the extinction by sulfide—turnover and recovery of a naturally eutrophic, meromictic seawater lake

www.elsevier.com/locate/jmarsys

Journal of Marine System

The extinction by sulfide—turnover and recovery of a naturally

eutrophic, meromictic seawater lake

I. Cigleneckia,*, M. Caricb, F. Krsinicb, D. Vilicicc, B. Cosovica

aRudjer Boskovic Institute, Center for Marine and Environmental Research, Zagreb, Bijenicka 54, 10001 Zagreb, CroatiabInstitute of Oceanography and Fisheries-Split, Laboratory of Plankton Ecology Dubrovnik, Kneza Damjana Jude 12,

20001 Dubrovnik, CroatiacUniversity of Zagreb, Faculty of Science, Division of Biology, Rooseveltov trg 6, 10001 Zagreb, Croatia

Received 27 June 2003; accepted 20 August 2004

Available online 10 December 2004

Abstract

Since 1994 seasonal variations of temperature, salinity and vertical distribution of dissolved oxygen, nutrients, dissolved

organic carbon (DOC), surface-active substances (SAS), reduced sulfur species (RSS), phyto- and zooplankton have been

investigated in water column of the Rogoznica Lake. During the thermohaline stratification (spring and summer), the surface

water is well-oxygenated while anoxia is occurring in the bottom layer. Anoxic deep water is characterized by high

concentrations of RSS (up to 10�3 M, mainly in the form of sulfide); nutrients (NH4+, up to 150 AM; PO4

3�, up to 22 AM; SiO44�,

up to 400 AM) and DOC (up to 6 mg/l) as a result of the pronounced remineralization of allochthonous organic matter produced

in the surface water. The depth position of anoxic water layer changes seasonally and is greatly influenced by rainfall, the

influence of which is visible from decreased salinity in deeper layers as well. As a result of autumn/winter mixing, bottom water

rich with nutrients is coming to the surface, supporting new phytoplankton and oxygen productions. Natural eutrophication of

the lake is strongly influenced by nutrient recycling under anaerobic conditions.

Turnover of lake water layers in September 1997 occurred so quickly that it resulted in the appearance of total anoxia and

sulfide presence throughout the water column. The aerobic flora and fauna died and added to the oxygen demand and the

production of nutrients. Those long-term datasets help in recognizing the general trend from unique cases, such as the anoxic

event in 1997, which was related to anthropogenic eutrophication.

D 2004 Elsevier B.V. All rights reserved.

Keywords: Karstic seawater lake; Rogoznica Lake; Eutrophication; Meromixis; Anoxia; Reduced sulfur species; Nutrients; Dissolved organic

carbon

Abbreviations: DOC, dissolved organic carbon; DOM, dis-

0924-7963/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.jmarsys.2004.08.006

solved organic matter; POC, particular organic matter; RSS, reduced

sulfur species; RTZ, reduction transition zone; SAS, surface-active

substances.

* Corresponding author.

E-mail address: [email protected] (I. Ciglenecki).

1. Introduction

Hypoxia or anoxia occur throughout the world in

coastal waters that receive high loading of nutrients

s 56 (2005) 29–44

I. Ciglenecki et al. / Journal of Marine Systems 56 (2005) 29–4430

from anthropogenic sources (Rabalais and Nixon,

2002). Besides high nutrient loadings, systems most

susceptible to oxygen depletion are typically char-

acterized by seasonally strong thermohaline strat-

ification, which inhibits vertical mixing of the water

layer rich in oxygen, so that bottom layer oxygen

demand due to organic matter decomposition can

cause bottom hypoxia or even anoxia. Nearly

permanent anoxia in bottom waters is well known

in deep isolated basins such as the deeps in the

Baltic Sea or smaller coastal basins and deep silled

fjords like those in Norway (Jacobs et al., 1985;

Klaveness, 1990; Fallesen et al., 2000; Strom and

Klaveness, 2003; Rasmussen et al., 2003). Gener-

ally, nutrient-stimulated increase in organic matter

load, or in other words eutrophication-related

phenomena, are often blamed for coastal anoxia

(Wu, 2002).

Lakes are natural laboratories in which biogeo-

chemical processes can be effectively studied. They

are much quicker to respond to environmental

pressures than ocean basins, and owing to their

smaller size of reservoirs, the biogeochemical

signals of such perturbations are amplified. In

many lakes, due to their topographic and hydro-

graphic conditions anoxia could exist in the bottom

layer seasonally or permanently (Nurnberg, 1995).

Oxygen or the lack thereof is the driving force for

many processes in lakes. It governs aqueous

chemistry via oxidation and reduction processes

and contributes to biological diversity and abun-

dance by creating different habitats. Many lakes

with anoxic hypolimnia are eutrophic, but some of

them are oligotrophic brown-water lakes.

Apart from temperature and salinity stratifications,

many lakes with anoxic deep waters are chemically

stratified. Meromixis is a condition of persistent

chemical stratification with incomplete mixing over

the course of a year: it usually results in anoxia and

accumulation of nutrients in permanently stagnant

layer (Romero and Melack, 1996). Numerous soft and

saline lakes worldwide are known to be meromictic,

e.g. Mono Lake in California (USA) (Romero and

Melack, 1996); Mahoney Lake in British Columbia

(Canada) (Overmann et al., 1991, 1996); Hall Lake in

Washington (USA) (Balistrieri et al., 1994); Pavin

Lake in France (Alberic et al., 2000) and Lake Cadano

in Switzerland (Schanz et al., 1998).

