r- and γ-HexachlorocyclohexaneMeasurements in the Brine Fractionof Sea Ice in the Canadian HighArctic Using a Sump-Hole TechniqueM . P U C K O , * , † , ‡ G . A . S T E R N , † , ‡

R . W . M A C D O N A L D , ‡ , § A N DD . G . B A R B E R ‡

Freshwater Institute, Department of Fisheries and Oceans, 501University Crescent, Winnipeg, Manitoba, R3T 2N6, Canada,Centre for Earth Observation Science, University of Manitoba,460 Wallace Building, 125 Dysart Road, Winnipeg, R3T 2N2,Canada, and Institute of Ocean Sciences, Department ofFisheries and Oceans, 9860 West Saanich Road, Sidney,British Columbia, Canada, V8L 4B2

Received July 6, 2010. Revised manuscript receivedSeptember 25, 2010. Accepted November 2, 2010.

We used holes augered partially into first-year sea ice(sumps) to determine R- and γ-HCH concentrations in sea-icebrine. The overwintering of the CCGS Amundsen in theCanadian western Arctic, as part of the Circumpolar FlawLead (CFL) System Study, provided the circumstances to allowbrine to accumulate in sumps sufficiently to test themethodology. We show, for the first time, that as much as50% of total HCHs in seawater can become entrapped withinthe ice crystal matrix. On average, in the winter first-yearsea ice HCH brine concentrations reached 4.013 ( 0.307 ng/Land 0.423 ( 0.013 ng/L for the R- and γ-isomer, respectively.In the spring, HCHs decreased gradually with time, with increasingbrine volume fraction and decreasing brine salinity. Thesedecreasing concentrations could be accounted for by both thedilution with the ice crystal matrix and under-ice seawater.We propose that the former process plays a more significantrole considering brine volume fractions calculated in thisstudy were below 20%. Levels of HCHs in the brine exceed under-ice water concentrations by approximately a factor of 3, acircumstance suggesting that the brine ecosystem has been,and continues to be, the most exposed to HCHs.

Introduction

Hexachlorocyclohexane (HCH) was used as an insecticidebetween 1950 and 2000 in two main formulations: lindane(pure γ isomer) and technical HCH (R isomer 60-70%, �isomer 5-12%, and γ isomer 10-15%). HCHs are transportedto Arctic regions by winds and ocean currents and can beremoved from the marine systems by bacterial degradation,outgassing, and Arctic Ocean outflow (1-4). Although levelsof HCHs in the Arctic declined rapidly over the last 20 years(5), they remain high enough to make HCH a “contaminantof concern” (6). HCH high concentrations and isomericdifferences in physical and chemical properties (7) provide

unique opportunities to develop insight into contaminantpathways in Arctic ecosystems (8).

There is a large body of research and modeling availableon the accumulation of HCHs in the Arctic Ocean (1-7). Incontrast, almost nothing is known about the accumulationof HCHs in sea ice. Sea ice has an intricate and highly variableinternal structure consisting of ice crystals and inclusions ofbrine and air, often referred to as total porosity (9). Brinepockets, with mean major axis lengths of hundreds ofmicrometers, are much smaller than air bubbles (millimeters).Brine inclusions are sandwiched between the ice crystals,and their size and degree of connectivity are stronglyassociated with temperature at the time of ice formation. Asice warms and its brine volume increases, the size distributionshifts toward larger brine pockets. As sea ice passes a criticalbrine volume fraction of 5%, a marked transition occurs (10).Above 5%, brine can move through the ice, individual brinepockets may coalesce, and gravity drainage and replenish-ment of brine with seawater becomes possible. The relativelysmall air volume typical of first-year sea ice that has notexperienced any significant periods of warming is usuallynegligible compared to brine volume (11). As sea ice warmsto about -3 °C in the spring, air volume may account forapproximately 20% of the total porosity (11).

In general, salts become entrapped in the brine fractionof sea ice as the ice forms (12), however almost nothing isknown about concurrent partitioning of HCHs. To the bestof our knowledge, there has been only one study of thesecontaminants in Arctic sea ice (13). The authors emphasizedthe importance of brine processes for accumulation/rejectionof HCHs during ice formation. They showed that HCH isgenerally associated with the brine fraction of sea ice;however, they also noted that as much as 40% of total HCHinventory can remain within the ice crystal matrix, whichcontains only 7% of the total salts, until the final stages ofmelt. However, to our knowledge, no direct measurementsof HCHs in the brine have ever been published.

