Organic Matter Chlorination Rates in Different Boreal Soils: The Roleof Soil Organic Matter ContentMalin Gustavsson,†,* Susanne Karlsson,† Gunilla Oberg,‡ Per Sanden,† Teresia Svensson,† Salar Valinia,§

Yves Thiry,∥ and David Bastviken†

†Department of Thematic Studies, Water and Environmental Studies, Linkoping University, 58183 Linkoping, Sweden.‡Institute for Resources, Environment and Sustainability, University of British Columbia, 428-2202 Main Mall Vancouver, BritishColumbia, Canada, V6T 1Z4.§Department of Aquatic Sciences and Assessment SLU, Swedish University of Agricultural Sciences, P.O. Box 7050, SE 75007Uppsala, Sweden.∥Andra, Research and Development Division, 1/7 rue Jean-Monnet, 92298 Chatenay-Malabry Cedex, France.

ABSTRACT: Transformation of chloride (Cl−) to organicchlorine (Clorg) occurs naturally in soil but it is poorlyunderstood how and why transformation rates vary amongenvironments. There are still few measurements of chlorina-tion rates in soils, even though formation of Clorg has beenknown for two decades. In the present study, we compareorganic matter (OM) chlorination rates, measured by 36Cltracer experiments, in soils from eleven different locations(coniferous forest soils, pasture soils and agricultural soils) anddiscuss how various environmental factors effect chlorination.Chlorination rates were highest in the forest soils and strongcorrelations were seen with environmental variables such assoil OM content and Cl− concentration. Data presentedsupport the hypothesis that OM levels give the framework forthe soil chlorine cycling and that chlorination in more organicsoils over time leads to a larger Clorg pool and in turn to a highinternal supply of Cl− upon dechlorination. This providesunexpected indications that pore water Cl− levels may becontrolled by supply from dechlorination processes and can explain why soil Cl− locally can be more closely related to soil OMcontent and the amount organically bound chlorine than to Cl− deposition.

■ INTRODUCTIONChloride participates in a complex biogeochemical cycle, anddetails regarding this cycle have been discussed for severalyears. In the late 1980s and early 1990s1−3 it was revealed thatlarge amounts of naturally formed chlorinated organic matter(Clorg) were present ubiquitously in the environment.2−6

Chlorinated organic compounds were earlier believed tooriginate from anthropogenic activities only.7 An increasingamount of evidence has confirmed that formation of Clorgoccurs naturally and that Clorg is as abundant as the chloride ion(Cl−) in organic soils.3,6,8,9 Even though it is nowadays widelyaccepted that Cl− can be transformed to Clorg, the underlyingprocesses are not well understood. However, microorganismsand enzymes appear be important for both the formation anddegradation processes of Clorg.

1,10,11 Imbalance between theseprocesses seems to favor the build up of a soil pool of Clorg.

12 Itseems however as if the balance between the two processes canshift, leading to that Cl− is retained during some conditions andreleased from soil during other conditions.13,14

Questions about how formation/mineralization of Clorg takesplace and how fast these processes occur are of interest for anumber of reasons. First, it is still often assumed that Cl− isinert and freely mobile in soil and Cl− has frequently been usedin hydrological research and biogeochemical modeling.Typically, Cl− is used as an inexpensive and easily measuredtracer for soil and groundwater movements when calculatingbudget and deposition estimates.15 Not accounting for the Cl−

cycling including retention and release of Clorg can lead to biaswhen doing hydrological modeling. Second, the widespreadperception that Cl− is inert in soil is also applied to radioactivechlorine (36Cl).16,17 36Cl is produced in the irradiated fuelassembly during reactor operations and has a half-life of 3.01 ×105 years. The mobility of Cl− in the environment and the long

Received: September 12, 2011Revised: December 7, 2011Accepted: December 22, 2011Published: December 22, 2011

Article

pubs.acs.org/est

© 2011 American Chemical Society 1504 dx.doi.org/10.1021/es203191r | Environ. Sci. Technol. 2012, 46, 1504−1510

half-life of 36Cl make it a critical radionuclide to account for insafety assessment of any repositories for radioactive waste.18

Formation of 36Clorg may result in increased residence time of36Cl and possible prolonged exposure to soil organisms andplants following contamination of soils. Knowledge aboutchlorination rates in various soil types is therefore important inhydrological as well as risk analysis models.Previous studies suggest that rates may vary between sites

and a large influence of environmental variables is indi-cated.19,20 However, direct measurements of chlorination ratesin different soil types with comparable methods are needed totest these indications and develop a more solid understandingof how various environmental variables influence chlorinationrates. In the present study we compare organic matterchlorination rates in soils from eleven different locations inconiferous forests, pastures or agricultural fields and relate therates to soil organic matter, pH, total Clorg and extractable Cl−.

