Calcium, magnesium and strontium cycling in stratified, hardwater lakes: Lake Kinneret (Sea of Galilee), Israel

Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

Geochimica et Cosmochimica Acta 105 (2013) 372–394

Calcium, magnesium and strontium cycling in stratified,hardwater lakes: Lake Kinneret (Sea of Galilee), Israel

Amitai Katz a,⇑, Ami Nishri b

a The Fredy and Nadine Herrmann Institute of Earth Sciences, The Hebrew University of Jerusalem, The Edmond Safra Campus,

Givat-Ram, Jerusalem 91904, Israelb Israel Oceanographic and Limnological Research, Yigal Allon Kinneret Limnological Laboratory, P.O. Box 447, Migdal 14950, Israel

Received 5 April 2011; accepted in revised form 15 November 2012; available online 27 December 2012

Abstract

In stratified hardwater lakes, calcite crystallization in the epilimnion and partial dissolution in the hypolimnion play impor-tant roles in the cycling of Ca2+ in the water column. Mg2+ and Sr2+ coprecipitate with this mineral, to be released togetherwith Ca2+ upon its dissolution. Here, we focus on Lake Kinneret (Sea of Galilee, Israel), a stratified, warm, hardwater lake,searching for an interpretable, high resolution picture of the processes that drive Ca2+, Mg2+ and Sr2+cycling in this and sim-ilar lakes elsewhere.

In total, 1428 water samples were collected from the lake, and another 81 samples were collected from the Jordan Riverand two small streams discharging into it, covering a full monomictic cycle from December 2001 through March 2003. Par-ticulate material was retrieved from sediment traps in the hypolimnion. The water samples were analyzed for Na+, K+, Mg2+,Ca2+, Sr2+, Cl�, SO4

2�, HCO3� and Br�, and the trapped particulates were examined under SEM and analyzed for their Ca,

Mg, and Sr content. Individual calcite crystals within the particulate material underwent LA-ICP-MS analysis for Ca, Mg,and Sr.

Downward transport of Ca2+ involves crystallization of calcite in the epilimnion, followed by its settling through the watercolumn, followed by its partial dissolution in the hypolimnion. Magnesium and Sr2+ follow Ca2+ by coprecipitation in calciteand are released upon its dissolution in the hypolimnion. Upward cycling of the three solutes occurs through an admixture ofthe hypolimnetic water into the epilimnion during thermocline deepening, terminated by overturn of the lake.

Removal rates of Mg2+ and Sr2+ from the lake, relative to that of Ca2+, and the water–calcite distribution coefficients, DMg

and DSr, were calculated from the analyses using: (1) epilimnion water and ‘bulk’ particulates; (2) epilimnion water only; and(3) epilimnion water and calcite crystals in the particulates. The DSr values obtained were internally consistent (0.194 ± 5.9%,0.22–0.28, and 0.204–0.232, respectively); however, the DMg value for the ‘bulk’ particulates (0.0477) was �6 times higher thanthat corresponding to calcite crystals included therein (0.00841). This difference is attributed to phytoplankton debris in the‘bulk’ particulates and should be of concern to geochemists using Mg/Ca ratios in limestone to reconstruct ancient aquaticenvironments.

The cycling of Ca2+, Mg2+ and Sr2+ in the stratified lake is monitored by Mg/Ca vs. Sr/Ca regression diagrams, whereeach line represents the water column composition on a specific date. The distance of the data points from the initial(mixed lake) coordinates reflects the fractionation of Ca2+, Mg2+ and Sr2+ after the onset of stratification. The regressionlines rotate in an orderly anticlockwise direction in response to the high calcite flux from the epilimnion in spring and thehigher rate of Mg2+ replenishment than that of Sr2+, compensating for their loss in calcite. The release of Mg2+ and Sr2+

from dissolving calcite in the hypolimnion, at an Mg/Sr ratio (7.5 eq/eq) much lower than that in the surrounding water(�160 eq/eq), lends additional support to the anticlockwise rotation. Reversal of rotation occurs in summer, when calcite

0016-7037/$ - see front matter � 2012 Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.gca.2012.11.045

⇑ Corresponding author. Tel.: +972 2 658 4620 (O)/+972 2 6791251 (H).

E-mail address: [email protected] (A. Katz).


mailto:[email protected]


A. Katz, A. Nishri / Geochimica et Cosmochimica Acta 105 (2013) 372–394 373

crystallization and the freshwater supply slow down, and brackish and saline water sources take control over the Mg/Srratio in the lake. Subsequent turnover and mixing of the lake in winter reset the Mg/Ca and Sr/Ca ratios to their originalvalues before the next stratification.� 2012 Elsevier Ltd. All rights reserved.

1 Abbreviations: LC, L. Constance; LO, L. Ontario; LL, L.Lugano; LS, L. Sempach; LH, L. Hagelseewli; R, Review onlacustrine calcite precipitation; LLU, L. Lucern.

1. INTRODUCTION

1.1. Overview

Stratified lakes are very common on Earth’s surface.They consist of three horizontal layers: an epilimnion,underlying the lake surface, a mesolimnion below it and ahypolimnion, overlying the lake’s bottom. Stratification iscaused by density differences between the epilimnion andhypolimnion waters; the intermediate mesolimnion displaysgradients in chemical composition and physical propertiesbetween these layers.

The processes responsible for these density differencesinclude seasonal warming or cooling of the lake surface,input of dense, saline water into fresher water layersor vice versa, crystallization of minerals in surface waterand their dissolution in deeper waters (Boehrer and Schu-ltze, 2008).

Stratified lakes are useful for tracing solute behavior innatural aqueous environments and provide vital informa-tion for the interpretation of ancient lacustrine environ-ments. Moreover, understanding their geochemistry isdirectly relevant to tackling the growing shortage of potablewater. The following facts should be noted:

(a) Compartmentalization of a lake by stratificationoffers for study three coexisting environments thatdiffer in their physical and chemical properties andin the factors that caused these differences.

(b) The ubiquity of stratified lakes permits the study ofspecific questions under very different climatic andgeomorphic conditions. Examples: Lake Baikal(Boehrer et al., 2000); Lake Vanda and several otherlakes in the McMurdo valley, Antarctica (Armitageand House, 1962); the Dead Sea and nearby evapora-tion ponds in Jordan and Israel (Katz and Starinsky,2009); Lake Constance in Switzerland/Germany(Hohman et al., 1997).

(c) The frequency of stratification events varies greatlybetween lakes, enabling a study of the processes thatoperate within them at different time scales. Mero-

mictic lakes have a stable configuration for periodslonger than 1 year. Polymictic lakes mix episodically,twice or more each year, while monomictic lakes mixonce a year closely following seasonal climaticperiodicities.

(d) The development of anoxic conditions is very com-mon in the hypolimnia of stratified lakes, impedingutilization of large fractions of their volume. Theseverity of this phenomenon may be recognizedrecalling that, in monomictic lakes, the timing ofstratification and the growing water demand are cou-pled by the season.

Detailed descriptions, modeling and interpretation ofthe dynamics, biology and geochemistry of stratified lakesare provided in the relevant, voluminous literature (e.g.,Hutchinson’s (1957) and Boehrer and Schultze’s (2008)comprehensive reviews).

Stratification of hardwater lakes exerts a particular effecton the cycling of Ca2+, Mg2+ and Sr2+ due to crystalliza-tion of calcite in the epilimnion and its dissolution in thehypolimnion. Being a major ion in such lakes and a majorcomponent of calcite, the effect on Ca2+ cycling is readilyunderstandable. Coprecipitation with the same mineralmay admit other alkaline earth solutes into the same track,specifically Mg2+ and Sr2+, thereby explaining their devia-tion from conservative behavior.

The present paper summarizes a high resolution study ofthe Ca2+, Mg2+ and Sr2+ behavior in the fresh, stratified,hardwater Lake Kinneret (Sea of Galilee, Israel) and is fo-cused on an evaluation of the role of crystallization, disso-lution and coprecipitation in their cycling.

1.2. Objectives

The objective of this study is to obtain a finely-resolvedpicture of the cycling of Ca2+, Mg2+ and Sr2+ in Lake Kin-neret, as an example of a subtropical, monomictic, hardwater

lake, during a full mixing-stratification cycle and to proposean explanation of the general applicability of the findings.

1.3. Approach

The crystallization and dissolution of calcite have beenrepeatedly demonstrated to play important roles in the cy-cling of Ca2+, a major solute in stratified hardwater lakes(e.g., Stabel, 1986 [LC]; Stabel et al., 1986 [LC]; Stabel,1989 [LC]; Kuchler-Krischun and Kleiner, 1990 [LC]; Ho-dell et al., 1998 [LO]; Ramisch et al., 1999 [LL and LS];Ohlendorf and Sturm, 2001 [LH]; Muller et al., 2003 [LS];Dittrich and Obst, 2004 [R]; Dittrich et al., 2004 [LLU];Muller et al., 2006 [LLU and LS]).1 Most studies, however,have focused on the behavior of calcite at or close to thesediment–water interface and on the estimation of the cal-cium and carbonate fluxes involved. With few exceptions(e.g., Muller, 1969; Stabel, 1986, 1989; Stabel et al., 1986),less attention has been directed to the complementary partof the Ca cycle in the water column and to its effect on thelinked Sr2+ and Mg2+ cycles that are known to coprecipi-tate with calcite (e.g., Morse and Bender, 1990 and refer-ences therein).

https://www.researchgate.net/publication/239356220_A_Treatise_on_Limnology_I?el=1_x_8&enrichId=rgreq-0632235d-1d8a-4c0e-84ec-1104590fa2cf&enrichSource=Y292ZXJQYWdlOzI1NjcxNTkzMDtBUzoxNTU3NTg2MTYzODc1ODRAMTQxNDE0NzE1NjEzMw==

https://www.researchgate.net/publication/226618508_High_strontium_contents_and_SrCa-ratios_in_Lake_Constance_waters_and_carbonates_and_their_sources_in_the_drainage_area_of_the_Rhine_River_Alpenrhein?el=1_x_8&enrichId=rgreq-0632235d-1d8a-4c0e-84ec-1104590fa2cf&enrichSource=Y292ZXJQYWdlOzI1NjcxNTkzMDtBUzoxNTU3NTg2MTYzODc1ODRAMTQxNDE0NzE1NjEzMw==

https://www.researchgate.net/publication/240439041_Similarities_and_Differences_in_the_Annual_Temperature_Cycles_of_East_German_Mining_Lakes?el=1_x_8&enrichId=rgreq-0632235d-1d8a-4c0e-84ec-1104590fa2cf&enrichSource=Y292ZXJQYWdlOzI1NjcxNTkzMDtBUzoxNTU3NTg2MTYzODc1ODRAMTQxNDE0NzE1NjEzMw==

https://www.researchgate.net/publication/30018517_Processes_of_deep-water_renewal_in_Lake_Baikal?el=1_x_8&enrichId=rgreq-0632235d-1d8a-4c0e-84ec-1104590fa2cf&enrichSource=Y292ZXJQYWdlOzI1NjcxNTkzMDtBUzoxNTU3NTg2MTYzODc1ODRAMTQxNDE0NzE1NjEzMw==

374 A. Katz, A. Nishri / Geochimica et Cosmochimica Acta 105 (2013) 372–394

Here, we attempt to bridge this gap by providing a suf-ficiently detailed database required to describe and explainthe relationship between Ca2+, Mg2+ and Sr2+ in the watercolumn and the response of these solutes to lake dynamics.

The (vertical) chemical heterogeneity of stratified lakesand the dependence of their dynamics on season dictatedthe deployment of a high-resolution sampling programalong a complete mixing-stratification cycle, which fre-quently required identification and quantization of very

small, temporal or spatial changes in the measured param-eters. Earlier work demonstrated that waters collected con-temporaneously (i.e., within a few days) during thestratified period at different locations in Lake Kinneret dis-play remarkable lateral homogeneity (Bar Giora-Magal,2001). This finding allowed frequent sampling of one site,located at the deepest point of the lake, at a much higherresolution.

Lake Kinneret (henceforth LK, Fig. 1), a hardwater,stratified monomictic water body, has been thoroughlyand continuously investigated since the middle of the lastcentury, providing a long and comprehensive geochemical,biological and limnological record. The size of the lake, itsgeometry and local climate sustain its stable monomicticbehavior, while its bathymetry permits high-resolution sam-pling even at maximum depth (�40 m).

Magnesium and Sr2+ were neither reported to formautochthonous minerals in LK nor in its sediments, makingtheir uptake during crystallization of calcite, and releaseupon dissolution important processes in their cycling.

Fig. 1. Location map of the study area. Star symbols mark sampling sitYa = Yahudiya tributary, Me = Meshushim tributary, and JR = Jordan(For interpretation of the references to color in this figure legend, the re

Knowledge of the effective distribution coefficients of Sr2+

and Mg2+ (DSr and DMg) between calcite and LK water

therefore became crucial for evaluating the fate of these sol-utes in the water column, the more so considering the largevariation in the DSr and DMg values reported in the litera-ture. Moreover, comparison of the DSr and DMg valueswas expected to allow testing of the proposal that a non-trivial fraction of ‘coprecipitated’ Mg2+ actually resideswithin organic debris rather than in the calcite crystal struc-ture proper (Stoll et al., 2001; Blanco-Ameijeiras et al.,2012). A special section in the paper (3.5.1) discusses thissubject and its implications.

