lable at ScienceDirect

Estuarine, Coastal and Shelf Science 133 (2013) 260e272

Contents lists avai

Estuarine, Coastal and Shelf Science

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

The impact of natural water column mixing on iodine and nutrientspeciation in a eutrophic anchialine pond (Rogoznica Lake, Croatia)q

Vesna �Zic a,*, Marina Cari�c b, Irena Ciglene�cki a

aRuCer Bo�skovi�c Institute, Division for Marine and Environmental Research, Bijeni�cka 54, 10001 Zagreb, CroatiabUniversity of Dubrovnik, Institute for Marine and Coastal Research, Kneza D. Jude 12, 20101 Dubrovnik, Croatia

a r t i c l e i n f o

Article history:Received 8 April 2013Accepted 10 September 2013Available online 25 September 2013

Keywords:iodine speciationnutrientsanchialine environmentsmeromictic lakesnatural mixingCroatiaRogoznica Lake

q Note e some of the results presented in thisextended abstract in a Special Issue of Natura CroaticaCukrov and V. �Zic/2nd International Symposium on A* Corresponding author. Current address: Hrvatsk

agement Laboratory, Uvala �Skar, 22001 �Sibenik, CroatE-mail addresses: vesna.zic@si.t-com.hr (V. �Zic), m

(M. Cari�c), irena@irb.hr (I. Ciglene�cki).

0272-7714/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.ecss.2013.09.008

a b s t r a c t

The effect of natural water column mixing on the inorganic iodine and nutrient speciation in an intenselyeutrophicated anchialine pond has been studied in late seasons of 2003 and 2004. As a result of anextremely warm and very dry European summer of 2003 this system attained isohaline and isothermalconditions in late-summer. Vertical mixing between highly reducing deep water with that in the upperlayers produced intermediate redox conditions and homogeneous nutrient distribution within the watercolumn. Nitrogen speciation additionally suggests a pronounced activity of nitrifying bacteria and/orarchaea arising from an influx of the ammonium nitrogen. Meanwhile, in 2004 the water columnattained nearly isohaline conditions late in November. Partial mixing of the water column resulted indownward transport of dissolved oxygen and moderate upward fluxes of nutrients into the surfacelayers, whereby the flux of ammonium was sufficiently high to promote nitrification. A response of theinorganic iodine to mixing of the water column was in accord with those of major constituents andnutrients. The results demonstrate that, in spite of marked changes in speciation, total inorganic iodineremained preserved within the system over each survey. Oxic to hypoxic redox conditions formedimmediately after the overturn in September 2003 favoured iodate reduction, while slow mixing inNovember 2004 promoted iodide oxidation. In this system dominated by the sedimentary influence adegree of oxygenation of deep water seems to control the inorganic iodine inventory for the pond.

� 2013 Elsevier Ltd. All rights reserved.

� �1

1. Introduction

Over the past several decades much of research effort has beenfocused to establish the general pattern of iodine distribution indifferent marine environments, as well as to define the principalcontrols of iodine speciation in seawater. While the former is ratherwell known, the later task still represents a challenge to iodinescientists, since, sometimes, the field studies provide contrastingresults, while the laboratory those do not always match fieldobservations.

Iodine is the most abundant biophilic and redox sensitive minorelement in seawater (Elderfield and Truesdale, 1980; Wong, 1991).In the oceans, the largest iodine reservoir on the earth, iodinemainly presents as its inorganic species; iodate ðIO�

3 Þ and iodide

study were published as an(2012, Vol. 21) (Guest Eds.: N.nchialine Ecosystems).e vode, Central Water Man-ia.arina.caric-gluncic@unidu.hr

All rights reserved.

(I ), at a total concentration of about 0.45 mmol l . While iodategenerally predominates in deep oceanic waters and surface watersaround the poles, both species may be found at comparable con-centrations in tropical and sub-tropical surface waters or shallowshelf seas (Luther et al., 1988; Truesdale, 1994a; Campos et al., 1996;Bluhm et al., 2011). Dissolved (Truesdale, 1975; Zheng et al., 2011)and particulate (Wong et al., 1976; Jickells et al., 1990) organic-Iattain maximal concentrations within the euphotic zone, butgenerally represent only a minor fraction of total iodine in openocean waters. Meanwhile, in aquatic environments which areexposed to a greater terrestrial influence, eutrophication, or morepronounced sedimentary influence, iodine speciation is morecomplex (Butler et al., 1988; Luther et al., 1991; Luther andCampbell, 1991; Edwards and Truesdale, 1997; Wong and Cheng,1998; Beck and Bruland, 2000; Truesdale et al., 2001a; Schwehret al., 2005; �Zic et al., 2011, 2012). In such systems not only thatorganic-I component may be substantial, but also the biogeo-chemical processes that affect the iodine system are usually farmore pronounced than those in open ocean waters.

From thermodynamics, iodate should predominate under oxicconditions, while iodide under anoxic ones (Sillen, 1961). There-fore, it is not surprising that numerous studies have been

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272 261

conducted in order to explain the presence of iodide in surfacewaters. It is now generally accepted that that the processesinvolved in nutrient dynamics also affect the iodine system andthat the presence of biota may result in iodate reduction to iodide.However, the contrary results of different studies do not allow usto be conclusive upon whether at natural iodine levels in seawateriodine speciation is more affected by phytoplankton activity duringproduction (Moisan et al., 1994; Wong et al., 2002; Chance et al.,2007) or by bacterial activity during regeneration (Waite andTruesdale, 2003; De la Cuesta and Manley, 2009; Bluhm et al.,2010). Abiotic processes (Spokes and Liss, 1996; Truesdale andUpstill-Goddard, 2003) have also been proposed as controls ofiodine speciation in oxic seawater, but, overall, biologically medi-ated reactions seem to be of greater importance. Meanwhile, inpermanently or seasonally stratified aquatic systems, where highbiological activity is coupled with insufficient vertical mixing,iodine speciation approaches thermodynamic predictions, withiodide predominating under anoxic conditions (Wong and Brewer,1977; Emerson et al., 1979; Ullman et al., 1990; Luther andCampbell, 1991; Stipani�cev and Branica, 1996; Rue et al., 1997;Truesdale et al., 2001a,b, 2013). In these environments a numberof inorganic and organic electron-donors may potentially reduceiodate (e.g. HS�, Mn2þ, Fe2þ, humic substances) (Zhang andWhitfield, 1986; François, 1987; Luther and Tsamakis, 1989),either at chemocline during stratification period, or during mixingof the water column during renewals.

Aside from being extensively studied for iodate reduction toiodide, these systems also represent valuable natural laboratoriesfor studying the opposite process of iodide oxidation to iodate,since any upward diffusion and/or advection flux across thechemocline would provide principally iodide (Truesdale et al.,2001b; �Zic and Branica, 2006; Truesdale et al., 2013). Thus, anyincrease in iodate concentration can more safely be ascribed toin-situ iodide oxidation. Albeit the process of iodide oxidation toiodate has received a substantially lower attention than thereversal one of iodate reduction, the bacterial mediation emergesas a preferred explanation (Butler et al., 1981; Luther et al., 1995;Edwards and Truesdale, 1997). Perhaps the most intriguing puz-zle in iodide oxidation to iodate is that none of the laboratorystudies, with isolated specific marine bacteria, was capable todemonstrate iodide oxidation up to iodate; the only oxidationproducts were either molecular iodine or organic iodine com-pounds (Gozlan, 1968; Fuse et al., 2003; Amachi, 2008). Incontrast, the results from the field studies (Edwards andTruesdale, 1997; Beck and Bruland, 2000; �Zic and Branica, 2006;�Zic et al., 2010; Truesdale et al., 2013), together with estimatesfrom the hydrographic modelling (Campos et al., 1996; Truesdaleet al., 2001b), suggest that iodide oxidation to iodate must occurin seawater.

This study examines the reactivity of the inorganic iodine sys-tem when it is exposed to a natural stress of mixing of oxic andanoxic/sulfidic waters in the water column of a small meromicticpond, Rogoznica Lake. The survey spans late seasons, when surfacecooling promotes favourable conditions for at least partial mixingbetween the mixolimnion and monimolimnion (e.g. Ciglene�ckiet al., 2005). Since the results of our earlier iodine study (�Zic andBranica, 2006), which was conducted about 9 months after thispond experienced the acute anoxic event in 1997 (Bari�c et al., 2003;Ciglene�cki et al., 2005), indirectly suggested that there might be alink between iodate formation and regenerated nitrate production,it was hypothesised that the intensive simultaneous sampling ofiodine, nutrient and sulfur speciation in this late season will pro-vide useful information upon the extent of iodate reduction/for-mation during such events and upon the eventual similaritybetween the iodine and the nitrogen system.

2. Study area

Rogoznica Lake (Zmajevo oko) can be considered as an anchia-line pond (�Zic et al., 2010) on the Gradina Peninsula, at the easterncoast of the Adriatic Sea (43�320 N, 15�580 E) (Fig. 1). The surfacearea of the pond is 10,276 m2 (Land Registry Office, �Sibenik),maximumdepth 15m, while the volume of water and the sedimentsurface area, estimated from bathymetric data (Mihel�ci�c et al.,1996), are about 85,700 m3 and 14,400 m2, respectively (�Zic andBranica, 2006). It is surrounded with high vertical cliffs thatgreatly reduce the extent of wind-driven mixing. Since there is noobvious surface connection to the sea, or any known extensive sub-surface connection, the seawater replenishment is restricted by the100e150 m of the karstified carbonate rock. The phase shift inpond’s damped tide to that outside in the bay is around 2 h.

Because the pond is under both terrestrial andmarine influence,yet sufficiently isolated from the surrounding sea, the salinitygradient is generally sufficiently high to prevent complete mixingof thewater column during winter over at least one annual cycle, sothat the pond belongs to meromictic environments. Stabile strati-fication also isolates the oxic mixolimnion, which is dominated byphotosynthesis, from the anoxic monimolimnion, where regener-ation and sedimentary influence prevail, and where high sulfide(up to 10 mmol l�1), ammonium (up to 250 mmol l�1), phosphate(up to 22 mmol l�1), silica (up to 370 mmol l�1) and iodide (up to2.3 mmol l�1) concentrations were registered (Ciglene�cki et al.,2005; �Zic and Branica, 2006). High nitrate concentrations (up to30 mmol l�1) during winter and spring are considered to be pri-marily generated by nitrification (Ciglene�cki et al., 2005; �Zic et al.,2010). High concentrations of dissolved (up to 500 mmol l�1) andparticulate (up to 120 mmol l�1) organic carbon and surface activesubstances (up to 0.05 mmol l�1) additionally indicate that the pondis highly productive (�Cosovi�c et al., 2000; Svensen et al., 2008;Penezi�c et al., 2010; Plav�si�c et al., 2011). The sediments in the pondhave been classified as authigenic carbonate sediments of mainlybiogenic origin, and are consisted primarily of calcite, aragonite,amorphous silica, pyrite and organic matter (Mihel�ci�c et al., 1996;Ciglene�cki et al., 2006).