■ EXPERIMENTAL SECTIONSoil Sampling. Soil samples from four different coniferous

forests (F1−F4, podsol type profiles), four pastures (P1−P4)were collected in September 2008, and agricultural soil samplesfrom three fields (A1-A3) in November 2008. The sites wereselected to be representative of Swedish forest and pasture soiltypes according to the criteria of the Swedish National ForestInventory.21 Another criterion was that all sites should belocated within a 100 km radius from Linkoping, Sweden, forlogistic reasons. The selection was done in consultation withthe Department of Forest Resource Management, The SwedishUniversity of Agricultural science (SLU), responsible for theSwedish national soil sampling program.22

The forested sites differed with respect to age of trees,canopy cover and soil texture (proportion of soil particle size)(Table 1). The age of the trees growing on the forest sitesvaried from 25 to 50 years. The canopy cover varied from 50 to80%. Pasture is here defined as land that is grazed yearly, notregularly ploughed, may carry trees, but not commercial timberproduction. The pastures had been grazed for 40 to over 100years. The canopy cover of the pastures varied from 5 to 20%.The agricultural soil samples from the three fields werecollected in the Lanna experimental farm (lat. 58°20′ N. long.13°06′ E) in Vastergotland, Sweden. The agricultural soils wereall collected in the same experimental area and they representone common agricultural soil type but the fields were chosen tohave different previous cropping systems or agricultural

practices. One of the soils had been used for organic farmingof wheat, another was collected where grassland had beengrown for four years and the third soil was collected in an areawhere grain was produced using artificial fertilizers (Table 1).All types of soils were collected with a spade from the topsoil

layer, 5−15 cm below ground level and transported inpolyethylene plastic bags to the laboratory. The soils werethen refrigerated at 4 °C until they were used in anyexperiments.

Determination of Soil Characteristics. A subsample ofeach soil was collected prior to the experiment and used todetermine soil−water content (by drying at 105 °C for 31 h),soil organic matter, pH, total Clorg and extractable Cl−. Soilorganic matter content was determined by loss of ignition(LOI) at 550 °C for 8 h, assuming that the carbon contentequaled 50% of LOI. pH was measured in extracts of water and1.0 M KCl according to ref 23. Any Cl− present in the sampleswas extracted according to the same procedure as for 36Cl−

described below, with the exception that the last twoextractions were done with 0.01 M KNO3 instead of KCl.The extracts were frozen and after thawing analyzed forchloride concentrations by ion chromatography with chemicalsuppression (MIC-2, Metrohm) according to Standardiza-tion.24 The residual soil was dried and milled, and 0.02 g wasincinerated and analyzed for total organic halogens (TOX)content according to Asplund et al.25 using an ECS3000analyzer (Euroglas).

Experimental Setup. The soils were sieved through a 2mm mesh, distributed in 50 mL plastic centrifugation tubes(Sarstedt, Germany), and incubated at 20 °C with addition of36Cl− according to the methodology of previous studies.9,26

Briefly, 2 g fresh soil was transferred to each tube (threereplicates for each soil). For coniferous and pasture soils,diluted 36Cl− (Amersham Biotech; specific activity 0.52 MBqmg Cl1−) solutions were added to each test tube giving a finalconcentration of 36Cl− corresponding to 350 000 disintegrationsper minute (DPM; 1 Bq = 60 DPM). For the agricultural soils,a different isotope batch from American RadiolabeledChemicals Inc. was used (specific activity 0.23 MBq mgCl1−) with a final addition of 200 000 DPM to each tube. Thevolume of added 36Cl− solution in all soils varied from 0.4 to 1.3mL depending on how much water was needed to be able todistribute the isotope homogeneously in each soil type. Even ifthe volume differed between soil types the total addedradioactivity was the same for each soil type as specified

Table 1. Soil and Vegetation Characteristics at the Sample Sites

sample land use soil parent materiala texture vegetation field layer canopy cover (%) stand age (years)

F1 coniferous forest till coarse siltyb high - low herb types 80 25F2 coniferous forest till fine sandyc high − low herb types 44 50F3 coniferous forest till fine sandy thin leaved grass types 48 50F4 coniferous forest till coarse silty bilberry 51 50P1 pasture till sandyd broad leaved grass types 12 74P2 pasture till coarse silty high − low herb types 9 55P3 pasture till sandy broad leaved grass types 19 105P4 pasture till sandy broad leaved grass types 5 37A1 agricultural till clayeye various 0A2 agricultural till clayey ley 0A3 agricultural till clayey barley 0

aSoil parent material comprises information on genesis and grain-size distribution of the parent material for soil formation. bSoil with a grain sizebetween 0.06 and 0.02 mm. cSoil with a grain size between 0.2 and 0.06 mm. dSoil with a grain size between 0.6 and 0.2 mm. eSoil with a grain size<0.002 mm.