Preliminary work toward this project was initiated in1997. The main research was accomplished betweenDecember 2001 and March 2003.

1.4. Paper layout

The next section (1.5) provides background informationon LK. The research methodology in the field and labora-tory is detailed in Chapter 2. The results obtained by usand their discussion follow in Chapter 3, covering datalocation and organization (3.1), thermal and chemical re-sponse to lake dynamics (3.2), evaporation/precipitation,freshwater throughput and saline water intrusion effectson cycling (3.3), the Ca2+–HCO3

� relationship (3.4) andthe effect of Mg2+ and Sr2+ coprecipitation in calcite ontheir cycling (3.5). The last section provides a discussionof the distribution of Mg2+ and Sr2+ between LK water

es, where station A = collection site of all the lake water samples,River. The red numbers correspond to site enumeration in Table 1.ader is referred to the web version of this article.)


and calcite (3.5.1) and of the consistency of the data ob-tained (Section 3.5.1.1). Section 3.5.2 describes and dis-cusses the Ca2+–Mg2+–Sr2+ relationship in the watercolumn and its significance for evaluating the cycling ofthe solutes in the lake. The conclusions derived from thestudy are detailed in Chapter 4.

1.5. Background

1.5.1. Lake physiography, structure and water balance

Lake Kinneret is a freshwater lake located in northeast-ern Israel (Fig. 1), within the Dead Sea Rift valley, atapproximately 210 mbsl. The lake is a warm, monomictic,hardwater body, has an area of �167 km2, a maximumdepth of �42 m and a volume of 4.0 (±0.3) � 109 m3. Themaximum water stand is usually attained in May, decliningto its minimum stand in fall or early winter.

The 22-year averaged annual H2O input to LK (October1986 through September 2008) was estimated to be662 � 106 m3 (Berger et al., 2010), yielding a residence timeof �6 years for its water. During our research, hydrologicalyear 2001/2002, the Jordan River contributed �62% of thetotal H2O input to the lake (Berger and Shaw, 2003). Directprecipitation contributed another 10% and on-shore andsubmerged saline springs constituted the balance. Roughly,some 60–70% of the annual H2O input is pumped out, andapproximately 30% evaporates.

Stratification of LK generally lasts for 9–10 months(Figs. 2–4). The thermocline appears at a depth of 13(±2) m and deepens toward fall or early winter when over-turn occurs. During our study, the epilimnion waters tem-perature reached �30 �C in high summer, compared withonly 15–17 �C in the hypolimnion (Figs. 2B and 3A).

Vertical mixing, limited to the lowermost, 10-m thickwater layer (Rimmer et al., 2006), allows some solute ex-change between the lake’s floor and the overlying hypolim-nion (Nishri et al., 2000; Lemckert et al., 2004). A laminarflow regime prevails above the bottom layer, and exchangeof solutes between the hypolimnion and epilimnion ismostly retarded by stratification (Yeats and Imberger,2003).

1.5.2. Saline and freshwater salt sources

The bulk of the salts, including over 90% of the Cl andthe Na, enters LK via submerged and on-shore saline, Ca-chloridic springs (Serruya, 1978; Klein-BenDavid et al.,2004, 2005), discharging mainly in the Fuliya and Tabghaarea (Kolodny et al., 1999; Nishri et al., 1999). Freshwatersources bring in the remainder.

1.5.3. Salt and water throughput

The chemistry and hydrogeology of LK and its vicinityhave been investigated since early last century (e.g., Fried-mann, 1913; Goldschmidt et al., 1967; Mazor and Mero,1969; Starinsky, 1974; Serruya, 1978; Simon and Mero,1992; Assouline, 1993; Kolodny et al., 1999; Nishri et al.,1999; Moise et al., 2000; Berger and Shaw, 2003; Stilleret al., 2009; Berger et al., 2010). These and other studies re-vealed large differences in the ion concentration ratios be-tween the saline and fresh water sources as well as

seasonal variations in their inputs. The freshwater inputgenerally peaks in late winter, while pumping and evapora-tion become dominant in summer (Fig. 2D). Both fresh andsaline water inputs and outputs mainly occur via the epilim-nion, resulting in a non-synchronized water and saltthroughput. During the same time hypolimnetic watergradually gets admixed into the epilimnion upon thermo-cline deepening.

The marked monomictic stratification that separates theepilimnion from the underlying hypolimnion allows the for-mer to channel the water throughput and the latter to pre-serve a relatively stable chemical and physical depiction ofthe preceding mixed lake.

2. METHODS

2.1. Sampling

2.1.1. Water

A total of 1509 water samples were collected during ourstudy. Of these, 1428 were recovered from 14 profiles at sta-tion A and 81 were recovered from the Jordan River andthe Yahudiya and Meshushim streams near their dischargepoints into the lake (Fig. 1).

Lake water was sampled using an all-plastic diaphragmpump (FloJet, model 4305) through a reinforced, 45-mlong, 3=4

00 I.D. Terraflexe non-collapsible PVC hose. A plas-tic-coated, 5-kg cylindrical aluminum weight hooked to theend of the PVC hose served to stabilize the assembly. Aplastic filter mesh (0.1 mm opening) was suspended acrossthe entrance port of the hose to exclude coarser particulatematerial. A flow rate of 12 L min�1 was applied with inter-mittent, 2-min flushing between samples to avoid cross-con-tamination. The water temperature was measured at theentrance port of the assembly with ±0.1 �C accuracy and±0.02 �C precision using a calibrated Fenwal Electronicsemodel 199-103 LAG-A01 thermistor. The water was storedin pre-washed 330- or 500-ml PET bottles until analysis.The Jordan River and stream samples were retrieved bylowering 5-L plastic buckets into the water from roadbridges on the northeastern LK routes #92 and #87.

Selected samples for alkalinity, sulfide and sulfate anal-ysis were collected from the lake with a Van-Dorn samplerand transferred into 300-ml BOD bottles. Sulfide was fixedon board as ZnS by adding 1 ml of a 2 N Zn-acetate and1 ml of 6 N NaOH to the sample. Fifty milliliters were with-drawn from the bottles within 48 h after sampling, followedby alkalinity titration in the laboratory. An aliquot of150 ml of the clear supernatant solution was used for gravi-metric SO4

2� analysis, while the precipitate and the residualsolution were reserved for iodometric titration.

2.1.2. Particulate matter

Particles settling through the water column were col-lected in sediment traps deployed in station A for periodsaveraging 2 weeks at water depths of �28 and �34 m. Eachtrap assembly consisted of four, 50-cm long � 5 cm ID, ver-tically aligned PVC tubes mounted in a plastic frame.

The particulate matter accumulated within a 15-cm long(Na-Azide poisoned) plastic vial, screwed to the bottom of

https://www.researchgate.net/publication/29456465_Physical_Properties_of_Turbulent_Benthic_Boundary_Layers_Generated_by_Internal_Waves?el=1_x_8&enrichId=rgreq-0632235d-1d8a-4c0e-84ec-1104590fa2cf&enrichSource=Y292ZXJQYWdlOzI1NjcxNTkzMDtBUzoxNTU3NTg2MTYzODc1ODRAMTQxNDE0NzE1NjEzMw==

https://www.researchgate.net/publication/230822286_The_Physical_Regime_and_Respective_Biogeochemical_Processes_in_the_Lower_Water_Mass_of_Lake_Kinneret?el=1_x_8&enrichId=rgreq-0632235d-1d8a-4c0e-84ec-1104590fa2cf&enrichSource=Y292ZXJQYWdlOzI1NjcxNTkzMDtBUzoxNTU3NTg2MTYzODc1ODRAMTQxNDE0NzE1NjEzMw==

01/12/2001 01/03/2002 01/06/2002 01/09/2002 01/12/2002 01/03/2003

0

25

50

75

250

300 D

Tota

l wat

er in

put &

out

put t

o LK

(m3 x1

06 x m

onth

-1)

Date

Total water input Total water output

0.52

0.56

0.60

0.64

0.68

0.72

0.76

0.80

0.84

0.88

0.92

0.96

1.001/1/2002 1/7/2002 1/1/2003 1/4/20031/4/2002 1/10/2002

50 100 150 200 250 300 350 400 4500

-10

-20

-30

F scale

Julian date

Dep

th (m

bel

ow s

urfa

ce)

A

01/12/2001 01/03/2002 01/06/2002 01/09/2002 01/12/2002 01/03/200314

16

18

20

22

24

26

28

30

32 Hypolimnion Epilimnion

Date

Tem

pera

ture

(°C

)

B

01/12/2001 01/03/2002 01/06/2002 01/09/2002 01/12/2002 01/03/2003

660

680

700

720

740

760

780

800 C

Hypolimnion Epilimnion

TDS

(mg/

L)

Date

Fig. 2. (A) Deepening of the thermocline (white diamond markers) and corresponding change in the volume fraction (F) of the epilimnion(colored contours) in LK during the 2002 stratified period. Black circle markers below the abscissa denote mixed lake. (B and C) Averagewater temperatures and salinity (in TDS units), respectively, in the upper (10 m layer) epilimnion and the deep (8 m layer) hypolimnion duringthe monomictic cycle. Each data point represents an averaged batch of 21–28 samples. The exact depth, temperature and chemicalcomposition of each sample in the averaged sections is provided in Table ES-2. (D) Total water input (solid blue line and rectangular markers)and output (dashed red line and circle markers) to and from LK during the research period (December 2001 through March 2003). Note theunsynchronized throughput. Data from Berger and Shaw (2003). (For interpretation of the references to color in this figure legend, the readeris referred to the web version of this article.)


A

B

C

D

E

F

Fig. 3. Isopleth diagrams of temperature (A), TDS (B), concentrations of Cl� (C), Na+ (D), K+ (E) and the Na/Cl ratio (F), plotted vs. waterdepth and time for Lake Kinneret. Temperature (A) and TDS (B) maps are included for reference.


each tube and was left idle for 48 h after retrieval to allowcomplete settling. Supernatant water was discarded, and theresidual slurry was frozen for storage and was freeze-driedbefore analysis.

2.2. Analysis

The analyses were performed in the geochemical labora-tory of the Institute of Earth Science at the Hebrew Univer-sity of Jerusalem, where most of the required analyticalequipment (ICP-OES, FIA, LA-ICP-MS) was available.SEM analysis was performed in the geochemical laboratoryof the Israel Geological Survey, Jerusalem. Analyses of se-lected samples of lake water for alkalinity and sulfide were

performed in the Yigal Allon IOLR Kinneret LimnologicalLaboratory.

2.2.1. Water

The samples were filtered through Whatmane #40 filterpaper and diluted with deionized water (18.3 MX cm�1) tooptimal analytical ranges. Sodium, Ca, Sr, Mg, K, and SO4

(as S) were analyzed in triplicate by ICP-OES on a fullyautomated Perkin-Elmer Optima-3000 radial ICP system,using the Na589.592, Ca422.673, Sr407.771, Mg280.270,K766.491 and S180.669 spectral lines. The internal preci-sion of the ICP analysis was typically better than ±0.5%(RSD) except for SO4 (±1–3%), with the blanks being neg-ligible relative to the sample concentrations. Because much

A

B

C

D

E

F

Fig. 4. Isopleth diagrams of the concentrations of Ca2+ (A), Mg2+ (B), Sr2+ (C), and HCO3� (D) and the Sr/Ca and Mg/Ca ratios. Note the

marked fractionation of Ca2+, Sr2+ and HCO3�, reflected also by the Sr/Ca and Mg/Ca ratios between the epilimnion and hypolimnion. We

propose crystallization of calcite in the epilimnion and its partial dissolution in the hypolimnion as the main processes responsible for theseobservations.


of the analytical instability originated in the sample introduc-tion system, the precision of the Sr/Ca and Mg/Ca ratios wasmuch better than that of the individual ions (compare theconcentration-depth profiles of Ca, Sr and Mg in the mixed

lake (Fig. 5A, C, and D) with their corresponding concentra-

tion ratio-depth profiles (Fig. 6A and B)).Automated flow injection analysis (FIA) was used, in

triplicate, for Cl, Br and bicarbonate determination,deploying a Lachat Instruments model QE system withoptical detection. Chloride and Br were analyzed by the fer-ric thiocyanate #4500-Cl� E and the phenol red #4500-Br�

methods, respectively (after Eaton et al., 1995). For HCO3

analysis, Lachat Instruments method #10-303-31-1-A wasapplied. Accordingly, alkalinity is related to the change incolor, induced by a measured volume of the sample, to

the poorly buffered solution (pH 3.1) of methyl orange. Ingeneral, the internal analytical precision of the FIA analysiswas approximately ±0.5% RSD, and the concentrations inthe blanks were insignificant relative to the sampleconcentrations.

Calibration solutions for ICP and FIA runs were pre-pared from Mercke single element standard solutions.Instrument drift was monitored by running calibrationstandard sets after every 10-sample batch and was correctedby an off-line, in-house program.