The climate of this area is Mediterranean, with mean annualtemperature of 16 �C and mean annual precipitation of 825 mm,which is very seasonal. In addition to intra-annual variations, inter-annual differences may also be pronounced (Ciglene�cki et al., 2005)as registered over this study (Fig. 2). The statistical analyses in-dicates that the year 2003 was extremely warm and very dry(610 mm yr�1), while the year 2004 warm and wet (950 mm yr�1)relative to the thirty-year average (1961e1990) (Statistical Year-books, Republic of Croatia e Central Bureau of Statistics, http://www.dzs.hr/default_e.htm, Croatian Meteorological and Hydro-logical Service, http://klima.hr/ocjene_arhiva.html). The beginningof December 2004 was extremely rainy and this area received160 mmwithin only couple of hours, and additional 100 mm by theend of a month.

3. Sampling and analysis

Depth profiles were taken from a small boat in the centre of thepond on August 26th and September 23rd, 26th and 29th 2003, andon November 2nd, 12th, 22nd and December 20th 2004. Thesampling in 2003 was conducted earlier, since the salinity structurein August suggested that the eventual mixing of the water columnwill be triggered by surface cooling. Meanwhile, the results fromspring and summer seasons of 2004 (�Zic et al., 2010) indicated thatfavourable conditions for mixing will probably occur later in theautumn. The sampling was performed by 5 l Niskin sampler, which,when brought to the surface, was attached to a nitrogen cylinder in

Fig. 1. Location and vertical transect (WesteEast) of the Rogoznica Lake in relation to the surrounding sea.

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272262

order to preserve hypoxic and/or anoxic conditions over the samplewhile it is transferred to sampling bottles. The bottles, previouslyflushed with nitrogen, were filled with the sample to overflowing,and were rapidly closed by a rubber septum held additionally byparafilm in order to avoid contamination with air. No headspaceand bubbles were left in the bottles.

Temperature and salinity were determined immediately uponsample collection with a mercury in-glass thermometer andrefractometer (Atago, Japan), respectively. Dissolved oxygen con-centration was determined by the standard manual Winkler’smethod (Strickland and Parsons, 1972).

The concentrations of nutrients were measured by spectrom-etry. Nitrite and nitrate þ nitrite were analysed after reaction withsulphanilamide and N-(1-naphtyl)-ethylenediamine dihydro-chloride solutions, before and after reduction of samples on col-umns filled with copper coated cadmium granules, respectively

0

10

20

30

J F M A M J J A S O N D J

tem

peratu

re/

oC

3002

Fig. 2. Annual variability of monthly means of temperature (white bars) and precipitation (gRepublic of Croatia e Central Bureau of Statistics, http://www.dzs.hr/default_e.htm). The ar

(Bendschneider and Robinson, 1952; Wood et al., 1967). The indo-phenol blue method was used for ammonium determination(Solorzano, 1969; Ivan�ci�c and Degobbis, 1984), while phosphatewas analysed according to Murphy and Riley (1962). The method isbased on formation of phosphomolybdate complex, which is sub-sequently reduced with ascorbic acid to form a strongly colouredblue molybdenum complex. Silicic acid was analysed with hepta-molybdate, using methol as a reducing agent in the presence ofoxalic acid (Mullin and Riley, 1955). Samples for ammonium ana-lyses were preserved with phenol, while the samples for nitrate,nitrite, phosphate and silicic acid analysis were frozen until analysisthat was performed within a month (Macdonald and McLaughlin,1982; Macdonald et al., 1986).

Reduced sulfur species (viz. total RSS and non-volatile RSS) wereanalysed by linear sweep voltammetry (LSV) (Ciglene�cki et al.,1996; Ciglene�cki and �Cosovi�c, 1997), within 8 h of sampling. Since

F M A M J J A S O N D

0

50

100

150

200

250

300

precip

itatio

n/m

m

4002

rey bars) at station closest to the sampling site (�Sibenik) (Source: Statistical Yearbooks,rows highlight sampling months.

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272 263

the sub-samples for sulfur speciation were collected under inertnitrogen atmosphere, as explained previously, and no bubbles werenoticed on the walls of specially designed glass sampling bottlesjust prior to measurements, our assumption is that the loss ofhydrogen sulfide was negligible. The concentration of volatile RSSwas calculated as the difference between the total (RSST) and non-volatile reduced sulfur species. In the Rogoznica Lake non-volatilefraction is principally consisted of S0 component of polysulfidesand S8, while the volatile fraction is chiefly free sulfide (H2S þ HS�)and any sulfide that is freed by acidification (Bura-Naki�c et al.,2009). Electrochemical measurements were performed with m-Autolab Electrochemical Instruments (Eco Chemie) connected with663 VA Stand Metrohm electrode. The working electrode was ahanging mercury drop electrode, while reference was an Ag/AgCland graphite as auxiliary electrode. Detection limit of the methodwas (1.0 � 0.5) nmol l�1.

The concentrations of iodate and iodide in unfiltered sampleswere measured using a PAR 384B Electrochemical Analyser inconjunction with PAR 303A static mercury drop electrode (SMDE),with an Ag/AgCl (sat. NaCl) reference electrode and a platinumwireas a counter electrode. Iodate and iodide were determined by dif-ferential pulse voltammetry (Herring and Liss, 1974) and cathodicstripping square wave voltammetry (Luther et al., 1988), respec-tively, and sodium sulfite was added to facilitate the removal ofdissolved oxygen (Wong and Zhang, 1992). To obtain concentrationrange optimal to iodide method samples were diluted with Milli-Qwater prior to analyses. The blank was typically lower than0.002 mmol l�1. Detection limit of the iodate method was0.02 mmol l�1. Peak currents were determined from the baseline-corrected voltammograms using a spline function implementedin a home-written software package, ECDSOFT (Omanovi�c andBranica, 1998; Pi�zeta et al., 1999; Omanovi�c, 2006) and calibrationwas by the standard addition method. Replicate sample analysestypically gave precision within �5% for iodate and iodide, so thatprecision for total inorganic iodine would be within �7%(TII ¼ iodate þ iodide). Samples were stored in the dark at 4 �C forno more than about a week before analysis. Potentially, some al-terations in inorganic iodine speciation during this short period ofstorage were feasible (Campos, 1997; Farrenkopf et al., 1997).

4. Results

4.1. Depth profiles in August and September 2003

4.1.1. Major constituentsThe profiles on Aug. 26th describe well the persistent structure

of the water column during the stratified period (Fig. 3a). Hightemperature (z28 �C) within the mixolimnion depths (0e9 m) andrather high salinity within the entire water column (38e40) reflectwarm and dry preceding seasons (Sec. 2, Fig. 2). Dissolved oxygenconcentrations within the upper half of the water column (over-saturation of up to 150%) indicate pronounced photosyntheticproduction. As a result of settling of this organic material, andbecause the volume of water up to 9m represents almost 90% of thepond’s volume, anoxic and sulfidic conditions developed in the 10%representing the deep water. The anoxic layer contained primarilysulfide, at concentration of up to 1200 mmol l�1. Only about 10% ofRSST was in the form of elemental sulfur and organo-sulfurcompounds.

In September 2003 the water column became nearly isothermaland isohaline on all three occasions (Fig. 3a). That would be inaccord with rather low amount of precipitation and significantlylower temperatures registered in September (Fig. 2).

On the first day of sampling in September (Sep. 23rd), dissolvedoxygen concentrations within the upper half of the water column

and those of RSST and sulfide in the anoxic layer were alreadyreduced for about 50%, 60% and 80%, respectively, relative to con-centrations obtained in August. This is well exemplified by Fig. 4that presents the whole-pond content of selected variables, calcu-lated from the available bathymetric data and the pond surface area(Sec. 2). This approach has been used in two previous studies (�Zicand Branica, 2006; �Zic et al., 2010), where arguments for lateralhomogeneity of the pond and uncertainties of these estimates arediscussed. Of course, in the close vicinity of the sediment or duringsubstantial runoff one may expect more pronounced differencesbetween the central and the lateral sampling points. The profiles inSeptember indicate that the mixing between the deep water andthat in the upper layers continued to be slow. Between Sep. 23rdand Sep. 26th dissolved oxygen concentration in the large volumeof water up to 8 m decreased from 140 mmol l�1 to 90 mmol l�1,while, with the exception of the deepest sample, the redox condi-tions in deepwater changed fromanoxic and sulfidic to hypoxic andnon-sulfidic. This pattern largely continued up to Sep. 29th, it is justthat the surface water became more oxygenated (Figs. 3a and 4).

Over the entire period hypoxic and oxic water contained pri-marily non-volatile reduced sulfur fraction, at concentration of upto 0.03 mmol l�1. Sulfide was generally low under oxic conditions(up to 0.004 mmol l�1) and was detected only on the first day ofsampling in September.

4.1.2. NutrientsDepth profiles of nutrients on Aug. 26th (Fig. 3a) are in

accordance with stratification and redox conditions within thewater column. Severely depleted nutrients within the upper halfof the water column reflect pronounced phytoplankton activity,while significantly higher concentrations of phosphate and silicicacid in the anoxic water would be in accord with regeneration ofsettling algae and diffusion from the anoxic sediment. Whilephosphate correlates closely with sulfide in the anoxic layer(rP ¼ 0.9998, p ¼ 0.001, N ¼ 4) (Fig. 5), the profile of silicic acid ismore closely associated with the density structure (Fig. 3a), andthe Pearson’s correlation coefficient (rP) of 0.94 for these twovariables between 7 and 13 m is statistically significant atp ¼ 0.01 (N ¼ 7). Consequently, the nutricline with respect tosilicic acid relative to that of phosphate is shifted upwards forabout 2 m, reflecting different geochemical reactivity and bio-logical utilisation of phosphate and silicic acid following thepositive redox-turnover.