Here, we report the concentrations of R- and γ-HCHmeasured directly in the brine using an ice sump. Reliabilityof this sampling technique for HCHs under winter and springice conditions is discussed. Based on total ice measurements,the partitioning of HCHs between the brine and ice crystalmatrix is also estimated. Finally, the potential for ice crystalmelting and seawater replenishment to dilute the HCH inbrine during spring is described with respect to the measureddecrease of HCH brine concentrations.

Experimental SectionSamples were collected from the CCGS Amundsen as a partof the International Polar Year (IPY) Circumpolar Flaw Lead(CFL) System Study in the Canadian High Arctic (BeaufortSea) between January 6 and 11 (winter, one station) and theApril 21 and May 14 (spring, 4 stations), 2008 (Figure 1).Winter samples consisted of a set of 10 measurements takenevery 12 h for 5 days. Brine samples were measured for R-and γ-HCH and salinity (90 cm depth); bulk ice sampleswere measured for R- and γ-HCH concentration (0-10 cmand 70-108 cm depth), ice bulk salinity and temperature;seawater beneath the ice was measured for R- and γ-HCHconcentration and salinity; and air was measured for R- andγ-HCH concentration (Figure 1). Spring measurementsincluded R- and γ-HCH brine concentration (80 and 100 cmdepth) and corresponding brine salinity, ice bulk salinityand temperature, and under-ice seawater R- and γ-HCHconcentration (Figure 1).

* Corresponding author e-mail: umpucko@cc.umanitoba.ca.† Freshwater Institute, Department of Fisheries and Oceans.‡ Centre for Earth Observation Science, University of Manitoba.§ Institute of Ocean Sciences, Department of Fisheries and Oceans.

Environ. Sci. Technol. 2010, 44, 9258–9264

9258 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 24, 2010 10.1021/es102275b 2010 American Chemical SocietyPublished on Web 11/15/2010

Field Sampling. Sea ice samples were obtained with a9-cm i.d., Mark II coring system (Kovacs Enterprises,Lebanon, NH). After collection, ice cores were cut into 4layers (0-10, 10-40, 40-70, and 70-108 cm based ondifferences in vertical ice texture) using a metal saw, placedin tightly closed containers, and melted. Ice cores were storedat -23 °C in horizontal position before cutting to minimizeartifacts related to vertical brine movement. Brine sampleswere obtained by coring a hole in the ice to a prescribeddepth, closing it tightly with a styrofoam cap, and pumpingthe brine out of it after 2-12 h using a syringe and 100 cmof Teflon tubing. Seawater was collected from beneath theice using a Wildco Kemmerer water sampler (Ben MeadowsCompany, Yulee, FL), through the hole immediately after icecore extraction. Sea ice temperature profiles were measuredimmediately after extracting the ice core by placing aTraceable digital thermometer (model 4000, Control Com-pany, Friendswood, TX ((0.05 °C)), in 5-mm diameter holesdrilled at 5-10 cm intervals starting from the bottom of theice core. Air was sampled by drawing about 2500 m3 (samplingduration about 7 h, flow rate about 7 m3/min) through aWhatman EPM 2000 glass fiber (GF) filter followed bypolyurethane foam (PUF) plugs of 7.5-cm i.d. and 6.5-cmlength (Tish Environmental) using the Gast Regenair Blower(Cole-Parmer, Vernon Hills, IL). The pump was mounted onthe bow of the ship at deck level (∼ 10 m above the sea level).The blower was calibrated using the Sierra 620 Mass FlowMeter (Sierra Instruments, Monterey, CA).

Laboratory Analysis. Salinity was calculated from con-ductivity and temperature using a HACH SENSION5 portableconductivity meter (Hach, Loveland, CO,( 0.01). Melted seaice samples, under-ice seawater samples (4 L), and brinesamples (2-4 L) were spiked with d6-R-HCH surrogate,pumped through 0.7-µm GFF filters (Whatman) to removeparticulate matter, and filtered through solid-phase extraction(SPE) cartridges (Oasis) at a flow rate <20 mL/min. Extractionof HCHs was done by elution of SPE cartridges with acetone.

Subsequently, they were transferred into iso-octane. Finally,the solution was cleaned up with concentrated sulphuricacid (14).

In the case of air samples, PUF filters were spiked withrecovery standard (PCB 30 congener), extracted using ASE2000 Accelerated Solvent Extractor (Dionex) with hexane,reduced to 1 mL volume by rotoevaporation, and cleaned upand fractionated on Florisil column chromatography.

Quantitative analysis of R- and γ-HCH was performedusing a Varian 3800 gas chromatograph-electron capture (GC-ECD) equipped with a 60-m, JW DB5 column (0.25-mm i.d.,0.25-µm film thickness). For details see description in ref 13.