Alkalinity titration and sulfide analysis of selectedwaters followed the American Public Health Association,APHA (1971) protocol with slight modifications. Alkalinitywas titrated with 0.1 N HCl to a pH of 4.3. The iodometrictitration followed the addition of a measured volume of a

40

30

20

10

0

0.0155 0.0160 0.0165 0.0170 0.0175 0.0180

23/1/2002 17/6/2002 22/11/2002 13/3/2003

Sr (meq/L)

Dep

th (m

)

C

45

40

35

30

25

20

15

10

5

0

2.70 2.75 2.80 2.85 2.90 2.95

23/1/2002 17/6/2002 22/11/2002 13/3/2003

Mg (meq/L)

Dep

th (m

)

D

40

30

20

10

0

1.5 2.0 2.5 3.0 3.5

23/1/2002 17/6/2002 22/11/2002 13/3/2003

Ca (meq/L)

Dep

th (m

)A

40

30

20

10

0

1.5 2.0 2.5 3.0 3.5 4.0

23/1/2002 17/6/2002 22/11/2002 13/3/2003

HCO3 (meq/L)

Dep

th (m

)

B

Fig. 5. Depth-concentration profiles sampled at station A during the mixed (23/1/2002 and 13/3/2003) and stratified (17/6/2002 and 22/11/2002) periods in Lake Kinneret. Figs. (A), (B), (C) and (D) represent Ca2+, HCO3

�, Sr2+ and Mg2+, respectively.


0.025 N KI solution, in excess of the expected HS� concen-tration, and dissolution of the resulting precipitate in 1.5 mlof concentrated HCl. A 0.025 N Na2S2O3 solution served asthe titrant.

2.2.2. Particulate matter

2.2.2.1. Chemical analysis. Approximately 100 mg of thesample powder was weighed to the nearest 0.1 mg and dis-solved in 3 ml diluted (�0.4 N) HCl in Teflon PFAe-cappedvessels. The HCl/sample mass ratio used limited the dissolu-tion of CaCO3 to less than 60 mg, minimizing leaching ofnon-carbonate material as much as possible during the pro-cess. The resulting slurry in the vessel was filtered throughWhatmane #40 filter into 25-ml PMPe volumetric flasksand diluted to volume. An aliquot of the solution was furtherdiluted to the optimal analytical range and analyzed for Ca,Sr and Mg, as described in Section 2.2.1.

2.2.2.2. SEM analysis. SEM examination was performed oncarbon-coated samples using a JEOL-840 SEM with anISIS Oxford EDS system.

2.2.2.3. LA-ICP-MS analysis. Six or seven calcite crystals ineach sample were selected under the system’s microscopeand analyzed for Ca, Mg, and Sr. An ArF-excimer laser

(new wave) was used, emitting 15-ns long pulses at a wave-length of 193 nm, with a repetition rate of 10 Hz, fluence of30 J cm�2 and variable spot sizes of 5–15 lm. The samplegrains were attached to an epoxy mount and positioned ina helium-flushed cell (at 1.1 L min�1) equipped with a 0.5-mm ID inlet nozzle. The surface of each calcite crystalwas continuously peeled off until complete ablation. Theelement count rate was determined by an AGILENT(7500 CX) ORS quadrupole mass spectrometer,using RF power of 1500 W, a 2.0-mm ID injection tube,and a nebulizer and auxiliary gas flows of 0.85 L Ar min�1

each. The samples were calibrated against SRM NIST610and Iceland spar standards. During the run, each samplewas bracketed between two standard duplicates. The accu-racy of the analysis was estimated at ±0.05% for Ca,±0.56% for Sr and ±3.3% for Mg. The analytical precisionrelevant to each calcite crystal is stated along with the ana-lytical data.

3. RESULTS AND DISCUSSION

3.1. Data location and organization

The chemical composition of Lake Kinneret water andits major brackish and freshwater feeders is summarized

40

30

20

10

0

0.0050 0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

23/1/2001 17/6/2002 22/11/2002 13/3/2003

Sr/Ca (eq/eq)

Dep

th (m

)A

40

30

20

10

0

0.90 0.95 1.00 1.05 1.10

23/1/2001 17/6/2002 22/11/2002 13/3/2003

Ca/HCO3 (eq/eq)

Dep

th (m

)

C

40

30

20

10

0

0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5

23/1/2002 17/6/2002 22/11/2002 13/3/2003

Mg/Ca (eq/eq)

Dep

th (m

)

B

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

3.4

3.6

2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4

123

4

5

6 78 9

1012 3

45

67

8

9

10

HCO3 vs. Ca in hypolimnion - quadratic fit HCO3 vs. Ca in epilimnion - linear fit HCO3 vs. Ca (analytical data) - hypolimnion HCO3 vs. Ca (analytical data) - epilimnion

HCO 3- C

a plot

Ca (meq/L)

HC

O3 (m

eq/L

)

DMixed lake

HCO 3- Ca plot

%calcite (ca

lculated)

-25

-20

-15

-10

-5

0

5

10

15

20

25

Affe

cted

cal

cite

(%)

Calcite dissolutionCalcite crystallization

Dissolution only (Ca=HCO 3)

Fig. 6. (A) Sr2+/Ca2+, (B) Mg2+/Ca2+, and (C) Ca2+/HCO3� concentration ratio vs. depth profiles, sampled at station A during the mixed (23/

1/2002 and 13/3/2003) and stratified (17/6/2002 and 22/11/2002) periods of Lake Kinneret. (D) The relationship between HCO3� and Ca2+ in

LK between January and November 2002. The filled markers represent analytical data; the dot-centered markers represent the affected

fraction of calcite (in %) removed or added to that which existed in the mixed lake before stratification (Ca = 2.67 meq L�1;HCO3 = 2.56 meq L�1, on 23/1/2002, Table 1). Crystallization of calcite in the epilimnion (filled red circles) and its partial dissolution inthe hypolimnion (filled blue rectangles), respectively, are the main processes responsible for the removal and addition of these solutes. Thesolid lines were calculated by regression analysis of the analytical data. The crystallization of calcite (solid red line) follows an almostlinear trend ([HCO3

�] = 0.08658 + 0.9189 � [Ca2+]; R2 = 0.9919), while its dissolution (solid blue line) follows a quadratic curve([HCO3

�] = �2.5338 + 3.0867 � [Ca2+] � 0.40943 � [Ca2+]2; R2 = 0.9882). Unlike the Ca2+ depletion in the epilimnion, which reached amaximum on 17/6/2002 (#5) before declining toward 22/11/2002 (#10), the buildup of HCO3

� in the hypolimnion is continuous, and itsexcess over Ca2+ increases throughout the stratified period. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)


in Table 1. The detailed, original analyses are presented inTables ES-1 (freshwater sources) and ES-2 (LK waters) inthe Electronic Supplement to this paper. Magnesium, Caand Sr analyses of bulk particulate matter (by ICP-OES)and of corresponding epilimnetic water are detailed in Ta-bles 2 and 3. LA-ICP-MS analyses of individual calcitecrystals in the particulate matter samples are provided inTable 4.

The upper portion of Table 1 contains three datapairs, representing LK samples collected before (23/1/2002), during (22/8/2002) and after (30/1/2003) the strat-ified period of the lake. Each pair includes the averagecomposition of the upper, 10-m thick surface layer andof the 8-m thick water body directly overlying the lake’sbottom. The averages represent batches of 21–28 samples,

sequentially retrieved from the water column on the spec-ified date. The in-layer thermal homogeneity is expressedby the standard deviation of the temperature about itsaverage, while the overall homogeneity of the lake is re-flected by the temperature and concentration differencesbetween the surface and bottom water layers. The tem-perature and concentration differences during the twobracketing mixed-lake periods were practicallyindistinguishable.

The second portion of Table 1 lists the chemical compo-sition reported for the main brackish and saline watersources (TDS > 1000 mg L�1) around LK. Additionalinformation may be found in Kolodny et al. (1999), Nishriet al. (1999), Moise et al. (2000) and Klein-BenDavid et al.(2004, 2005).

Table 1Chemical analyses of Lake Kinneret waters and of its saline and freshwater sources. Concentrations are given in meq L�1, TDS in mg L�1.

Site # (Fig. 1) Site and field information Na K Mg Ca Sr Cl SO4 HCO3 Br TDS Source

Lake Kinneret (station A)

1 Mixed – 23/1/2002

Surface (0–8 m); N = 25; T = 15.66 ± 0.15 �C 6.14 0.19 2.85 2.66 0.0175 7.99 1.20 2.55 0.0254 736 1Bottom (10 m); N = 26; T = 15.51 ± 0.02 �C 6.15 0.19 2.85 2.67 0.0175 8.02 1.20 2.58 0.0255 739 1

1 Stratified – 22/8/2002

Surface (0–8 m); N = 23; T = 29.95 ± 0.36 �C 6.13 0.19 2.83 2.16 0.0165 7.94 1.19 2.05 0.0258 693 1Bottom (0–8 m); N = 28; T = 16.45 ± 0.06 �C 6.09 0.19 2.85 3.12 0.0179 7.89 1.07 3.12 0.0244 769 1

1 Mixed – 30/1/2003

Surface (0–8 m); N = 25; T = 16.43 ± 0.16 �C 6.32 0.19 2.97 2.69 0.0180 8.25 1.18 2.48 0.0258 746 1Bottom (10 m); N = 21; T = 16.24 ± 0.03 �C 6.34 0.19 2.99 2.71 0.0181 8.31 1.18 2.53 0.0258 752 1

Brackish and saline sources

2 Tiberias (West) – spring 296 8.08 53.5 166 1.44 507 16.4 2.30 2.88 30297 23 Fuliya-5 (West) – spring 16.3 0.31 5.37 7.78 0.0331 21.3 2.04 5.13 0.0562 1779 24 Fuliya-6 (West) – spring 15.3 0.29 5.27 7.53 0.0306 20.3 1.93 5.15 0.0489 1710 25 Fuliya-6/2 (West) – spring 19.6 0.38 6.00 8.88 0.0406 26.5 2.52 5.08 0.0715 2093 26 Druzi (NW) – spring 25.2 0.55 5.75 12.4 0.114 36.0 2.23 4.41 0.156 2590 27 Kinneret 7 (NW) – well 13.8 0.32 3.46 9.43 0.0765 20.9 1.15 4.69 0.0867 1653 28 En Sheva (NW) – spring 21.4 0.49 5.12 12.4 0.1105 31.4 1.81 4.77 0.133 2329 29 Ma’ayan Matok (NW) – spring 34.0 0.75 6.97 15.8 0.163 49.9 2.92 4.36 0.199 3410 210 Sartan Iver (NW) – spring 61.6 1.32 12.0 25.0 0.269 90.3 5.40 4.98 0.332 5917 211 Ha’On (SE) – well 160 10.87 74.8 27.9 1.01 239 15.9 8.98 1.20 15520 212 Gofra (East) – spring 57.3 2.17 18.8 11.2 0.256 79.9 2.60 6.87 n.d. 5241 313 Bet Zaida (NE) artesian well (Bz in Fig. 1) 156 0.90 70.5 36.1 0.233 273 0.227 3.03 1.54 15205 2

Freshwater sources

14 Jordan R. 4/2/1997–28/3/1999; N = 28 (JR in Fig. 1) 0.538 0.0411 0.838 3.32 0.00412 0.466 0.720 3.37 BDL 356 1±RSD (%) 13.2 19.2 13.5 14.2 13.4 21.0 47.0 5.5 BDL 13.0

15 Meshushim R. 29/7/1997–23/2/1999; N = 27 (Me in Fig. 1) 1.42 0.0792 1.52 1.45 0.00524 0.736 0.184 3.50 BDL 340 1±RSD (%) 16.8 19.1 10.4 5.8 5.4 16.9 23.8 10.2 BDL 11.7

16 Yahudiya R. 29/7/97–23/2/1999; N = 27 (Ya in Fig. 1) 1.61 0.0549 1.90 1.74 0.00695 0.924 0.236 4.14 BDL 398 1±RSD (%) 21.4 22.5 15.6 7.30 7.97 21.3 24.6 16.3 BDL 15.0

(1) Present study; (2) Klein-BenDavid et al. (2004); (3) Moise et al. (2000).“BDL” = below detection limit (<0.5 mg Br/L); n.d. = not determined.Detailed analyses of the freshwater riverine samples are given in Table ES-1 and those of the lake waters in Table ES-2 in the Electronic Supplement.

A.

Katz,

A.

Nish

ri/G

eoch

imica

etC

osm

och

imica

Acta

105(2013)

372–394381

Table 2Distribution coefficients of Sr2+ and Mg2+ between calcite and LK epilimnetic water, calculated from analyses of the ‘bulk’ material trappedin the hypolimnion (Eq. (3)).