On Sep. 23rd the concentrations of ammonium, phosphorus andsilicic acid within the upper half of the water column increased forabout five times, as a result of intrusion of nutrient rich deep water,where phosphate and silicic acid concentrations decreased forabout 50%. Again, the Pearson’s correlation between phosphate andsulfide in anoxic water is high (rP¼ 0.9997, p¼ 0.02, N¼ 3), it is justthat the slope of the linear regression line of sulfide vs. phosphatedecreased from 82 (�1) to 44 (�1) mmol l�1/mmol l�1, suggestingfast sulfide oxidation in deep anoxic water (Fig. 5). The profiles ofnutrients in deep water generally support slow-mixing regime ofthe pond between Sep. 23rd and Sep. 29th.

Nitrogen speciation in September indicates a pronounced ni-trate formation within the entire water column, particularly sobetween Sep. 23rd and Sep. 26th, when about 70 moles of nitratewere formed in the pond in three days. This corresponds to a netnitrate formation rate of 270 nmol l�1 day�1. While other redox-couples, such as Mn(III, IV)/Mn(II) or Fe(III)/Fe(II), may potentiallyoxidise ammonium, as has been observed in marine sediments (e.g.Luther et al., 1997; Anschutz et al., 2005), it is far more likely that anincrease in nitrate concentration along the entire depth profile is aresult of nitrification that was promoted by an ammonium influxinto oxygenated waters, as observed elsewhere (e.g. Ward, 2008).

0

2

4

6

8

10

12

14

0 1 2 3 4

de

pth

/m

c(nitrate + nitrite)/ mol l-1

0

2

4

6

8

10

12

14

0.0 0.2 0.4 0.6 0.8

dep

th

/m

c(iodate)/ mol l-1

0

2

4

6

8

10

12

14

15 20 25 30

dep

th

/m

t /oC

0

2

4

6

8

10

12

14

0.0 0.5 1.0 1.5 2.0 2.5 3.0

dep

th

/m

c(iodide)/ mol l-1

0

2

4

6

8

10

12

14

30 35 40 45

dep

th

/m

S

0

2

4

6

8

10

12

14

0 100 200 300 400

de

pth

/m

c(O2)/ mol l

-1

0

2

4

6

8

10

12

14

22 24 26 28 30

dep

th

/m

t/kg m

-3

0

2

4

6

8

10

12

14

0 5 10 15 20

de

pth

/m

c(phosphate)/ mol l-1

0

2

4

6

8

10

12

14

0 50 100 150 200

de

pth

/m

c(silicic acid)/ mol l-1

0

2

4

6

8

10

12

14

0 500 1000 1500

de

pth

/m

c(RSST, sulfide)/ mol l

-1

0

2

4

6

8

10

12

14

0.0 0.5 1.0 1.5 2.0 2.5 3.0

dep

th

/m

c(TII)/ mol l-1

0

2

4

6

8

10

12

14

0 20 40 60

de

pth

/m

c(ammonium)/ mol l-1

Aug. 26 Sep. 23 Sep. 26 Sep. 29 a

Fig. 3. a. Profiles of temperature, salinity, density, dissolved oxygen, reduced sulfur species, phosphate, nitrate þ nitrite, ammonium, silicic acid, iodate, iodide and total inorganiciodine in the Rogoznica Lake in 2003. The smaller symbols with dotted lines on oxygen plot refer to theoretical dissolved oxygen concentrations at saturation (with respect totemperature and salinity) at given depths. The smaller symbols with dotted lines on RSS plot refer to volatile RSS. b. Same as for Fig. 3a in 2004. Dotted lines on iodine plots refer toconcentrations rationalised to S ¼ 38 on Dec. 20th.

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272264

0

2

4

6

8

10

12

14

0 2 4 6 8

dep

th

/m

c(nitrate + nitrite)/ mol l-1

0

2

4

6

8

10

12

14

0.0 0.2 0.4 0.6

de

pth

/m

c(iodate)/ mol l-1

0

2

4

6

8

10

12

14

0.0 0.5 1.0 1.5 2.0

de

pth

/m

c(iodide)/ mol l-1

0

2

4

6

8

10

12

14

5 10 15 20 25 30

de

pth

/m

t/oC

0

2

4

6

8

10

12

14

20 25 30 35 40

dep

th

/m

S

0

2

4

6

8

10

12

14

0 100 200 300 400

de

pth

/m

c(O2)/ mol l

-1

0

2

4

6

8

10

12

14

15 20 25 30

de

pth

/m

t/kg m

-3

0

2

4

6

8

10

12

14

0 5 10 15 20

de

pth

/m

c(phosphate)/ mol l-1

0

2

4

6

8

10

12

14

0 100 200 300 400

dep

th

/m

c(silicic acid)/ mol l-1

0

2

4

6

8

10

12

14

0 500 1000 1500

de

pth

/m

c(RSST, sulfide)/ mol l

-1

0

2

4

6

8

10

12

14

0.0 0.5 1.0 1.5 2.0

de

pth

/m

c(TII)/ mol l-1

0

2

4

6

8

10

12

14

0 5 10 15 20 25

dep

th

/m

c(ammonium)/ mol l-1

Nov. 2 Nov. 12 Nov. 22 Dec. 20

b

Fig. 3. (continued).

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272 265

3300

20300

6100

5400

18+

4

3200

10100

2600

1000

44+

13

3200

6700

15

0

10

0

112+

15

3200

8100

1 0 115+

16

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

salt/t DO RSS sulfide nitrate+nitrite

m(s

alt)/t o

r n

/m

ol

73 (75)

30 (30)

43 (45)

84

13

71

85

9 76

81

8 73

0

20

40

60

80

100

TII iodate iodide

n/m

ol

3000

16100

2600

2400

21+

4

3000

17200

2900

2700

89+

13

3100

19100

2600

2300

121+

4

2800

17600

2500

2200

31

8+

24

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

salt/t DO RSS sulfide nitrate+nitrite

m(s

alt)/t o

r n

/m

ol

55 (51)

21 (19)

34 (33)

59 (56)

22 (20)

37 (36)

54 (53)

24 (23)

30 (29)

54 (48)

27 (22)

27 (26)

0

20

40

60

80

100

TII iodate iodide

n/m

ol

Aug. 26th

Sep. 23rd

Sep. 26th

Sep. 29th

2003

Nov. 2nd

Nov. 12th

Nov. 22nd

Dec. 20th

2004

Fig. 4. Salt content and molar amounts of dissolved oxygen, total reduced sulfur species, sulfide, nitrate þ nitrite (each specified), total inorganic iodine, iodate and iodide in theRogoznica Lake in 2003 (upper graphs) and 2004 (bottom graphs). Iodine concentrations were rationalised to S ¼ 38, while values in brackets refer to estimates with commonconcentrations. Error-bars refer to �10%. Note that the scales on the left hand side are logarithmic.

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272266

4.1.3. IodineDepth profiles of iodate and iodide on Aug 26th exemplify well

iodine behaviour in two opposite redox compartments (Fig. 3a). Up toaround 7 m iodate was more abundant than iodide, while in anoxicwater it was essentially absent. Meanwhile, iodide concentrationswere high in deep water, ranging between 1.8 mmol l�1 and2.5 mmol l�1. These exceptionally high iodide concentrations clearlyindicateapronouncedsedimentary influenceon iodideenrichment indeepwater and suggest that the upwardsdiffusion of free iodide fromthe sediment pore water is facilitated by anoxic conditions in wateroverlying the sediment. The most pronounced gradient in iodideconcentration (1.07 mmol l�1 m�1) is placed at the redox-cline (9e10 m). One may expect that if iodate had not been reduced at thisdepth range for 0.20 mmol l�1 this gradient would be about 20%steeper. Upon the simplest presumption that onlymolecular diffusion

0

200

400

600

800

1000

1200

1400

0 5 10 15 20

c (phosphate)/ mol l-1

c(s

ulf

ide

)/

mo

l l-

1

Aug. 26Sep. 23Sep. 26

Fig. 5. Co-variations between sulfide and phosphate in the Rogoznica Lake in 2003 (left siexplained in the text.

was active during stable stratification, this gradient in iodide con-centration, together with a molecular diffusion coefficient for iodideof 1.72$10�4 m2 day�1 (Robinson and Stokes, 1955), yields to an up-ward iodide flux of �0.20 mmol m�2 day�1. Given that this flux rep-resents onlyabout3%of potentiallyavailable total iodideflux fromthesediment pore water in the anoxic zone (�Zic and Branica, 2006), it isnot surprising that iodide is highlyenriched indeepwater.While tidalforcing most likely induces some lateral advection, the profiles sug-gest that it is less likely that it will unduly affect the vertical exchangeat these depths. The profiles of both iodide and TII share high geo-metric similarity with the profile of silicic acid (Fig. 3a), and in bothcases the Pearson’s correlation coefficient of rP > 0.97 is statisticallysignificant at p ¼ 0.001 (N ¼ 10). However, even if a complex role oforganic-I is put aside, the inter-conversions between iodate and io-dide in this biogeochemically reactive system make total inorganic

0

200

400

600

800

1000

1200

1400

0 5 10 15 20

c (phosphate)/ mol l-1

c(s

ulf

ide

)/

mo

l l-

1

Nov. 2Nov. 12Nov. 22

de graph) and 2004 (right side graph). The regression lines on the left-side graph are

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272 267

iodine to be more appropriate variable for correlation (Fig. 6), asalready suggested by Truesdale (1994b).

On the first day of sampling on September 23rd, iodate concen-trations within the oxic water column (Fig. 3a) and the iodate in-ventory for the pond (Fig. 4) were about 50% lower than those inAugust. This trend continues up to Sep. 26th, so that over the entireperiod iodate and dissolved oxygen (Sec. 4.1.1.) co-varied ratherclosely (rP ¼ 0.89, p ¼ 0.0001, N ¼ 32) (Fig. 6). Meanwhile, iodidedisplays thesame featureasnutrients. Indeepwater (10e13m) iodideconcentrations decreased, while those in surface layers (0e7 m)increased for about two and three times, respectively. Note thatapproximately one-third of the latter figuremay be ascribed to iodatereduction to iodide and not to the mixing process itself. The overallresult is that on Sep. 23rd iodide concentrations, unlike those of nu-trients, were essentially uniform with depth (mean (�s.d.) ¼ 0.9(�0.1) mmol l�1). The discrepancy between iodide and nutrients issimply a result of far greater enrichment of phosphate and silicic acidin deep water relative to surface layers than is the case with iodine.The same applies to total inorganic iodine (mean (�s.d.) ¼ 1.0(�0.1) mmol l�1). Consequently, all the profiles of iodide and totalinorganic iodine in September are comparable and the inventories ofthese variables change only slightly (Fig. 4). The exception is low io-dide concentration at 7 m on Sep. 29th (Fig. 3a). Although that resultseems to deviate from the trend we have decided not to treat it as anoutlier because analysis in triplicate,with variousdilutions, yielded toan essentially the same value of 0.430 (�0.002) mmol l�1. Since adecrease in iodide concentration was not accompanied by anyappreciable increase in iodate concentration, it is hard to offer anyreasonable explanation but the one that involves a third iodine form.

