate is virtually absent in fresh water (Wong, 1991; Wong and Zhang, 2003a; Wong and Cheng, 2001, 2008), its interfering effect should be confined only to marine waters. Hydrogen peroxide is a well known reactive trace species that is present ubiquitously in both surface freshwater and marine waters and its concentration can reach several μM (Cooper et al., 1988). In the presence of a catalyst, it is known to react readily with iodide to form iodine under acidic conditions. Here, its possible interference in the determination of dissolved oxygen in natural waters by using the Winkler's reaction scheme is examined and its implications on several types of studies are assessed. 2. Methods The interference of hydrogen peroxide was determined by measuring the concentration of apparent oxygen in de-ionized water, artificial seawater and surface seawater with 0 to 500 or 0 to 50 μM of added hydrogen peroxide. (The concentration of apparent oxygen is the sum of the concentration of dissolved oxygen and the oxygen-equivalent of hydrogen peroxide.) The samples were allowed to equilibrate with the laboratory air for at least 2 h in polyethylene bottles before the addition of hydrogen peroxide. Sub-samples were drawn at least in duplicate from each bottle for the determination of the concentrations of apparent oxygen spectrophotometrically by using the Winkler's reaction scheme according to the method of Pai et al. (1993, 1998). Sodium azide was used to remove any interference caused by the presence of nitrite. The average precision of triplicate analyses of 12 samples with concentrations ranging from 212 to 303 μmol kg−1 was ±0.3 μmol kg−1 or about ±0.1% (Wong and Li, 2009). These values are about the same as those claimed in the literature (Pai et al., 1993; Labasque et al., 2004). When multiple sub-

samples of artificial seawater and pure water were exposed to the laboratory air simultaneously for the same amount of time, the resulting bottle to bottle variations in the concentration of apparent oxygen, ±0.2%, were similar to the analytical uncertainty (Wong and Li, 2009), suggesting that any variability in the exchange of oxygen between the solution and the atmosphere among the bottles was negligible. Thus, changes in the concentration of apparent oxygen upon the addition of hydrogen peroxide were the result of the presence of hydrogen peroxide. While the concentrations of added hydrogen peroxide used in these experiments were significantly higher than those found in natural waters in most cases, the lowest concentration added, 10 μM, approached the higher end of the observed concentrations (Cooper and Zika, 1983). The effect of hydrogen peroxide on the determination of community respiration rate and primary production were studied by measuring changes in the concentration of apparent oxygen by the method of Pai et al. (1993, 1998) upon dark and light incubation by following the scheme of Carignan et al. (1998) while monitoring the accompanying changes in the concentration of hydrogen peroxide by the method of Zhang and Wong (1999). In the scheme, multiple 60 ml BOD bottles were filled with samples of freshly collected coastal seawater. Two to three were used for the immediate duplicate or triplicate determinations of apparent oxygen. The concentration of hydrogen peroxide in the sample of water was also determined immediately in duplicate or triplicate. In the dark incubation, the remaining bottles were kept in the dark at laboratory temperature (22 ± 2 °C). In the light incubation, the remaining bottles were incubated in an incubation chamber under light with an irradiance of 130 to 140 μE m−2 s−1 at a constant temperature of 22 ± 1 °C. After a specified amount of time, three or four bottles were removed. Two or three of them were used for the duplicate or triplicate determinations of apparent oxygen. The concentration of hydrogen peroxide was determined in duplicate or triplicate in the remaining bottle. The maximum incubation time varied between 4 and 24 h. The average relative standard deviations in the duplicate or triplicate determinations of apparent oxygen and hydrogen peroxide were usually around ±0.1%, and ±1% respectively. 3. Results and discussion 3.1. The interference of hydrogen peroxide Hydrogen peroxide reacts with iodide under acidic conditions to form molecular iodine such that: þ



H2 O2 þ 2H þ 2I →I2 þ2H2 O

ð5Þ

This is a well studied reaction (Wong and Zhang, 2007). In pure solutions, the reaction is notoriously slow (Morgan, 1954). However, under acidic condition, and especially in the presence of a catalyst, such as molybdate, tungstate, copper or iron, the reaction can be greatly accelerated (Kolthoff and Stenger, 1942). In the presence of excess iodide, I2 is then converted to I− 3 as shown in reaction (4). The combination of reactions (4) and (5) in the presence of added molybdate has been used as a standard method for the titrimetric determination of hydrogen peroxide with thiosulfate (Bassett et al., 1978). However, it is unclear whether the presence of Mn (OH)2 or Mn+ 2 in the Winkler's reaction scheme may likewise promote reaction (5). If reaction (5) reaches completion, 1 mol of hydrogen peroxide leads to the formation of 1 mol of molecular iodine. Since 1 mol of oxygen results in the formation of 2 mol of molecular iodine in the Winkler's reaction scheme according to reactions (2) and (3), the presence of 1 mol of hydrogen peroxide is equivalent to the apparent presence of 0.5 mol of oxygen. In samples of de-ionized water, artificial seawater and natural surface seawater that had been equilibrated with the atmosphere in the laboratory, the concentration of the apparent oxygen increased linearly with increasing concentration of added hydrogen peroxide.