Epilimnetic water – “bulk” material (Eq. (3))

Sample Elevation (m) Started on Duration(days)

Mid-date Mg(meq L�1)

Ca(meq L�1)

Sr(meq L�1)

Sr/Ca(eq/eq)

Mg/Ca(eq/eq)

Bulk trap material

1 1.5 02/04/2002 18 11/04/2002 0.225 3.02 0.00397 0.00132 0.07462 9 02/04/2002 18 11/04/2002 0.218 3.26 0.00409 0.00125 0.06693 1.5 25/04/2002 19 04/05/2002 0.224 5.36 0.00721 0.00135 0.04174 1.5 09/05/2002 8 13/05/2002 0.303 6.30 0.00922 0.00146 0.04815 1.5 18/05/2002 10 23/05/2002 0.286 5.60 0.00883 0.00158 0.05116 1.5 04/06/2002 12 10/06/2002 0.331 4.99 0.00781 0.00156 0.06637 9 04/06/2002 12 10/06/2002 0.309 4.86 0.00746 0.00154 0.0636

Sampled on Sr(meq L�1)

Mg(meq L�1)

Ca(meq L�1)

Sr/Ca(eq/eq)

Mg/Ca(eq/eq)

DSr DMg

Epilimnetic water

08/04/2002 0.741 33.8 51.4 0.006595 1.08 0.199 0.068808/04/2002 0.741 33.8 51.4 0.006595 1.08 0.190 0.061707/05/2002 0.709 33.2 44.3 0.007322 1.24 0.184 0.033807/05/2002 0.709 33.2 44.3 0.007322 1.24 0.200 0.039007/05/2002 0.709 33.2 44.3 0.007322 1.24 0.215 0.041417/06/2002 0.693 33.5 38.0 0.008343 1.45 0.188 0.045617/06/2002 0.693 33.5 38.0 0.008343 1.45 0.184 0.0438

Average 0.194 0.0477RSD (±%) 5.9 26.7


The third section in Table 1 lists the average chemicalcomposition of the Jordan River and two smaller streams,Meshushim and Yahudiya (Fig. 1) that drain the basalticterrain northeast of the lake.

3.2. The thermal and chemical expression of lake dynamics

The thermal development of the water column structureduring our study is presented in Fig. 2A by the relation be-

Table 3ICP-OES analyses of epilimnion water collected on different dates duringmethod #2 (Eq. (4)). The change in Mg2+ concentration was too small f

Period Date Sr(meq L�1)

Ca(meq L�1)

Mg(meq L�1)

Srf

23/01/2002

0.0175 2.74 2.84 Fu

1 22/11/2002

0.0168 2.37 2.88 0.9

08/04/2002

0.0169 2.56 2.78 8/4

2 07/05/2002

0.0162 2.21 2.73 0.9

3 17/06/2002

0.0158 1.90 2.75 0.9

17/06/2002

0.0158 1.90 2.75 Sum

4 22/11/2002

0.0168 2.37 2.88 1.0

tween thermocline depth, epilimnion/lake volume ratio, F,and time. The contemporaneous changes in temperatureand TDS in the epilimnion and hypolimnion are illustratedin Fig. 2B and C. The response of Cl�, Na+ and K+ to sea-sonal lake dynamics is displayed in contour maps in Fig. 3and that of the alkaline earth ions and bicarbonate is dis-played in Fig. 4. Four detailed depth-concentration profilesof Ca2+, HCO3

�, Sr2+ and Mg2+ sampled before, duringand after the stratified stage are presented in Fig. 5A–D,

the stratified period of LK and applied in the calculation of DSr byor reliable calculation of DMg by the same approach.

/Sri Caf/Cai ln(Srf/Sri) ln(Caf/Cai) DSr DMg

ll cycle integration (23/1/02 through 22/11/02)

63 0.864 �0.0372 �0.147 0.254 n/a

/2002 data taken as starting composition

57 0.862 �0.0441 �0.149 0.297 n/a

35 0.739 �0.0670 �0.302 0.222 n/a

Average: 0.259 n/aRSD (%): 20.5 n/a

mer time integration (17/6/02 through 22/11/02)

65 1.250 0.063 0.223 0.282 n/a

Table 4LA-ICP-MS Mg, Ca and Sr analyses of individual calcite crystals in material trapped in the hypolimnion of LK, and calculated DMg and DSr.values (Eq. (5)).

Crystal # Mg (ppm) Mg (Std.) Ca (wt%) Ca (Std.) Sr (ppm) Sr (Std.) Mg/Ca (M) Sr/Ca (M) DMg DSr

TRAP AU156-L, water depth: 35 m, collection period: 6/5/2009–12/5/2009

1 2274 9.5 40.0 0.0020 1305 3.2 0.00937 0.00149 0.00784 0.2232 2441 7.3 40.0 0.0021 1181 6.5 0.01006 0.00135 0.00841 0.2023 2648 7.1 40.0 0.0019 980 4.4 0.01091 0.00112 0.00912 0.1674 2940 6.0 40.0 0.0024 1428 4.9 0.01212 0.00163 0.01013 0.2445 2273 9.5 40.0 0.0018 1172 3.8 0.00937 0.00134 0.00783 0.2006 2768 5.2 40.0 0.0021 1387 5.2 0.01141 0.00159 0.00954 0.2377 2514 9.0 40.0 0.0022 1279 4.7 0.01036 0.00146 0.00866 0.218

Average 0.00879 0.213RSD (%) 9.8 12.2

TRAP AU156-U, water depth: 28 m, collection period: 6/5/2009–12/5/2009

1 2812 5.5 40.0 0.004663 1342 4.3 0.01159 0.00154 0.00969 0.2292 1087 26.7 40.0 0.006464 929 5.9 0.00448 0.00106 0.00375 0.1593 2895 6.8 40.0 0.007258 1178 5.8 0.01193 0.00135 0.00998 0.2014 2851 32.6 40.0 0.002535 1305 6.0 0.01175 0.00149 0.00983 0.2235 2149 31.5 40.0 0.005666 1356 4.6 0.00886 0.00155 0.00741 0.2326 2326 9.6 40.0 0.005645 1228 6.8 0.00959 0.00140 0.00802 0.2107 2040 69.5 40.0 0.007557 1378 7.9 0.00841 0.00158 0.00703 0.235

Average 0.00796 0.213RSD (%) 27.9 12.6

TRAP AU157-L, water depth: 35 m, collection period: 12/5/2009–19/5/2009

1 2735 8.7 40.0 0.0044 1245 2.6 0.01127 0.00142 0.00943 0.2122 3677 8.3 40.0 0.0017 1127 6.3 0.01516 0.00129 0.01267 0.1923 2025 6.4 40.0 0.0032 1177 6.3 0.00835 0.00135 0.00698 0.2014 2453 8.5 40.0 0.0011 1234 4.6 0.01011 0.00141 0.00845 0.2115 2755 8.0 40.0 0.0031 1261 4.5 0.01136 0.00144 0.00949 0.2156 2985 7.4 40.0 0.0037 1130 3.2 0.01230 0.00129 0.01029 0.193

Average 0.00955 0.204RSD (%) 19.9 4.9

TRAP AU157-U, water depth: 28 m, collection period: 12/5/2009–19/5/2009

1 888 5.7 40.0 0.000668 1326 5.9 0.00366 0.00152 0.00306 0.2262 2451 45.7 40.0 0.000665 1379 8.1 0.01010 0.00158 0.00845 0.2353 2190 25.8 40.0 0.002688 1544 7.7 0.00903 0.00177 0.00755 0.2644 2721 35.5 40.0 0.004488 1267 6.7 0.01122 0.00145 0.00938 0.2165 2211 6.8 40.0 0.002476 1263 5.7 0.00912 0.00144 0.00762 0.2166 2951 9.8 40.0 0.002788 1497 23.0 0.01217 0.00171 0.01017 0.2567 1811 15.8 40.0 0.003488 1231 5.9 0.00747 0.00141 0.00624 0.210

Average 0.00750 0.232RSD (%) 31.2 9.0


and their corresponding concentration ratio – depth plotsare provided in Fig. 6A (Sr/Ca), 6B (Mg/Ca), and 6C(Ca/HCO3).

The temperature, TDS and individual solute diagrams inthese figures indicate that complete or nearly complete mix-ing of LK was achieved sometime during January 2002 andagain in January 2003. A short, isolated dilution event ofthe bottom waters by early winter floods occurred nearthe 2002/2003 transition, as indicated by TDS (Fig. 3B),Cl� (3C), Na+ (3D), K+ (3E), Mg2+ (4B), and Sr2+ (4C).The homogeneity of LK was sustained by mixing duringwinter until the onset of stratification in the spring. The ini-tially shallow thermocline that appeared in April deepened

during the stratified period, inducing a growth of the epi-limnion volume fraction (F) in the lake from approximately0.5 in the spring to 0.9 in late December, 2002 (Fig. 2A).The mixing that must have been caused by such a volumechange (Table 5, #1) is evident in the epilimnion thermalstructure (Fig. 3A). In contrast, the hypolimnion retainedits preceding ‘winter temperature’ (between 15 and 17 �C)throughout the monomictic cycle but, during the same time,underwent pronounced changes in the concentrations ofseveral solutes, notably those displayed by Ca2+, Sr2+,Mg2+ and HCO3

� (Fig. 4).The disaccord between the thermal and the chemical

evolution and the conspicuous differences between the var-

Table 5Mass transfer processes in LK and other hardwater monomictic lakes.

# Process description Relevance and effects

1 Admixture of hypolimnetic water into the epilimnion uponthermocline deepening

Operative in any monomictic lake, but always along with other processes

2 Direct evaporation/precipitation from or on the lake Concentration/dilution of solutes w/o change in ion ratios. Mixing lines passthrough origin

3 Flushing by freshwater influx (runoff and rivers) Epilimnion dilution, with little effect on hypolimnion. Mixing lines of solutepairs are linear with non-zero intercepts on ordinate

4 Intrusion of saline water from on-shore and/or submergedsprings

Epilimnion: Saline water sinks into hypolimnion under its elevated densityHypolimnion: Increase in TDS;Linear ion-ion mixing plots between saline and hypolimnetic watersMixed lake stage: Increase in concentration of brine-related solutes (Cl, Na,Sr, etc.) with time

5 Precipitation of CaCO3 from the epilimnion with orwithout partial dissolution in the hypolimnion

Epilimnion: Linear Ca2+ vs. HCO3� plots extrapolating through origin

Increased Sr/Ca and Mg/Ca ratios in the water due to calcite crystallization(DSr and DMg < 1)Hypolimnion: Decreased Sr/Ca and Mg/Ca ratios due to dissolution ofcalcite settling from the epilimnionCa2+ vs. HCO3

� plot non-linearly due to BSR and nitrification under anoxicconditions (see #6)

6 Hypolimnetic bacterial sulfate reduction (BSR) Increase in H2S, HS� and crystallization of elemental S at the expense ofSO4

2+ in hypolimnion

Fig. 7. The total variation (in %) of the Na+, K+, Mg2+, Ca2+,Sr2+, Cl�, HCO3

� and SO42� concentrations relative to their

reference values in the mixed lake on 23/1/2002. Note the small (8–10%) variation in K+, Na+, Mg2+, and Cl relative to the variationin HCO3

�, Ca2+, SO42� and Sr2+, for which calcite crystallization/

dissolution and BSR are responsible. The concentration of Mg2+ inthe lake is sufficiently buffered by inflow to mask its removal incalcite.


iation ranges of the solute concentrations in the stratifiedlake (8–70%, Fig. 7) cannot be explained by mixing alone.Possible solute transfer processes other than mixing arelisted in Table 5, #2–6 and are discussed further below.

3.3. Effects of evaporation/precipitation, freshwater

throughput and saline water intrusion on solute cycling

The main input of freshwater into LK occurred duringwinter and early spring (Fig. 2D). During much of thistime, the lake was mixed, and the addition of freshwaterby rivers, floods and direct precipitation resulted mainlyin dilution rather than affecting the ion concentration ratios(Table 5, #2); evaporation from the lake also did not alterthese ratios. Furthermore, because much of the lake’s fresh-water throughput occurs via the epilimnion, freshwaterflushing has little effect on the hypolimnion water composi-tion (Table 5, #3).

The effects of saline water intrusion on the compositionof LK (Table 5, #4) can be revealed by constructing feasiblemixing models between the freshwater input and each salinesource feeding the lake (Table 1) and testing the actual com-position of the lake against the model predictions. Accord-ingly, we plotted full mixing paths of selected ions betweeneach saline water source and freshwater end-member(Fig. 8). Ion concentration ratios were used rather thanion concentrations to cancel out any evaporation and dilu-tion effects (Table 5, #2). Because of its conservative behav-ior and major role in the electrical charge balance of thesaline waters, the chloride ion was selected as the denomi-nator in the first four ratios plotted (Fig. 8A–D). The frac-tion (R) of saline water in the mixture was derived fromdata provided by Berger and Shaw (2003). Accordingly,during the hydrological year 2002, the Jordan River con-

tributed �62% by volume of the total H2O input to thelake. Another �10% H2O was added to LK by direct pre-cipitation, with a negligible effect on the ion ratios therein.Hence, the fraction R of the ‘saline’ end member in the in-put mixture was �0.31 (v/v).