R- and γ-HCH concentrations presented in this study werecorrected for percent recovery but not for mean blank.Average recovery was 80% for brine, 83% for the seawater,73% for ice, and 73% for air. Limits of detection (LOD), definedas 3 standard deviations of the mean field blank, were 0.019and 0.007 ng/L for R- and γ-HCH, respectively. Limits ofquantification (LOQ), defined as 10 standard deviations ofthe mean field blank value, were estimated at 0.047 and 0.011ng/L for R- and γ-HCH, respectively.

Calculations and Data Analysis. Brine volume fraction,υBRINE, was calculated as follows (15):

where υBRINE is brine volume fraction (%) ((0.06%), SICE is icesalinity, and TICE is ice temperature (°C).

Mixing ratios (V2/V1) of brine and seawater, or brine andmelting ice crystal matrix, of salinity S2 and salinity S1, requiredto produce the volume of liquid (V1 + V2) of salinity S3, werecalculated as follows:

Concentrations of HCHs in the brine (CBRINE1) resultingfrom previously calculated mixing ratios of brine (VBRINE0)

FIGURE 1. Sampling locations in the winter (D14) and spring (D41, D43, F2, F3) 2008 (A) and scheme of sampling at winter and springstations (B).

υBRINE ) SICE(49.185/|TICE| + 0.532) · 0.1

(-0.5 °C>TICE> - 22.9 °C) (1)

V2/V1 ) (S1 - S3)/(S3 - S2) (2)

VOL. 44, NO. 24, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 9259

and seawater (VSW) or melting ice crystal matrix (VIM)accounting for the observed decrease in salinity between thesampling periods were calculated as follows:

where CBRINE0 is the HCH concentration in the brine measuredin the previous sampling period, CIM is the HCH concentrationin the ice crystal matrix calculated in this paper (0.162 and0.017 ng/L for R-HCH and γ-HCH, respectively), and CSW isthe HCH concentration in the under-ice seawater measuredin the previous sampling period.

Relative brine water to air concentrations at equilibriumpartitioning (fugacity ratio, fw/fa ) 1) in the sump werecalculated for both HCH isomers (1):

where Cw and Ca are the concentrations of dissolved andgaseous HCHs in the brine water and air, respectively, R isthe gas constant (8.314 Pa m3/deg mol), T is the airtemperature (249 K), and HLC is the Henry’s Law Constantat the closest available temperature and salinity (0.104 and0.0627 Pa m3/mol for 0.5 °C seawater for R- and γ-HCH,respectively) (16).

Coefficient of variation (CV) was used as a measure ofHCH variability in the brine (17):

where SD is the standard deviation.Statistical analyses include Pearson’s correlation matrix

with Bonferroni probability test, and two-sample t-test(SYSTAT 11, Systat Software Inc.). Prior to statistical analyses,R- and γ-HCH concentrations, ice bulk salinity, ice tem-perature, brine salinity, and brine volume fraction were log-transformed to meet the normal distribution criterion.

Results and DiscussionReliability of the Sumps for HCH Sampling. To ourknowledge, sumps have not previously been used to measureHCH concentrations directly in sea-ice brine. The sumptechnique allows for the collection of sufficient volumes ofbrine such that LOQs for HCH are exceeded. However, themethod might be vulnerable to seepage of brine from upperlayers of ice or from the underlying seawater, to melting ofthe ice crystal matrix in direct contact with the brine in thesump hole, and to air-brine gas exchange (Figure 2).

In the winter, the mean brine volume fraction for thebulk sea ice layer corresponding to the depth of sump holeswas 4.3%. Thus, based on conservative partitioning of saltsbetween brine and the ice crystal matrix, salinity measured

in the sump holes should be roughly 23 times moreconcentrated than that of the corresponding bulk ice samples(12). Our samples exhibited a ratio of 29, implying that thebrine from a 90-cm deep hole contained a portion of moreconcentrated brine from the upper core layers. Consideringthe most extreme mixing scenario, brine from the top 10 cmof ice with salinity of 293.6 (based on a mean ice bulk salinityand brine volume fraction of 8.5 and 3.0%, respectively), andthe requirement to increase the brine salinity from 101.6(23-fold increase) to 128.1 (29-fold increase), eq 2 impliesthat 0.16 parts of brine from the top 10 cm would have tomix with one part of brine from the 90-cm sump hole. ForHCHs, this would result in a maximum increase of 13% and8% for the R- and γ-isomers, respectively. Seeping ofunderlying seawater, with average salinity of 28.8 (Table 1),or the melting of the ice crystal matrix into the sump hole,would result in a significant decline in brine salinity andtherefore need not be considered further. Using eq 4, and anaverage air temperature of-24 °C during the sampling period,the sump-hole ratios of HCH concentrations in the brinewater to air are estimated to be approximately 20 × 103 and33 × 103 for theR- and γ-isomers at equilibrium, respectively.Assuming a rough air/brine volume ratio in the sump holeof 11.5, brine-air gas exchange could not decrease brineHCH concentrations by more than 0.1%.