G.T.F. Wong et al. / Marine Chemistry 122 (2010) 83–90

The results are summarized in Table 1 and a typical relationship in one of the experiments is shown in Fig. 1. The correlation coefficients, r2, all exceeded 0.999. Thus, a linear relationship could represent the results well. The average value of the slopes of these relationships, 0.499 ± 0.005 mol mol−1, was indistinguishable from the theoretical value of 0.5 mol mol−1 within the statistical uncertainty. Thus, the presence of hydrogen peroxide can cause an over-estimation in the concentration of oxygen, and this interference can be explained by reaction (5) as it reaches completion. The presence of the other added chemical reagents in the Winkler's reaction scheme, such as Mn+ 2 and sodium azide, does not interfere with reaction (5). To further test whether the reaction between hydrogen peroxide and iodide may be promoted by the presence of Mn+ 2, Mn(OH)2, and/or natural occurring trace metals in the sample, similar experiments were conducted in de-ionized water and surface seawater by adding the alkaline iodide solution and the acid in the Winkler's reaction scheme (Pai et al., 1993) before the addition or without the addition of Mn+ 2 to the samples. The concentration of hydrogen peroxide was expressed as the concentration of apparent oxygen. In this analytical scheme, since the solution had been acidified before the addition of Mn+ 2, Mn(OH)2 was not formed to sequester the dissolved oxygen. Thus, oxygen was not included in the analyses. The results are shown in Fig. 2 and Table 2. In all cases, the concentration of apparent oxygen increased linearly with increasing concentration of added hydrogen peroxide. The correlation coefficients all exceeded 0.99. The intercepts were all close to zero, indicating that oxygen was not involved in the reactions. In the absence of added Mn+ 2, when the samples were analyzed immediately after the addition of the reagents, the slope of the relationship between the concentration of apparent oxygen and added hydrogen peroxide in deionized water, 0.0106±0.0004 mol mol−1, was equivalent to only 2% of the theoretical value of 0.5 mol mol−1, indicating that, as expected, little hydrogen peroxide underwent reaction (5) in the absence of a catalyst (Morgan, 1954). Upon addition of Mn+ 2, the slope increased to 0.0127 ± 0.0004 mol mol−1, or about 3% of the theoretical value. The slope in natural seawater was 0.0182±0.0004 mol mol−1, or 4% of the theoretical value, but it was about twice of that in de-ionized water. Similar trends could be demonstrated more clearly when the reaction was allowed to proceed for 15 min before the samples were catalyzed. In de-ionized water, the slope was 0.1468±0.0009 mol mol−1, or 29% of the theoretical value. In the presence of added Mn+ 2 and in natural surface seawater, the slope increased to 0.2851±0.0003 mol mol−1 or 57% of the theoretical value and 0.204±0.003 mol mol−1, or 41% of the theoretical value respectively. The higher slope in natural seawater suggests that the reaction between hydrogen peroxide and iodide was promoted by some constituents in seawater. A possible catalyst of the reaction in seawater is molybdate. At a concentration of 0.1 μM (Brewer, 1975), it is one of the more abundant trace elements in seawater, and it is a known catalyst of reaction (5) (Kolthoff and Stenger, 1942). The higher slope in the presence of Mn+ 2 indicates that Mn+ 2 can indeed promote the reaction. However, Mn+ 2 is apparently not a particularly effective catalyst since the resulting slopes were still significantly lower than the

85

Fig. 1. Relationship between apparent oxygen and added hydrogen peroxide in aged coastal seawater. There are 3 data points at each concentration level of added hydrogen peroxide. The best fit line is shown.

theoretical slope of 0.5 mol mol−1, indicating that the reaction had reached completion. Since reaction (5) goes to completion when (OH)2 is formed, a possible reaction scheme that may account for different behaviors in the presence and absence of Mn(OH)2 is oxidation of Mn(OH)2 by hydrogen peroxide such that:

not Mn the the

H2 O2 þ 2MnðOHÞ2 →2MnðOHÞ3

ð6Þ

Reaction (6) is an exothermic reaction with a free energy change of −36 kcal mol−1, and Mn(OH)2 is present to react with hydrogen peroxide in excess in the Winkler's reaction scheme. If reaction (6) goes to completion, the Mn(OH)3 formed may then undergo reactions (3) and (4) to yield I− 3. In the standard method for the Winkler's determination of dissolved oxygen in natural waters (Strickland and Parsons, 1972), the reagent blank and the strength of the titrant, sodium thiosulfate, are determined

Table 1 Linear relationships between the concentrations of apparent oxygen and added hydrogen peroxide. Medium

De-ionized water

Artificial seawater Surface seawater

Average

H2O2 added

Slope

Intercept −1

μM

mol mol

0–500 0–400 0–50 0–400 0–50 0–400 0–400 0–50

0.500(± 0.001) 0.495(± 0.001) 0.502(± 0.003) 0.507(± 0.001) 0.501(± 0.003) 0.492(± 0.001) 0.495(± 0.001) 0.497(± 0.004) 0.499 ± 0.005

N

r2

18 15 18 15 18 15 15 18

1.0000 1.0000 0.9994 1.0000 0.9994 0.9999 1.0000 0.9992

μM 285.4(± 0.2) 290.3(± 0.1) 229.5(± 0.1) 236.6(± 0.1) 238.2(± 0.1) 295.8(± 0.3) 290.3(± 0.1) 256.4(± 0.1)

Fig. 2. Relationships between apparent oxygen and added hydrogen peroxide when the sample was acidified before the addition of alkaline I− and/or Mn+ 2 in order to preclude the formation of Mn(OH)2. There are 2 or 3 data points at each concentration level of added hydrogen peroxide. The best fit lines are shown. Open symbols — right axis; samples analyzed immediately. Solid symbols — left axis; hydrogen peroxide allowed to react with I− for 15 min before the analyses. ◊, ♦ — in Milli-Q reagent grade water without the addition of Mn+ 2; ○, ● — in Milli-Q reagent grade water with the addition of Mn+ 2; Δ, ▲ — in aged coastal seawater without the addition of Mn+ 2.

86

G.T.F. Wong et al. / Marine Chemistry 122 (2010) 83–90

Table 2 Linear relationships between the concentrations of apparent oxygen and added hydrogen peroxide in acidified water. Medium

Waiting time#

H2O2 added

Slope

min

μM

mol mol−1

%*

Intercept

N

r2

μM

added after the addition of I− and acid No Mn De-ionized 0 0–500 0.0106(± 0.0004) water 15 0–500 0.1468(± 0.0009) Surface 0 0–500 0.0182(± 0.0004) seawater 15 0–500 0.204(± 0.003)

2 29 4 41

−0.4(±0.1) −0.4(±0.3) 0.03(± 0.1) −1.3(±0.9)

12 12 12 12

0.99 1.00 0.99 1.00

Mn+ 2 added after the addition of I− and acid De-ionized 0 0–500 0.0127(± 0.0004) water 15 0–500 0.2851(± 0.0003)

3 57

0.03(± 0.1) 0.4(±0.1)

18 18

0.99 1.00

+2

⁎% of theoretical slope of 0.5 mol mol−1. Amount of time waited after all the reagents had been added before the absorbance was read.