Finally, we inserted the respective, analytically deter-

mined ion ratios that existed in the mixed lake before theonset of stratification into the diagrams and verified theircompliance with the curves at the calculated R = 0.31 value

FuliyaTiberiasBet ZaidaSartan IverGofraHa'OnEn ShevaDruziMatokR=0.31

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.9 1.00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.71.01.11.2

Mg/

Ca

(eq/

eq)

R

E

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.9 1.00.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

Mg/

Cl (

eq/e

q)

R

B

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.9 1.00.000

0.002

0.004

0.006

0.008

0.010

Sr/C

l (eq

/eq)

R

C

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.9 1.00.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

Na/

Cl (

eq/e

q)

R

A

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.9 1.00.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.06

7

Ca/

Cl (

eq/e

q)

R

D

Fig. 8. Mixing diagrams between the main freshwater source of LK (Jordan River) and nine brackish and saline water sources on or near thelakeshores. Ionic ratios, rather than concentrations, are used to remove the dilution/concentration effects by precipitation/evaporation. (A)Na/Cl; (B) Mg/Cl; (C) Sr/Cl; (D) Ca/Cl; (E) Mg/Ca. The red, center-dotted circle in each diagram represents the respective ratio calculatedfrom the analysis of the mixed lake waters on 23/1/2002 and placed at R = 0.31 (the fraction of freshwater (Jordan River) in the total H2Oinput to the lake). The diagrams demonstrate that the ratios of Na+, Mg2+, and Sr2+ to Cl� are in agreement with a Fuliya-Jordan Rivermixture (Table 1, #s 3–5 and 14). The Mg/Ca and Ca/Cl ratios, which are incompatible with any saline end member mixture with the JordanRiver, can readily be explained by removal of Ca2+ from the lake in calcite. (For interpretation of the references to color in this figure legend,the reader is referred to the web version of this article.)


(center-dotted circles in Fig. 8). The resultant mixing curvesof the Na/Cl, Mg/Cl, Ca/Cl, Sr/Cl and Mg/Ca ion concen-tration ratios are depicted in Fig. 8A–E.

The following conclusions may be derived from the mix-ing diagrams:

1. The brackish Fuliya springs (Table 1, #3–5) standout as the most likely saline end member in themixture (Fig. 8A–C), in agreement with the conclusionsreported by Kolodny et al. (1999) and Nishri et al.(1999).

2. No combination of two or more of the saline watersources listed in Table 1 would produce the Na/Cl ratiomeasured in LK (Fig. 8A) by mixing with the waters ofthe Jordan River.

3. The Ca/Cl ratio that prevailed in the mixed lake is muchlower than that required by mixing (Fig. 8D).

4. The high Mg/Ca ratio in the mixed lake (Fig. 8E) cannotbe explained by mixing with any known saline source ofLK. Rather, we attribute it to loss of Ca2+ from thelake’s water column (Table 5, #5).

3.4. The Ca–HCO3 relationship

The relationship between Ca2+ and HCO3� during the

2002 monomictic cycle in LK was evaluated based onBSE-SEM imaging of calcite crystals trapped in the hypo-limnion, chemical analyses of epilimnetic and hypolimnetic

waters, and thermodynamic calculation of the saturation ofthe water with respect to calcite, as discussed below.

A. BSE-SEM petrography: Particulate material trappedduring early summer in the hypolimnion at approxi-mately 28 m below the lake’s surface illustrates that(Fig. 9): (1) the bulk of the material is composed ofcalcite crystals. (2) The crystals preserved their idio-typic, rhombohedral shape. (3) Most crystals are ofsimilar size, not exceeding �10–15 lm in edge length.(4) All calcite crystals display severe dissolutionpitting.

B. Chemical analyses: Fig. 5A and B presents four Ca2+

and HCO3� concentration-depth profiles sampled

before, during and after the stratified period of themonomictic cycle in LK. The corresponding Ca/HCO3 concentration ratio profiles are presented inFig. 6C. After the onset of stratification, the follow-ing changes occurred: (1) Depletion of Ca2+ andHCO3

� in the epilimnion and enrichment of thesesolutes in the hypolimnion already in early summer(Fig. 5A and B, 17/6/2002 profiles). (2) A decreasein the epilimnetic depletion of Ca2+ and HCO3

�

but continuation of their enrichment in the hypolim-nion toward mid- and late-summer (22/11/2002 pro-files). (3) A decline of the Ca/HCO3 ratio in the entirewater column in early summer (Fig. 6C, 17/6/2002profile), continuing this trend in the hypolimnion intothe fall (22/11/2002).

Fig. 9. BSE-SEM image of calcite crystals that accumulated in asediment trap suspended 9 m above the bottom of LK in station Afor a period of 12 days as of 4/6/2002 (scale bar = 20 lm). Theidiotypic rhombohedral crystal habit is readily discernible despitesevere pitting caused by dissolution during settling of the crystalsthrough the hypolimnion. The textural evidence is in agreementwith our chemical analyses and thermodynamic calculations,demonstrating enrichment of dissolved Ca2+ and HCO3

� in thehypolimnion, while the epilimnion water is oversaturated withrespect to calcite, and the hypolimnetic water is undersaturatedwith respect to the same mineral (see also Fig. 10B).

1.45 1.50 1.55 1.60 1.651.4

1.6

1.8

2.0

2.2

2.4

DIC

-Ca

(mm

ole/

L)

Ca (mmole/L)

DIC - Ca = -4.60 + 4.162 * CaR2 = 0.910

C

01/01/2002 01/04/2002 01/07/2002 01/10/2002 01/01/20030

2

4

6

8

10

12

14

calc

iteDate

EpilimionHypolimnion

-----B

01/12/2001 01/03/2002 01/06/2002 01/09/2002 01/12/2002 01/03/20031.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

3.4

Epilimnion Hypolimnion

Ca

(meq

/L)

Date

Spring

SummerA

Ω

Fig. 10. (A) A concentration–time slice of the Ca2+-depth-timeisopleth diagram presented in Fig. 4A. Calcite crystallization in theepilimnion is responsible for the steep decline of the Ca2+

concentration in the epilimnion from early March through June.Partial dissolution of this mineral in the hypolimnion accounts forthe increase in the Ca2+ concentration therein during the sameperiod (“Spring”-labeled arrow). Later in summer, the decreasingcalcite flux through the hypolimnion constrains dissolution to thesediment–water interface, as expressed in the lower slope of theCa2+-time plot (“Summer”-labeled arrow). (B) The degree ofsaturation with respect to calcite (Xcalcite) of the epilimnion andhypolimnion waters in Lake Kinneret in 2002. The epilimnionwater is oversaturated throughout the year, increasing steeply to itsmaximum value (X � 13–14) during the crystallization period ofthis mineral. Supersaturation declines (but remains high) sometimeduring May, decreasing to 2–3 X units at the end of the year. Thehypolimnion experiences slight undersaturation as of April, asreflected by the dissolution of calcite until the lake’s overturn inwinter. (C) The relationship between [DIC-Ca] and the Caconcentration during the anoxic stage of the hypolimnion, betweenMay and November. The DIC was calculated using the PHREEQCprogram. [DIC-Ca] is assumed to represent the fraction of DICthat directly formed via the biodegradation of organic matter andnot from calcite dissolution.


C. Thermodynamics: The PHREEQC (online) inputdata and the thermodynamically calculated Xcalcite

and (DIC-Ca2+) results are provided in Tables ES-3–ES-5 in the Electronic Supplement. A graphic pre-sentation of the results is provided in Fig. 10A–C.Supersaturation with respect to calcite prevailed inthe epilimnion throughout the research period, whilethe hypolimnion became progressively undersatu-rated from April through October–November, 2002.

Findings A–C can be interpreted as follows:The calcite crystals that accumulated in the traps are pri-

mary crystallization products formed in the epilimnionwater, as indicated by their idiotypic, rhombohedral shapes,their narrow size distribution, and their upper edge size lim-it (Fig. 9). Chemical evidence for the epilimnetic origin ofcalcite is provided by the marked depletion of Ca2+ andHCO3

� in the epilimnion in early summer (Fig. 5A andB) at a DHCO3/DCa slope of 0.9189 (R2 = 0.9919,Fig. 6D), coinciding with the timing of the calculatedXcalcite peak in the epilimnion water (Fig. 10B).

The partial dissolution of calcite in the hypolimnion isrevealed by the deep pits that litter the calcite crystal faces(Fig. 9). Coeval hypolimnion waters display increasing con-centrations of Ca2+ and HCO3

� at a Ca/HCO3 ratio < 1(Figs. 5A, B and 6C) as well as an increasing under-satura-tion with respect to calcite (Fig. 10B) from April (Xcalcite -� 1) through October–November 2002 (Xcalcite = 0.72–0.78). Calcite dissolution alone cannot explain the drop inthe Ca/HCO3 ratio from 1.05 (that of the mixed lake) toapproximately 0.9 in the hypolimnion in fall 2002(Fig. 6C) or account for the deviation of the hypolimnetic


samples from the ‘Dissolution only’ line (Fig. 6D). Thus, anadditional source of bicarbonate is required for thehypolimnion.

Similar phenomena have been reported from otherlakes, notably in pre-alpine lakes (e.g., Muller et al., 2003,2006). In their paper on the fate of calcite and organic mat-ter in Lake Lucerne and Lake Sempach in Switzerland,Muller et al. (2006) investigated in detail the pore waterconcentration profiles of Ca2+, CO3

2� pH and O2 acrossthe water–sediment interface. Their conclusion, that degra-dation of organic matter is a major source of bicarbonate inanoxic pore water (beyond that produced via calcite disso-lution), is in agreement with our findings in the water col-umn of LK (Fig. ES-1). Hence, organic matter settlingthrough the hypolimnion degrades by oxidation, resultingin the complete consumption of oxygen therein. This oxicstage is followed by denitrification (Eq. (1)) and then by sul-fate reduction (Eq. (2)) under anoxic conditions (Hadas andPinkas, 1995).

Accordingly, oxidation of one mole of Redfield-type or-ganic matter would produce 124 and 106 mol of H+ andHCO3

�, respectively, with subsequent production of someH2CO3 via carbonate equilibria, where both H+ andH2CO3 are capable of dissolving calcite.

Denitrification and sulfate reduction are described by:

C106H263O110N16Pþ 84:8NO3�

¼ 106HCO3� þ 16NH4

þ þHPO42� þ 42:4N2

þ 7:2Hþ þ 42:4H2O ð1Þ

and by

C106H263O110N16Pþ 53SO42�

¼ 106HCO3� þ 16NH4

þ þHPO42� þ 53HS�

þ 39Hþ: ð2Þ

Evidently, less H+ is produced by biodegradation under an-

oxic conditions than during the preceding oxic stage, allow-ing more HCO3

� to remain in solution without formationof H2CO3 and accounting for the HCO3

� excess we ob-served in the hypolimnion.

Our calculations confirm growth of CO2 and HCO3�

concentrations as well as a decrease in pH and Xcalcite inthe hypolimnion during the anoxic stage (Tables ES3 andES4).

While calcite dissolution in the hypolimnion continuedthroughout the stratified period (Figs. 4A and 5A), theCa2+ concentration–time slope in spring (February–May)was steeper than that characterizing the later summer andfall (Fig. 10A). It is significant that the steep Ca2+-timeslope we observed coincides with an elevated calcite fluxthrough the hypolimnion (Koren and Ostrovsky, 2000). Itis possible that this correlation between the Ca2+ concentra-tion–time slope and calcite flux implies that calcite dissolu-tion in the hypolimnion is constrained by the availability ofcalcite surface area, which in turn depends on the rate ofcalcite production in the epilimnion. The deceleration ofthis process later in summer and in fall left the calcite-richsediment–water interface to take over as the main dissolu-tion site (Figs. 4A and 5A). The fraction of DIC (dissolved

inorganic carbon) that originated from the degradation oforganic matter in the hypolimnion can be estimated fromthe difference between the DIC and Ca2+ concentrations.By plotting this difference vs. Ca2+ (Fig. 10C), one obtainsthe ratio between DIC contributed by degradation overthat originating from calcite dissolution (D(DIC-Ca)/DCa). During the anoxic stage, this ratio in the hypolim-nion was 4.16, meaning that only 19–20% of the DIC orig-inated from calcite dissolution. In contrast, thecorresponding ratio for the oxic stage, extending from theend of February into May, was 1.26, indicating a muchhigher contribution (44%) of dissolved calcite to the DIC.It is also noteworthy that the thermodynamically calculatedD(DIC-Ca)/DHS� ratio corresponding to the anoxic stagewas 1.9, very near the ratio of 2 (=106/53) predicted fromEq. (2).

3.5. The effect of coprecipitation on the cycling of Ca2+, Sr2+

and Mg2+

The effect of calcite crystallization in the epilimnion andits dissolution in the hypolimnion on the cycling of Ca2+,Sr2+ and Mg2+ can be assessed from the relationship be-tween the concentration ratios of Mg/Ca and Sr/Ca(Fig. 11).

In an Mg/Ca–Sr/Ca coordinated diagram, the path ofpure calcite crystallization or dissolution would be depictedby a perfectly linear plot crossing the origin, with a slopeequal to the Mg/Sr concentration ratio in the surroundingsolution. In impure, real systems, however, calcite, as anyother mineral, coprecipitates Sr2+, Mg2+ and a wealth ofother ions. Hence, interpretation of the data presented inFig. 11 requires knowledge of the distribution coefficients(DSr and DMg) between calcite and the water from whichit crystallized. The voluminous literature relevant to thisquestion (e.g., Holland et al., 1964; Holland, 1966; Kitanoet al., 1971; Katz et al., 1972; Katz, 1973; Lorens, 1981;Mucci and Morse, 1983; Tesoriero and Pankow, 1996; Cur-ti, 1999; Huang and Fairchild, 2001; Tang et al., 2008) con-tains a wide scatter of these coefficients even for the samesolute and highlights the problems involved in their directapplication to natural systems (e.g., Morse and Bender,1990; Rimstidt et al., 1998). To circumvent this problem,we used our water and calcite analyses to derive ‘effective’DSr and DMg values operative in LK. Our approach andcalculations, which can readily be applied to monomicticlakes elsewhere, are provided in the following section.