In the spring, average brine salinity was 16 times greaterthan in the corresponding bulk ice sample compared to theexpected ratio of 10 based on an average brine volume fractionof 9.7% (Table 1). In this case, we calculated that 0.35 partsof brine from the top 10 cm, with salinity of 168.1, wouldneed to have mixed with one part of brine from the 90-cmsump hole to increase the final salinity from 50.9 (expected)to 81.5 (measured). As in the winter season, this would notbe expected to have any significant impact on HCH brineconcentrations.

These calculations suggest, therefore, that the sumptechnique provides a reliable method to measure HCHconcentrations in the brine under both winter and spring iceconditions.

Distribution of HCHs between the Brine Fraction andIce Crystal Matrix. HCH levels in the brine in the winterwere approximately 13 times higher than in bulk ice (Table1), reaching 4.013 ( 0.307 (SE) ng/L for R-HCH, and 0.423( 0.013 (SE) ng/L for γ-HCH. Assuming conservativepartitioning of HCHs between the ice crystal structure andbrine, and 1 to 0.16 volume mixing of brine from 90 cm andthe top 10 cm of ice, HCH brine concentrations should equal7.660 and 0.882 ng/L for the R- and γ-isomers, respectively.Accordingly, up to 48% of R-HCH and 52% of γ-HCH mustbe entrapped within the ice crystal matrix. If we neglect theair volume (11), attribute 95.7% of the total ice volume to theice crystal matrix, and apply a 10% correction for ice matrixversus brine density (18), the associated R-HCH and γ-HCHconcentrations are calculated to be 0.162 and 0.017 ng/L,respectively. This result compares very well with those of anindependent brine drainage experiment (13), where theauthors simulated the spring ice melt conditions by gradualwarming of the ice cores, and determined HCH levels andsalinity in the progressing meltwater fractions (Figure 1S,Table 1S). In this experiment, roughly 40% of the total HCHinventory was associated with the last three meltwaterfractions containing only 7% of the total salts. The R- andγ-HCH concentrations measured in each of these meltwaterfractions averaged 0.151 and 0.015 ng/L, respectively.

Gradual Decrease of HCH Brine Concentrations in theSpring. In winter, HCH brine concentrations were notsignificantly correlated with ice and brine physical charac-teristics including brine and bulk ice salinity, ice temperature,and brine volume fraction (Pearson’s correlations, data log-transformed prior to analysis, p > 0.05). The mean coefficients

FIGURE 2. Schematic diagram of potent methodological issuesassociated with a sump-hole technique in the case of HCHmeasurements.

CBRINE1 ) (VBRINE0CBRINE0 + VIM/SWCIM/SW)/(VBRINE0 + VIM/SW) (3)

Cw/Ca ) RT/HLC (4)

CV(%) ) 100 SD/mean (5)

9260 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 24, 2010

TABL

E1.

Valu

es,

Mea

ns,

Stan

dard

Erro

rs(S

Es),

and

Coef

ficie

nts

ofVa

riabi

lity

(CVs

)fo

rr

-an

dγ-

HCH

Conc

entra

tion

and

Corre

spon

ding

Phys

ical

Varia

bles

inth

eBr

ine,

Bulk

Ice,

Unde

r-Ice

Seaw

ater

,and

Air

inth

eW

inte

r20

07/2

008

and

Sprin

g20

08a

win

ter

2007

/200

8

date

(sta

tion)

vari

able

6Ja

n(D

14)

6Ja

n(D

14)

7Ja

n(D

14)

7Ja

n(D

14)

8Ja

n(D

14)

8Ja

n(D

14)

9Ja

n(D

14)

9Ja

n(D

14)

10Ja

n(D

14)

10Ja

n(D

14)

mea

nSE

CV(%

)

R-H

CH

BR

INE

90cm

(ng

/L)

4.71

55.

322

3.34

35.

165

4.63

74.

473

2.43

33.

169

3.30

33.

572

4.01

30.

307

24γ-

HC

HB

RIN

E90

cm(n

g/L

)0.

398

0.46

90.

421

0.45

60.

381

0.49

10.

427

0.39

10.

435

0.36

40.

423

0.01

310

SB

RIN

E90

cm13

2.5

124.

012

2.0

126.

513

1.0

118.

015

4.0

125.

011

7.5

130.

012

8.1

3.3

8R

-HC

HIC

E90

cm(n

g/L

)0.

275

0.32

50.