#

in solutions prepared from laboratory reagent grade water or seawater. As variable amounts of hydrogen peroxide may be present in both deionized water and seawater, their presence will lead to variable positive blanks and underestimations in the concentration of the titrant sodium thiosulfate solution. (About 10−1 μM, and occasionally even higher concentrations, of hydrogen peroxide can be found routinely in freshly prepared distilled de-ionized Milli-Q water in our laboratory.) These errors are likely to be small since the concentrations of hydrogen peroxide are low in comparison to the concentrations of the reagents, and, as discussed previously, the reaction between hydrogen peroxide and iodide does not go to completion in the absence of a suitable catalyst. Nevertheless, for good laboratory practice, it is still advisable to use water that is essentially free of hydrogen peroxide in the preparation of these solutions. Laboratory reagent grade water and natural water that are substantially free of hydrogen peroxide may be prepared by stirring the former over manganese dioxide and by storing the latter in the dark for about a week (Zhang and Wong, 1994). 3.2. Oceanographic implications 3.2.1. Over-estimation of the concentration of oxygen The occurrence of hydrogen peroxide in natural waters is fairly well known. Its primary source in surface waters is through photochemical production by using dissolved organic matter as the chromophore (Cooper and Zika, 1983; Cooper et al., 1994) and its primary sink is through biologically mediated decomposition (Cooper and Zepp, 1990; Moffett and Zafiriou, 1990; Petasne and Zika, 1997; Wong et al., 2003). As a reactive transient, the concentration of hydrogen peroxide in natural waters is highly dynamic. It varies diurnally, daily and seasonally (Zika et al., 1985; Cooper et al., 1988, 1989; Cooper and Lean, 1989; Fujiwara et al., 1993). Spatially, in the open oceans, the concentration is the highest, usually around 0.06 to 0.3 μM, at the surface and it decreases rapidly with depth to less than 0.01 μM below the surface mixed layer (Zika et al., 1985; Johnson et al., 1989; Moore et al., 1993; Miller and Kester, 1994; Hanson et al., 2001; Yuan and Shiller, 2001, 2004, 2005). The concentrations in surface coastal seawater, inshore marine waters and freshwater tend to extend from the high end of this range upwards. Values as high as 3 μM have been reported in some samples of river water (Cooper and Zika, 1983; Cooper et al., 1988, 1989; Cooper and Lean, 1989; Kieber and Helz, 1995). Estimations of the average concentrations of hydrogen peroxide in the mixed layer, the upper thermocline and the deep water below the euphotic zone in the open ocean, and in inshore water and freshwater and their corresponding concentrations of apparent oxygen are listed in Table 3. The values in the upper thermocline are the averages of the values in the mixed layer and the deep water. While the contribution of apparent oxygen from hydrogen peroxide is probably negligible in the deep water, in the mixed layer, it is obviously significant relative to the claimed precision of

up to ±0.04 μmol kg−1 in the determination of dissolved oxygen. If the samples for oxygen determinations are not kept in the dark after sample collection, additional hydrogen peroxide may be formed photochemically during storage (Cooper et al., 1994). Production rates of 10−1 to 10−2 μmol kg−1 h−1 have been commonly found in natural waters (Yuan and Shiller, 2001; Clark et al., 2009), but rates as high as 100 μmol kg−1 h−1 have also been reported (Cooper et al., 1988, 1994). While ultraviolet light is more effective in inducing the formation of hydrogen peroxide, visible light may still account for a fairly large fraction of the total production (Scully et al., 1996). Wong and Wong (2001) reported that visible light was responsible for one third of the production of hydrogen peroxide in samples of lake water. At these production rates, analytically significant quantities of apparent oxygen from hydrogen peroxide, at 10−2 to 10−1 μmol kg−1, can be readily accumulated in a few hours of storage. The effect of this production of hydrogen peroxide after sample collection is probably small because, as a general practice, dissolved oxygen samples are usually pickled soon after sample collection and then stored away from direct sunlight until they are analyzed Nevertheless, storing the samples in the dark before sample analyses should be the recommended practice. Biological dark production of hydrogen peroxide during storage has also been reported (Cooper et al., 1994). However, this process is probably unimportant here. The added chemicals used in the pickling of the oxygen sample should have greatly impeded, if not terminated, most biological activities. The spatial distributions of hydrogen peroxide and iodate follow contrasting patterns. As a result, the over-estimation in the concentration of oxygen in the Winkler's method caused by their presence will also follow contrasting patterns. Thus, in the open oceans, the concentration of hydrogen peroxide decreases while that of iodate increases with depth. Shoreward, the concentration of iodate may drop to undetectable levels (about b0.02 μM) in inshore water and freshwater (Luther et al., 1991; Wong, 1995; Wong and Zhang, 1992, 2003a, b; Abdel-Moati, 1999; Cook et al., 2000; Truesdale and Jones, 2000; Truesdale et al., 2001; Wong and Cheng, 2001, 2008; Wong et al., 2004) while that of hydrogen peroxide increases. Wong and Li (2009) estimated the average concentrations of iodate and their corresponding concentrations of apparent oxygen in the different types of water and their results are also listed in Table 3. In inshore and freshwater, the interference due to the presence of hydrogen peroxide far exceeds that of iodate. In the open ocean surface water, the former is equivalent to about 15% of the latter, a small but not necessarily negligible additional over-estimation. The total over-estimation from both sources of interference is about 0.6 μmol kg−1 in most waters (Table 3).

3.2.2. Effect on the choice of the solubility equation of oxygen, and the estimation of the air–sea exchange flux and the biological contribution to the saturation anomaly of oxygen Wong and Li (2009) pointed out that the over-estimation of dissolved oxygen in the Winkler's reaction scheme caused by the presence of natural occurring iodate in seawater may have affected the reliability of the equations used for estimating the solubility of oxygen in natural water, in the estimation of the air–sea exchange flux of oxygen, and in the estimation of the biological contribution to the saturation anomaly of oxygen. The presence of hydrogen peroxide will result in an additional over-estimation and exacerbate all these problems. Thus, the solubility equations of oxygen (Weiss, 1970; García and Gordon, 1992), which were based on the equilibrium concentrations determined by using the Winkler's method (Carpenter, 1966; Murray and Riley, 1969), would have yielded overestimated solubilities of oxygen. The presence of hydrogen peroxide will increase the overestimation, and further argues for using, as the equation of choice, that of García and Gordon (1992), which was derived from the results of Benson and coworkers (Benson et al., 1979; Benson and Krause, 1984), since these investigators used a manometric method, that was free from the interference of both iodate and hydrogen peroxide, in the determination of dissolved oxygen.

G.T.F. Wong et al. / Marine Chemistry 122 (2010) 83–90

87

Table 3 Estimated concentration ranges of hydrogen peroxide and iodate in aquatic sub-environments and their contribution to apparent oxygen. Sub-environment

Open ocean Mixed layer Upper thermocline Deep water Coastal ocean Inshore and fresh water

Hydrogen peroxide

Iodate

Total

Conc.

App. O2

Conc.