3.5.1. Distribution of Sr2+ and Mg2+ ions between lake water

and calcite

The distribution coefficients DMg and DSr were calcu-lated from the changes in the Ca2+, Mg2+ and Sr2+ concen-trations in the epilimnion water and their respectiveconcentrations in calcite collected from sediment traps sus-pended in the underlying hypolimnion.

Three procedures were applied:

1. Division of the Sr/Ca and of the Mg/Ca ratios in the‘bulk’ trapped material by their corresponding ratios inthe epilimnion water (Eq. (3)).

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5 26/12/2001

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

A

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5 23/1/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

B

Cluster average coordinates:Sr/Ca= 0.00655 ±0.000014Mg/Ca= 1.072 ±0.00874Number of samples = 88

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5 28/2/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

C

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.58/4/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

D

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.57/5/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

E

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.517/6/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

F

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5 15/7/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

G

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5 22/8/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

H

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.519/9/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

I

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5 22/10/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

J

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.522/11/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

K

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.523/12/2002

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

L

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5 30/1/2003

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

M

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.513/3/2003

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

N

0.0055 0.0060 0.0065 0.0070 0.0075 0.0080 0.0085

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

26/12/2001 23/1/2002 28/2/2002 8/4/2002 7/5/2002 17/6/2002 15/7/2002 22/8/2002 19/9/2002 22/10/2002 22/11/2002 23/12/2002 30/1/2003 13/3/2003

Mg/

Ca (e

q/eq

)

Sr/Ca (eq/eq)

O

Fig. 11. The relationship between the Mg/Ca and Sr/Ca ratios in the water column of LK during the period of study. The 14 profiles sampledbetween 26/12/2001 and 13/3/2003 are represented by separate, chronologically ordered and alphabetically labeled diagrams. Each diagramcontains the data points of all the samples collected on the listed date, except for Fig. 11O, which contains the complete data set. Linearregression statistics of the individual diagrams are summarized in Table 6. Fig. 11B presents the data collected in the mixed LK on 23/1/2002,which served to define a reference, or starting point, against which the evolution of the lake during its subsequent stratification was compared.Fig. 11A reflects the decay of stratification toward the end of the 2001 cycle, and Fig. 11M and N represents the mixed lake that started thenext cycle in 2003.


2. (Sr only): Calculation of the mass of Sr2+ and of Ca2+

lost per unit volume of epilimnion water during a givenperiod (Eq. (4)). The corresponding mass of calcite wasthen obtained from the loss of Ca2+, and the mass ofcoprecipitated Sr2+ was obtained from the drop in theconcentration of this solute during the same period. Thisprocedure is not applicable to Mg2+ because of its highambient concentration in the lake, which was almostunaffected by calcite crystallization (Table 1).

3. The same as (1) above but using the Sr/Ca and Mg/Caratios in individual calcite crystals analyzed by LA-ICP-MS rather than in the ‘bulk’ trapped material (Eq. (5)).

Procedures (1) and (3) necessitate the assumption thatthe analyzed calcite formed in the water was similar in com-position to that sampled in the epilimnion, a reasonableassumption considering the short crystallization periods un-der review and the rapid settling of the crystals to the traps.

Method (2) required water analyses only and is thustransparent to errors arising from mineral impurities inthe traps. Integration over different time segments providedverification of the consistency of the distribution coeffi-cients obtained.

Definitions of the distribution coefficients correspondingto the three procedures are provided below, where “Trace”,“Bulk”, “Water” and “Cryst.” designate, respectively, Sr2+

or Mg2+ concentration, ‘bulk’ trap material, epilimnionwater, and individual crystals of calcite analyzed by LA-ICP-MS. The subscripts i and f denote the subscripted ra-tios in the initial (i) and final (f) crystallization solution.Subscripts 1, 2, and 3 refer to method number.

D1 ¼ðTrace

Ca2þ ÞBulk

ðTraceCa2þ ÞWater

; ð3Þ

D2 ¼lnðTracef

TraceiÞWater

lnðCa2þf

Ca2þiÞWater

; ð4Þ

as defined by Doerner and Hoskins (1925) for heteroge-neous crystallization, and

D3 ¼ðTrace

Ca2þ ÞCryst:

ðTraceCa2þ ÞWater

: ð5Þ

The data required for the calculations, along with the re-sults, are provided in Tables 2–4.

3.5.1.1. Consistency of DSr and DMg. The DSr results ob-tained (Tables 2–4) are well within the wide range of datareported in the literature (Section 3.5) and show a reason-able internal consistency. However, the DMg values calcu-lated by method [1] (‘bulk’ material analysis) equal0.0477 ± 27%, which is much higher than the0.00841 ± 23.4% obtained using method [3] (individualcrystal analysis by LA-ICP-MS). Moreover, irrespectiveof the calculation method applied, the DSr results exhibita smaller scatter than those of DMg. If the differences be-tween the DMg values derived using methods 1 and 3 arosefrom erroneous interpretation of the crystallization systemin the epilimnion or of the dissolution process of the crys-tals en route to the traps, one would expect similar inconsis-tencies to appear in the DSr values calculated using the samemethods; however, this effect was not observed. Rather, we


attribute the inconsistent DMg results to the lower purity ofthe ‘bulk’ particulate matter used in method #1 relative tothat of the individually selected, clean calcite crystals ana-lyzed by LA-ICP-MS (method #3). Albeit clean, transpar-ent calcite crystals constitute the majority of the ‘bulk’material, it also contains organic debris visible in theSEM images, including algal and perhaps other chloro-phyll-bearing remains. These may contribute non-calcitic

Mg to the analysis (Stoll et al., 2001; Blanco-Ameijeiraset al., 2012). Notably, algae have been proposed to triggernucleation of calcite in stratified Lake Constance (Stabel,1986, 1989). The lower concentration of Sr in phytoplank-ton (Blanco-Ameijeiras et al., 2012) would mask a similareffect for this solute. Our interpretation does not encourageusing method #1 for the evaluation of the distribution ofMg2+ between calcite crystals and fresh lake water. Rather,it emphasizes the danger in application of Mg/Ca ratios inlacustrine or marine carbonate sediments and rocks for pa-leo-limnological or paleo-oceanographic studies. We be-lieve, however, that LA-ICP-MS analysis of carefullyselected clean crystals as performed via method #3 may re-duce intercrystalline contamination. This conclusion isbased on the following results: (a) the LA results are lowerby a factor of �6 relative to the ‘bulk’ results; (b) the aver-age of the DMg values derived from LA analyses(0.0084 ± 0.0020) is much lower than that obtained undercarefully controlled laboratory conditions using high purityreactants (e.g., 0.057 and 0.068 at 25 and 35 �C, respec-tively; Katz, 1973); (c) contamination by particulate organ-ic (or silicate) linked Mg of the calcite crystals selected forLA analysis would be in conflict with the fact thateven lower DSr values (0.204–0.232) were obtained by LAanalysis than those obtained from the epilimnetic wateranalyses (0.254), integrated over the entire stratified periodin 2002.

Our DSr results (0.194 ± 5.9%, 0.22–0.28, and 0.204–0.232 for methods #1, 2 and 3, respectively) are very similarto Stabel’s (1989) DSr data (DSr = 0.19) for Lake Constancedespite the significant differences in chemical compositionand physical and limnological properties between the twolakes (Muller, 1969; Stabel et al., 1986). This similaritymay be related to Stabel’s (1989) observation that coprecip-

Table 6Linear regression statistics of Mg/Ca–Sr/Ca plots (Fig. 11).

Fig. 11 Date N Intercept

A 26/12/2001 99 �0.18B 23/01/2002 88 1.15C 28/02/2002 88 �0.12D 08/04/2002 97 �0.30E 07/05/2002 103 �0.33F 17/06/2002 117 �0.33G 15/07/2002 110 �0.32H 22/08/2002 123 �0.29I 19/09/2002 125 �0.28J 22/10/2002 122 �0.26K 22/11/2002 100 �0.24L 23/12/2002 85 0.00M 30/01/2003 83 0.36N 13/03/2003 88 0.30

itation with calcite is the principal removal mechanism ofSr2+ from the lake’s water. Our data for LK strongly sup-ports this conclusion.

The similarity between the DSr values obtained from‘bulk’ material and pure calcite crystals indicates the pau-city of Sr-rich aragonite particles in the lake center. Suchgrains, which are quite abundant in the littoral areas ofLK (Rosenthal and Katz, 1989) must have settled well be-fore reaching station A. The same should also be true forother particles of similar hydraulic properties, notably cal-cite detritus.

The consistent distribution coefficients of Sr2+ obtainedusing methods 1–3 and those of Mg2+ using method 3emphasize the usefulness of these routines for estimatingthe role of coprecipitation with calcite in the water column

of LK and in any hardwater lake elsewhere.

3.5.2. The significance of the Ca–Mg–Sr relationship in the

water column

The relationship between the Mg/Ca and Sr/Ca ratios inLK during our study is plotted in Fig. 11A–N. The firstprofile (A) belongs to the previous cycle (2001), while thesecond one (B) represents the homogeneous lake fromwhich the stratification under study evolved. Diagrams11M and 11N belong to the next (2003) cycle, and 11O dis-plays the entire data set. The relevant regression statisticsare presented in Table 6. The position of a data point ona regression line reflects the progress of Ca2+ addition/re-moval to or from the water column and thus depends ontime and water depth (Fig. 4E (Sr/Ca) and 4F (Mg/Ca)).The total lengths of the Mg/Ca–Sr/Ca plots reflect the frac-tionation of Ca2+ between the upper layer of the lake fromwhich calcite precipitated and the bottom waters in which itpartially dissolved.

The migration, in opposite directions, of the data pointsfrom their original position before stratification (i.e., coor-dinates 0.00655/1.072, Fig. 11B) reflects the overall effect ofcalcite crystallization in the epilimnion (shifting upwardsand to the right) and of its dissolution in the hypolimnion(shifting downwards and to the left). Fig. 12A and Table 6illustrate the relation between line extension and time in2002.

Error Slope Error R2

0.0015 185 0.2 0.99990.4534 (�13) 69.2 0.00040.0568 179 8.8 0.82800.0590 209 9.3 0.84180.0028 214 0.4 0.99960.0016 214 0.2 0.99990.0015 212 0.2 0.99990.0076 209 1.1 0.99640.0097 209 1.5 0.99400.0081 207 1.2 0.99580.0088 203 1.3 0.99570.0331 172 5.0 0.93360.2412 112 36.1 0.10570.2520 115 41.0 0.0834


Two features of the Mg/Ca–Sr/Ca relationship during thestratified period in LK stand out (Figs. 11, 12 and Tables 6, 7):

(a) A good linear correlation (R2 > 0.993) between theMg/Ca and Sr/Ca ratios in eight out of the nine pro-files sampled during the stratified period (Fig. 11D–L).

(b) All the regression lines that correspond to the strati-fied lake have intercepts 6�0.23 ordinate units anddisplay an orderly, time-dependent change of their

01/01/2002 01/04/2002 01/07/2002 01/10/2002 01/01/2003

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40calcite crystallization calcite dissolution

Exte

nsio

n (

(Mg/

Ca)

2 +(S

r/Ca)

2 )1/2

Date

A

-

-

-

-

-

-

-

Inte

rcep

t

Mg/

Ca

Sr/Ca

CMAX. SLOPE

(17/6/2002)

MIN. SLOPE

(22/11/2002)

MIXED LAKE

(23/1/2002)

2b

1bHPCD

EPCC

2a

1a

Mg/

Ca

(eq/

eq)

ΔΔ

Fig. 12. (A) Calcite crystallization in the epilimnion (blue dashed line andsolid line and circular data markers) vs. time, as reflected by extension ofCrystallization peaked in mid-June 2002. (B) Rotation of the Mg/Ca–Sr/Cuntil June–July, reversing thereafter until turnover of LK, is expressed bytime. (C) A conceptual scheme of processes responsible for the rotationevolution during the cycle is defined by the lake’s composition before thethe epilimnion in spring follows the EPCC black, dotted arrow, and its dHPCD arrow. While these processes are occurring, anticlockwise rotationSr2+, which were removed from the epilimnion in calcite (dotted arrow 1ain the hypolimnion at a Mg/Sr ratio of 7.5 into the ambient water, in wmaximum slope in mid-June (solid, orange line). At that point, the crystalbegins (arrows 2a and 2b), controlled now by the Mg/Ca and Sr/Ca ratiofreshwater input that dwindles during summer (see (D) below). Finally,(solid, purple arrow). (D) The Mg/Ca–Sr/Ca ratios in mixed LK before strThe blue solid line across the LK marker represents a (hypothetic) puresprings (#s 3–5) and the Jordan River (#14) fall along the extension ofadmixture of the saline springs #s 2, 6, 8, 9, 10 and 12 would support a closlow down. The freshwater supply to the lake from the small Meshushimnegligible in summer. (For interpretation of the references to color in this fi

slopes (Table 6), increasing from the onset of stratifi-cation to June–July and decreasing thereafter towardNovember 2002 (Fig. 12B).