Environmental Science & Technology Article

dx.doi.org/10.1021/es203191r | Environ. Sci. Technol. 2012, 46, 1504−15101505

above. This represented a mass of 11.2 μg Cl− per test tube inconiferous forest and pasture soils and 14.4 μg Cl− per test tubein the agricultural soil. After addition of 36Cl− solution, thesamples were dried under a fan at room temperature to reachoriginal fresh weight (less than 2 h). A set of samples (threereplicate tubes from each soil) were taken out and put in thefreezer when the correct weight was reached. The rest of thesamples were placed in a dark room and the tubes were aeratedusing aquaria pumps (4 × 15 min each day). The aquariapumps pumped moisturized air through the headspace of thetubes by first purging the air through water in a separatecentrifugation tube, to minimize drying of the soil. Weeklycontrols of water content were performed by weighing thetubes. If the water content reached a weight five percent lower/higher than original weight, this was adjusted by either wateradditions or removal of water by drying the soil with an openlid until it reached the initial weight. The air flow wascontrolled on a weekly basis. Upon sampling, three replicatetubes per treatment were removed from the experimental setupand immediately frozen until further analyses. The coniferousforest and pasture soils were sampled on five occasions (day 0,23, 55, 82, and 138) and the agricultural soil was sampled on sixoccasions (day 0, 30, 58, 108, 136, and 169).Extraction of 36Cl−. Any 36Cl− in the samples was removed

by a series of four subsequent extractions; two with water andtwo with KCl (0.01 M). To ensure release of intracellular 36Cl−,the samples were frozen (24 h, −18 °C), dried, and sonicated(45 s, 50% intensity; Bandelin Sonorex RK510H). Theextraction procedure was as follows: after freezing the samples,they were thawed at room temperature (2 h), 20 mL Milli-Qwater was added to each tube, placed on an end-overend shaker

for 30 min, centrifuged (6000g for 10 min), and thesupernatant was transferred by pipet to new centrifuge tubes(extract no. 1). The soil was then dried (60 °C, 24 h), milled,rewetted with 5 mL water, and sonicated. Subsequently, 15 mLwater was added followed by shaking, centrifugation andsupernatant removal according to above yielding extract no. 2.Then a procedure with addition of 20 mL 0.01 M KCl, shaking,centrifugation, and supernatant removal was repeated twiceproducing extracts nos. 3 and 4. Tests with soil amended withknown amounts of 36Cl− confirmed that the extraction protocolefficiently extracted inorganic chloride from the soil in line withprevious approaches.9,26 The extracts were finally frozen andthe residual soil was dried at 60 °C for 24 h and stored untilfurther analyses.

Organic 36Cl in Extracts. The amounts 36Cl bound toorganic matter in the extracts (36Clorgex) were determined inselected samples to check its abundance compared to the36Clorg in the residual soil after the extractions. A volume of 1mL of acidified nitrate solution (0.2 M KNO3, 0.02 M HNO3),and approximately 4 drops of concentrated HNO3 (yielding apH <2) were added to 10 mL of the extracts. The mixture wasshaken with 50 mg of activated carbon for 60 min, and filteredthrough a polycarbonate filter (Millipore, 0.45 μm mesh). Thefilter with the activated carbon and the adsorbed 36Clorgex wasrinsed with acidic nitrate solution (6 × 3 mL, 0.01 KNO3, 0.001M HNO3), followed by acidified Milli-Q water (6 × 3 mL, pH2 by acidification with HNO3), to remove any remaining Cl−.The filter with the 36Clorgex was combusted at 1000 °C under astream of O2 gas according to the procedure for analyzingAOX.25 The H36Cl gas formed during combustion was thentrapped in 0.1 M NaOH.27 This procedure leading the gas

Figure 1. I−III. Fraction of added 36Cl− recovered as 36Clorg over time in forest F1−F4 (panel I), pasture P1−P4 (panel II) and agricultural soilsamples A1-A3 (panel III) (average ±1SD, n = 3) (see text for details). Generally all the different soils for each soil type was sampled the same day,but data points have in the visual graphs been shifted up to 2 days to separate error bars from each other. This does not affect calculations based onthis data.