App. O2

App. O2

μM

μmol kg−1

μM

μmol kg−1

μmol kg−1

0.15 ± 0.1 0.075 ± 0.05 0.005 ± 0.003 0.25 ± 0.15 1 ± 0.8

0.075 ± 0.05 0.038 ± 0.03 0.003 ± 0.002 0.13 ± 0.08 0.5 ± 0.4

0.35 ± 0.1 0.39 ± 0.07 0.43 ± 0.03 0.15 ± 0.15 0.05 ± 0.05

0.52 ± 0.15 0.57 ± 0.1 0.63 ± 0.05 0.23 ± 0.23 0.08 ± 0.08

0.60 ± 0.20 0.61 ± 0.13 0.63 ± 0.05 0.36 ± 0.31 0.58 ± 0.48

App. O2 — apparent oxygen.

The % deviation of the apparent air–sea exchange flux, as a result of the over-estimations in the concentration of oxygen, from the actual flux, %D, is given by the equation (Wong and Li, 2009):   n h  io ′ × 100 %D = δF = F ×100= ðδΔÞ = Δap 1− δΔ = Δap

ð7Þ

where δF is the overestimated flux, F′ is the actual flux, δΔ is the overestimated % saturation anomaly, and Δap is the % apparent saturation anomaly (Wong and Li, 2009). At the average temperature and salinity of 20 °C and 35 in open ocean surface seawater, the equilibrium concentration of oxygen is 225 μmol kg−1. At the average concentration of hydrogen peroxide of 0.15 ± 0.1 μM, the corresponding δΔ is 0.033 ± 0.022%. At the global average Δap oxygen of 3%, (Broecker and Peng, 1982), the presence of hydrogen peroxide would constitute 1.1 ± 0.7% of this super-saturation and lead to an over-estimation of the global efflux of oxygen from the ocean to the atmosphere of 1.1 ± 0.7%. Since the presence of iodate in surface seawater can lead to an overestimation of 8 ± 2%, the total over-estimation is 9 ± 3%. Regionally, in areas where Δap is small, between −1 and 1%, such as in the warm waters in the tropical Atlantic and Pacific and in the high latitude waters in the Pacific, Wong and Li (2009) showed that the effect can be much larger and, at times, even the direction of the sea–air exchange flux may be reversed. They reported that, at its average concentration in the surface mixed layer (Table 3), the presence of iodate alone can result in % D ranging between −36% and +54% in the Western Atlantic and between −106% and +394% in the Central Pacific. The additional presence of hydrogen peroxide at its average concentration of 0.15 μM (Table 3) will lead to %D ranging between −39% and +67% in the Western Atlantic and between −105% and +1029% in the Central Pacific. In inshore water and freshwater, at a concentration of hydrogen peroxide of 1 ± 0.8 μM, an equilibrium concentration of oxygen of 285 μmol kg−1 at 20 °C and salinity 0 and a 3% apparent saturation of oxygen, hydrogen peroxide may lead to an over-estimation in the efflux of oxygen to the atmosphere of 6 ± 5% while the effect of iodate may become negligible as its concentration drops to undetectable levels. Similarly, the δΔ, that results from the presence of hydrogen peroxide, will add to that caused by the presence of iodate and contribute directly to an over-estimation in the biological contribution to the saturation anomaly of oxygen, Δj, in the mixed layer (Wong and Li, 2009). Thus, in the open ocean, the presence of hydrogen peroxide adds around 0.03% to Δj. The effect is small but may not be negligible in regions with low Δj, such as in the subarctic Pacific where Δj as low as 0.0% had been reported (Emerson et al., 2001). 3.2.3. Estimations of community respiration rate and primary production Because of the transient nature of hydrogen peroxide, an intriguing possibility is how changes in its concentration during incubation may affect the estimations of community respiration rate and primary production in mixed layer water. Community respiration rate is estimated almost exclusively from the decrease in the concentration of oxygen in a sample of water before and after a dark incubation for a

specified period of time at a specified temperature (Carignan et al., 1998; Riser and Johnson, 2008). In highly productive coastal waters, the rate may vary between 100 to 101 μmol kg−1 d−1 (York et al., 2001). It drops to mostly 10−l to 100 μmol kg−1 d−1 in the mesotrophic East China Sea (Chen et al., 2003). In the oligotrophic open ocean, the rates are mostly less than 1 μmol kg−1 d−1 and values as low as 0.04 μmol kg−1 d−1 have been reported (Duarte and Agusti, 1998; Williams et al., 2004; Maixandeau et al., 2005; Riser and Johnson, 2008). Even during a bloom, the rate is only several μmol kg−1 d−1 (Robinson et al., 2002). The incubation time may vary between 4 and 24 h (Carignan et al., 1998; Williams et al., 2004). Even at the maximum incubation time of 24 h, the concentration change in dissolved oxygen during the incubation period would mostly be b1 μmol kg−1 in the oligotrophic open ocean. Thus, any error of the order of 10−2 μmol kg−1 can become significant. Hydrogen peroxide can introduce errors into the estimation of respiration rate in several ways. First, its presence in the water sample at the time of collection will lead to an over-estimation of the initial concentration of oxygen in the sample. Secondly, the concentration of hydrogen peroxide may change during the preparation of the sample for the dark incubation, and during the dark incubation itself. Any inadvertent exposure of the sample to light during sample preparation may lead to the formation of hydrogen peroxide. In the dark, both the biologically mediated production (Palenik et al., 1987) and decomposition of hydrogen peroxide (Zepp et al., 1987; Petasne and Zika, 1997; Wong et al., 2003) may occur. The dark biological production rates are typically of the order of 10−2 to 10−3 μmol kg−1 h−1 (Palenik and Morel, 1988, 1990). Overall, decomposition dominates in surface waters where higher initial concentrations of hydrogen peroxide are found. Decomposition half lives ranging from several hours in coastal waters to several days in oligotrophic open ocean waters have been reported (Cooper et al., 1994; Petasne and Zika, 1997). Thus, a significant fraction of the hydrogen peroxide in these waters may disappear during incubation and it will appear as an apparent oxygen consumption by respiration. This will lead to an over-estimation of the community respiration rate. The maximum over-estimation in the community respiration rate occurs when all the hydrogen peroxide present initially in the seawater sample disappears completely upon incubation. The magnitude of this source of error would then be limited by the amount of hydrogen peroxide present, or 0.075 ± 0.05 μmol kg−1 of apparent oxygen in open ocean surface waters, but it can exceed 1 μmol kg−1 in inshore water and freshwater. In sub-surface waters where the initial concentration of hydrogen peroxide is low, biological dark production may exceed the decomposition of hydrogen peroxide. This will lead to a net accumulation of hydrogen peroxide (Palenik and Morel, 1988), and an under-estimation of respiration rate. If this net accumulation of hydrogen peroxide exceeds the respiratory consumption of oxygen, then, a theoretically impossible negative respiration rate will be found. Indeed, while a linkage to the biological dark production of hydrogen peroxide remains to be shown, the occurrence of small unexplained negative respiration rates in sub-surface waters is not uncommon (Williams et al., 2004). Given the complexity of the situation, it is not possible to predict a priori precisely what the overall effect of the decomposition and