The changes in the slopes and intercepts are antitheticand reflect rotation rather than lateral displacement of thedata points (Fig. 12C). We attribute the good linearity ofthe plots to the major fractionation process operating inLK, i.e., calcite crystallization and dissolution in the water

01/04/2002 01/06/2002 01/08/2002 01/10/2002 01/12/20020.34

0.32

0.30

0.28

0.26

0.24

0.22Intercept

Date

B

200

202

204

206

208

210

212

214

Slope

Slo

pe0.000 0.005 0.010 0.015 0.020 0.025

0.0

0.5

1.0

1.5

2.0D

Sr/Ca (eq/eq)

LK (23/1/2002)

12

4,3,5

14

1615

10

9

6

2

8

rectangular data markers) and dissolution in the hypolimnion (redthe Mg/Ca–Sr/Ca lines presented in Fig. 11 and listed in Table 6.a regression lines vs. time. The direction of rotation, anticlockwise

the orderly and continuous changes in the intercepts and slopes withof the Mg/Ca–Sr/Ca lines: the starting point of the Mg/Ca–Sr/Caonset of stratification (red, filled circle). Crystallization of calcite inissolution in the hypolimnion occurs along the oppositely directedof the line is caused by faster replenishment of Mg2+ than that of

), as well as by the release of Mg2+ and Sr2+ from calcite dissolvinghich this ratio is �160 eq/eq. These processes bring the line to its

lization of calcite slows down, and rotation in the opposite directions in the saline and brackish water sources of LK, rather than by therotation is halted upon turnover and mixing of the lake in winteratification (red full circle) and in its various fresh and saline sources.calcite crystallization/dissolution path. The position of the Fuliyathe LK line and would have little effect on its slope. However, anckwise rotation of the lines once calcite production and dissolutionand Yahudiya streams (#15 and 16, respectively, Fig. 1) becomes

gure legend, the reader is referred to the web version of this article.)

Tab

le7

Th

efr

acti

on

atio

no

fM

g2+

and

of

Sr2

+,

rela

tive

toC

a2+

,in

the

wat

erco

lum

no

fL

ake

Kin

ner

et,

asca

lcu

late

dfr

om

the

regr

essi

on

so

fM

g/C

ao

ver

Sr/

Ca

rati

os

inea

chp

rofi

leb

etw

een

26/1

2/20

01an

d13

/3/2

003.

Th

em

inim

aan

dm

axim

aar

ed

efin

edas

the

low

est

and

hig

hes

tM

g/C

aan

dS

r/C

ara

tio

s,re

spec

tive

ly,

fou

nd

ina

give

np

rofi

le.

Th

eep

ilim

net

ican

dh

ypo

lim

net

icex

ten

sio

ns

wer

e

calc

ula

ted

asE

xten

sion¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

DS

rC

a

��

�� 2þ

DM

gC

a

��

�� 2

q,

wh

ere

the

Dva

lues

des

ign

ate

the

abso

lute

diff

eren

ceb

etw

een

the

resp

ecti

veb

rack

eted

rati

ofo

un

din

agi

ven

pro

file

and

that

wh

ich

pre

vail

edin

the

mix

edla

ke

(Jan

uar

y23

rd,

2002

pro

file

).

Dat

eR

2(M

g/C

a)m

inim

um

(Sr/

Ca)

min

imu

m(M

g/C

a)m

axim

um

(Sr/

Ca)

max

imu

mE

pil

imn

ion

exte

nsi

on

Hyp

oli

mn

ion

exte

nsi

on

To

tal

exte

nsi

on

Ext

ensi

on

rati

o

26/1

2/20

011.

000

0.86

20.

0056

81.

090.

0069

00.

018

0.21

00.

228

0.08

623

/01/

2002

0.00

01.

072

0.00

655

1.07

0.00

655

0.00

00.

000

0.00

0n

/a28

/02/

2002

0.82

80.

964

0.00

606

1.06

0.00

649

0.01

50.

108

0.12

30.

140

08/0

4/20

020.

842

1.01

00.

0062

91.

090.

0066

40.

017

0.06

20.

078

0.27

107

/05/

2002

1.00

00.

950

0.00

598

1.26

0.00

748

0.19

00.

122

0.31

21.

549

17/0

6/20

021.

000

0.92

00.

0058

81.

450.

0083

30.

380

0.15

20.

531

2.50

315

/07/

2002

1.00

00.

912

0.00

582

1.31

0.00

763

0.24

30.

160

0.40

21.

521

22/0

8/20

020.

996

0.91

10.

0056

91.

300.

0076

60.

232

0.16

10.

393

1.43

819

/09/

2002

0.99

40.

831

0.00

539

1.31

0.00

750

0.23

70.

241

0.47

80.

985

22/1

0/20

020.

996

0.85

50.

0054

21.

270.

0073

80.

200

0.21

70.

417

0.92

422

/11/

2002

0.99

60.

831

0.00

531

1.20

0.00

707

0.13

20.

241

0.37

30.

550

23/1

2/20

020.

934

1.01

90.

0060

11.

160.

0066

80.

092

0.05

30.

145

1.73

830

/01/

2003

0.10

61.

107

0.00

668

1.11

0.00

668

0.03

50.

035

0.07

01.

000

13/0

3/20

030.

083

1.00

80.

0061

61.

010.

0061

60.

064

0.06

40.

127

1.00

0

Th

eli

nea

rco

rrel

atio

nco

effici

ent

bet

wee

nth

eM

g/C

aan

dS

r/C

ain

each

pro

file

isgi

ven

inth

ese

con

d(“

R2”

)co

lum

n.


column. An admixture of hypolimnetic water into the epi-limnion (Table 5, #1) would not affect the linearity northe slope or intercept of the regression lines. Rather, theywere caused by two processes that operated independentlyin the epilimnion (arrows 1a and 2a in Fig. 12C) and inthe hypolimnion (arrows 1b and 2b, same figure). We relatethe line rotation in the epilimnion to the unbalanced ratesof loss (due to coprecipitation in calcite) and supply (by in-flux into the lake) of Sr2+ and of Mg2+, where both ratesare normalized relative to those of Ca2+, as describedhenceforth.

The supply of Sr2+ to the epilimnion from the watersources listed in Table 1 varies from 0.0012 to 0.036 eqSr2+ per eq Ca2+ and that of Mg2+ varies from 0.25 to2.7 eq Mg2+ per eq Ca2+. Similarly, the rates of loss of bothmetals from the epilimnion in calcite can be estimated fromthe Sr:Mg:Ca proportion in this mineral. Taking 1200 and2500 ppm as representative concentrations of Sr2+ andMg2+ in the crystals, respectively (Table 4), one obtains re-moval rates of 0.00137 eq Sr2+ and 0.0103 eq Mg2+ per eqCa2+. Division of the supply rate by the loss rate of Sr2+

and of Mg2+ reveals that the compensation of Mg2+ lostin calcite by the contemporaneous Mg2+ influx into LK is5–40 times larger than that of Sr2+, depending on the spe-cific water source under consideration.

It has been demonstrated (Koren and Ostrovsky, 2000)that the main flux of calcite in LK occurs in spring. Thesame conclusion is also supported by the drop in the epilim-nion Ca2+ concentration and the degree of saturation of itswater with respect to calcite (Fig. 10A and B). Because thecoefficients DMg and DSr are <1 (Tables 2–4), both the Mg/Ca and Sr/Ca ratios in the epilimnion water should increaseupon calcite crystallization. The DMg/DSr ratio, however, isas low as 0.04 (Table 4), implying that a much larger frac-tion of the epilimnetic Sr2+ than of epilimnetic Mg2+ is re-moved in calcite, thereby explaining the initiallyanticlockwise rotation and the corresponding decrease inintercept values during that period (Fig. 12B).

The April through June anticlockwise rotation of theMg/Ca–Sr/Ca lines due to coprecipitation of Mg2+ andSr2+ with calcite in the epilimnion (Fig. 12C, arrow 1a) iscomplemented by the concomitant release of these ions intothe hypolimnion water upon calcite dissolution (Fig. 12C,arrow 1b). The Mg/Sr ratio in this mineral is approximately7.5 eq/eq, which is much lower than that in the surroundingwater (�160 eq/eq), and dissolution must therefore drag thehypolimnetic section of the regression lines (HPCD inFig. 12C) downwards during the same period.

The decreasing production of calcite in the epilimnion asof July (Fig. 12A) decelerated the loss of Mg2+ and Sr2+

from the epilimnion due to coprecipitation and, thereby,also decelerated their addition to the hypolimnion waterby dissolution. Simultaneously, the freshwater supply de-creased (Fig. 2D), allowing the saline springs, almost allof which had Mg/Sr ratios lower than the maximum Mg/Sr ratio attained in June, to assume control of the Mg/Ca–Sr/Ca line slopes. Fig. 12D demonstrates that the con-tribution of Mg2+ and Sr2+ from saline sources 2, 6, 8, 9and 12 (Table 1) to the epilimnion would result in aclockwise rotation of the slope. The Fuliya saline springs


(#s 3–5) and the Jordan River waters (#14) fall almost pre-cisely on the initial (mixed lake) slope, and their inflowwould thus exert little effect on line rotation.

4. CONCLUSIONS

The following conclusions, relevant to the cycling of sol-utes in monomictic, hardwater lakes, were derived from thebehavior of Ca2+, Mg2+ and Sr2+ in Lake Kinneret (the Seaof Galilee) during a complete limnological cycle (fall 2001through spring 2003):

1. Starting from a homogeneous, well-mixed lake in winter,the subsequent cycling of dissolved solutes is controlledby stratification and separation of the water column intoan epilimnion and an underlying hypolimnion.

2. The downward transport of Mg2+ and Sr2+ in the watercolumn is facilitated by coprecipitation with calcite dur-ing its (biologically-induced) crystallization in the epi-limnion in spring and early summer.

3. Upon arrival in the anoxic hypolimnion, the settling cal-cite crystals partially dissolve, releasing their Ca2+, Mg2+

and Sr2+ load into the ambient water.4. Upward cycling of solutes from the hypolimnion into the

epilimnion is driven by the deepening of the thermoclinethroughout the stratified period of the lake, resulting inan admixture of hypolimnetic waters into the down-wards expanding epilimnion, but leaving the shrinkinghypolimnion less affected.

5. The role of coprecipitation with calcite in the cycling ofMg2+ and Sr2+ in a monomictic lake can be assessed bycalculating their effective distribution coefficients (DMg

and DSr) between calcite and the lake water, rather thanresorting to published data, which display large varia-tions, depending on the experimental conditions usedfor their derivation. Effective DMg and DSr values canbe obtained by perceiving the epilimnion as a crystalliza-tion vessel and the hypolimnion as a collecting flask ofthe settling crystals. Chemical analyses of the epilimnionwater and of coeval, trapped calcite crystals are thenused to calculate the desired D values.

6. While the DSr values obtained are internally consistent, irre-spective of the calculation and materials analyzed, meaning-ful DMg values could be obtained only by comparing theMg/Ca ratios in the epilimnetic waters with the Mg/Caratios in clean calcite crystals, individually selected underthe LA-ICP-MS microscope from the trapped particulates.Utilization of bulk analyses instead of analyzing selectedcrystals resulted in excessively high, inconsistent DMg values,most likely caused by the presence of Mg-bearing phyto-plankton debris intermingled between the calcite crystals.The non-existence of this problem with Sr2+ is attributedto its much lower concentrations in organic matter. Evalua-tion of the geochemical evolution of ancient, aqueous depo-sitional environments based on ‘bulk’ Mg/Ca ratiosanalyses of limestone beds should therefore be expectedto suffer from the same problem.

7. The relationship between the Ca2+, Mg2+ and Sr2+ con-centrations in the water column during the stratified per-iod of Lake Kinneret and similar hardwater lakes can be

described by high-quality regression lines (R2 > 0.99)plotted in Mg/Ca–Sr/Ca coordinates. Each line repre-sents the water column of the lake at a specific time.The Mg/Ca and Sr/Ca ratios that prevailed in the mixedlake before stratification define a starting point fromwhich fractionation of Ca2+, Mg2+ and Sr2+ in the watercolumn evolved after the onset of stratification. The sub-sequent crystallization of calcite in the epilimnion anddissolution of this mineral in the hypolimnion arereflected by data points shifted upwards or downwards,respectively, along the line. The Mg/Ca–Sr/Ca regres-sion lines display an orderly rotation in time duringthe stratified period, rotating in an anticlockwise direc-tion during the high calcite flux period in spring. Therotation stems from the higher rate of Mg2+ replenish-ment to the epilimnion, relative to that of Sr2+, to com-pensate for the removal of these solutes in calcite. Theanticlockwise rotation is further enhanced by the addi-tion of Mg2+ and Sr2+ to the hypolimnion water fromthe dissolving calcite crystals, in which the Mg/Sr ratio(7.5 eq/eq) is much lower than in the surrounding water(�160 eq/eq). Upon the slowdown of calcite crystalliza-tion and the dwindling freshwater supply in summer,the direction of rotation is reversed under the dominantcontrol of the brackish and saline water sources of thelake. Subsequent turnover and mixing of the lake in win-ter reset the Mg/Ca and Sr/Ca ratios to their valuesbefore the 2002 cycle.