Environmental Science & Technology Article

dx.doi.org/10.1021/es203191r | Environ. Sci. Technol. 2012, 46, 1504−15101506

stream through two scintillation vials in series, each holding 10mL of 0.1 M NaOH, yielded a recovery of >98% of the 36Clpresent in the sample prior to combustion.26

Inorganic 36Cl in Extracts. The amount of inorganic 36Clin the extracts was determined by analyzing filtrates afterremoval of 36Clorgex, as described above. An aliquot of 10 mLwas transferred to scintillation vials for liquid scintillationcounting (LSC).Organic 36Cl in the Residual Soil. The dried residual soil

(after the four extractions mentioned above) from the 36Clamended treatments was milled and approximately 0.2 g soil(exact amount noted to the third digit) was combusted todetermine the amount of organically bound 36Cl (36Clorg) asdescribed above for determination of 36Clorgex. Previous testshave confirmed that the chlorine associated with the residualsoil and detected this way is organically bound and associatedwith humic and fulvic acids.26

Liquid Scintillation Counting (LSC). The solutionscontaining trapped 36Cl (NaOH solutions for 36Clorg and36Clorgex, and water solution for the 36Cl−) were analyzed for36Cl by LSC (Beckman LX 6300). The analysis was correctedfor quench using standard quench curves prepared fromsolutions with the same matrix composition as the samples(e.g., 0.1 M NaOH). Before analyzing the samples ascintillation cocktail (Ultima Gold XR, Chemical InstrumentsAB) was added to all 36Cl samples and also to blank controls(Milli-Q water and scintillation cocktail). All radioactivemeasurements were corrected for background radiation bysubtracting radioactivity in the blank controls.Chlorination Rates. The amount of 36Clorg was plotted

over time (Figure 1 I−III). The specif ic chlorination rate wasexpressed on a daily basis (d−1), that is, the fraction of thestanding stock Cl− that became organically bound per day, andwas determined by the slope of the regression line over time indays versus the fraction of added 36Cl− recovered as 36Clorg. Thespecific chlorination rate multiplied with the total content of

Cl− in the original soil gave the chlorination rate expressed asng Cl g dw soil−1 d−1.

■ RESULTS AND DISCUSSION

Soil Characteristics. Soil characteristics for the studiedsoils are given in Table 2. The soil−water content, Clorg, andCl− were significantly higher in the forest soils as compared tothe pasture and agricultural soils. The molar based chlorine-to-carbon ratios were significantly higher in the agricultural soils ascompared to the two other soil types (Kruskal−Wallis, p =0.038). The pH was lowest in the forest soil (Table 2; Kruskal−Wallis, p = 0.01 pHH2O, p = 0.02 for pHKCl).

Formation of Clorg. The transformation of Cl− to Clorg atthe termination of the experiment varied between 14 and 25%of the added 36Cl− in the coniferous forest soils (Figure 1 I−III), which was considerably higher as compared to the othersoil types. The specific chlorination rate (fraction of the Cl−

pool that gets converted to Clorg per day: unit d−1) was 2−3

times higher in the samples from coniferous forest than in thepasture and agricultural soils (Kruskal−Wallis, p = 0.037; Table2). The highest chlorination rate in terms of ng Cl g−1 soil d.w.d−1 was detected in one of the forest soils (90 ng Cl g−1 d.w.d−1) (Table 2). The average chlorination rate in the forest soilswas 54 ng Cl g−1 d.w. d−1 which was significantly higher thanthe average rate in pasture and agricultural soil (Kruskal−Wallis, p = 0.009). The slowest chlorination rate was detectedin agricultural soils, which had an average value of 2.6 ng Cl g−1

d.w. d−1, while the average chlorination rate in the pasture soilswas 3.7 ng Cl g−1 d.w. d−1.The chlorination rates observed in the present study are in

the same range as previously conducted studies (Table 2,Figure 2). Bastviken et al.9 studied a coniferous forest soil withapproximately the same amount of organic matter content andthe total amount formed 36Clorg, varied from 10 to 37% and thechlorination rate ranged from 12 to 78 ng g−1 d.w. d−1. Similarchlorination rates can also be found in other studies,28,29 but

Table 2. Soil Characteristics, Chlorination Rates and Cl:C Ratio for Soils in This Study and in Other Previously PublishedStudies

soil typesoil name/source

water content(fraction of f.w.)

LOI (%d.w.)