88

G.T.F. Wong et al. / Marine Chemistry 122 (2010) 83–90

production of hydrogen peroxide may be in the estimation of the community respiration rates. However, it would not be unreasonable to expect a non-trivial effect. In three samples of coastal seawater, the changes in the concentration of oxygen and hydrogen peroxide upon dark incubation were followed and the results are shown in Table 4. A decrease in the concentration of hydrogen peroxide with time accompanied the decrease in the concentration of oxygen in all the samples. The overestimations in the oxygen consumption rates as a result of this dark decomposition of hydrogen peroxide varied between 0.5 to 4.8%. However, it is not difficult to envision how errors exceeding 10% and even reaching tens of percents may be introduced in other combinations of the initial concentration and the decomposition half life of hydrogen peroxide, and the respiration rate. For example, Carignan et al. (1998) reported a respiration rate of 6.7 mg O2 m−3 h−1 (or 0.21 μM h−1) in the surface water of Croche Lake by using a dark incubation time of 4 h. Assuming a concentration of hydrogen peroxide of 1 ± 0.8 μM and a decomposition half life of 6 h in this inland freshwater, the change in the total concentration of apparent oxygen and in the concentration of apparent oxygen due to the decomposition of hydrogen peroxide during the incubation time would be 0.84 and 0.19 ± 0.15 μM respectively. This would correspond to an over-estimation in the respiration rate of 28%, with a range of 5 to 66%. An error of this magnitude would have been quite significant. Primary production is sometimes estimated from the increase in the concentration of oxygen in a sample of natural water before and after it has been incubated under light for a specified period of time at a specified temperature (Carignan et al., 1998; York et al., 2001; Maixandeau et al., 2005). As in the estimation of the respiration rate, the presence of hydrogen peroxide may lead to an over-estimation of the initial concentration of oxygen in the sample. However, a more significant effect is likely the photochemical production of hydrogen peroxide in the light bottle during sample preparation and incubation and it may lead to a net accumulation of hydrogen peroxide during incubation, and a corresponding over-estimation in primary production. The changes in the concentration of oxygen and hydrogen peroxide in two samples of coastal seawater that were incubated under light were also followed and the results are shown in Table 4. In both cases, both the concentration of hydrogen peroxide and that of oxygen increased with time. The production of hydrogen peroxide led to over-estimations in the production of oxygen by 1 to 3%. The over-

Table 4 Changes in the concentration of oxygen and hydrogen peroxide upon dark and light incubation. Sample

Time

O2

H2O2

O2 change

Over-estimation*

h

μM

μM

μM

%

Dark incubation 1 0 18 24 2 0 18 3 0 18

239.44 237.57 237.11 229.51 220.69 256.81 253.45

0.259 0.138 0.118 0.886 0.073 0.076 0.043

– −1.87 −2.33 – −8.82 – −3.36

– 3.3 3.1 – 4.8 – 0.5

Light incubation 1 0 2.5 4 6 2 0 1 2 3 4

236.61 237.30 238.78 240.27 277.16 278.90 280.43 282.13 280.14

0.163 0.182 0.216 0.237 0.085 0.159 0.181 0.232 0.260



– 1.4 1.2 1.0 – 2.2 1.5 1.5 3.0

0.69 2.17 3.67 – 1.74 3.27 4.97 2.98

*Over-estimation in O2 consumed or produced due to decomposition or formation of H2O2. Sample 1 — off Ya Liu; sample 2 — Tan Shui River mouth; sample 3 — off Wan Li.

estimations were small in these relatively productive coastal waters. Carignan et al. (1998) reported production rates of apparent oxygen in Croche Lake waters of 18.9 mg O2 m−3 h−1 (or 0.59 μM h−1) under an irradiance of 151 μE m−2 s−1. Cooper et al. (1994) reported photochemical production rates of hydrogen peroxide in lake waters of mostly 0.2 to 0.5 μM h−1 under natural sunlight condition. Taking a value of 0.3 μM h−1 and that the efficiency is reduced by two thirds in the incubation bottle (Wong and Wong, 2001), then, the apparent oxygen production rate due to the photochemical production of hydrogen peroxide would be 0.05 μM h−1. This would result in an over-estimation in primary production of 9%. Thus, the photochemical production of hydrogen peroxide does lead to an over-estimation in the determination of primary production by the oxygen method. Whether the effect is negligible or not remains to be seen.

3.4. Correcting for the interference of hydrogen peroxide in the determination of the concentrations of oxygen in natural waters It would be ideal if hydrogen peroxide and apparent oxygen can both be determined in a sample. Then, a correction for the presence of hydrogen peroxide may be made by subtracting 0.5 times its concentration from the apparent concentration of oxygen. This may be the only viable approach in obtaining reliable estimations in the rate of community respiration and primary production by measuring changes in the concentration of oxygen during incubation. For the existing historical field data, an accurate correction cannot be made without knowing the exact protocol in how the samples were handled as hydrogen peroxide could have been produced or consumed during sample storage. However, given that oxygen samples are usually drawn and pickled soon after sample collection and the pickled samples are stored away from direct sunlight, changes in the concentration of hydrogen peroxide after sample collection are likely to be small. If this were the case, corrections may need to be made only for the original concentration of hydrogen peroxide in the sample. Since the range of concentration of hydrogen peroxide in various types of natural waters is relatively well known (Table 3), this correction can be made readily. Thus, 0.075 ± 0.05 and 0.038 ± 0.05 μmol kg−1 of apparent oxygen may be subtracted from the observed values in the surface mixed layer and the upper thermocline respectively. The corresponding correction in the deep water, 0.003 ± 0.002 μmol kg−1, is probably small enough to be ignored. When the over-estimations due to the presence of iodate (Wong and Li, 2009) are also taken into account, the concentration of oxygen will be reduced by 0.60 ± 0.20, 0.61± 0.15 and 0.63 ± 0.05 μmol kg−1 in the surface mixed layer, the thermocline layer and the deep water respectively (Table 3). Considering that the concentration of oxygen in the surface mixed layer ranges from 350 μmol kg−1 in the cold polar regions to 200 μmol kg−1 in the tropics, these corrections for the presence of hydrogen peroxide and iodate will result in an increase in the uncertainty in the concentration of oxygen by an additional ±0.05 to ±0.1% in these waters. At the presently widely claimed precision of ±0.1%, the uncertainties introduced by these corrections will degrade it by 50 to 100%. At the highest precision claimed, ±0.016%, these corrections can increase the uncertainty by more than six fold. In coastal seawater, and inshore water and freshwater, again, if changes in the concentration of hydrogen peroxide after sample collection can be assumed to be negligible, then, from the known distributions of hydrogen peroxide and iodate in these waters, the correction can be estimated to be 0.36± 0.31 and 0.58 ± 0.48 μmol kg−1 respectively (Table 3), and, at a concentration of oxygen of 250 μmol kg−1, these corrections will increase the uncertainty in the determination of oxygen in these waters by ±0.12 and ±0.19%. Nevertheless, while these corrections can be readily made, in order to avoid confusion, such corrections on the published values should be discouraged until an internationally accepted standard practice can be put in place.