Rogoznica Lake is a typical example of meromictic

saline lake. The lake is a shallow (15 m deep),

naturally intensely eutrophicated and intermittently

anoxic karstic seawater lake situated on the eastern

coast of the Adriatic Sea (Mihelcic et al., 1996;

Ciglenecki et al., 1996). During the thermohaline

stratification the surface water is well oxygenated

(oxygen saturation up to 300%), while hypoxia/anoxia

occurs in the bottom layer (Stipanicev and Branica

1996; Ciglenecki et al., 1998). Anoxic deep water

(below 10 m depth) is characterized by high concen-

trations of reduced sulfur compounds (RSS up to 1

mM, mainly in the form of sulfide) (Ciglenecki et al.,

1998), iodine species (up to 1 AM) (Stipanicev and

Branica 1996) and nutrients (NH4+, up to 150 AM;

PO43�, up to 22 AM; SiO4

4�, up to 400 AM) (Krsinic et

al., 2000), as well as dissolved organic carbon (DOC

up to 2–3 mg l�1) (Cosovic et al., 2000), indicating

the pronounced remineralization of allochthonous

organic matter in this water layer, produced in the

surface water. The eutrophication of the lake is

strongly influenced by nutrient recycling under

anaerobic conditions.

In September 1997 total anoxia occurred in the

Rogoznica Lake, followed by massive death of

planktonic and benthic organisms (Krsinic et al.,

2000; Baric et al., 2003). Due to the decomposition of

dead organisms, concentrations of nutritious salts

tremendously increased in the entire water column

and were maintained in the lake several months after

the disastrous anoxia, which points out to very

complex and slow lake regeneration processes

(Krsinic et al., 2000; Baric et al., 2003). Due to the

extreme ecological conditions which prevail in this

lake, phyto- and zooplankton populations are repre-

sented by a relatively small number of species, some

of them, however, in the populations denser then

those in the surrounding sea (Krsinic et al., 2000;

Vilicic et al., in press). After the period of total

anoxia, the number of species and their abundance

were significantly reduced in comparison with the

pre-anoxia values. Diatoms were found to be the

dominant microphytoplankton group while copepods,

the heterotrophic zooplankton organisms, play an

important role and control the biochemical processes

in the lake, especially in the post-anoxic period

(Krsinic et al., 2000; Baric et al., 2003; Vilicic et

al., in press).

I. Ciglenecki et al. / Journal of Marine Systems 56 (2005) 29–44 31

Seasonal variations of temperature, salinity and

oxygen, as well as vertical distribution of reduced

sulfur species (RSS), nutrients, dissolved organic

matter (DOM), surface-active substances (SAS), and

phyto- and zooplankton populations have been inves-

tigated in the Rogoznica Lake since 1994. The aim of

this paper is to present the importance of long-term

investigations in studying eutrophication processes of

the Rogoznica Lake. Special attention was paid to the

conditions before and after the disastrous anoxia

occurred in September 1997.

2. Method

2.1. Study site

Rogoznica Lake is a karstic depression filled with

seawater, situated on the eastern Adriatic coast, 40 km

south from Sibenik, Croatia (43832VN 15858VE). Thelake is located on the Gradina peninsula, in the

vicinity of the village of Rogoznica and is nowadays

very close to the nautical center (still under con-

struction) ( Fig. 1). It has an area of 10,276 m2 and a

maximum depth of 15 m. The lake has circular shape

and is surrounded with sheer, karstic cliffs (4–23 m

high), which completely protect the lake from the

influence of the wind (Fig. 1). Detailed bathymetric

Fig. 1. View from the air of the Gradina peninsula indicating the position

(eastern Adriatic coast, Croatia).

features of the lake have been described previously

(Mihelcic et al., 1996; Baric et al., 2003). There is no

visible connection with the surrounding sea but lake

tides are detectable on the cliffs, although with a

certain delay, indicating that an underground water

connection through porous karst exists.

The Gradina peninsula where the lake is situated

is a part of the SW wing of an overturn Upper-

Cretaceous anticline and consists of bulky and

rarely layered Senonian limestone, with dolomite

lenses. The origin of the Rogoznica Lake is related

to the collapse of the roof of a primary underground

cavity, which was formed by corrosive and erosive

actions of underground paleostreams. During the

Holocen, a postglacial rise of the sea level

(Flandrian transgression), this cavity was filled with

seawater passing through the underground passages

and cracks, thus acquiring its present appearance

(Mihelcic et al., 1996).

2.2. Sampling and analyses

Water samples from the Rogoznica Lake were

collected with 5 l Niskin sampling bottles in April

and July 1994, April 1995, and at approximately

monthly intervals between October 1995 and

December 2000, along the vertical profile of the

lake. Water samples for oxygen and sulfide

of the Rogoznica Lake, nautical center and the village of Rogoznica


measurements were always the first to be taken

from the Niskin bottles, immediately after the

sampler was back on board. To preserve anoxic

conditions, Niskin sampler was attached to a N2

cylinder to maintain N2 over the sample while it is

transferred through a tube into glass bottles from

the Niskin sampler. The bottles had been previously

flushed with N2 and filled with samples to over-

flowing. Temperature and salinity were measured in

situ with Hg-thermometer and refractometer (Atago,

Japan).

Samples for sulfur analyses were measured unfil-

tered and freshly, within 24 h.

Samples for surface-active substances (SAS) and

dissolved organic carbon (DOC) measurements were

stored in dark glass bottles. Non-filtered lake water

samples were analysed for SAS within 24 h. DOC

analyses were performed with the samples filtered

immediately after the sampling (Whatman GF/F

filters-pore size 0.7 m), poisoned with HgCl2, and

stored in cold dark place prior to measurements.

Samples for phytoplankton (300 ml) and zooplank-

ton (1000 ml) analyses were preserved in 2% neutral-

ized formaldehyde solution (final concentration).