TOX(μg/g)

Cl− (μg gdw−1)

average chlorination rate(ng Cl− g−1 d.w. d−1)

specificchlorinationa

(d−1)pH

(KCl)pH (RO-water)

Cl-to-Cratiob (‰)

forest F1 0.35 14.6 68 46.9 44.1 0.00094 4.0 5.0 0.32forest F2 0.54 41.4 191 82.5 90.2 0.00109 2.9 4.4 0.31forest F3 0.36 16.8 102 36.8 37.0 0.00101 3.2 4.7 0.41forest F4 0.47 26.0 66 33.2 46.6 0.00140 3.8 4.8 0.17pasture P1 0.25 7.2 35 6.6 4.9 0.00076 4.2 5.6 0.33pasture P2 0.33 14.9 40 12.6 3.5 0.00027 5.3 5.7 0.18pasture P3 0.24 8.4 39 6.7 4.9 0.00074 4.2 5.2 0.32pasture P4 0.36 11.0 59 6.3 1.4 0.00021 4.3 5.4 0.36agricultural A1 0.25 6.1 38 6.6 2.7 0.00041 5.2 6.6 0.42agricultural A2 0.24 6.3 41 8.1 2.6 0.00032 4.7 6.3 0.44agricultural A3 0.23 5.1 40 4.7 2.6 0.00055 5.4 7.0 0.53forest Bastviken et

al 20070.65 51.0 314 17.4 5.0 0.00055 no

datano data 0.42

forest Bastviken etal 2009

0.33 27.0 133 21.0 32.2 0.00362 nodata

no data 0.33

sandy loam Lee et al2001

no data 4.0 no data no data no data 0.00138 nodata

4.5 no data

peat Silk et al1997

0.88 94.5 137 137.0 89.8 0.00066 nodata

no data 1.69

aSpecific chlorination rate denote the fraction of Cl− being chlorinated per time unit, i.e., specific chl. rate = average chlorination rate/Cl− conc.bNumber of chlorine atoms per carbon atoms multiplied by 1000 to yield ‰ units.

Environmental Science & Technology Article

dx.doi.org/10.1021/es203191r | Environ. Sci. Technol. 2012, 46, 1504−15101507

those studies used different soil types. Previous studies ofconversion rates in soil indicate a similar span (4−40%.19What Environmental Variables Influence Formation

of Cl− to Clorg? Correlations between chlorination rates anddifferent environmental variables in Table 2 (Cl− concentration,LOI, TOX, pH, and water content) showed that thechlorination rate was significantly associated with all of thestudied environmental variables, that is, organic matter content,water, pH and concentration of Cl−, and that the variables werestrongly cross-correlated (Table 3). This is in line with previousstudies, which indicate that organic matter,20,30 and Cl−31 arerelated to formation of Clorg,

8,32 and strongly cross-correlatedwith each other. Further evaluation is therefore needed, andcarried out below, to assess the relative importance of thedifferent studied environmental factors on soil chlorinationrates.The Cl− concentrations in the present study were 3−10

times higher in the forest soils as compared to the two othersoil types (Table 2, Figure 3). This is in line with some previousobservations which suggest that Cl− generally is higher in forest

top-soils than in other soil types.8 To our knowledge, noplausible explanation to the observed differences has to datebeen provided. Differences in chloride deposition cannotexplain the 3−10-fold difference as it is only twice as high orless as compared to the two other soil types (Figure 3 and ref33). Weathering of bedrock cannot explain the differenceeither, since the bedrock at the forest sites as well as thepastures is dominated by acidic, slow weathering minerals(granitic), as illustrated by the low levels of inorganic ions inthe soil−water in most Swedish coniferous soils.34 Theagricultural areas in Sweden are characterized by higheralkalinity and buffer capacity involving higher, rather thanlower inorganic ion content. The agricultural soils may alsohave received some additional Cl− through fertilizers andassociated irrigation. Altogether, this would suggest higher, notlower Cl− concentrations in the agricultural soils, which standsin contrast to what was found.

Figure 2. Chloride (Cl−) concentration (diamonds) and chlorinationrate (μg Cl− g soil−1 d−1) (triangles) versus soil organic matter (% ofd.w.) here measured as loss of ignition (LOI) data from this study andthe literature.9,26 Note that the open symbol represent data from aprevious study26 on soil with the highest water content presumablyreducing chlorination rates by reducing the levels of O2 and therebyreactive oxygen used in chlorination (see Table 2).

Table 3. Correlation Coefficients for Pearson Correlations between Chlorination Rates, Cl−, TOX, LOI, and Water Content,Data from This Study (a) (n = 11) and Data from This Study and the Literature (b) (7,23 (n = 13)

average chl. rate (ngCl− g−1 d.w. d−1)

specific chl.Rate (d−1)

H2O (partof f.w.)

LOI (%of dw.)