G.T.F. Wong et al. / Marine Chemistry 122 (2010) 83–90

Acknowledgments This work was supported in part by the National Science Council, Taiwan through grant number NSC98-2611-M-001-004-MY3, and by the Academia Sinica through a thematic research grant titled “Atmospheric Forcing on Ocean Biogeochemistry (AFOBi)”. References Abdel-Moati, M.A.R., 1999. Iodine speciation in the Nile River estuary. Mar. Chem. 65, 211–225. Bassett, J., Denny, R.C., Jeffery, G.H., Mendham, J., 1978. Vogel's Textbook of Quantitative Inorganic Analysis including Elementary Instrumental Analysis. Longman, Essex, England. 925 pp. Benson, F.B., Krause Jr., D., 1984. The concentration and isotopic fractionation of oxygen dissolved in freshwater and seawater in equilibrium with the atmosphere. Limnol. Oceanogr. 29, 620–632. Benson, B.B., Krause Jr., D., Peterson, M.A., 1979. The solubility and isotopic fractionation of gases in dilute aqueous solution. I. Oxygen. J. Solution Chem. 8, 655–690. Brewer, P.G., 1975. Minor elements in sea water, In: Riley, J.P., Skirrow, G. (Eds.), Chemical Oceanography, 2nd ed. Academic Press, London, pp. 415–496. Broecker, W.S., Peng, T.H., 1982. Tracers in the Sea. ELDIGIO Press, Palisades, New York. 690 pp. Broenkow, W.W., Cline, J.D., 1966. Colorimetric determination of dissolved oxygen at low concentrations. Limnol. Oceanogr. 14, 450–454. Carignan, R., Blais, A.-M., Vis, C., 1998. Measurement of primary production and community respiration in oligotrophic lakes using the Winkler method. Can. J. Fish. Aquat. Sci. 55, 1078–1087. Carpenter, J.H., 1965a. The accuracy of the Winkler method for dissolved oxygen analysis. Limnol. Oceanogr. 10, 135–140. Carpenter, J.H., 1965b. The Chesapeake Bay Institute technique for the Winkler dissolved oxygen method. Limnol. Oceanogr. 10, 141–143. Carpenter, J.H., 1966. New measurements of oxygen solubility in pure and natural water. Limnol. Oceanogr. 11, 264–277. Chen, C.-C., Shiah, F.-K., Gong, G.-C., Chiang, K.-P., 2003. Planktonic community respiration in the East China Sea; importance of microbial consumption of organic carbon. Deep Sea Res. II 50, 1311–1325. Clark, C.D., De Bruyn, W.J., Jones, J.G., 2009. Photochemical productionof hydrogen peroxide in size-fractionated Southern California coastal waters. Chemosphere 76, 141–146. Clesceri, L.S., Greenberg, A.E., Eaton, A.D., 1998. Standard Methods for the Examination of Water and Wastewater, 20th Edition. American Public Health Association, Washington DC, pp. 4-129–4-136. Cook, P.L.M., Carpenter, P.D., Butler, E.C.V., 2000. Speciation of dissolved iodine in the waters of a humic-rich estuary. Mar. Chem. 69, 179–192. Cooper, W.J., Lean, D.R.S., 1989. Hydrogen peroxide concentration in a northern lake: photochemical formation and diel variability. Environ. Sci. Technol. 23, 1425–1430. Cooper, W.J., Zepp, R.G., 1990. Hydrogen peroxide decay in waters with suspended soils: evidence for biologically mediated processes. Can. J. Fish. Aquat. Sci. 47, 888–893. Cooper, W.J., Zika, R.G., 1983. Photochemical formation of hydrogen peroxide in surface and ground waters exposed to sunlight. Science 220, 711–712. Cooper, W.J., Lean, D.R.S., Carey, J.H., 1989. Spatial and temporal patterns of hydrogen peroxide in lake waters. Can. J. Fish. Aquat. Sci. 46, 1227–1231. Cooper, W.J., Zika, R.G., Petasne, R.G., Plane, J.M.C., 1988. Photochemical formation of H2O2 in natural waters exposed to sunlight. Environ. Sci. Technol. 22, 1156–1160. Cooper, W.J., Shao, C., Lean, D.R.S., Gordon, A.S., Scully Jr., S.E., 1994. Factors affecting the distribution of H2O2 in surface waters. In: Baker, L.A. (Ed.), Environmental Chemistry of Lakes and Reservoirs. : American Chemical Society Advances in Chemistry Series, 237. American Chemical Society, Washington D.C, pp. 391–422. Craig, H., Hayward, T., 1987. Oxygen supersaturation in the ocean: biological versus physical contributions. Science 235, 199–202. del Giorgio, P.A., Duarte, C.M., 2002. Respiration in the open ocean. Nature 420, 379–384. del Giorgio, P.A., Coe, J.J., Cimbleris, A., 1997. Respiration rates in bacteria exceed phytoplankton production in unproductive aquatic systems. Nature 385, 148–151. Duarte, C.M., Agusti, S., 1998. The CO2 balance of unproductive aquatic ecosystems. Science 281, 234–236. Duarte, C.M., Agusti, S., del Giorgio, P.A., Cole, J.J., 1999. Regional carbon imbalances in the oceans. Science 284, 1735b. Emerson, S., Quay, P.D., Stump, C., Wilbur, D., Shudlich, R., 1995. Chemical tracers of productivity and respiration in the subtropical Pacific Ocean. J. Geophys. Res. 100 (C8), 15873–15887. Emerson, S., Mecking, S., Abell, J., 2001. The biological pump in the subtropical North Pacific Ocean: nutrient sources, Redfield ratios, and recent changes. Global Biogeochem. Cycles 15 (3), 535–554. Emerson, S., Stump, C., Johnson, B., Karl, D.M., 2002. In situ determination of oxygen and nitrogen dynamics in the upper ocean. Deep Sea Res. I 49, 941–952. Fujiwara, K., Ushiroda, T., Takeda, K., Kumamoto, Y., Tsumota, T., 1993. Diurnal and seasonal distribution of hydrogen peroxide in seawater of the Seta Inland Sea. Geochem. J. 27, 103–115. García, H.E., Gordon, L.I., 1992. Oxygen solubility in seawater: better fitting equations. Limnol. Oceanogr. 37, 1307–1312.