2.2.1. Chemical analyses

Electroanalytical methods of cathodic stripping

alternating current (CSACV)- (Ciglenecki et al.,

1996) and linear sweep voltammetry (LSV) (Luther

et al., 1985; Ciglenecki and Cosovic, 1997) have been

used for the determination of reduced sulfur species

(RSS) in the lake water. Electrochemical measure-

ments were performed with a Metrohm Polarographic

Analyzer E-506 connected with HMDE (Metrohm) as

the working electrode (for CSACV), and with Autolab

Electrochemical Instruments (Eco Chemie) connected

with 663 VA Stand Metrohm electrode for LSV,

respectively.

The content of surface-active organic matter was

determined by A.C. voltammetry (out-of-phase

mode), with a Metrohm E-506 polarecorder using a

HMDE as the working electrode and was expressed as

the equivalent amount of Triton-X-100, as described

previously (Cosovic and Vojvodic, 1998).

The reference electrode in all electrochemical

measurements was an Ag/AgCl (3M KCl) electrode,

and a platinum wire served as the auxiliary

electrode.

DOC measurements were performed by using a

Shimadzu TOC-500 Analyzer provided with high-

temperature catalytic oxidation.

Nutrient concentrations were measured by standard

methods (Strickland and Parsons, 1972; Ivancic and

Degobbis, 1984).

2.2.2. Phytoplankton and zooplankton counts

Phytoplankton cell counts were obtained by the

inverted microscope method (Utermohl, 1958).

Subsamples of 50 ml were analysed microscopically

after 24 h sedimentation time. Microphytoplankton

cells (MICRO, cells longer than 20 m) were counted

under the magnifications of 400� (1–2 transects)

and 100� (transects along the rest of counting-

chamber base plate).

Zooplankton subsamples were allowed to settle in

the laboratory until reaching 20 ml from the original

volume, the process taking 72 h (Krsinic, 1980). A

glass cell of 7�4.5�0.5 cm was used for counting.

Adult and juvenile stages of copepods were identified

and counted using an Olympus inverted microscope

under magnification of 100�, examining the entire

catch. Copepod faecal pellets in subsamples repre-

senting a quarter of the original sample volume were

counted under magnification of 400�.

3. Results and discussion

3.1. Hydrography and water column stratification

During the investigation period (from 1994 to

2000) variations in vertical gradients of salinity and

temperature as well as those in depth position of

the reduction transition zone (RTZ) have been

observed in the Rogoznica Lake. Seasonal varia-

tions of salinity and temperature in the surface (0–2

m) and bottom layers (12 m depth) of the

Rogoznica Lake are presented in Fig. 2A and B.

As shown in Fig. 2A, salinity values of the surface

layer in the Rogoznica Lake ranged between 24

and 37 psu (depending on meteorological condi-

tions), and were the lowest during winter and

spring months, with the values between 24 and 29

psu. This is not surprising since autumn/early

winter and spring are rainy periods in this part of

the Adriatic coast, with up to 180 mm month�1

Fig. 2. Seasonal variations from 1994 to 2000 of: (A) salinity and (B) temperature at the surface (0–2 m) and bottom layers (12 m depth) of the

Rogoznica Lake. The turnover of the water layers in 1997 is denoted by arrows; individual years were highlighted by the lines drawn vertically

through the figures at each January.


rain (average year rainfall value for this area is

between 700 and 900 mm) (Gajic-Eapka et al.,

2003). It is important to stress that rainfall and

surface runoff are the only sources of freshwater

inflow in the Rogoznica Lake (Ciglenecki et al.,

1998). The influence of freshwater was also visible

from the salinity decrease even in the bottom layer,

especially during the autumn of 1996 when

relatively high average rainfall value of 115 mm

month�1 (the average for September, October and

November) was detected in this area, in comparison

with 53.6 mm month�1 in 1995 and 65.5 mm

month�1 in 1997. Similarly good correlation

between low salinity in the bottom layer and the

levels of rainfall was observed also for the autumn

of 1998, and the spring of 1996, 1997, 2000. In the

mentioned periods relatively low salinity of 34 psu

was detected in the bottom water layer, where it

usually varied from 36 to 38 psu.

The position of halocline during winter and early

spring is situated in surface water (between 2 and 5

m depths), transferring in deeper water during

summer, when it is positioned between 8 and 10 m

depths (Fig. 3).

In parallel with halocline, the lake is permanently

thermally stratified. The position of thermocline

mainly follows the position of halocline, enhancing

the stability of the water column stratification (Fig. 3).

Figs. 2B and 3 show that stable thermocline with

lower temperature in the surface water exists in winter

and spring, while in summer period it intensifies, with

lower temperature observed in the bottom water as

well. These changes occur seasonally, which can be

well noticed from temperature-inverse peaks shown in

Fig. 3. Vertical profiles of temperature, salinity and oxygen saturation in the Rogoznica Lake, in different seasons.



Fig. 2B. The temperature in surface water varied from

9 8C (winter and spring) to 29 8C (summer) while in

bottom water those changes ranged between 9 and

26 8C (Fig. 2B).

The bottom water layer in Rogoznica Lake is

dense, with the values of rt between 25 and 27 kg m-3

and between 18 and 27 kg m-3, respectively (Baric et

al., 2003). During the year the difference in density

between the surface and bottom layers is changing

from 5 kg m�3 (stable stratification) to almost 0 kg

m�3 value (holomictic conditions, such as the turn-

over event in September 1997 (Fig. 4A), when the rt

difference was 0.11 kg m�3 (Baric et al., 2003). The

turnover of the lake in 1997 and appearance of anoxic

conditions in the entire water column is denoted in

Figs. 2 and 8 by the arrows.