413

0.28

50.

420

n/a

0.25

40.

220

0.28

70.

267

0.30

50.

023

23γ-

HC

HIC

E90

cm(n

g/L

)0.

032

0.03

30.

034

0.03

40.

030

0.03

80.

029

0.02

90.

045

0.03

20.

034

0.00

214

SIC

E90

cm4.

24.

24.

64.

15.

64.

24.

24.

14.

24.

24.

40.

111

TIC

E90

cm(°

C)

-4.

4-

5.8

-6.

1-

6.4

-5.

2-

4.9

-5.

5-

5.0

-5.

3-

4.8

-5.

30.

212

υ BR

INE

90cm

(%)

4.9

3.8

3.9

3.3

5.6

4.4

3.9

4.3

4.1

4.5

4.3

0.2

15R

-HC

HIC

E5

cm(n

g/L

)0.

288

0.29

90.

416

0.37

70.

284

0.33

40.

314

0.29

8n

/a0.

253

0.31

80.

017

16γ-

HC

HIC

E5

cm(n

g/L

)0.

037

0.04

00.

057

0.04

30.

038

0.04

70.

039

0.04

70.

046

0.04

60.

044

0.00

214

SIC

E5

cm9.

58.

79.

08.

38.

78.

67.

68.

18.

58.

18.

50.

26

TIC

E5

cm(°

C)

-11

.2-

18.0

-17

.7-

17.3

-17

.3-

18.9

n/a

-19

.1-

18.0

-18

.7-

17.4

0.8

14υ B

RIN

E5

cm(%

)4.

72.

83.

02.

82.

92.

7n

/a2.

52.

82.

63.

00.

222

R-H

CH

WA

TE

R(n

g/L

)1.

144

1.08

41.

091

1.03

31.

092

0.91

71.

047

0.95

00.

927

0.95

51.

024

0.02

68

γ-H

CH

WA

TE

R(n

g/L

)0.

101

0.10

60.

112

0.09

60.

089

0.10

40.

115

0.09

90.

117

0.11

30.

105

0.00

39

SW

AT

ER

31.7

29.1

20.9

29.0

27.5

29.0

30.0

31.5

30.8

28.7

28.8

1.0

11R

-HC

HA

IR(p

g/m

3 )10

.726

10.8

559.

875

n/a

6.77

56.

755

5.65

7n

/a9.

128

n/a

8.53

90.

800

25γ-

HC

HA

IR(p

g/m

3 )9.

362

6.85

19.

759

n/a

6.76

03.

194

n/a

n/a

7.45

9n

/a7.

231

0.96

033

spri

ng20

08

date

(sta

tion)

vari

able

21A

pr(D

4180

cm)

21A

pr(D

4110

0cm

)28

Apr

(D43

80cm

)28

Apr

(D43

100

cm)

9M

ay(F

280

cm)

9M

ay(F

210

0cm

)14

May

(F3

80cm

)14

May

(F3

100

cm)

mea

nSE

CV(%

)

R-H

CH

BR

INE

(ng

/L)

3.19

03.

624

2.80

93.

027

2.09

52.

012

1.84

81.

827

2.55

40.

245

27γ-

HC

HB

RIN

E(n

g/L

)0.

478

0.57

00.

411

0.34

90.

271

0.25

70.

223

0.21

50.

347

0.04

637

SB

RIN

E10

5.0

92.5

93.5

93.0

73.5

71.5

60.0

63.0

81.5

5.8

20S

ICE

5.3

3.7

4.7

4.7

5.9

5.6

5.1

5.9

5.1

0.3

15T

ICE

(°C

)-

5.4

-3.

7-

4.5

-3.

1-

3.0

-1.

9-

2.2

-1.

7-

3.2

0.5

41υ B

RIN

E(%

)5.

05.

15.

37.

710

.114

.711

.817

.69.

71.

749

R-H

CH

WA

TE

R(n

g/L

)0.

885

0.91

20.

877

0.88

30.

889

0.00

82

γ-H

CH

WA

TE

R(n

g/L

)0.

122

0.12

60.

107

0.10

40.

115

0.00

510

an

/a,

no

tav

aila

ble

;S

,sa

linit

y;T

,te

mp

erat

ure

;υ B

RIN

E,

bri

ne

volu

me

frac

tio

n.

VOL. 44, NO. 24, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 9261

of variability (CVs) were estimated at 24% for R-HCH and10% for γ-HCH (Table 1), which was most probably associatedwith the small-scale randomness of distribution in theenvironment. However, in spring ice, which becomes per-meable to liquid transport, as its mean brine volume fractionapproached 9.7% ( 4.7% (SD), both R- and γ-HCH con-centrations gradually decreased as a function of increasingbrine volume fraction and decreasing brine salinity (Figure3). Both of these ice physical parameters could independentlyexplain over 80% of HCH variability in the brine in the spring(R2 > 0.80, p < 0.05). Spring sea-ice processes that couldpotentially be responsible for gradual dilution of brine HCHinclude melting of the ice crystal matrix and replenishmentof brine with seawater.