TOX(μg/g)

Cl− conc.(μg Cl−

g−1 d.w. n = 2)Cl-to-C

ratio (‰)bμg Cl/gOC−1 d−1

pH(KCl)

pH (RO-water)

specific chl.rate

0.77aa

0.72ba

H2O 0.88aa 0.63aa

0.44b 0.09bLOI 0.93aa 0.65aa 0.96aa

0.57ba 0.31b 0.95ba

TOX 0.91aa 0.53a 0.82aa 0.90aa

0.39b 0.21b 0.88ba 0.94ba

Cl− conc. 0.99aa 0.68aa 0.85aa 0.91aa 0.93aa

0.85ba 0.26b 0.56ba 0.58ba 0.42bCl-to-Cratio(‰)b

−0.31a −0.36a −0.54a −0.49a −0.11a −0.27a−0.30b −0.19b −0.24b −0.20b 0.11b −0.28b

μg Cl/gOC−1 d−1

0.83aa 0.78aa 0.61aa 0.60aa 0.62aa 0.83aa −0.20a0.74ba −0.36b 0.22b 0.20b 0.07b 0.81ba −0.25b

pH (KCl) −0.808aa −0.736aa −0.717aa −0.732aa −0.788aa −0.78aa 0.254a −0.726pH (RO-water)

−0.749aa −0.688aa −0.777aa −0.755aa −0.673aa −0.734aa 0.594 −0.686aa 0.904aa

aCorrelation is significant at the 0.05 level (2-tailed). bNumber of chlorine atom per carbon atoms multiplied by 1000 to yield ‰ units.

Figure 3. Chloride (Cl−) concentrations in the three different soiltypes (forest, pasture and agricultural), plotted against Cl− deposition.Comparison to the 1:1 line illustrate that pore water Cl− concentrationcan be independent of Cl− deposition, most notably for the forest soilswith high amounts of organic matter and organochlorines in this study(see text for details).

Environmental Science & Technology Article

dx.doi.org/10.1021/es203191r | Environ. Sci. Technol. 2012, 46, 1504−15101508

We hypothesize that the higher concentrations of Cl− inforest soils is a result of higher turnover rates between Cl− andClorg, that is, that not only the formation rate but also the Clorgmineralization rate (release of Cl−, for example, transformationof Clorg to Cl− and not overall OM mineralization), is higher inthese soils. Our hypothesis infers that the large pool of Clorg incombination with a high Clorg mineralization rate, renderincreased concentrations of Cl− in soil−water, of which themajority is reconverted to Clorg rather than leached from thesystem. The proposed process would explain the observedcorrelations between organic matter, Cl− and Clorg.It seems reasonable to assume that the comparably larger

storage of Clorg in forest soils has the potential of serving as arelatively more important source of Cl− through Clorgmineralization (release of Cl−). This implies that high organicmatter content in combination with high chlorination rates andlarge storage of Clorg acts as a regulator of Cl

− levels.In summary, the outlined hypothesis is in line with previous

studies, which suggests that the presence of organic matter is akey factor for soil Clorg formation.20 Given appropriateavailability of organic matter and microbial activity, chlorinationproceeds at rates allowed by the Cl− concentrations. A largerpool of organic matter will, over time, allow a morepronounced build-up of Clorg, which in turn will release moreCl− upon Clorg mineralization, raising Cl− levels (this Cl− sourceis decoupled from chlorine input from deposition andweathering) and further stimulating chlorination leading tohigher rates (Figure 4).

The here proposed hypothesis aligns well with a recent studydemonstrating that the activity of chloroperoxidases, which arebelieved to be important for soil organic matter chlorination,are regulated by a number of interrelated factors, such as pH,Cl− and redox.35 Taken together, this suggests that part of theregulatory framework for chlorination is driven by a number ofinterrelated factors influencing (i) the enzymatic activity (suchas substrates including availability of reactive oxygen species(ROS) being necessary for formation of reactive chlorine whichreacts with the organic matter), in combination with (ii) factorsinfluencing the production of enzymes (such as microbialactivity, microbial community composition), and35 where theamount and type of organic matter play a key role (Figure

5).7,36 This framework explains the difficulty in identifying asimple regulation of natural formation of Clorg.

19

The here suggested framework (Figure 4 and 5) provide apromising basis to develop models of chlorination rates and soilCl− and Clorg levels. Altogether, the data presented here supportthe hypothesis that soil organic matter levels provide theframeworks for the soil chlorine cycle. This means that Cl−

levels in many soils could be regulated by chlorine cycling(chlorination and dechlorination given enough soil organicmatter) and be largely independent of the short-term (yearly)external chlorine supply by deposition or weathering.

■ AUTHOR INFORMATION

Corresponding Author*Phone: +46 13 28 21 79; e-mail: malin.gustavsson@liu.se.

Author ContributionsM.G., S.K., T.S., and S.V. performed most of the practical work.P.S. and Y.T. participated in planning and provided valuableinput throughout the study. G.O. and D.B. were project leadersand planned and supervised the study. M.G. was primary authorand carried out primary statistical and conceptual analyses. D.B.assisted with data analysis and writing of the draft version. Allauthors contributed in the revision process. The authors arelisted in alphabetical order except for the first and the lastauthor.