The results of this study repeated once more that

meteorological conditions, which include the quantity

of rainfall and cold or mild winter, greatly influenced

the stratification and mixing of the water column in

the Rogoznica Lake (Ciglenecki et al., 1998; Cosovic

et al., 2000). This is in accordance with the fact that

Rogoznica Lake is a typical saline lake, which is lying

Fig. 4. Vertical profiles of: (A) oxygen, temperature and salinity; (B) RSS

1997.

in hydrologically closed basin. It is known that in

such basins the balance between inputs of freshwater

and evaporation from the lake surface determines the

changes in the size, salinity and ultimately the stability

of the water column (Romero and Melack, 1996).

Consequently, water mixing in most of the saline

lakes is likely to be influenced by and result to from

climatic variations (Romero and Melack, 1996).

Temperature-inverse peaks and salinity data shown

in Fig. 2A and B indicate that mixing of the lake water

in the investigation period occurred every year in

autumn as a consequence of lake water-cooling

followed by pycnocline dissolution. It is important

to emphasise that in this period holomictic event with

total anoxic conditions in the entire water column was

noticed only once, in September 1997. This fact

suggests that water mixing in September 1997 was

influenced by the reasons other than water-cooling.

Although there exists no direct evidence, we believe

that the complete lake water turnover and the

disastrous anoxia in 1997 were of anthropogenic

origin. Namely, during spring and summer months in

1997 culminated the series of construction works in

in the Rogoznica Lake during the disastrous anoxia on October 2,


the course of nautical center building (Fig. 1), which

were performed in the vicinity of the lake. Mining and

deep drilling in karstic rocks could probably have

caused vertical vibration of anyway weakly stable

water column (the salinity difference between surface

and bottom waters was 1 g kg�1) and led to quick

overturn of water layers.

3.2. Seasonal variations of anoxic conditions

Depending on the meteorological and thermoha-

line conditions, stratification in the lake is the most

stable during winter and spring, with an inverse

thermocline with warm and sulfide rich bottom water

built up during winter (situation observed in January

1998, Fig. 3). RTZ then transfers upward, and can be

situated between 8 and 9 m depths. RTZ in the

Rogoznica Lake during the rainy spring is charac-

terized by a strong chemocline, when the minimum

difference between the surface and bottom layer

salinity was N4 g kg�1 (Romero and Melack, 1996).

At the chemocline, depending on the presence of a

strong gradient of environmental conditions (light,

oxygen, temperature, H2S), the existence of an

usually 0.1–0.5 m thick, dense pink coloured water

layer, conferred by bacteriochlorophylls of purple

sulfur bacteria could be found. The dominant species

was Chromatium sp. with the concentration up to

3.8�108 cells ml�1, which corresponded to 51% of

total bacteria present (7.4�108 cells ml�1) (Ciglenecki

et al., 1997).

Phytoplankton activity and oxygen production in

the Rogoznica Lake gradually move downward the

position of RTZ during the late spring and summer

period, when it is usually situated at 12 m depth or

even below (see the situation observed in June 1997

and July 1998, presented in Fig. 3).

During summer months, due to surface evaporation

and low local precipitation, the meromictic events in

the Rogoznica Lake, with minimum salinity differ-

ence between the surface and the bottom water

generally up to 2 g kg�1, could be classified as

weakly and moderately stable.

In autumn, due to the lake surface water-cooling

followed by pycnocline dissolution, holomictic con-

ditions are established. Since there is not any physical

and chemical resistance, RSS freely diffuse from the

bottom to the surface and cause oxygen depletion

owing to oxidation. In Fig. 4A and B holomictic

conditions with anoxia and presence of RSS in the

entire water column (detected in the lake in September

1997) were shown. The considerable release of

hydrogen sulfide following the oxygen depletion

made the water look milky yellowish due to the

generation of free sulfur.

Seasonal variations in oxygen and RSS concen-

trations of the surface and bottom waters are presented

in Fig. 5A and B. RSS and oxygen concentrations in

the bottom water layer are inversely proportional and

varied between 10�8 and 10�3 M, and between 0 and

0.34 mM, respectively. During the period of very low

oxygen concentration (below 0.05 mM) in the spring

of 1994 and 1995, and in the spring and summer of

1998 and 1999, the concentration of RSS was very

high (between 10�4 and 10�2 M). Small peaks of

oxygen concentrations recorded in fall and winter of

1995, 1997, 1998, 1999 and 2000 (Fig. 5A) coincide

with the decrease in RSS concentration as well as with

thermohaline conditions (Fig. 2A and B), being a

result of water layers mixing.

Lower, hypoxic values of 0.15 mM (i.e. oxygen

saturation below 50%) in the entire water column

were usually found during autumn mixing (such as in

October 1995, November 1999 and November 2000)

as presented for November 1999 in Fig. 3. The values

equal or even higher than 0.33 mM (i.e. oxygen

saturation between 120% and 150%) were usually

detected in the surface water layer during the

pronounced spring phytoplankton bloom. An excep-

tionally high value of oxygen concentration in the

surface water (0.33 mM) and oxygen saturation of

137% (Fig. 5A) were observed in November 1997, 1

month after the disastrous anoxia occurred, when

owing to the huge release of nutrients, winter

phytoplankton bloom was detected (Krsinic et al.,

2000; Baric et al., 2003).

Extremely high concentration of oxygen (0.58

mM, Fig. 5A) and oxygen saturation of 250% (Fig.

3) were detected in the surface water in February

2000, when the entire water column was oxic.

Namely, as regards the salinity and rainfall data

(Fig. 2A), the winter of 1999/2000 (December,

January, February) and the spring of 2000 (March,

April, May) were characterized by relatively low

levels of rainfall, i.e. 221 mm for those 6 months, in

comparison to 310 mm in the same period of 1998/

Fig. 5. Seasonal variations from 1994 to 2000 of: (A) oxygen and (B) RSS at the surface (0–2 m) and bottom layers (12 m depth) of the

Rogoznica Lake. The turnover of the water layers in 1997 is denoted by arrows; individual years were highlighted by the lines drawn vertically

through the figures at each January.