4.2. Depth profiles in November and December 2004

4.2.1. Major constituentsTemperature and salinity profiles in November illustrate well the

appreciable heat lost andmoderate evaporation from a small body ofwater in late season (Figs. 3b and 4b). The combined effect of atemperature decrease (Dtz 6 �C) and an increase in salinity (DSz 2)within the upper half of the water column between Nov. 2nd andNov. 22nd is that on Nov. 22nd the density stratification decreased toonly about 0.1 kg m�3 m�1 between the surface and the bottom. Astep-like salinity structure additionally suggests that double diffusiveprocesses,which are common inmeromictic lakes (von Rohden et al.,2009), could have also been active. The completemixing of thewatercolumn was prevented by a heavy rain that followed shortly after(Sec. 2, Fig. 2), so that on Dec. 20th the surface layers were stratifiedby a pronounced halocline (Fig. 3b). Direct precipitation could haveaffected salinity only for about 30% and the rest of freshwater inputmay be ascribed to run-off, as suggested by �Cosovi�c et al. (2000).

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 50 100 150 200 250

c (silicic acid)/ mol l-1

c(T

II)/µ

mo

l l-

1

Aug. 2003Nov. 2004

Fig. 6. Co-variations between total inorganic iodine and silicic acid (left side graph) an

Over the entire period the top 5 m were slightly over-saturatedwith respect to dissolved oxygen (Fig. 3b). A slow dissolution of adensity gradient in November clearly affected mid-water distribu-tion of dissolved oxygen and the position of redox-cline. Overall,between Nov. 2nd and Nov. 22nd a large volume of water between5 and 10 m became more oxygenated, while the inventory of dis-solved oxygen for the pond increased for about 20% (Fig. 4). It is ofnote that both on Nov. 2nd and Nov. 12th hypoxic samples con-tained appreciable concentration of sulfide (up to 94 mmol l�1).Deep anoxic water contained primarily sulfide (z90%), at con-centration of up to 1260 mmol l�1. Although these concentrationsare comparable to those registered in August 2003, the whole RSSTin the pond was about twice lower (Fig. 4).

4.2.2. NutrientsThe profiles of nutrients in November (Fig. 3b) are comparable

to those obtained a year earlier, before mixing occurred (Fig. 3a).Ammonium, phosphate and silicic acid were depleted in the upperhalf of the water column, but highly enriched in deep anoxic water.Similarly, the phosphate nutricline is placed approximately 2e3 mdeeper than that of silicic acid and the Pearson’s correlation yieldsstatistically significant correlation between phosphate and sulfidein deep anoxic water (rP ¼ 0.956, p ¼ 0.0001, N ¼ 11) (Fig. 5). Be-tween Nov. 2nd and Nov 22nd the concentration of both phosphateand silicic acid increased within the surface layers for about0.1 mmol l�1 and 10 mmol l�1, respectively. This would be in accordwith the general pattern of major variables and slow convectivemixing of the water column. The same should apply to ammoniumtoo, but it appears that nitrification additionally affected nitrogenspeciation in the surface layers. A net nitrate þ nitrite formationrate for this period equals 60 nmol l�1 day�1.

The differences in distribution of nutrients in deep water up toDec. 20th were only minor. Meanwhile, higher ammonium andnitrate concentrations in surface layers (0e2 m) seem to be asso-ciated with unusually high precipitation and appreciable terrestrialrunoff (Sec. 4.2.1.). A simple calculation based on inorganic nitrogenatmospheric deposition (Statistical Yearbooks of the Republic ofCroatia, http://www.dzs.hr/hrv/publication/stat_year.htm) sug-gests that the amount of direct precipitation could have caused anincrease in nitrate and ammonium concentrations within the top2 m for about 3 mmol l�1, but because of the heavy runoff thisexternal supply to the pond was possibly even greater.

4.2.3. IodineWhile salinity changes during the mixing event in September

2003 were rather negligible, during the survey in 2004 bothevaporation and dilution by rainwater affected the salinity of thewater column. This slightly complicates the interpretation of the

0.0

0.2

0.4

0.6

0.8

0 100 200 300 400c (DO)/ mol l

-1

c(io

da

te

)/

mo

l l-

1

2003Nov. 2004

d between iodate and dissolved oxygen (right side graph) in the Rogoznica Lake.

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272268

iodine profiles, when temporal changes are of interest, since iodineconcentrations depend on salinity. To overcome this, as to focusonly to biogeochemical processes, the concentrations were alsonormalised to a salinity of the surrounding seawater (S ¼ 38), bymultiplying the concentration by the ratio of surrounding seawaterto pond water salinities. In the iodine literature the term “ration-alised concentrations” (Truesdale, 1994a) is used more frequently.These concentrations should not be viewed as common those, sincethey represent hypothetical concentrations that would be achievedby either evaporation or dilution of a sample, depending onwhether the initial salinity is lower or higher than 38, respectively.

Depth distributions of iodate and iodide in November 2004were in accord with the redox conditions within the water column(Fig. 3b) and generally match those obtained a year earlier duringthe stratification period (Fig. 3a). Both species were, however,reduced in concentration, so that the molar amount of total inor-ganic iodine in the pond was about 25% lower (Fig. 4). Up to around5 m iodate concentration (z0.30 mmol l�1) was double the con-centration of iodide, while in sulfidic water iodide was the onlydetectable form, present in concentrations of up to 1.58 mmol l�1.Total inorganic iodine varied only little with depth within the top7 m. Although the differences between the samplings were onlyminor, a slight increase in total inorganic iodine concentrationbetween Nov. 2nd and Nov 22nd would be in agreement with thecorresponding changes in density structure and with the profiles ofnutrients (Sec. 4.2.2.). At greater depths the differences weresomewhat more pronounced and principally affected by changes iniodide concentration. The positive redox turnover at chemoclinedoes not seem to have affected the upward iodide flux unduly(Fig. 3b). It appears then that, unlike with phosphate, the impact ofiron or manganese oxyhyroxides at chemocline (Helz et al., 2011)on inorganic iodine scavenging was only minor.

Throughout the water column total inorganic iodine displaysessentially the same feature as in Aug. 2003; a high degree of positivecorrelation with silicic acid (rP ¼ 0.973, p ¼ 0.0001, N ¼ 29) (Fig. 6)(Note that the deepest point on Nov. 12th (Fig. 3b) was omitted fromcalculation, as it clearly deviated from the rest of the pairs.). Mean-while, iodate is closely associated with dissolved oxygen (rP ¼ 0.972,p ¼ 0.0001, N ¼ 20) (Fig. 6). As with dissolved oxygen, the mostprominent feature of iodate distribution between Nov. 2nd and Nov22nd is that iodate concentrations increasedwithin thewater column,particularly at mid-depths (Fig. 3b), largely at the expense of those ofiodide. When transferred to iodate and iodide inventories for theentire pond (calculated from rationalised concentrations), this equalsto a net iodate generation rate of about 65 nmol l�1 month�1 and to anet iodidedepletion rate of�75nmol l�1month�1.When calculationsare performed with common concentrations, the rates are80 nmol l�1 month�1 and �60 nmol l�1 month�1, respectively.

Up to December 20th the concentrations and the distributions ofiodate and iodide changed only little below 7 m, but within thesurface layers both species experienced dilution by rainwater(Fig. 3b). Meanwhile, the rationalised concentrations indicate thatwithin the top 2 m iodate concentration increased, while that ofiodide decreased for about 0.11 mmol l�1 and 0.08 mmol l�1,respectively. Accordingly, the profiles (Fig. 3b) and molar amounts(Fig. 4) of rationalised total inorganic iodine on Nov. 22nd and Dec.20th are highly comparable and suggest that, aside from dilution,there was little gain or loss of total inorganic iodine in the system.

5. Discussion

5.1. A response of major constituents to meteorological conditions

Unlike with freshwater lakes, where temperature is the prin-cipal determinant of density, in this anchialine system the density

structure primarily depends upon the salinity gradient. Therefore,as to achieve conditions needed for overturn of the whole watercolumn this system also requires a particular sequence of meteo-rological events that will shift the balance between evaporationand precipitation toward the former. These conditions were ful-filled over the extremely warm and very dry summer of 2003 (Sec.4.1.1), which represents one of the meteorological extremes formany of the European regions. It is interesting to note, as tocontrast, that the 2003 heat wave prolonged and increased thermalstratification in two freshwater lakes in Switzerland and causedsevere hypolimnetic oxygen depletions (Jankowki et al., 2006). Inthe Rogoznica Lake rapid and complete mixing events withappearance of anoxic conditions in the entire water column wereregistered in 1997 (Bari�c et al., 2003; Ciglene�cki et al., 2005), and2011 (Bura-Naki�c et al., 2012), while similar hypoxic condition to2003 was recorded in November of 2009 (Helz et al., 2011). Allthese events may be related to meteorologically warmer and/ordrier years (Katu�sin, 1998; Croatian Meteorological and Hydrolog-ical Service, http://klima.hr/ocjene_arhiva.html). On all of theseoccasions the surface salinities over spring and summer were highenough to contribute to dissolution of a density gradient in latesummer. Meanwhile, the year 2004 was also warm but relative wet(Sec. 2, Fig. 2), so that the low surface salinities persisted both inspring and summer (�Zic et al., 2010). Slow mixing of the watercolumn therefore beganwith a temporal shift of about two months(Sec. 4) and the complete mixing of the water column was inter-rupted by rain, which is common in this late season. In recentwarmer history, since 1997, this anchialine pond experienced evenfour severe holomitic events. Since these were extremely rarebefore that period, we may well expect that an increase in fre-quency of warm and, perhaps more importantly, dry years wouldcontribute to a shift in mixing regime of this pond frommeromictic(Bari�c et al., 2003; Ciglene�cki et al., 2005, 2011) to monomictic; inan essentially opposite direction to the one predicted for thefreshwater lakes (Adrian et al., 2009).