■ ACKNOWLEDGMENTS

We are grateful for the financial support from the SwedishResearch Council (VR; Project no. 2006-5387). We also thankMats Walheim (SLU) for the help with selecting forest andpasture soils, Johan Roland at Lanna Experimental Farm (SLU)for assisting us with the sampling of agricultural soils, MonicaPetersson for assisting us with soil sampling and pH analysis,

Figure 4. Illustration of possible chlorine cycling in soil. This studyindicate an important role of soil organic matter (OM) forchlorination of organic material and soil chloride (Cl−) levels asfollows. Given appropriate availability of organic matter chlorinationproceed at rates allowed by the Cl− concentrations and otherinfluencing variables, and a larger pool of chlorinated organic matter(Clorg) will build up over time. This pool will subsequently releasemore Cl− upon dechlorination raising pore water Cl− levels (regardlessof chlorine input from deposition and weathering) and furtherstimulating chlorination leading to higher rates. By these mechanismspore water Cl− levels can be more influenced by OM and Clorg levelsover time than by the yearly chlorine deposition.

Figure 5. Illustration of possible relationships between factors shownto influence chlorination in soil by this and other studies.9,19,35 Thefigure emphasize the dominating role of soil organic matter (OM) ashypothesized in this study, and that OM can affect chlorideconcentrations (Cl−), pH, soil−water content (water), organismcommunities, and the levels of reactive oxygen species (ROS) andchloroperoxidase-type enzymes (CPO) in ways important for thechlorination in soils.35.

Environmental Science & Technology Article

dx.doi.org/10.1021/es203191r | Environ. Sci. Technol. 2012, 46, 1504−15101509

and Laura Marang from EDF (Electricite de France) forvaluable discussions.

■ REFERENCES(1) Hoekstra, E. J. on the Natural Formation of Chlorinated OrganicCompounds in soil, PhD thesis, TNO Institute of EnvironmentalSciences, Free University of Amsterdam, Delft, 1999.(2) Muller, G.; Nkusi, G.; Scholer, H. F. Natural organohalogens insediments. J. Prakt. Chem./Chem.-Ztg. 1996, 338, 23−29.(3) Gribble, G. W. The diversity of naturally produced organo-halogens. Chemosphere 2003, 52, 289−297.(4) Asplund, G.; Grimvall, A. Organohalogens in nature − morewidespread than previously assumed. Environ. Sci. Technol. 1991, 25,1346−1350.(5) Haselman, K.; Ketola., R.; Laturnus., F.; Lauritsen., F.; Grøn., C.Occurence and formation of chloroform at Danish forest sites. Atmos.Environ. 2000, 34, 187−193.(6) Svensson, T.; Sanden, P.; Bastviken, D.; Oberg., G. Chlorinetransport in a small catchment in southeast Sweden during two years.Biogeochemistry. Biogeochemistry 2006, 82 (2), 181−199.(7) Oberg, G. The natural chlorine cycle − fitting the scatteredpieces. Appl. Microbiol. Biotechnol. 2002, 58, 565−581.(8) Johansson, E.; Sanden., P.; Oberg., G. Organic chlorine indeciduous and coniferous forest soils in southern Sweden. Soil science2003, 168 (5), 347−355.(9) Bastviken, D.; Svensson, T.; Karlsson, S.; Sanden, P.; Oberg., G.Temperature sensitivity indicates that chlorination of organic matter inforest soil is primarily biotic. Environ. Sci. Technol. 2009, 43 (10),3569−3573.(10) Oberg, G. Chloride and organic chlorine in soil. Acta Hydrochim.Hydrobiol. 1998, 26 (3), 137−144.(11) van Pee, K. H., S. Unversucht. Biological dehalogenation andhalogenation reactions. Chemosphere 2003, 52, 299−312.(12) Clarke, N.; Fuksova, K.; Gryndler, M.; Lachmanova, Z.; Liste,H.-H.; Rohlenova, J.; Schroll, R.; Schroder, P.; Matucha, M. Theformation and fate of chlorinated organic substances in temperate andboreal forest soils. Environ. Sci. Pollut. Res. 2009, 16, 127−143.(13) Bastviken, D.; Sanden, P.; Svensson, T.; Stahlberg, C.;Magounakis, M.; Oberg, G. Chloride-retention and release in a borealforest soil: Effects of soil water residence time and nitrogen andchloride loads. Environ. Sci. Technol. 2006, 40 (9), 2977−2982.(14) Svensson, T.; Lovett, G. M.; Likens, G. E. Is chloride aconservative ion in forest ecosystems? Biogeochemistry 2010,DOI: 10.1007/s10533-010-9538-y.(15) Kirchner, J. W.; Feng., X. H.; Neal, C. Fractal stream chemistryand its implications for contaminant transport in catchments. Nature2000, 403, 524−527.(16) Kashparov, V.; Colle, C.; Zvarich, S.; Yoschenko, V.; Levchuk,S.; Lundin, S. Soil-to-plant halogens transfer studies 2. Root uptake ofradiochlorine by plants. J. Environ. Radioact. 2005, 79, 233−253.(17) Levchuk, S.; Kashparov, V.; Lazarev, N.; Colle, C.; Howard, B.;Yoschenko, V.; Yoschenko, L. Transfer of Cl from herbage into tissuesand milk products of dairy cattle and pigs. Radiat Environ Biophys.2008, 47, 111−119.(18) Frechou, C.; Degros, J.-P. Measurement of 36Cl in nuclearwastes and effluents: Validation of a radiochemical protocol with an in-house reference sample. J. Radioanal. Nucl. Chem. 2005, 263 (2), 333−339.(19) Oberg, G.; Bastviken, D. Transformation of chloride to organicchlorine in terrestrial environments: variability, extent and implica-tions. In press (DOI: 10:1080/10643389.2011.592753).(20) Redon, P.-O.; Abdesselam, A.; Bastviken, D.; Cecchini, S.;Nicolas, M.; Thiry, Y. Chloride and organic chlorine in forest soils:storage, residence times, and influence of ecological conditions.Environ. Sci. Technol. 2011, 45, 7202−7208.(21) Swedish National Forest Inventory Website; http://www.slu.se/nfi (accessed 11 April 2011).(22) The Swedish University of Agricultural science Website; http://www.slu.se/en/ (accessed 7 April 2011).