To quantify how much each of these processes couldcontribute to the dilution of HCH, we calculated the volumesin which the brine would have to mix with seawater or theice matrix to reproduce the salinity observations using eq 2.The calculations were made for the period transitionsbetween each of the 4 sampling stations characterized by aprogressive increase in brine volume fraction, and decreasein brine salinity (Table 2; season progressions 1-2, 2-3, 3-4).For example, the observed decrease in brine salinity of 32between periods 1 and 2 could be accounted for by mixing1 part of brine with 0.48 parts of seawater or with 0.33 partsof the melting ice crystal matrix. The resulting calculatedconcentrations of R-HCH and γ-HCH in the brine (eq 3) arepresented in Table 2 and range between 3-19% and 2-31%of their measured values, respectively. Thus, both dilutionwith the ice crystal matrix and replenishment of brine withseawater could account for the gradual decrease of HCHconcentrations as the season progressed (Figure 4). Althoughthe relative importance of these two diluting mechanismscannot be established with the present data, it is most likelythat ice crystal matrix melting is primarily responsible for

the decrease in the HCH brine concentrations until brinepockets grow large enough to coalesce. Once coalescencehas occurred, the elongated brine channels from the bottomof the ice to the depth of brine sampling would provide aplausible means to permit brine replenishment with seawater.However, replenishment by seawater has been shown tooccur only when brine volume fractions exceed 20% (9), athreshold not reached in the current study.

γ-HCH Levels in the Brine at the Beginning of Spring.In the early spring (stations D41 and D43, 21-28 Apr, 2008)the levels of γ-HCH in the brine did not decrease comparedto the winter sampling period as might have been predictedfrom brine salinity (Figure 4B). Average γ-HCH brineconcentration in the early spring, 0.452 ( 0.047 ng/L (SE),was not significantly different from the winter value of 0.423( 0.013 ng/L (t-test, p ) 0.60). A similar phenomenon wasobserved to occur for γ-HCH concentrations in ice samplescollected from the same sampling locations (13). The authorsreported that relative to the R-HCH concentrations, theγ-HCH levels in the bulk sea ice samples were harder topredict based on the long-term dependence on sea ice salinity(R2 ) 0.70, p < 0.01, and R2 ) 0.40, p < 0.01, respectively). Thisdiscrepancy was ascribed to differences in atmosphere-snowprocesses in the spring and/or spatial variability. Correctingfor this discrepancy (as estimated based on the differencesin HCH concentrations in the ice samples between the R-and γ-isomers, 36%), the γ-HCH brine concentration for theearly spring period was calculated to be 0.377 ng/L, whichis much closer to the value calculated assuming dilution withthe melting ice crystal matrix (Figure 4B).

Implications for Biological Exposures. Brine containedwithin sea ice currently exhibits the highest HCH concentra-tions in any abiotic Arctic environment, exceeding under-ice water concentrations by a factor of approximately 3 inthe spring (Table 1). Sea-ice brine pockets comprise an

FIGURE 3. Pearson’s correlations between HCH concentrations and brine volume fraction, and brine salinity in the spring; resultslog-transformed prior to analysis.

TABLE 2. Brine/Seawater and Brine/Ice Matrix Mixing Ratios That Would Account for the Decrease in the Brine Salinity, andMeasured Brine HCH Concentrations versus Calculated Ones Based on the Brine Dilution with Sea Water or Ice Crystal MatrixAs the Sampling Season Progressed

season progressiona

variable 1 (D14) 2 (D41, D43) 3 (F2) 4 (F3)

number of samples 10 4 2 2average brine volume fraction (%) 4.3 5.8 12.4 14.7average brine salinity 128.1 96.0 72.5 61.5brine/seawater mixing ratio 1/0.48 1/0.57 1/0.34brine/ice matrix mixing ratio 1/0.33 1/0.32 1/0.18measured R-HCH concentration (ng/L) 4.013 3.163 2.054 1.838calculated R-HCH concentration (brine dilution with seawater) (ng/L) 3.044 2.341 1.753calculated R-HCH concentration (brine dilution with ice matrix) (ng/L) 3.057 2.435 1.765measured γ-HCH concentration (ng/L) 0.423 0.452 0.264 0.219calculated γ-HCH concentration (brine dilution with seawater) (ng/L) 0.320 0.333 0.224calculated γ-HCH concentration (brine dilution with ice matrix) (ng/L) 0.322 0.346 0.226

a 1 ) (D14) Jan 6-10, 2008; 2 ) (D41 and D43), Apr 21-28, 2008; 3 ) (F2), May 9, 2008; 4 ) (F3), May 14, 2008.