The principal difference between the mixing events in 1997(Ciglene�cki et al., 2005) and 2011 (Bura-Naki�c et al., 2012) and thosein 2003 (Sec. 4.1.1) and 2009 (Helz et al., 2011) is that in the formercase the entirewater column became anoxic, while in the later oxic/hypoxic conditions were preserved. These differences can beattributed to different ratios between dissolved oxygen concen-trations in mixolimnion and the amount of built-up sulfide inmonimolimnion, and to velocity and intensity of mixing processwhich are influenced by meteorological conditions of the atmo-sphere. While during stratification periods (Fig. 3a and b) we mayexpect that the flux of sulfide from the anoxic zone and its fate atthe chemocline was principally mediated by purple phototrophicsulfide oxidising bacteria from the genus Chromatium (Kamyshnyet al., 2011), presuming the fragility of this bacterial consortium itseems more likely that sulfide oxidation in Sep. 2003 (Fig. 3a) wasdriven abiotically, catalysed by iron and/or manganese (Lutheret al., 2011; Helz et al., 2011).

5.2. Recycled nutrients

The nutrient results are comparable to those reported previ-ously for the pond and add additional argument that high pro-ductivity of this system heavily relies on the availability of therecycled nutrients (Ciglene�cki et al., 2005). Anoxic conditions,which commonly prevail in deep water, facilitate free upwardnutrient diffusion from the sediments (Beutel, 2006; Middelburgand Levin, 2009), but the mixing events may be considered as be-ing principally responsible for the enrichment of surface waters innutrients. A high degree of positive correlation between phosphateand sulfide (Fig. 5) exemplifies that well, since during stratification

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272 269

period the flux of phosphate from the anoxic zone and its avail-ability for the phytoplankton seems to be also affected by phos-phate scavenging onto a metal oxide phase (Scranton et al., 2006).Similarly, the surface nitrate pool in this system appears to beclosely related to biologically mediated process of nitrification,which represents the latest stage of nitrogen regeneration (Ward,2005). High nitrate generation rates of up to 270 nmol l�1 day�1

within the entire water column (Sec. 4.1.2. and 4.2.2.) may beascribed to the enhanced ammonium supply during mixing. Theserates are comparable to nitrification rates reported for the openoceans or oxygen minimum zones (Ward, 2008).

The external nutrient input to the pond is, however, by nomeans negligible, andmaywell explain sustained eutrophication ofthis system. In particular, the annual nitrate and ammonium at-mospheric loading in this area appears to be sufficiently high tocause an increase in inorganic nitrogen concentration within theentire water column for almost 6 mmol l�1 yr�1 (Statistical Year-books of the Republic of Croatia, http://www.dzs.hr/hrv/publication/stat_year.htm). Given that the internal nutrient inputfrom the anoxic zone during mixing events (Fig. 3a) is only about50% higher, this atmospheric loading appears to be rather sub-stantial, particularly if accompanied with heavy runoff. The com-bined effect of high internal loading caused by water columnturnover in 2003 and the above mentioned external loading (Fig. 2)resulted in atypical features within the water column over thespring and summer months of 2004 (�Zic et al., 2010). High con-centrations of nutrients in April 2004, especially those of nitrate (upto 30 mmol l�1), promoted phytoplankton growth, which, in turn,heavily over-saturated the water in the pondwith dissolved oxygen(z250%). The anoxic layer, which is commonly considerably thicker(Ciglene�cki et al., 1998, 2005), wasmissing (�Zic et al., 2010), and thisapparently decreased upward fluxes of nutrients from the sedi-ments (Beutel, 2006; Rigaud et al., 2013). At the same time, both thesuspended biomass of particulate organic carbon in the watercolumn (POC) (z100 mmol l�1), and the settling flux of particulateorganic carbon at 5 and 10 m (z62.5 mmol m�2 day�1) were high(Svensen et al., 2008). When coupled together, these two fluxesresulted in lower nutrient concentrations in November 2004.

5.3. Inorganic iodine

5.3.1. On the constancy of total inorganic iodine during mixing ofthe water column

Important information gained from this study is that, relative tomarked changes in inorganic iodine speciation (viz. iodate reduc-tion to iodide in Sep. 2003 and iodide oxidation to iodate in Nov.2004), total inorganic iodine remained preservedwithin the systemon each occasion (Fig. 4). This suggests fast kinetics of a six-electrontransfer and little organic iodine formation in a potential reactionbetween reactive iodine intermediates (I2 or HIO) and naturalorganic matter (Truesdale, 1974; Truesdale and Luther, 1995;Schlegel et al., 2006; Xu et al., 2013). That is not to say that therole of organic iodine in this eutrophic system is only minor. On thecontrary, the amount of autochthonous or allochthonous organiciodine component and its regeneration in deep water ultimatelycontributes to the inventory of inorganic iodine in the pond. It isjust that in-situ formation of organic iodine in this system domi-nated by photosynthesis shows a seasonal pattern (Stipani�cev andBranica, 1996) and appears to be more closely related to biologicalactivity than to fast chemical reactions that involve reactive iodineintermediates. With an experience gained from subterraneananchialine systems, where organic iodine component is also sub-stantial, but does not seem to affect the inorganic iodine system (�Zicet al., 2011), we may also presume that a portion of organic-I in thisanchialine pond (Stipani�cev and Branica, 1996; �Zic et al., 2010) is of

a terrestrial origin (Schwehr et al., 2005; Schlegel et al., 2006),particularly since higher surface organic-I concentrations are foundto be associated with lower salinities. At the moment we can onlyspeculate upon the exact source of this allochthonous organic-Icomponent, but it may well be that the aerosol organic-I (seeSaiz-Lopez et al., 2012 for a detailed review), emitted from the sea-surface during bubble bursting (Seto and Duce, 1972; Truesdaleet al., 2012), is introduced to the pond over the terrestrialenvironment.

5.3.2. The inventories of inorganic iodine for the pondThe iodine results (Fig. 4) clearly indicate marked differences

between late seasons of 2003 and 2004 with respect to the molaramount of total inorganic iodine in pond water. While higheramounts in 2003 are comparable to those obtained in 1998 and1999, the lower values in 2004 agree well with a set of data from1993/1994, 2000 and 2004 (Stipani�cev and Branica, 1996; �Zic andBranica, 2006; �Zic et al., 2010). These higher amounts are perhapsbest explained by prolonged and enhanced internal iodide loadingfrom the sediment pore waters (�Zic and Branica, 2006), facilitatedby the euxinic conditions in the overlying water (Ciglene�cki et al.,2005), as suggested by Price and Calvert (1978). This would be inaccord with iodine behaviour in some other aquatic environments,where redox conditions in deep water seem to determine deep-water iodide concentrations (Butler et al., 1988; Ullman et al.,1988; Gilfedder et al., 2009; Chapman and Truesdale, 2011). Sincea shift from anoxic to oxic conditions in the overlying water maygreatly reduce the iodine flux from the sediments and pore waters,due to iodate adsorption ontometal oxides (Ullman and Aller,1985;Anschutz et al., 2000), we may expect that, as with nutrients (Sec.5.2.), the lack of a considerably thicker anoxic layer over spring andsummer 2004 (�Zic et al., 2010) largely contributed to a decrease iniodide concentration in deep water in Nov. 2004. This would be inaccord with a response of nutrients to various degrees of oxygen-ation of the water column, as observed elsewhere (e.g. Rigaud et al.,2013). An opposite situation was, for example, registered after theacute anoxic event in 1997 (in 1998 and possibly 1999), whenanoxic and sulfidic conditions were preserved in deep water andcontributed to high inorganic iodine concentrations for a substan-tially longer period (�Zic and Branica, 2006). These results exemplifywell how particular sequences of meteorological events may resultin a highly specific iodine response in semi-isolated aquatic envi-ronments dominated by the sedimentary influence.

Meanwhile, high sediment accumulation rate in this system(0.093 g cm�2 yr�1; z4.5 mm yr�1) (Mihel�ci�c et al., 1996), togetherwith high POC settling flux (z62.5 mmol m�2 day�1) (Svensen et al.,2008), suggest that theverticalfluxofparticulateorganic iodinemightalso be substantial. This would explain the enrichment of iodine inpore waters of the sediments and the iodide flux at the sedimentewater interface (up to �6.1 mmol m�2 day�1) (�Zic and Branica,2006). Although the magnitude of this flux is somewhat lower thanthose reported for estuarine or near-shore sediments (upto�41mmolm�2day�1) (Wong,1991), it is sufficient to affect thedeepwater iodine concentration in rather short period. Given that POC/PON atomic ratio of about 7 in suspended material of the RogoznicaLake (Svensen et al., 2008) meets the Readfield ratio for the marinephytoplankton, we may presume that I/C atomic ratio in particulateorganic matter is comparable to I/C assimilation ratio for oceanic andcoastal seston (z1.2$10�4) (Elderfield and Truesdale, 1980; Chanceet al., 2010). This yields to a downward particulate organic iodineflux of about 7.5 mmol m�2 day�1. While some sorption of iodate andiodide onto solid-phases is feasible, albeit less likely due to slightlybasic conditions (UllmanandAller,1985;Nagata and Fukushi, 2010), aremoval of iodine from the surface waters is probably principallycontrolled by the export of particulate organic iodine.

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272270

5.3.3. Changes in inorganic iodine speciation during water columnmixing

The response of the iodate-iodide redox couple to mixing of thewater column in late season of 2003 was essentially opposite to theone registered in 2004. Thus, while overturn of thewater column inSep. 2003 resulted in iodate reduction to iodide (Sec. 4.1.3), slowmixing in Nov. 2004was accompanied by iodide oxidation to iodate(Sec. 4.2.3.). Interesting though, highly comparable and highlysignificant positive correlations between iodate and dissolved ox-ygen are found for both surveys (Fig. 6), although the processes thatwere operating in the water column and were affecting iodate anddissolved oxygen concentrations in 2003 were apparently mark-edly different to those in 2004.

Since Fig. 6 incorporates both spatial (depth) and temporal scale,the results for 2003 mixing event suggest that the reducing agentsaffected (relatively) iodine and dissolved oxygen speciation in asimilar way. Although a number of potential electron donors (e.g.HS�, Mn2þ, Fe2þ, NHþ

4 ) were re-distributed within the entire watercolumn, only sulfide could have been abundant enough to affect thepool of dissolved oxygen to such an extent (Fig. 4), most likely in atrace metal catalysed process (Konovalov et al., 2003; Luther et al.,2011). The profiles of Mn and Fe in the pond over 2009 (Helz et al.,2011) support this interpretation, as both elements are found topersist in the water column under oxic/hypoxic conditions, at con-centrations of up to 0.1 mmol l�1. These concentrations are suffi-ciently high to directly affect IO�

3 =I� redox couple. For example, at

slightly basic conditions measured here (data not presented) theeventual presence of Mn2þ in the water column may potentiallyreduce iodate (e.g. Truesdale et al., 2001b). Iodate reduction bysulfide is also feasible, and perhaps even more likely, since sulfidereadily reduces iodate to either molecular iodine (Zhang andWhitfield, 1986) or iodide (Luther and Tsamakis, 1989). The resultsof this study suggest that when oxic and anoxic waters are mixed,the extent of iodate reduction to iodide by sulfide would heavilydepend upon the kinetics of sulfide oxidation by oxygen and showthat reduction may not necessarily involve formation of substantialquantities of reactive iodine intermediates and/or organic-I.