(23) Determination of pH, ISO 10390; International Organization forStandardization, Soil Quality: Geneva, 1994.(24) Water QualityDetermination of Dissolved Fluoride, Chloride,Nitrite, Orthophosphate, Bromide, Nitrate and Sulphate Ions, UsingLiquid Chromatography of IonsPart 1: Method for Water with LowContamination, ISO 10304−1:1992; Standardization, E. C. f.: Brussels,1995.(25) Asplund, G. Determination of the total and leachable amountsof organohalogens in soil. Chemosphere 1994, 28, 1467−1475.(26) Bastviken, D.; Thomsen, F.; Svensson, T.; Karlsson, S.; Sanden,P.; Shaw, G.; Matucha, M.; Oberg, G. Chloride retention in forest soilby microbial uptake and by natural chlorination of organic matter.Geochim. Cosmochim. Acta 2007, 71, 3182−3192.(27) Laniewski, K.; Dahlen., J.; Boren., H.; Grimvall, A.Determination of group parameters for organically bound chlorine,bromine, and iodine in precipitation. Chemosphere 1999, 3, 771−782.(28) Lee, R. T.; Shaw, G.; Wadey, P.; Wang, X. Specific association of36Cl with low molecular weight humic substances in soils.Chemosphere 2001, 43, 1063−1070.(29) Silk, P. J.; Lonergan, G. C.; Arsenault, T. L.; Boyle, C. D.Evidence of organochlorine formation in peat bogs. Chemosphere1997, 35 (12), 2865−2880.(30) Johansson, E.; Sanden., P.; Oberg., G. Spatial patterns of organicchlorine and chloride in Swedish forest soil. Chemosphere 2003, 52,391−397.(31) Rodstedth, M.; Stahlberg, C.; Sanden, P.; Oberg., G. Chlorideimbalances in soil lysimeters. Chemosphere 2003, 52, 381−389.(32) Matucha, M.; Gryndler, M.; Forczek, S. T.; Schroder, P.;Bastviken, D.; Rohlenova, J.; Uhlírova, H.; Fuksova, K. A chlorine-36and carbon-14 study of the role of chlorine in the forest ecosystem. J.Labelled Compd. Radiopharm. 2007, 50, 437−439.(33) Swedish Environmental Institute Homepage: www.krondroppsnatet.ivl.se (accessed 3 March 2011).(34) Swedish EPA Website; http://www.naturvardsverket.se/en/(accessed 28 February 2011).(35) Manoj, K. M.; Hager, L. P. Chloroperoxidase, a Janus Enzyme.Biogeochemistry 2008, 47 (9), 2997−3003.(36) Bengtsson, P.; Bastviken, D.; de Boer, W.; Oberg, G. Possiblerole of reactive chlorine in microbial antagonism and organic matterchlorination in terrestrial environments. Environ. Microbiol. 2009, 11(6), 1330−1339.

Environmental Science & Technology Article

dx.doi.org/10.1021/es203191r | Environ. Sci. Technol. 2012, 46, 1504−15101510