1999. Such situations resulted in very unstable salinity

stratification in the lake during winter and spring,

which enabled the transfer of oxygen from the surface

to lower layers. Consequently, anoxic conditions with

high concentrations of RSS in the year 2000 were

found only at the bottom (14 m depth).

3.3. Nutrients

Regularly higher concentrations of nutritious salts:

ammonium (between 1.20 and 250 AM), silicate

(between 8.7 and 370 AM) and phosphate (between

0.51 and 22 AM) were detected in the Rogoznica Lake

only at the bottom layer as a result of persistent

chemical stratification with incomplete mixing over the

course of the year, which enables the accumulation of

nutrients and other substances (DOC, RSS, iodine

species) derived from biogeochemical processes occur-

ring in the bottom water layer and sediment. Enhanced

biogeochemical activity such as particulate organic

matter (POC) decay and nutrient demineralisation in

stagnant water layer is a typical characteristic of

meromictic lakes (Romero andMelack, 1996; Hongve,

1997).

After the disastrous anoxia in September 1997,

concentrations of all nutrients increased in the entire

water column, even in the surface layer, and remained

that high several months after the total anoxia due to

mass mortality, decomposition of large amount of POC

and lack of intensive primary production (Krsinic et al.,

2000; Baric et al., 2003). Seasonal variations of

ammonium and nitrate concentrations in the water

column of the Rogoznica Lake, between 1996 and

2000, are presented in Fig. 6. In this presentation, for

Fig. 6. Seasonal variations from 1996 to 2000 of: (A) ammonium and (B) nitrate in the water column of the Rogoznica Lake. The turnover of the

water layers in 1997 is denoted by arrows; individual years were highlighted by lines drawn vertically through the figures at each January. Data

are shown for the water layers of 0–2, 5–9, 10–13 and 14 m depth.


better illustration, water column was divided in four

different layers: surface layer (0–2 m), middle (5–9 m

depth), near-bottom (10–13 m depth) and bottom layer

(14 m depth). As shown in Fig. 6A, distribution of

ammonium concentrations in the lake water column

shows periodic changes during the year (the same is

valid for phosphate and silicate). Decreasing values

were recorded in the bottom layer during the winter,


while increasing concentrations were noticed in the

middle and surface layers at that time. During spring

and summer, the situation was opposite, with increas-

ing concentrations in the bottom and decreasing

concentrations in the middle and surface layers.

Contrary to ammonium as well as phosphate and

silicate, vertical distribution of nitrates in the entire

investigation period was rather uniform, except in

winter time, when relatively high values (up to 19 AM)

were detected in the surface and middle water layers

(Fig. 6B). These observed nitrate peaks correlate very

well with temperature-inverse peaks presented in Fig.

2B, indicating the lake water mixing as well. The

presented nitrate results confirm that water layers

mixing occur every year, in autumn. Namely, during

the autumn, once the uniform density has been

established, the breakdown of meromixis occurs and,

parallel with RSS, a large pulse of nutrients is created as

the nutrient-rich bottom water is mixed throughout the

water column. Simultaneously, oxygen from the sur-

face is transferred to the lower depths.

The observed enrichment in nitrates throughout the

water column results from nitrification processes and

is a major source for a bnewQ primary production

(Wassmann, 1986). Such behaviour is typical for

upwelling areas and highly eutrophic marine–lake

systems, where new organic matter production pre-

vails over regenerated production (Vollenweider et al.,

1996). Besides, new production leads to an increase of

suspended biomass in the euphotic zone while

regenerated production tends to maintain a constant,

or decreasing biomass level.

3.4. Organic matter distribution related to phyto-

plankton and zooplankton activity before and after the

disastrous anoxia

Seasonal variations of DOC and SAS in surface

(0.5–2 m) and bottom water layers (10–13 m depth) of

the Rogoznica Lake in the entire investigation period

are presented in Fig. 7A and B.

The concentration range of DOC between 1 and 5

mg l�1, with an average concentration of 1.82 mg l�1

for the entire investigation period, is typical for

eutrophic and seasonally anoxic marine and fresh-

water systems. Similar ranges of DOC concentrations

have been reported for the water columns of

meromictic and seasonally anoxic lakes, such as Hall

Lake in the USA (Balistrieri et al., 1994); Pavin Lake

in France (Alberic et al., 2000) and Esthwaite Water in

the U.K. (Hamilton-Taylor et al., 1996).

Enhanced concentrations of DOC in the aforemen-

tioned lakes have been detected in anoxic bottom

water as a result of anaerobic decomposition pro-

cesses. In the Rogoznica Lake, as regards DOC

distribution in surface and bottom layers, we could

distinguish the period before and after the disastrous

anoxia. In the time-period before total anoxia, higher

concentrations of DOC were always detected at the

chemocline, in the bottom layer, due both to the

phytoplankton exudates and to the decomposition of

POC (Cosovic et al., 2000). In the post-anoxic time,

vertical distribution of DOC changed seasonally, and

in the summer of 1998 and winter/spring period of

2000, higher values of DOC were detected in the

surface layer (Fig. 7A). In the time before the anoxia

60% of the samples contained DOC concentrations in

the range between 1 and 2 mg l�1, and 40% of the

samples contained those concentrations in the range

between 2 and 3 mg l�1 (Cosovic et al., 2000). In

post-anoxic time these ranges were changed in the

sense of increased number of samples with lower

DOC values, i.e. 75% of the samples contained DOC

values between 1 and 2 mg l�1, pointing to certain

changes in the organic matter composition.