9262 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 24, 2010

extreme-environment habitat, the function of which in polarmarine environments remains unclear (19). The present highlevels in brine, which imply former concentrations duringthe period of peak loading (1980-1990) of perhaps 10 timesgreater (5, 6), therefore pose a risk to the sea ice biotainhabiting the bottom of the ice in spring, or living withinbrine channels throughout winter (20). In the central Arctic,ice algae are thought to contribute up to 57% of the primaryproduction (water column+ sea ice) (21), again emphasizingthe potential importance of the brine concentration processin amplifying lower food web exposure, which then translatesinto higher biological exposures all the way up through theArctic marine food web.

We expect that observed and projected change in Arcticsea-ice climate will make the brine concentration pathwaymore widespread in the future Arctic Ocean. Recent obser-vations show that multiyear ice area in the Arctic Ocean isdwindling, with seasonal ice expanding to fill the space. Thisshift will also affect the structure of the Arctic marine foodweb (22). Open water regions of enhanced primary produc-tivity (polynyas and marginal ice zones, MIZ) will occur inareas that until now have been characterized by a perennialice cover and an extremely low biological productivity. Thisshift will most likely decrease the annual percentage con-tribution to the food web by sea-ice primary producers infavor of pelagic producers. Taken together, we could see amore widespread late winter-early spring exposure to brinecontaminant amplification followed by a late-spring pelagicdilution within the food web.

AcknowledgmentsWe thank the crew of CCGS Amundsen for the field workassistance without which this study could never have beenaccomplished. Also, we express our gratitude to CFLparticipantssDebbie Armstrong, Amanda Chaulk, Elizabeth

Shadwick, and Laura Simmssfor sharing their time andknowledge in the field. Our special thanks goes to JasonPavlich and Wojciech Walkusz for the unbelievable amountof hard physical field work they put into ice sampling for thisstudy, without which the samples collection for this experi-ment would never have been possible. We also thank BrunoRosenberg for help and advice regarding laboratory analysis.Finally, we thank the Canadian program office of theInternational Polar Year, the Natural Sciences and Engineer-ing Research Council (NSERC), Canada Foundation forInnovation (CFI), Canada Research Chairs (CRC), the De-partment of Fisheries and Oceans Canada, and ArcticNet forfunding.

Supporting Information AvailableTable with the sea-ice brine drainage experiment results(24-26 Jan, 2008) and a figure showing salinity, R-HCHconcentration, and γ-HCH concentration in progressingmeltwater fractions in the sea-ice brine drainage experiment(after ref 13). This material is available free of charge via theInternet at http://pubs.acs.org/.

Literature Cited(1) Jantunen, L. M.; Bidleman, T. F. Reversal of the Air-Water Gas

Exchange Direction of Hexachlorocyclohexanes in the Beringand Chukchi Seas: 1993 versus 1988. Environ. Sci. Technol. 1995,29, 1081–1088.

(2) Harner, T.; Kylin, H.; Bidleman, T. F.; Strachan, W. M. J. Removalof R- and γ-Hexachlorocyclohexane and Enantiomers ofR-Hexachlorocyclohexane in the Eastern Arctic Ocean. Environ.Sci. Technol. 1999, 33, 1157–1164.

(3) Bidleman, T. F.; Kylin, H.; Jantunen, L. M.; Helm, P. A.;Macdonald, R. W. Hexachlorocyclohexanes in the CanadianArchipelago. 1. Spatial Distribution and Pathways of R-, �-, andγ-HCHs in Surface Water. Environ. Sci. Technol. 2007, 41, 2688–2695.

FIGURE 4. Measured versus calculated brine r-HCH (A) and γ-HCH (B) concentrations as a function of decreasing brine salinity andprogressing sampling season. * Season progession 1 ) 6-10 Jan, 2008; 2 ) 21-28 Apr, 2008; 3 ) 9 May, 2008; and 4 ) 14 May,2008.

VOL. 44, NO. 24, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 9263

(4) Jantunen, L. M.; Helm, P. A.; Kylin, H.; Bidleman, T. F.Hexachlorocyclohexanes (HCHs) In the Canadian Archipelago.2. Air-Water Gas Exchange of R- and γ-HCH. Environ. Sci.Technol. 2008, 42, 465–470.