89

Gattuso, J.-P., Frankignoulle, M., Smith, S.V., 1999. Measurement of community metabolism and significance in the coral reef CO2 source-sink debate. Proc. Nat. Acad. Sci. U.S.A. 96, 13017–13022. Grasshoff, K., Kremling, J., Ehrhardt, M., 1999. Methods of Seawater Analysis. WileyVCH, Weinheim. 600 pp. Hanson, A.K., Tindale, N.W., Abdel-Moati, M.A.R., 2001. An Equatorial Pacific rain event: influence on the distribution of iron and hydrogen peroxide in surface waters. Mar. Chem. 75, 69–88. Johnson, K.S., Willason, S.W., Wiesenburg, D.A., Lohrenz, S.E., Arnone, R.A., 1989. Hydrogen peroxide in the western Mediterranean Sea: a tracer for vertical advection. Deep-Sea Res. 36, 241–254. Karl, D.M., Laws, E.A., Morris, P., Williams, P.J.leB., Emerson, S., 2003. Metabolic balance of the open sea. Nature 426, 32. Keeling, R.F., Garcia, H.E., 2002. The change in oceanic O2 inventory associated with recent global warming. Proc. Nat. Acad. Sci. U.S.A. 99, 7848–7853. Kieber, R.J., Helz, G.R., 1995. Temporal and seasonal variations of hydrogen peroxide levels in estuarine waters. Estuar. Coast. Shelf Sci. 40, 195–503. Kolthoff, I.M., Stenger, V.A., 1942. Volumetric Analysis. : Theoretical Fundamentals, vol. I. Interscience, New York. 309pp. Knap, A.J., Michaels, A.F., Steinberg, D., Bahr, R., Bates, N., Bell, S., Countway, P., Close, A.R., Doyle, A.P., Dow, R.L., Howse, F.A., Gundersen, K., Johnson, R.J., Kelly, R., Little, R., Orcutt, K., Parson, R., Rathbun, C., Sanderson, M., Stone, S., 1997. BATS Methods Manual Version 4. U.S. JGOFS Planning and Coordination Office, Woods Hole, Massachusetts, Woods Hole, Massachusetts. 136 pp. Kuss, J., Roeder, W., Wlost, K.-P., DeGrandpre, M.D., 2006. Time-series of surface water CO2 and oxygen measurements on a platform in the Central Arkona Sea (Baltic Sea): seasonality of uptake and release. Mar. Chem. 101, 220–232. Labasque, T., Chaumery, C., Aminot, A., Kergoat, G., 2004. Spectrophotometric Winkler determination of dissolved oxygen: re-examination of critical factors and reliability. Mar. Chem. 88, 53–60. Luther III, G.W., Ferdelman, T., Culberson, C.H., Kostka, J., Wu, J.-F., 1991. Iodine chemistry in the water column of the Chesapeake Bay: evidence for organic iodine forms. Estuar. Coast. Shelf Sci. 32, 267–279. Maixandeau, A., Lefèvre, D., Fernández, I.C., Sempéré, R., Sohrin, R., Ras, J., Wambeke, F.V., Caniaux, G., Quéguiner, B., 2005. Mesoscale and seasonal variability of community production and respiration in the surface waters of the N.E. Atlantic Ocean. Deep Sea Res. I 52, 1663–1676. Miller, W.L., Kester, D.R., 1994. Peroxide variations in Sargasso Sea. Mar. Chem. 48, 17–29. Moffett, J.W., Zafiriou, O.C., 1990. An investigation of hydrogen peroxide chemistry in 18 surface waters of Vineyard Sound with H18 O2. Limnol. Oceanogr. 35, 2 O2 and 1221–1229. Moore, C.A., Farmer, C.T., Zika, R.G., 1993. Influence of the Orinoco River on hydrogen peroxide distribution and production in the eastern Caribbean. J. Geophys. Res. 98, 2289–2298. Morgan, K.J., 1954. Some reactions of inorganic iodine compounds. Q. Rev. 8, 123–146. Murray, C.N., Riley, J.P., 1969. The solubility of gases in distilled water and sea water – II. Oxygen. Deep-Sea Res. 16, 211–220. Pai, S.-C., Gong, G.-C., Liu, K.-K., 1993. Determination of dissolved oxygen in seawater by direct spectrophotometry of total iodine. Mar. Chem. 41, 343–351. Pai, S.-C., Kuo, T.-Y., Chung, S.-W., Su, T.-T., 1998. Azide-modified shibala colorimetric method for the determination of dissolved oxygen and an assessment of its applicability to environment studies. Chemistry (The Chinese Chem. Soc., Taipei) 56, 173–185. Palenik, B., Morel, F.M.M., 1988. Dark production of H2O2 in the Sargasso Sea. Limnol. Oceanogr. 33, 1606–1611. Palenik, B., Morel, F.M.M., 1990. Amino acid utilization by marine phytoplankton: a novel mechanism. Limnol. Oceanogr. 35, 260–269. Palenik, B., Zafiriou, O.C., Morel, F.M.M., 1987. Hydrogen peroxide production by a marine phytoplankter. Limnol. Oceanogr. 32, 1365–1369. Petasne, R.G., Zika, R.G., 1997. Hydrogen peroxide lifetimes in south Florida coastal and offshore waters. Mar. Chem. 56, 215–225. Platt, T., Harrison, W.G., Lewis, M.R., Li, W.K.W., Sathyendranath, S., Smith, R.E., Vezina, A.F., 1989. Biological production of the oceans: the case for a consensus. Mar. Ecol. Prog. Ser. 52, 77–88. Plattner, G.-K., Joos, F., Stocker, T.F., 2002. Revision of the global carbon budget due to changing air–sea oxygen fluxes. Global Biogeochem. Cycles 16, 1096. doi:10.1029/2001GB001746. Reinthaler, T., Bakker, K., Manuels, R., van Ooijen, J., Herndl, G.J., 2006. Fully authomated spectrophotometric approach to determine oxygen concentration in seawater via continuous-flow analysis. Limnol. Oceanogr: Methods 4, 358–366. Riser, S.C., Johnson, K.S., 2008. Net production of oxygen in the subtropical ocean. Nature 451, 323–325. Robinson, C., Widdicombe, C.E., Zubkov, M.V., Tarran, G.A., Miller, A.E.J., Rees, A.P., 2002. Plankton community respiration during a coccolithophore bloom. Deep Sea Res. II 49, 2929–2950. Scully, N.M., McQueen, D.J., Lean, D.R.S., Cooper, W.J., 1996. Hydrogen peroxide formation: the interaction of ultraviolet radiation and dissolved organic carbon in lake water along a 43–75°N gradient. Limnol. Oceanogr. 41, 540–548. Strickland, J.D.H., Parson, T.R., 1972. A Practical Handbook of Seawater Analysis. Fisheries Research Board of Canada Ottawa. 310 pp. Truesdale, V.W., Jones, K., 2000. Steady-state mixing of iodine in shelf seas off the British Isles. Cont. Shelf Res. 20, 1889–1905. Truesdale, V.W., Nausch, G., Baker, A., 2001. The distribution of iodine in the Baltic Sea during summer. Mar. Chem. 74, 87–98. Weiss, R.F., 1970. The solubility of nitrogen, oxygen and argon in water and seawater. Deep-Sea Res. 17, 721–735.