Moreover, similar changes were noticed in SAS

distribution/composition in the post-anoxic time when

higher values of SAS in the lake were detected in the

surface layer (winter/spring and autumn of 1998, Fig.

7B) as distinct from the time before total anoxia, when

systematically higher values of SAS were found in the

bottom layer.

The comparison of normalized surfactant activity

(comparable to Triton-X-100 when normalized to

DOC, Cosovic and Vojvodic, 1998) of the Rogoznica

Lake samples before and after the anoxic event

suggested significant changes, from more hydro-

phobic to the more hydrophilic character of dissolved

organic matter (DOM) in the post-anoxic time.

The distributions of both, SAS and DOC which are

derived only from phytoplankton and bacterial activity

in the Rogoznica Lake, in the time before total anoxia

were very well correlated with enhanced microphyto-

plankton cell-density, oxygen saturation and chloro-

phyll a concentration at the chemocline, in the bottom

waters (10–12 m depth) (Cosovic et al., 2000). The

Fig. 7. Seasonal variations from 1996 to 2000 of: (A) DOC and (B) SAS concentrations at the surface (0–2 m) and bottom layers (10–13 m

depth) of the Rogoznica Lake. The extremely high concentration of SAS (c0.5 mg l�1) was determined in the surface layer in February 2000,

which is marked by open circle and dotted lines. The turnover of the water layers in 1997 is denoted by arrows; individual years were

highlighted by the lines drawn vertically through the figures at each January.


changes in the organic matter distribution (ulti-

mately in its composition as well) in the post-anoxic

time could be probably resulting from grazing-

induced DOM production, which became more

expressed in the lake after the total anoxia (Krsinic

et al., 2000).

Due to the very extreme conditions, which prevail

in the lake, phytoplankton and zooplankton popula-

tions are represented by relatively small number of

species, some of them, however, in very dense

populations (104–106 cell l�1). The microplankton

was composed of about 30 species, among which the

most frequent and the most abundant are micro-

flagellates: Prorocentrum arcuatum and Hermesinum

adriaticum and diatoms: Chaetoceros curvisetus and

Eunotia sp. (Ciglenecki et al., 1998; Vilicic et al.,

1997, in press). Moderate microphytoplankton abun-

dance, with the values larger than 106 cell l�1 in not

more than 32% of 25 case studies (along with

relatively high concentrations of nutrients and organic

matter as well as anoxic conditions) provided

evidence that the Rogoznica Lake is a naturally


eutrophic system (Vilicic et al., in press). The

phytoplankton population in the lake after the anoxic

event was changed in respect of species abundance

and composition. Exceptionally low microphyto-

plankton density (V103–104 cell l�1) was found after

Fig. 8. Seasonal changes from 1997 to 1999 in average densities (average v

copepodites and adult A. italica and microphytoplankton (PHY) and (B) fae

of the water layers is denoted by arrows; individual years were highlighte

the anoxic event in September 1997, when the

dominance of small phytoplankton cells (cell size

2–40 Am) was recorded (Krsinic et al., 2000; Baric et

al., 2003; Vilicic et al., in press). Such phytoplankton

succession was governed by nutrient composition

alues from the surface to the 14 m depth) of: (A) zooplankton (zoo),

cal pellets, in the water column of the Rogoznica Lake. The turnover

d by the lines drawn vertically through the figures at each January.


after the anoxic event. Namely, ammonium, the

dominant nitrogen form in the post-anoxic period

(Fig. 6) favours small phytoplankton cells, while larger

ones are favoured by nitrate (Baric et al., 2003). Four

to six weeks after mass mortality, the biomass of larger

species became larger than the biomass of nano- and

pico-plankton taken together (Baric et al., 2003).

In post-anoxic time zooplankton density in the

Rogoznica Lake has also been completely changed.

Namely, after the anoxia, Acartia italica, endemic

to the Mediterranean Sea and the only planktonic

copepod species in the lake, re-established very

quickly in comparison with phytoplankton, the

recovery of which took longer, almost 1 year. The

population structure of A. italica was completely

different before and after total anoxia (Krsinic et

al., 2000). Before the anoxia, the population

dynamics of the A. italica depended primarily on

predator–prey relationship, i.e. top-down regulation,

while in the post-anoxic period, due to the absence

of predation pressure on naupli or post-naupliar

copepods, nutrient–phytoplankton–copepod relation-

ships, i.e. bottom-up regulation, became more

important.

Relatively good matching was observed in sea-

sonal changes of an average density of copepodites

and adult A. italica, microphytoplankton and faecal

pellets in the water column of the Rogoznica Lake

before and after the disastrous anoxia, as presented in

Fig. 8A and B. It is visible that the extremely high

density of post-naupliar individuals exerting high

grazing pressure, maintained a very low phytoplank-

ton density through winter/spring and summer/autumn

seasons of 1998 (Fig. 8A and B), which reflected

directly on the oxygen production and spreading of

anoxic zone from the bottom to the middle layers. At

the same time, as a direct consequence of grazing

process, there were recorded increases first in SAS

concentrations and further on in DOC concentrations,

coinciding with the increase of faecal pellets density,

which are known to be a significant source of DOM

(Carlson, 2002) (Figs. 7 and 8B).

Due to a huge release of nutrients after the lake

water mixing, 1 month after the disastrous anoxia

(November 1997), intensive phytoplankton activity of

mixotrophic dinoflagellates, reflecting on the oxygen

and RSS concentrations, was detected (Figs. 5A,B and

8A). Hence, the increased surface SAS concentra-

tions, detected during winter/spring season of 1998

(Fig. 7B), are directly related to both, the increased

phytoplankton activity detected in November 1997

and the high grazing pressure.