(5) Li, Y.-F.; Macdonald, R. W.; Ma, J. M.; Hung, H.; Venkatesh, S.Historical alpha-HCH budget in the Arctic Ocean: The ArcticMass Balance Box Model (AMBBM). Sci. Total Environ. 2004,324, 115–139.

(6) Macdonald, R. W.; Barrie, A. L.; Bidleman, T. F.; Diamond, M. L.;Gregor, D. J.; Semkin, R. G.; Strachan, W. M. J.; Li, Y.-F.; Wania,F.; Alaee, M.; Alexeeva, L. B.; Backus, S. M.; Bailey, R.; Bewers,J. M.; Gobeil, C.; Halsall, C. J.; Harner, T.; Hoff, J. T.; Jantunen,L. M. M.; Lockhart, W. L.; Mackay, D.; Muir, D. C. G.; Pudykiewicz,J.; Reimer, K. J.; Smith, J. N.; Stern, G. A.; Schroeder, W. H.;Wagemann, R.; Yunker, M. B. Contaminants in the CanadianArctic: Five years of progress in understanding sources, oc-currence and pathways. Sci. Total Environ. 2000, 254, 93–236.

(7) Xiao, H.; Li, N.; Wania, F. Compilation, Evaluation, and Selectionof Physical-Chemical Property Data for R-, �-, and γ-Hexachlo-rocyclohexane. J. Chem. Eng. Data 2004, 49, 173–185.

(8) Li, Y. F.; Macdonald, R. W. Sources and pathways of selectedorganochlorine pesticides to the Arctic and the effect of pathwaydivergence on HCH trends in biota: A review. Sci. Total Environ.2005, 342, 87–106.

(9) Perovich, D. K.; Gow, A. J. A quantitative description of sea iceinclusions. J. Geophys. Res. 1996, 101 (C8), 18,327–18,343.

(10) Golden, K. M.; Ackley, S. F.; Lytle, V. I. The Percolation PhaseTransition in Sea Ice. Science 1998, 282, 2238–2241.

(11) Richter-Menge, J. A.; Jones, K. F. The tensile strength of first-year sea ice. J. Glaciol. 1993, 39, 609–618.

(12) Weeks, W. F.; Ackley, S. F. The growth, structure and propertiesof sea ice. In The Geophysics of Sea Ice; Untersteiner, N., Ed.;Matrinus Nijhoff: Dordrecht, 1986 (NATO ASI B146); pp 164.

(13) Pucko, M.; Stern, G. A.; Barber, D. G.; Macdonald, R. W.;Rosenberg, B. The International Polar Year (IPY) CircumpolarFlaw Lead (CFL) System Study: The importance of brineprocesses for R- and γ-hexachlorocyclohexane (HCH) ac-cumulation or rejection in sea ice. Atmosphere-Ocean 2010, 48,(doi:10.3137/OC318.2010).

(14) Mukherjee, I.; Gopal, M. Chromatographic techniques in theanalysis of organochlorine pesticide residues. J. Chromatogr. A1996, 754, 33–42.

(15) Frankenstein, G. E.; Garner, R. Equations for Determining theBrine Volume of Sea Ice from-0.5 to-22.9 C. J. Glaciol. 1967,6, 943–944.

(16) Kucklick, J. R.; Hinckley, D. A.; Bidleman, T. F. Determinationof Henry’s law constants for hexachlorocyclohexanes in distilledwater and artificial seawater as a function of temperature. Mar.Chem. 1991, 34, 197–209.

(17) Sokal, R. R.; Rohlf, F. J. Biometry, 2nd, ed.; W.H. Freeman andCompany: New York, 1980; pp 860.

(18) Timcorint, G. W.; Frederking, R. M. W. A review of sea ice density.Cold Reg. Sci. Technol. 1996, 24, 1–6.

(19) Deming, J. Psychrophiles and polar regions. Curr. Opin.Microbiol. 2004, 5, 310–309.

(20) Horner, R.; Ackley, S. F.; Dieckmann, G. S.; Gulliksen, B.; Hoshiai,T.; Legendre, L.; Melnikov, I. A.; Reeburgh, W. S.; Spindler, M.;Sullivan, C. W. Ecology of sea ice biota. 1. Habitat, terminology,and methodology. Pol. Biol. 1992, 12, 417–427.

(21) Gosselin, M.; Levasseur, M.; Wheeler, P. A.; Horner, R. A.; Booth,B. C. New measurements of phytoplankton and ice algalproduction in the Arctic Ocean. Deep-Sea Res. II 1997, 44, 1623–1644.

(22) Gradinger, R. Climate change and biological oceanography ofthe Arctic Ocean. Philos. Trans. R. Soc. London, Ser. A. 1995,352, 277–286.

ES102275B

9264 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 24, 2010