Meanwhile, during partial mixing of the water column in Nov.2004 a marked increase in iodate concentration in a large volume ofwater up to around 10 m (Sec. 4.2.3) should have been essentiallyunaffectedbydownward transportofdissolvedoxygen, since iodide iskinetically meta-stable in oxygenated waters (e.g. Wong and Brewer,1977;Wong,1980). Theprocessof iodideoxidation thus requiresmorereactive oxygen species, such as hydrogen peroxide, or biologicalmediation (Luther et al., 1995; Anschutz et al., 2000). By the sameargument stated in a previous paragraph, oxidation of iodide by Mn(III, IV) does not seem likely, since at depth range of interest (0e10m)the pH ranged between 7.5 and 8.1 (data not presented).While iodidemay readily beoxidisedbyMnO2underacidicpH(e.g. Foxet al., 2010),this pH favours iodate reduction by Mn2þ. Thus, any upward Mn2þ

flux from the anoxic zone, wheremanganese attains concentration ofup to around 10 mmol l�1 (Helz et al., 2011), will actually tend todiminish the net-rate of iodide oxidation to iodate.

The results presented here add weight to studies conducted onthis (�Zic and Branica, 2006; �Zic et al., 2010) and other environments(Gilfedder et al., 2009; �Zic et al., 2011, 2012; Yi et al., 2012;Truesdale et al., 2013), which seem to support deduction that io-dide oxidation possibly runs in parallel with nitrification (Truesdaleet al., 2001b). An increase in iodate concentration in December2004 also corroborates this, since the atmospheric inorganic iodinedry- and wet-deposition fluxes in coastal areas of up to9.2 mmol m�2 yr�1 (Baker et al., 2001) are too low to affect theconcentrations within the water column, while the atmosphericammonium deposition flux is substantial enough to promotenitrification (Sec. 5.2.). The reader should be aware that none of the

above mentioned field studies, including this one, provided suffi-cient proof that the organisms that perform iodide oxidation toiodate are the nitrifying organisms. Nevertheless, whatever theprecise control of iodide oxidation to iodate is, this mounting evi-dence that iodide oxidation and nitrogen regeneration are tempo-rally and/or spatially linked is important information in itself forthe marine chemistry of iodine. Since nitrate production in Sep.2003was farmore pronounced than in Nov. 2004 (Fig. 4), this raisesan interesting possibility that the inorganic iodine profiles duringoverturn incorporate a net-effect of simultaneous iodate reductionand iodide oxidation. This is a common issue in field studies, whereopposing reactions may mimic the actual rates of single reactions,particularly in dynamic systems such as Rogoznica Lake. The resultsof this study, combined with those described earlier (�Zic andBranica, 2006; �Zic et al., 2010), yield surprisingly similar net-ratesof iodate formation in the pond for three different surveys(670� 120 nmol l�1 yr�1). Briefly though, other two estimates werecalculated from the amount of iodate formed in the pond on twooccasions; between the acute anoxic event in 1997 and July 1998,and between the hypoxic event in September 2003 and April 2004.Since they span longer periods (9 and 6 months, respectively), theestimate from the survey in November 2004 (780 nmol l�1 yr�1) isthe most reliable one, because it reduces more effectively thepossibility that the exchange of waters between the pond and thesurrounding sea contributed to an increase in iodate concentration.This rate is for about an order of magnitude higher than thoseobserved in a Scottish Loch (Edwards and Truesdale, 1997) or BalticSea (Truesdale et al., 2013), but highly comparable to estimates ofiodide oxidation rate in Atlantic and Pacific gyres (270 nmol l�1 yr�1

and 560 nmol l�1 yr�1, respectively) (Campos et al., 1996).

6. Conclusions

The data presented show how particular sequences of meteo-rological events, some of which can be regarded as extreme, haveaffected the internal processes in the anchialine pond, RogoznicaLake. An integration of these and previously published results re-veals an interesting possibility that this and similar anchialine en-vironments may well potentially serve as valuable sentinels ofclimate change, as has been recognised for the freshwater lakeenvironments (Adrian et al., 2009).

Although this study covers only short periods in late seasons, thenutrient results support previous suggestions that high productiv-ity of this system is fuelled by recycled nutrients (Ciglene�cki et al.,2005). The external nutrient loading appears to be appreciabletoo, and may well explain sustained eutrophication of the pond.

Iodate and iodide results show high reactivity of the inorganiciodine system when it is exposed to a natural stress of verticalmixing between the water compartments of essentially differentredox conditions. The conservative behaviour of total inorganiciodine over each survey suggests that the kinetics of redox trans-formations must have been fast during both iodate reduction andiodide oxidation, as otherwise substantial quantities of organic-Icould have been formed. Chemical reactions with sulfide and pro-cesses related to the last step of nitrogen regeneration are consid-ered to be principal controls of iodate reduction to iodide andiodide oxidation to iodate, respectively. Differences between thesurveys related to total amount of the inorganic iodine in the sys-tem seem to be a result of variable redox conditions in deep waterthat overlies the sediments.

Acknowledgements

The financial support of the Ministry of Science, Education andSports of the Republic of Croatia, under Projects 098-0982934-2720

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272 271

and 098-0982934-2717, as well as by Norwegian Research Foun-dation (NFR) as part of the South-Eastern European Programme:“Ecosystem dynamics, marine chemistry, aquaculture and man-agement in the Adriatic and north-Norwegian coastal zone” isgratefully acknowledged. We thank two anonymous reviewers fortheir constructive suggestions.

References

Adrian, R., O’Reilly, C.M., Zagarese, H., Baines, S.B., Hessen, D.O., Keller, W.,Livingstone, D.M., Sommaruga, R., Straile, D., Van Donk, E., Weyhenmeyer, G.A.,Winder, M., 2009. Lakes as sentinels of climate change. Limnol. Oceanogr. 54,2283e2297.

Amachi, S., 2008. Microbial contribution to global iodine cycling: volatilization,accumulation, reduction, oxidation, and sorption of iodine. Microb. Environ. 23,269e276.

Anschutz, P., Sundby, B., Lefrançois, L., Luther III, G.W., Mucci, A., 2000. Interactionsbetween metal oxides and species of nitrogen and iodine in bioturbated marinesediments. Geochim. Cosmochim. Acta 64, 2751e2763.

Anschutz, P., Dedieu, K., Desmazes, F., Chaillou, G., 2005. Speciation, oxidation state,and reactivity of particulate manganese in marine sediments. Chem. Geol. 218,265e279.

Baker, A.R., Tunnicliffe, C., Jickells, T.D., 2001. Iodine speciation and depositionfluxes from the marine atmosphere. J. Geophys. Res. 196, 28743e28749.

Bari�c, A., Grbec, B., Ku�spili�c, G., Marasovi�c, I., Nin�cevi�c, �Z., Grubeli�c, I., 2003. Massmortality event in a small saline lake (Lake Rogoznica) caused by unusualholomictic conditions. Sci. Mar. 67, 129e141.

Beck, N.G., Bruland, K.W., 2000. Diel biogeochemical cycling in a hyperventilatingshallow estuarine environment. Estuaries 23, 177e187.

Bendschneider, K., Robinson, R.J., 1952. A new spectrophotometric method for thedetermination of nitrite in seawater. J. Mar. Res. 11, 87e96.

Beutel, M.W., 2006. Inhibition of ammonia release from anoxic profundal sedimentsin lakes using hypolimnetic oxygenation. Ecol. Eng. 28, 271e279.

Bluhm, K., Croot, P., Wuttig, K., Lochte, K., 2010. Transformation of iodate to iodidein marine phytoplankton driven by cell senescence. Aqua. Biol. 11, 1e15.

Bluhm, K., Croot, P., Huhn, O., Rohardtand, G., Lochte, K., 2011. Distribution of iodideand iodate in the Atlantic sector of the southern ocean during austral summer.Deep-Sea Res. II 58, 2733e2748.

Bura-Naki�c, E., Helz, G.R., �Cosovi�c, B., Ciglene�cki, I., 2009. Reduced sulfur species in astratified seawater lake (Rogoznica Lake, Croatia); seasonal variations and ev-idence for organic carriers of reactive sulfur. Geochim. Cosmochim. Acta 73,3738e3751.

Bura-Naki�c, E., Margu�s, M., Orli�c, S., Ciglene�cki, I., 2012. Zmajevo oko e a uniqueexample of anchialine system on the Adriatic coast (Croatia) during spring-summer stratification and autumn mixed period. Natura Croatica 21, 17e20.

Butler, E.C.V., Smith, J.D., Fisher, N.S., 1981. Influence of phytoplankton on iodinespeciation in seawater. Limnol. Oceanogr. 26, 382e386.

Butler, E.C.V., Burton, H.R., Smith, J.D., 1988. Iodine distribution in an Antarcticmeromictic saline lake. Hydrobiologia 165, 97e101.

Campos, M.L.A.M., 1997. New approach to evaluating dissolved iodine speciation innatural waters using cathodic stripping voltammetry and a storage study forpreserving iodine species. Mar. Chem. 57, 107e117.

Campos, M.L.A.M., Farrenkopf, A.M., Jickells, T.D., Luther III, G.W., 1996.A comparison of dissolved iodine cycling at the Bermuda Atlantic time-seriesstation and Hawaii ocean time series station. Deep-Sea Res. Part II 43, 455e466.

Chance, R., Malin, G., Jickells, T.D., Baker, A.R., 2007. Reduction of iodate to iodide bycold water diatom cultures. Mar. Chem. 105, 169e180.

Chance, R., Weston, K., Baker, A.R., Hughes, C., Malin, G., Carpenter, L.,Meredith, M.P., Clarke, A., Jickells, T.D., Mann, P., Rossetti, H., 2010. Seasonal andinterannual variation of dissolved iodine speciation at a coastal Antarctic site.Mar. Chem. 118, 171e181.