According to our results, the surface DOC

maximum, which followed after SAS maximum,

showed very good matching with faecal pellets

density (see the situation in June 1998, when DOC

concentrations were increased in the entire water

column, Figs. 7A,B and 8B).

It is important to notice that extremely high SAS

concentrations (z0.5 mg l�1) were determined in the

surface layer of the Rogoznica Lake in February 2000.

Since these concentrations were very close to the

upper detection limit of the method, and for better

presentation of Fig. 7B, they have only been marked

in the figure (open circle and dotted lines) and the

value for filtered sample (0.137 mg l�1) was included

instead. This SAS peak matched very well with the

surface DOC maximum concentration detected in

February 2000 (Fig. 7A) and they are both closely

related to the extremely high photosynthetic activity

of nanoplanktonic cells, already shown in Fig. 3 as an

increase in oxygen saturation. Similarly pronounced

activity of nanoplanktonic green algae and picoplank-

tonic cyanobacteria, followed by an increased value of

oxygen saturation (300%), as well as those of DOC

and SAS, has been already detected in the bottom

water layer of the Rogoznica Lake, in the spring of

1996 (Cosovic et al., 2000).

The results obtained here are in accordance with the

studies showing that zooplankton grazers within

planktonic assemblages can affect the concentration

and the composition of DOM directly by releasing

DOM, and indirectly, via their predatory activities on

the bacteria and microalgae which can take up or

release DOM (Carlson, 2002; Kujawinski et al., 2002).

4. Conclusions

This study provides long-term datasets on the bio-

geochemistry of a meromictic, seawater lake, showing

how an interplay and balance between physical,

chemical and biological processes can influence

spatial distributions of organic and inorganic com-

pounds in one relatively isolated marine environment

such as the Rogoznica Lake.


Those long-term datasets show that holomictic

conditions in the Rogoznica Lake accompanied with

the appearance of total anoxia throughout the water

column in September 1997, followed by subsequent

changes in flora and fauna populations as well as DOM

distribution and their likely composition, were unusual

and of anthropogenic origin. Namely, mining and deep

drilling in karstic rocks as a part of construction works,

which took place in the vicinity of the lake in the course

of nautical center building, most probably caused

vertical vibration of water column and led to the quick

overturn of water layers. Due to its very peculiar

ecological characteristics, unique for the Adriatic as

well as for the Mediterranean area, as a natural habitat

for the endemic A. italica, this lake deserves to be

included in one of the protected natural environment

categories.

Owing to all the presented characteristics the

Rogoznica Lake can be regarded as a btest tubeQ forhypoxia/anoxia that might occur in other coastal

marine environments as a result of current nutrient

pollution trends, which are nowadays considered to be

the most pressing water pollution problem in the world.

Acknowledgements

The authors thank B. Iljadica, Z. Roman and Z.

Zovko for their technical assistance in the field work

and samples collection. Valuable suggestions and

comments of anonymous reviewers that improved

the manuscript are gratefully acknowledged.

The study was financially supported by the

Ministry of Science and Technology of the Republic

of Croatia, grant nos. 098122, 119144 and through

stimulating project for young researchers, no. 098445.

References

Alberic, P., Viollier, E., Jezequel, D., Grosbois, C., Michard, G.,

2000. Interaction between trace elements and dissolved organic

matter in the stagnant anoxic deep layer of a meromictic lake.

Limnology and Oceanography 45 (5), 1088–1096.

Balistrieri, L.S., Murray, J.W., Paul, B., 1994. The geochemical

cycling in a biogenic meromictic lake. Geochimica et Cosmo-

chimica Acta 58 (19), 3993–4008.

Baric, A., Grbec, B., Kuspilic, G., Marasovic, I., Nincevic, Z.,

Grubelic, I., 2003. Mass mortality event in a small saline lake

(Lake Rogoznica) caused by unusual holomictic conditions.

Scientia Marina 67 (2), 129–141.

Carlson, C.A., 2002. Production and removal processes. In: Hansell,

D.A., Carlson, C.A. (Eds.), Biogeochemistry of Marine Dis-

solved Organic Matter. Academic Press, pp. 91–153.

Ciglenecki, I., Cosovic, B., 1997. Electrochemical determination of

thiosulfate in seawater in the presence of elemental sulfur and

sulfide. Electroanalysis 9 (10), 1–7.

Ciglenecki, I., Kodba, Z., Cosovic, B., 1996. Sulfur species in

Rogoznica lake. Marine Chemistry 53, 101–111.

Ciglenecki, I., Kodba, Z., Vilicic, D., Caric, M., Baldi, F., Minacci,

A., Cosovic, B., 1997. Characterization of a Boundary Between

Oxic and Anoxic Conditions in a Small Sea–Lake, Rogoznica

Lake (Croatia). European Research Conferences on Natural

Waters and Water Technology. San Feliu de Guixols-Girona,

Spain.

Ciglenecki, I., Kodba, Z., Vilicic, D., Cosovic, B., 1998. Seasonal

variation of anoxic conditions in the Rogoznica Lake. Croatica

Chemica Acta 71 (2), 217–232.

Cosovic, B., Vojvodic, V., 1998. Voltammetric analysis of

surface-active substances in natural seawater. Electroanalysis

10, 429–435.

Cosovic, B., Ciglenecki, I., Vilicic, D., Ahel, M., 2000.

Distribution and seasonal variability of organic matter in a

small eutrophicated salt lake. Estuarine, Coastal and Shelf

Science 51, 705–715.

Fallesen, G., Andersen, F., Larsen, B., 2000. Life, death and revival

of the hypertrophic Mariager Fjord, Denmark. Journal of Marine

Systems 25, 313–321.