ACKNOWLEDGMENTS

The authors are grateful to the technical staff of the Institute ofEarth Sciences at the Hebrew University, A. Agranat, E. Kasher,Y. Sherer, and S. Meiri, for their assistance in the chemical analysisof the water samples and for preparing and maintenance of the fieldand lab equipment. Hebrew University Geology students N. Wald-man, A. Haliva-Cohen, A. Torfstein, N. Morag, M. Finzi-Schos-berger, and E. Shelef, and M. Hatab and J. Easton from theYigal Allon Kinneret Laboratory, were of great assistance in thefield work. The LA-ICP-MS analysis was performed by O. Dvir.Invaluable information and data were provided by B. Teltch andD. Berger from the Mekorot Water Co. and by the Lake KinneretYigal Allon Laboratory scientific and technical staff. Partial fund-ing of the project was provided by Israel Water Authority, Grant#65/01 to A.K.

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.gca.2012.11.045.

REFERENCES

American Public Health Association, APHA (1971) Standard

Methods for the Examination of Water and Wastewater, 14thed. APHA, Washington, DC.

Armitage K. B. and House H. B. (1962) A limnological reconnais-sance in the area of McMurdo Sound, Antarctica. Limnol.

Oceanogr. 7, 36–41.




Assouline S. (1993) Estimation of lake hydrologic budget termsusing the simultaneous solution of water, heat, and saltbalances and a Kalman filtering approach: application to LakeKinneret. Water Resour. Res. 29, 3041–3048.

Bar Giora-Magal E. (2001) Characterization and location ofinternal salinity sources in Lake Kinneret. M.Sc. thesis, TheHebrew University of Jerusalem (in Hebrew, English abstract),154p.

Berger D. and Shaw M. (2003) Water, salt and energy balances ofLake Kinneret for year 2001/2002. Ann. Report, MekorotWater Company, Jordan District, Lake Kinneret WatershedUnit.

Berger D., Rom M. and Teltch B. (2010) Water, salt and energybalances of Lake Kinneret for year 2001/2002. Ann. Report,Mekorot Water Company, Jordan District, Lake Kinneretwatershed Unit.

Blanco-Ameijeiras S., Lebrato M., Stoll H. M., Iglesias-Rodri-gueza M. D., Mendez-Vicente A., Sett S., Muller M. N.,Oschlies A. and Schulz K. G. (2012) Removal of organicmagnesium in coccolithophore calcite. Geochim. Cosmochim.

Acta 89, 226–239.

Boehrer B., Matzinger A. and Schimmele M. (2000) Similaritiesand differences in the temperature cycles of East Germanmining lakes. Limnologica 30, 271–279.

Boehrer B. and Schultze M. (2008) Stratification of lakes. Rev.

Geophys. 46, RG2005.

Curti E. (1999) Coprecipitation of radionuclides with calcite:estimation of partition coefficients based on a review oflaboratory investigations and geochemical data. Appl. Geo-

chem. 14, 433–445.

Doerner H. A. and Hoskins W. M. (1925) Coprecipitation ofradium and barium sulfates. J. Am. Chem. Soc. 47, 662–675.

Dittrich M. and Obst M. (2004) Are picoplankton responsible forcalcite precipitation in lakes?. Ambio 33 559–564.

Dittrich M., Kurz Ph. and Wehrli B. (2004) The role of autotrophicpicocyanobacteria in calcite precipitation in an oligotrophiclake. Geomicrobiol. J. 21, 45–53.

Eaton A. D., Clesceri L. S. and Greenberg A. E. (1995) Standard

Methods for the Examination of Water and Wastewater, 19th ed.American Public Health Association, Washington.

Friedmann A. (1913) Beitrage zur Chemischen und PhysikalischenUntersuchung der Thermen Palastinas. Gesselschaft Palastina-

Forschung, Publ. #4, S. Calvary, Berlin, 94p.Goldschmidt M. J., Arad A. and Neev D. (1967) The mechanism of

saline springs in the Lake Tiberias depression. Geol. Surv. Isr.

Bull. 45.

Hadas O. and Pinkas R. (1995) Sulfate reduction processes in thehypolimnion and sediments of Lake Kinneret, Israel. Freshw.

Biol. 33(1), 63–72.

Hodell D., Schelske C. L., Fahnenstiel G. L. and Robbins L. L.(1998) Biologically induced calcite and its isotopic compositionin Lake Ontario. Limnol. Oceanogr. 43, 187–199.

Hohman R., Kipfer R., Peeters F., Piepke G. and Imboden D. M.(1997) Processes of deep-water renewal in Lake Baikal. Limnol.

Oceanogr. 42, 841–855.

Holland H. D. (1966) The coprecipitation of metallic ions withCaCO3. Final Report, Department of Geology, PrincetonUniversity, New Jersey, 53p.

Holland H. D., Holland H. J. and Munoz J. (1964) Thecoprecipitation of cations with CaCO3: II. The coprecipitationof Sr2+ with calcite between 90 and 100 �C. Geochim. Cosmo-

chim. Acta 28, 1287–1302.

Huang Y. and Fairchild I. J. (2001) Partitioning of Sr2+ and Mg2+

into calcite under karst-analogue experimental conditions.Geochim. Cosmochim. Acta 65, 47–62.

Hutchinson G. E. (1957) A treatise on limnology. In Geography

Physics and Chemistry, vol. 1. John Wiley, New York.

Katz A. (1973) The interaction of magnesium with calcite duringcrystal growth at 25–90 �C and one atmosphere. Geochim.

Cosmochim. Acta 37, 1563–1586.

Katz A., Sass E., Starinsky A. and Holland H. D. (1972) Strontiumbehavior in the aragonite–calcite transformation: an experi-mental study at 40 and 90 �C. Geochim. Cosmochim. Acta 36,

481–496.

Katz A. and Starinsky A. (2009) Geochemical history of the DeadSea. Aquat. Geochem. 15, 159–194.

Kitano Y., Kanamori N. and Oomori T. (1971) Measurements ofdistribution coefficients of strontium and barium betweencarbonate precipitate and solution – abnormally high valuesof distribution coefficients measured at early stages of carbon-ate formation. Geochem. J. 4, 183–206.

Klein-BenDavid O., Gvirtzman H. and Katz A. (2005) Geochem-ical identification of fresh water sources in brackish ground-water mixtures; the example of Lake Kinneret (Sea of Galilee),Israel. Chem. Geol. 214, 45–59.

Klein-BenDavid O., Sass E. and Katz A. (2004) The evolution ofmarine evaporitic brines in inland basins: the Jordan-Dead Searift valley. Geochim. Cosmochim. Acta 68, 1763–1775.

Kolodny Y., Katz A., Starinsky A., Moise T. and Simon E. (1999)Chemical tracing of salinity sources in Lake Kinneret (Sea ofGalilee), Israel. Limnol. Oceanogr. 44, 1035–1044.

Koren N. and Ostrovsky I. (2000) Sedimentation in a stratifiedsubtropical lake. Verh. Int. Ver. Limnol. 27, 2636–2639.

Kuchler-Krischun J. and Kleiner J. (1990) Heterogeneously nucle-ated calcite precipitation in Lake Constance, a short resolutionstudy. Aquat. Sci. 52, 176–197.

Lemckert C., Antenuchi J., Saggio A. and Imberger J. (2004)Physical properties of turbulent benthic boundary layersgenerated by internal waves. J. Hydraul. Eng. 130, 58–69.

Lorens R. B. (1981) Sr, Cd, Mn and Co distribution coefficients incalcite as a function of calcite precipitation rate. Geochim.


Mazor E. and Mero F. (1969) Geochemical tracing of mineral andfresh water sources in the Lake Tiberias basin, Israel. J. Hydrol.

7, 276–317.

Moise T., Starinsky A., Katz A. and Kolodny Y. (2000) Raisotopes and Rn in brines and ground waters of the Jordan-Dead Sea Rift Valley: enrichment, retardation and mixing.Geochim. Cosmochim. Acta 64, 2371–2388.

Morse J. W. and Bender M. L. (1990) Partition coefficients incalcite: examination of factors influencing the validity ofexperimental results and their application to natural systems.Chem. Geol. 82, 265–267.

Mucci A. and Morse J. W. (1983) The incorporation of Mg2+ andSr2+ into calcite overgrowths: influences of growth rate andsolution composition. Geochim. Cosmochim. Acta 47, 217–233.

Muller G. (1969) High strontium contents and Sr/Ca-ratios in LakeConstance waters and carbonates and their sources in thedrainage area of the Rhine River (Alpenrhein). Mineral.

Deposita (Berl.) 4, 75–84.

Muller B., Wang Y. and Wehrli B. (2006) Cycling of calcite in hardwater lakes of different trophic states. Limnol. Oceanogr. 51,

1678–1688.

Muller B., Wang Y., Dittrich M. and Wehrli B. (2003) Influence oforganic carbon decomposition on calcite dissolution in surficialsediments of a freshwater lake. Water Res. 37, 4524–4532.

Nishri A., Imberger J., Eckert W., Ostrovsky I. and Geifman Y.(2000) The physical regime and the respective biogeochemicalprocesses in the Lower Water Mass of Lake Kinneret. Limnol.

Oceanogr. 45, 972–981.


Nishri A., Stiller M., Rimmer A., Geifman Y. and Krom M. (1999)Lake Kinneret (the Sea of Galilee): the effects of diversion ofexternal salinity sources and the probable chemical compositionof the internal salinity sources. Chem. Geol. 158, 37–52.

Ohlendorf C. and Sturm M. (2001) Precipitation and dissolution ofcalcite in a Swiss high Alpine lake. Arc. Antarc. Alpine Res. 33,

410–417.

PHREEQC program (online) Aqueous Geochemical modeling.Dept. Geosciences, North Dakota State University. Availablefrom: <http://www.ndsu.nodak.edu/webphreeq/>.

Ramisch F., Dittrich M., Mattenberger C., Wehrli B. and Wuest A.(1999) Calcite dissolution in two deep eutrophic lakes. Geochim.


Rimmer A., Eckert W. and Nishri A. (2006) Evaluation ofhypolimnetic diffusion parameters in thermally stratified lakes.Limnol. Oceanogr. 51, 1906–1914.

Rimstidt J. D., Balog A. and Webb J. (1998) Distribution of traceelements between carbonate minerals and aqueous solutions.Geochim. Cosmochim. Acta 62, 1851–1863.

Rosenthal Y. and Katz A. (1989) The applicability of traceelements in freshwater shells for paleogeochemical studies.Chem. Geol. 78, 65–76.

Serruya, C. (ed.) (1978) Lake Kinneret Monographiae Biologicae.Junk Publishers, The Hague, 502p.

Simon E. and Mero F. (1992) The salinization mechanism of LakeKinneret. J. Hydrol. 138, 327–343.

Stabel H.-H. (1986) Calcite precipitation in Lake Constance:chemical equilibrium, sedimentation and nucleation by algae.Limnol. Oceanogr. 31, 1081–1093.

Stabel H.-H. (1989) Coupling of strontium and calcium cycles inLake Constance. Hydrobiologia 176/177, 323–329.

Stabel H.-H., Kuchler-Krischum J., Kleiner J. and Merkel P.(1986) Removal of strontium by coprecipitation in LakeConstance with calcite. Naturwissenschaften 73, 551–553.

Starinsky A. (1974) Relationship between Ca-chloride brines andsedimentary rocks in Israel. Ph.D. thesis, The Hebrew Univer-sity of Jerusalem.

Stiller M., Rosenbaum J. M. and Nishri A. (2009) The origin ofbrines underlying Lake Kinneret. Chem. Geol. 262, 309–325.

Stoll H. M., Elcinar J. R., Alonso J. I. G., Rosenthal Y., Probert I.and Klaas C. (2001) A first look at paleotemperature prospectsof Mg in coccolith carbonate: cleaning techniques and culturemeasurements. G3 Electronic J. 2.

Tang J., Koehler S. J. and Dietzel M. (2008) Sr2+/Ca2+ and44Ca/40Ca fractionation during inorganic calcite formation: I.Sr incorporation. Geochim. Cosmochim. Acta 72, 3718–3732.

Tesoriero A. J. and Pankow J. F. (1996) Solid solution partitioningof Sr2+, Ba2+, and Cd2+ to calcite. Geochim. Cosmochim. Acta

60, 1053–1063.

Yeats P. S. and Imberger J. (2003) Pseudo two-dimensionalsimulations of internal and boundary fluxes in stratified lakesand reservoirs. Intl. J. River Basin Manage. 1, 297–319.

Associate editor: Robert H. Byrne

http://www.ndsu.nodak.edu/webphreeq

Calcium, magnesium and strontium cycling in stratified, hardwater lakes: Lake Kinneret (Sea of Galilee), Israel

Documents