90

G.T.F. Wong et al. / Marine Chemistry 122 (2010) 83–90

Williams, P.J.leB., Bowers, D.G., 1999. Regional carbon imbalances in the oceans. Science 284, 1735b. Williams, P.J.leB., Morris, P.J., Karl, D.M., 2004. Net community production and metabolic balance at the oligotrophic ocean site, station ALOHA. Deep Sea Res. I 51, 1563–1578. Wong, G.T.F., 1991. The marine geochemistry of iodine. Rev. Aquat. Sci. 4, 45–74. Wong, G.T.F., 1995. Dissolved iodine across the Gulf Stream Front and in the South Atlantic Bight. Deep Sea Res. I 42, 2005–2023. Wong, G.T.F., Brewer, P.G., 1974. The determination and distribution of iodate in South Atlantic waters. J. Mar. Res. 32, 25–36. Wong, G.T.F., Cheng, X.-H., 2001. Dissolved organic iodine in marine waters: role in the estuarine geochemistry of iodine. J. Environ. Monit. 3, 257–263. Wong, G.T.F., Cheng, X.-H., 2008. Dissolved inorganic and organic iodine in the Chesapeake Bay and adjacent Atlantic waters: speciation changes through an estuarine system. Mar. Chem. 111, 221–232. Wong, G.T.F., Li, K.-Y., 2009. Winkler's method overestimates dissolved oxygen in seawater: iodate interference and its oceanographic implications. Mar. Chem. 115, 86–91. Wong, A.Y.L., Wong, G.T.F., 2001. The effect of spectral composition on the photochemical production of hydrogen peroxide in lake water. Terr. Atmos. Ocean. Sci. 12, 695–704. Wong, G.T.F., Zhang, L.-S., 1992. Changes in iodine speciation across coastal hydrographic fronts in southeastern United States continental shelf waters. Cont. Shelf Res. 12, 717–733. Wong, G.T.F., Zhang, L.-S., 2003a. Geochemical dynamics of iodine in marginal seas: the southern East China Sea. Deep Sea Res. II 50, 1147–1162. Wong, G.T.F., Zhang, L.-S., 2003b. Seasonal variations in the speciation of dissolved iodine in the Chesapeake Bay. Estuar. Coast. Shelf Sci. 56, 1093–1106.

Wong, G.T.F., Zhang, L.-S., 2007. The kinetics of the reactions between iodide and hydrogen peroxide in seawater. Mar. Chem. 111, 22–29. Wong, G.T.F., Dunstan, W.M., Kim, D.-B., 2003. The decomposition of hydrogen peroxide by marine phytoplankton. Ocean. Acta 26, 191–198. Wong, G.T.F., Hung, C.-C., Gong, G.-C., 2004. Dissolved iodine species in the East China Sea — a complementary tracer for upwelling water on the shelf. Cont. Shelf Res. 24, 1465–1484. York, J.A., Witer, Z., Labudda, S., Ochocki, S., 2001. Comparison of primary production and pelagic community respiration rates in the coastal zone of the Gulf of Gdańsk. Oceanologia 43, 365–370. Yuan, J., Shiller, A.M., 2001. The distribution of hydrogen peroxide in the southern and central Atlantic ocean. Deep Sea Res. II 48, 2947–2970. Yuan, J., Shiller, A.M., 2004. Hydrogen peroxide in deep waters of the North Pacific Ocean. Geophys. Res. Lett. 31 (L01310), 1–4. doi:10.1029/2003GL018439. Yuan, J., Shiller, A.M., 2005. Distribution of hydrogen peroxide in the northwest Pacific Ocean. Geochem. Geophys. Geosyst. 6 (Q09M02), 1–13. doi:10.1029/2004GC000908. Zepp, R.G., Skurlatov, Y.I., Pierce, J.T., 1987. Algal-induced decay and formation of hydrogen peroxide in water: its possible role in oxidation of anilines by algae. In: Zika, R.G., Cooper, W.J. (Eds.), Photochemistry of environmental aquatic systems. : American Chemical Society Symposium Series, 327. American Chemical Society, Washington D.C, pp. 215–224. Zhang, L.-S., Wong, G.T.F., 1994. Spectrophotometric determination of H2O2 in marine waters with leuco crystal violet. Talanta 41, 2137–2145. Zhang, L.-S., Wong, G.T.F., 1999. Optimal conditions and storage for the determination of H2O2 in marine waters by the scopoletin-HRP fluorometric method. Talanta 48, 1031–1038. Zika, R.G., Moffett, J.W., Petasne, R.G., Cooper, W.J., Saltzman, E.S., 1985. Spatial and temporal variations of hydrogen peroxide in Gulf of Mexico waters. Geochim. Cosmochim. Acta 49, 1173–1184.