Gajic-Capka, M., Patarcic, M., Percec Tadic, M., Srnec, L.,

Zaninovic, K., 2003. Mean annual air temperature and

precipitation spatial distribution in Croatia. 3rd Croatian

Conference on Waters.

Hamilton-Taylor, J., Davison, W., Morfett, K., 1996. The bio-

geochemical cycling of Zn, Cu, Fe, Mn, and dissolved organic C

in seasonally anoxic lake. Limnology and Oceanography 41 (3),

408–418.

Hongve, D., 1997. Cycling of iron, manganese, and phosphate

in a meromictic lake. Limnology and Oceanography 42 (4),

635–647.

Ivancic, I., Degobbis, D., 1984. An optimal manual procedure for

ammonia analysis in natural waters by the indophenol blue

method. Water Research 18, 1143–1147.

Jacobs, L., Emerson, S., Skei, J., 1985. Partitioning and transport of

metals across the O2/H2S interface in a permanently anoxic

basin: Framvaren Fjord, Norway. Geochimica et Cosmochimica

Acta 49, 1433–1444.

Klaveness, D., 1990. Size structure and potential food value of the

plankton community to Ostrea edulis L. in a traditional

Norwegian basterspollQ. Aquaculture 86, 231–247.

Krsinic, F., 1980. Comparison of methods used in micro-zooplank-

ton research in neritic waters of the Eastern Adriatic. Nova

Thalassia 4, 91–106.

Krsinic, F., Caric, M., Vilicic, D., Ciglenecki, I., 2000. The calanoid

copepod Acartia italica Steuer, phenomenon in the small saline

Lake Rogoznica (Eastern Adriatic coast). Journal of Plankton

Research 22 (8), 1441–1464.


Kujawinski, E.B., Farrington, J.W., Moffett, J.W., 2002. Evidence

for grazing-mediated production of dissolved surface-active

material by marine protists. Marine Chemistry 77, 133–142.

Luther III, G.W., Giblin, A.E., Varsolona, R., 1985. Polarographic

analysis of sulfur species in marine porewaters. Limnology and

Oceanography 30 (4), 727–736.

Mihelcic, G., Surija, B., Juracic, M., Branica, B., 1996. History of

accumulation of trace metals in sediments of the Rogoznica lake

(Croatia). Science of the Total Environment 182, 105–115.

Nqrnberg, G.K., 1995. Quantifying anoxia in lakes. Limnology and

Oceanography 40, 1100–1111.

Overmann, J., Beatty, J.T., Krouse, H.R., Pfenning, N., Northcote,

T.G., 1991. Characterization of a dense, purple sulfur bacterial

layer in a meromictic salt lake. Limnology and Oceanography

36, 846–859.

Overmann, J., Beatty, J.T., Krouse, H.R., Hall, K.J., 1996. The

sulfur cycle in the chemocline of a meromictic salt lake.

Limnology and Oceanography 41 (1), 147–156.

Rabalais, N.N., Nixon, S.W., 2002. Nutrient over-enrichment in

coastal waters: global patterns of cause and effect (Special

Issue). Estuaries 25.

Rasmussen, B., Gustafsson, B.G., Stockenberg, A., Ertebjerg, G.,

2003. Nutrient loads, advection and turnover at the entrance to

the Baltic Sea. Journal of Marine Systems 39, 43–56.

Romero, J.R., Melack, J.M., 1996. Sensitivity of vertical mixing in

a large saline lake to variations in runoff. Limnology and

Oceanography 41 (5), 955–965.

Strfm, T.-E., Klaveness, D., 2003. Hunnebotn: a seawater basin

transformed by natural and anthropogenic processes. Estuarine,

Coastal and Shelf Science 56, 1177–1185.

Schanz, F., Fischer-Romero, C., Bachofen, R., 1998. Photosynthetic

production and photoadaptation of phototrophic sulfur bacteria

in Lake Cadgano (Switzerland). Limnology and Oceanography

43 (6), 1262–1269.

Stipanicev, V., Branica, M., 1996. Iodine speciation in the water

column of the Rogoznica Lake (Eastern Adriatic Coast). Science

of the Total Environment 182, 1–9.

Strickland, J.D.H., Parsons, T.R., 1972. A practical handbook of

seawater analysis. Bulletin of the Fisheries Research Board of

Canada 167, 1–310.

Utermfhl, H., 1958. Zur Vervollkommung der quantitativen

Phytoplankton Methodik. Mitteilungen-Internationale Vereini-

gung fqr Theoretische und Angewandte Limnologie 9, 1–38.

Vilicic, D., Marasovic, I., Kuspilic, G., 1997. The heterotrophic

ebridian microflagellate Hermesinum Adriaticum Zach. In the

Adriatic Sea. Archiv fur Protistenkunde 147, 373–379.

Vilicic, D., Cetinic, I., Ciglenecki, I., Cosovic, B., Caric, M., Buric,

Z. Seasonality of abundant microflagellates Prorocentrum

arcuatum Issel and Hermesinum adriaticum Zacharias in the

marine lake Rogoznica. Journal of Eukaryotic Microbiology, in

press.

Vollenweider, R.A., Rinaldi, A., Viviani, R., Todini, E., 1996.

Assessment of the state of eutrophication in the Mediterranean

Sea. MAP Technical Reports Series, vol. 106. UNEP, Athens.

455 pp.

Wassmann, P., 1986. Benthic nutrient regeneration as related to

primary productivity in the west-Norwegian coastal zone.

Ophelia 26, 443–456.

Wu, R.S.S., 2002. Hypoxia: from molecular responses to ecosystem

responses. Marine Pollution Bulletin 45, 35–45.

the extinction by sulfide—turnover and recovery of a naturally eutrophic, meromictic seawater lake

Documents