Chapman, P., Truesdale, V.W., 2011. Preliminary evidence for iodate reduction inbottom waters of the Gulf of Mexico during a hypoxic event. AquaticGeochemistry 17, 671e695.

Ciglene�cki, I., �Cosovi�c, B., 1997. Electrochemical determination of thiosulfate inseawater in the presence of elemental sulfur and sulfide. Electroanalysis 9, 1e7.

Ciglene�cki, I., Kodba, Z., �Cosovi�c, B., 1996. Sulfur species in Rogoznica lake. Mar.Chem. 53, 101e110.

Ciglene�cki, I., Kodba, Z., Vili�ci�c, D., �Cosovi�c, B., 1998. Seasonal variation of anoxicconditions in the Rogoznica lake. Croat. Chem. Acta 71, 217e232.

Ciglene�cki, I., Cari�c, M., Kr�sini�c, F., Vili�ci�c, D., �Cosovi�c, B., 2005. The extinction bysulfide-turnover and recovery of a naturally eutrophic, meromictic seawaterlake. J. Mar. Syst. 56, 29e44.

Ciglene�cki, I., Pichler, S., Prohi�c, E., �Cosovi�c, B., 2006. Distribution of redox-sensitive elements in bottom waters, porewaters and sediments of Rogozn-ica Lake (Croatia) in both oxic and anoxic conditions. Water Air Soil Pollut.Foc. 6, 173e181.

Ciglene�cki, I., Bura-Naki�c, E., Cari�c, M., Kr�sini�c, F., Vili�ci�c, D., Buri�c, Z., �Cosovi�c, B.,2011. Seasonal variation of oxic-anoxic conditions in the small, seawater lake(Rogoznica Lake, East Adriatic Coast) related to the global climate changes. In:Gökçekus, Hüseyin, Türker, Umut, LaMoreaux, James W. (Eds.), Water in theMediterranean Basin, Springer Earth Sciences Book (Environmental Earth

Sciences). Springer-Verlag Berlin and Heidelberg GmbH & Co. K, ISBN3642203345, pp. 230e240.

�Cosovi�c, B., Ciglene�cki, I., Vili�ci�c, D., Ahel, M., 2000. Distribution and seasonalvariability of organic matter in a small eutrophicated salt lake. Estuar. Coast.Shelf Sci. 51, 705e715.

De la Cuesta, J.L., Manley, S.L., 2009. Iodine assimilation bymarine diatoms and otherphytoplankton in nitrate-replete conditions. Limnol. Oceanogr. 54, 1653e1664.

Edwards, A., Truesdale, V.W., 1997. Regeneration of inorganic iodine species in LochEtive, a natural leaky incubator. Estuar. Coast. Shelf Sci. 45, 357e366.

Elderfield, H., Truesdale, V.W., 1980. On the biophilic nature of iodine in seawater.Earth Planet Sci. Lett. 50, 105e114.

Emerson, S., Cranston, R.E., Liss, P.S., 1979. Redox species in a reducing fjord:equilibrium and kinetic considerations. Deep-Sea Res. 26A, 859e878.

Farrenkopf, A.M., Luther III, G.W., Truesdale, V.W., van der Weijden, C.H., 1997. Sub-surface iodide maxima: evidence for biologically catalyzed redox cycling inArabian Sea OMZ during the SW intermonsoon. Deep-Sea Res. II 44, 1391e1409.

Fox, P.M., Kent, D.B., Davis, J.A., 2010. Redox transformations and transport of ce-sium and iodine (�1, 0, þ5) in oxidizing and reducing zones of a sand andgravel aquifer. Environ. Sci. Technol. 44, 1940e1946.

François, R., 1987. The influence of humic substances on the geochemistry of iodinein nearshore and hemipelagic marine sediments. Geochim. Cosmochim. Acta51, 2417e2427.

Fuse, H., Inoue, H., Murakami, K., Takimura, O., Yamaoka, Y., 2003. Production of freeand organic iodine by Roseovarius spp. FEMS Microbiol. Lett. 229, 189e194.

Gozlan, R.S., 1968. Isolation of iodine-producing bacteria from aquaria. Antonie vanLeeuwenhoek 34, 226.

Gilfedder, B.S., Petri, M., Biester, H., 2009. Iodine speciation and cycling in freshwaters: a case study from a humic rich headwater lake (Mummelsee). J. Limnol.68, 396e408.

Helz, G.R., Bura-Naki�c, E., Mikac, N., Ciglene�cki-Ju�si�c, I., 2011. New model for mo-lybdenum behavior in euxinic waters. Chem. Geol. 284, 323e332.

Herring, J.R., Liss, P.S., 1974. A new method for the determination of iodine speciesin seawater. Deep-Sea Res. 21, 777e783.

Ivan�ci�c, I., Degobbis, D., 1984. An optimal manual procedure for ammonia analysisin natural waters by the indophenol blue method. Water Res. 18, 1143e1147.

Jankowski, T., Livingstone, D.M., Heinrich, B., Richard, F., Pius, N., 2006. Conse-quences of the 2003 European heat wave for lake temperature profiles, thermalstability, and hypolimnetic oxygen depletion: implications for a warmer world.Limnol. Oceanogr. 51, 815e819.

Jickells, T.D., Deuser, W.G., Belastock, R.A., 1990. Temporal variations in the con-centrations of some particulate elements in seawaters of the Sargasso Sea andtheir relationship to Deep-Sea fluxes. Mar. Chem. 29, 203e219.

Kamyshny Jr., A., Zerkle, A.L., Mansaray, Z.F., Ciglene�cki, I., Bura-Naki�c, E.,Farquhar, J., Ferdelman, T.G., 2011. Biogeochemical sulfur cycling in the watercolumn of a shallow stratified sea-water lake: speciation and quadruple sulfurisotope composition. Mar. Chem. 127, 144e154.

Katu�sin, Z., 1998. Climate Temperature and Precipitation Anomalies in Croatia for1997. Reviews No. 6. Republic of Croatia, Meteorological and HydrologicalService, Zagreb, p. 26 (in Croatian).

Konovalov, S.K., Luther III, G.W., Friederich, G.E., Nuzzio, D.B., Tebo, B.M.,Murray, J.W., Oguz, T., Glazer, B., Trouwborst, R.E., Clement, B., Murray, K.J.,Romanov, A.S., 2003. Lateral injection of oxygen in highly stratified marinesystems: direct evidence for the Black Sea. Limnol. Oceanogr. 48, 2369e2376.

Luther III, G.W., Campbell, T., 1991. Iodine speciation in the water column of theBlack Sea. Deep-Sea Res. 38, S875eS882.

Luther III, G.W., Tsamakis, E., 1989. Concentration and form of dissolved sulfide inthe oxic water column of the ocean. Mar. Chem. 27, 165e177.

Luther III, G.W., Swartz, C.B., Ullman, W.J., 1988. Direct determination of iodide inseawater by cathodic stripping square wave voltammetry. Anal. Chem. 60,1721e1724.

Luther III, G.W., Church, T.M., Powell, D., 1991. Sulfur speciation and sulfide oxida-tion in the water column of the Black Sea. Deep-Sea Res. 38, 1121e1137.

Luther III, G.W.,Wu, J., Cullen, J.B.,1995. Redoxchemistryof iodine in seawater revisited:frontier molecular orbital theory considerations. Adv. Chem. Ser. 244, 135e155.

Luther III, G.W., Sundby, B., Lewis, B.L., Brendel, P.J., Silverberg, N., 1997. Interactionsof manganese with the nitrogen cycle: alternative pathways to dinitrogen.Geochim. Cosmochim. Acta 61, 4043e4052.

Luther III, G.W., Findlay, A.J., MacDonald, D.J., Owings, S.M., Hanson, T.E.,Beinart, R.A., Girguis, P.R., 2011. Thermodynamics and kinetics of sulfideoxidation by oxygen: a look at inorganically controlled reactions and biologi-cally mediated processes in the environment. Front. Microbiol. 2, 1e9.

Macdonald, R.W., McLaughlin, F.A., 1982. The effect of storage by freezing on dis-solved inorganic phosphate, nitrate and reactive silicate for samples fromcoastal and estuarine waters. Water Res. 16, 95e104.

Macdonald, R.W., McLaughlin, F.A., Wong, C.S., 1986. The storage of reactive silicatesamples by freezing. Limnol. Oceanogr. 31, 1139e1142.

Middelburg, J.J., Levin, L.A., 2009. Coastal hypoxia and sediment biogeochemistry.Biogeosciences 6, 1273e1293.

Mihel�ci�c, G., �Surija, B., Jura�ci�c, M., Bari�si�c, B., Branica, M., 1996. History of accu-mulation of trace metals in sediments of the saline Rogoznica Lake (Croatia).Sci. Tot. Environ. 182, 105e115.

Moisan, T.A., Dunstan, W.M., Udomkit, A., Wong, G.T.F., 1994. The uptake of iodateby marine phytoplankton. J. Phycol. 30, 580e587.

Mullin, J.B., Riley, J.P., 1955. The colorimetric determination of silicate with specialreference to sea and natural waters. Anal. Chim. Acta 12, 162e176.

V. �Zic et al. / Estuarine, Coastal and Shelf Science 133 (2013) 260e272272

Murphy, J., Riley, J.P., 1962. A modified single solution method for the determinationof phosphate in natural waters. Anal. Chim. Acta 27, 31e36.

Nagata, T., Fukushi, K., 2010. Prediction of iodate adsorption and surface speciationon oxides by surface complexation modeling. Geochim. Cosmochim. Acta 74,6000e6013.

Omanovi�c, D., 2006. ECDSOFT e ElectroChemistry Data SOFTware. http://www.irb.hr/en/str/zimo/laboratoriji/lfkt/djelatnici/DarioOmanovic/Software/.

Omanovi�c, D., Branica, M., 1998. Automation of voltammetric measurements bypolarographic analyser PAR 384B. Croat. Chem. Acta 71, 421e433.

Penezi�c, A., Ga�sparovi�c, B., Buri�c, Z., Frka, S., 2010. Distribution of marine lipidclasses in salty Rogoznica Lake (Croatia). Estuar. Coast. Shelf Sci. 86, 625e636.

Pi�zeta, I., Omanovi�c, D., Branica, M., 1999. The influence of data treatment on theinterpretation of experimental results in voltammetry. Anal. Chim. Acta 401,163e172.

Plav�si�c, M., Ciglene�cki, I., Strme�cki, S., Bura-Naki�c, E., 2011. Seasonal distribution oforganic matter and copper under stratified conditions in a karstic, marine,sulfide rich environment (Rogoznica Lake, Croatia). Estuar. Coast. Shelf Sci. 92,277e285.

Price, N.B., Calvert, S.E., 1978. The geochemistry of iodine in oxidized and reducedrecent marine sediments. Geochim. Cosmochim. Acta 37, 2149e2158.

Rigaud, S., Radakovitch, O., Couture, R.M., Deflandre, B., Cossa, D., Garnier, C.,Garnier, J.-M., 2013. Mobility and fluxes of trace elements and nutrients at thesediment-water interface of a lagoon under contrasting water columnoxygenation conditions. Appl. Geochem. 31, 35e51.

Robinson, R.A., Stokes, R.H., 1955. Electrolyte Solutions. Butterworths, London,p. 495.

Rue, E.L., Smith, G.L., Gregory, A.C., Bruland, K.W., 1997. The response of traceelement redox couples to sub-oxic conditions in the water column. Deep-SeaRes. I 44, 113e134.

Saiz-Lopez, A., Plane, J.M.C., Baker, A.R., Carpenter, L.J., von Glasow, R., GómezMartín, J.C., McFiggans, G., Saunders, R.W., 2012. Atmospheric chemistry ofiodine. Chem. Rev. 112, 1773e1804.

Schlegel, M.L., Reiller, P., Mercier-Bion, F., Barré, N., Moulin, V., 2006. Molecularenvironment of iodine in naturally iodinated humic substances: insight from X-ray absorption spectroscopy. Geochim. Cosmochim. Acta 70, 5536e5551.

Schwehr, K.A., Santschi, P.H., Elmore, D., 2005. The dissolved organic iodine speciesof the isotopic ratio of 129I/127I: a novel tool for tracing terrestrial organiccarbon in the estuarine surface waters of Galveston Bay, Texas. Limnol. Oce-anogr. Meth. 3, 326e337.

Scranton, M.I., McIntyre, M., Taylor, G.T., Muller-Karger, F., Fanning, K., Astor, Y.,2006. Temporal variability in the nutrient chemistry of the Cariaco Basin. In:Neretin, L.N. (Ed.), Past and Present Marine Water Column Anoxia, NATO Sci-ence Series: IV, vol. 64. Springer Press, pp. 139e160.

Seto, F.Y.B., Duce, R.A., 1972. A laboratory study of iodine enrichment on atmo-spheric sea-salt particles produced by bubbles. J. Geophys. Res. 77, 5339e5349.

Sillen, L.G., 1961. The physical chemistry of sea water. In: Sears, M. (Ed.), ChemicalOceanography. Publication of Am. Assoc. Adv. Sci, pp. 549e582.

Solorzano, L., 1969. Determination of ammonia in natural waters by phenol hypo-chlorite method. Limnol. Oceanogr. 14, 799e801.

Spokes, L.J., Liss, P.S., 1996. Photo chemically induced redox reactions in seawater: II.Nitrogen and Iodine. Mar. Chem. 54, 1e10.

Stipani�cev, V., Branica, M., 1996. Iodine speciation in the water column or theRogoznica Lake (Eastern Adriatic Coast). Sci. Tot. Environ. 182, 1e9.

Strickland, J.D.H., Parsons, T.R., 1972. A Practical Handbook of Sea Water Analyses,second ed. Fisheries Research Board of Canada, p. 167.

Svensen, C., Wexeles Riser, C., Cetini�c, I., Cari�c, M., 2008. Vertical flux regulation andplankton composition in a simple ecological system: snapshots from the smallmarine Lake Rogoznica (Croatia). Acta Adriatica 49, 37e51.

Truesdale, V.W., 1974. The chemical reduction of molecular iodine in seawater.Deep-Sea Res. I 21, 761e766.

Truesdale, V.W., 1975. ‘Reactive’ and ‘unreactive’ iodine in seawater e a possibleindication of an organically bound iodine fraction. Mar. Chem. 3, 111e119.

Truesdale, V.W., 1994a. Distribution of dissolved iodine in the Irish sea, a temperateshelf sea. Estuar. Coast. Shelf Sci. 38, 435e446.

Truesdale, V.W., 1994b. A re-assessment of Redfield correlations between dissolvediodine and nutrients in oceanic waters, and a strategy for further investigationsof iodine. Mar. Chem. 48, 43e56.

Truesdale, V.W., Upstill-Goddard, R., 2003. Dissolved iodate and total iodine alongthe British east coast. Estuar. Coast. Shelf Sci. 56, 261e270.

Truesdale, V.W., Luther III, G.W., 1995. Molecular iodine reduction by natural andmodel organic substances in seawater. Aqua. Geochem. 1, 89e104.

Truesdale, V.W., Nausch, G., Baker, A., 2001a. The distribution of iodine in the BalticSea during the summer. Mar. Chem. 74, 87e98.

Truesdale, V.W., Watts, S.F., Rendell, A.R., 2001b. On the possibility of iodideoxidation in the near-surface of the Black Sea and its implications to iodine inthe general ocean. Deep-Sea Res. I 48, 2397e2412.

Truesdale, V.W., �Zic, V., Garnier, C., Cukrov, N., 2012. Circumstantial evidence insupport of org-I as a component of the marine aerosol arising from a study ofmarine foams. Estuar. Coast. Shelf Sci. 115, 388e398.

Truesdale, V.W., Nausch, G., Waite, T.J., 2013. The effects of the 2001 Barotropicintrusion of bottom-water upon the vertical distribution of inorganic iodine inthe Gotland Deep. Continent. Shelf Res. 55, 155e167.

Ullman, W.J., Aller, R.C., 1985. The geochemistry of iodine in near-shore carbonatesediments. Geochim. Cosmochim. Acta 49, 967e978.

Ullman, W.J., Luther III, G.W., Aller, R.C., Mackin, J.E., 1988. Dissolved iodinebehaviour in estuaries along the east coast of the United States. Mar. Chem. 25,95e106.

Ullman, W.J., Luther III, G.W., de Lange, G., van der Sloot, H., 1990. Iodine chemistryin deep anoxic basins and overlying waters of the Mediterranean Sea. Mar.Chem. 31, 153e170.

von Rohden, C., Boehrer, B., Ilmberger, J., 2009. Double diffusion in meromictic lakesof the temperate climate zone. Hydrol. Earth Syst. Sci. Discuss. 6, 7483e7501.

Waite, T.J., Truesdale, V.W., 2003. Iodate reduction by Isochrysis galbana is relativelyinsensitive to de-activation of nitrate reductase activity are phytoplanktonreally responsible for iodate reduction in seawater? Mar. Chem. 81, 137e148.

Ward, B.B., 2005. Temporal variability in nitrification rates and related biogeochem-ical factors in Monterey Bay, California, USA. Mar. Ecol. Prog. Ser. 292, 97e109.

Ward, B.B., 2008. Nitrification. In: Capone, D.G., Bronk, D.A., Mulholland, M.R.,Carpenter, E.J. (Eds.), Nitrogen in the Marine Environment. Elsevier, Amsterdam,pp. 199e262.

Wong, G.T.F., 1980. The stability of dissolved inorganic species of iodine in seawater.Mar. Chem. 9, 13e24.

Wong, G.T.F., 1991. The marine geochemistry of iodine. Rev. Aqua. Sci. 4, 45e73.Wong, G.T.F., Brewer, P.G., 1977. The marine chemistry of iodine in anoxic basins.

Geochim. Cosmochim. Acta 41, 151e159.Wong, G.T.F., Cheng, X.-H., 1998. Dissolved organic iodine in marine waters:

determination, occurrence and analytical implications. Mar. Chem. 59,271e281.

Wong, G.T.F., Zhang, L.-S., 1992. Chemical removal of oxygen with sulfite for thepolarographic or voltammetric determination of iodate of iodide in seawater.Mar. Chem. 38, 109e116.

Wong, G.T.F., Brewer, P.G., Spencer, D.W., 1976. The distribution of particulate iodinein the Atlantic Ocean. Earth Planet Sci. Lett. 32, 441e450.

Wong, G.T.F., Piumsomboon, A.U., Dunstan,W.M., 2002. The transformation of iodateto iodide in marine phytoplankton cultures. Mar. Ecol. Prog. Ser. 237, 27e39.

Wood, E.D., Armstrong, F.A.J., Richards, F.A., 1967. Determination of nitrate inseawater by cadmium e copper reduction to nitrite. J. Mar. Biolog. Asso. U.K. 47,23e31.

Xu, C., Chen, H., Sugiyama, Y., Zhang, S., Li, H.-P., Ho, Y.-F., Chuang, C.-y,Schwehr, K.A., Kaplan, D.I., Yeager, C., Roberts, K.A., Hatcher, P.G., Santschi, P.H.,2013. Novel molecular-level evidence of iodine binding to natural organicmatter from Fourier transform ion cyclotron resonance mass spectrometry. Sci.Tot. Environ. 449, 244e252.

Yi, P., Aldahan, A., Possnert, G., Hou, X.L., Hansen, V., Wang, B., 2012. 127I and 129Ispecies and transformation in the Baltic Proper, Kattegat, and Skagerrak basins.Environ. Sci. Technol. 46, 10948e10956.

Zhang, J.-Z., Whitfield, M., 1986. Kinetics of inorganic redox reactions in seawater. I.The reduction of iodate by bisulfide. Mar. Chem. 19, 121e137.

Zheng, J., Yamada, M., Yoshida, S., 2011. Sensitive iodine speciation in seawater bymulti-mode size-exclusion chromatography with sector-field ICP-MS. J. Analyt.Atom. Spectromet. 26, 1790e1795.

�Zic, V., Branica, M., 2006. The distributions of iodate and iodide in Rogoznica Lake(East Adriatic Coast). Estuar. Coast. Shelf Sci. 66, 55e66.

�Zic, V., Cari�c, M., Viollier, E., Ciglene�cki, I., 2010. Intensive sampling of iodine andnutrient speciation in naturally eutrophicated anchialine pond (Rogoznica Lake)during spring and summer seasons. Estuar. Coast. Shelf Sci. 87, 265e274.

�Zic, V., Truesdale, V.W., Cuculi�c, V., Cukrov, N., 2011. Nutrient speciation and hy-drography in two anchialine caves in Croatia: tools to understand iodinespeciation. Hydrobiologia 677, 129e148.

�Zic, V., Truesdale, V.W., Garnier, C., Cukrov, N., 2012. The distribution of iodine in theCroatian Marine Lake, Mir e the missing iodate. Estuar. Coast. Shelf Sci. 115,377e387.