Environ Monit Assess (2012) 184:1231–1241 DOI 10.1007/s10661-011-2035-5

Reliable test methods for the determination of a natural production of chloroform in soils Christian Grøn · Frank Laturnus · Ole Stig Jacobsen

Received: 1 October 2010 / Accepted: 16 March 2011 / Published online: 12 April 2011 © Springer Science+Business Media B.V. 2011

Abstract Chloroform is one of the most frequently found anthropogenic groundwater contaminants. Recent investigations, however, suggested that chloroform in groundwater may also originate from a natural production in soils. As societies response to the occurrence of chloroform in groundwater may depend upon its origin as anthropogenic or naturally produced, test methods are needed to measure the potential of natural soil chloroform production. Field measurements of ambient air and soil air, and field and laboratory incubation studies were evaluated for measurement of relative soil chloroform production at a site with four different vegetation types (spruce forest, beech forest, grassland, and grain field) on comparable geological soil. All test methods showed varying soil production of chloroform with spruce forest soil being most

C. Grøn DHI Water, Environment, Health, 2970 Hørsholm, Denmark F. Laturnus (B) Institute for Biogeochemistry and Marine Chemistry, University of Hamburg, Bundesstrasse 55, 20146 Hamburg, Germany e-mail: [email protected] O. S. Jacobsen Geological Survey of Denmark and Greenland, 1350 Copenhagen, Denmark

productive and grain field soil being least productive. Field measurements of the ratio of soil air to ambient air chloroform concentrations exhibited the smallest difference between high production and low production areas, whereas laboratory incubation studies showed the largest difference. Thus, laboratory incubation studies are suggested as most efficient for estimating relative chloroform production in soil. The study indicated that soil samples should be tested not more than 14 days after sampling. Furthermore, it was found that potentially limiting compounds, such as chloride or nitrate, are not needed to be added in spike experiments to obtain reliable production results. However, it should be recognized that the processes of soil chloroform production are not known yet in all details. Other factors than those studied here may affect the test methods for soil chloroform production too. Keywords Chloroform · Soil production · Test methods · Groundwater contamination · Natural production

Introduction The organic solvent and chlorination byproduct chloroform (CHCl3 , trichloromethane) is one of the most frequently found groundwater contaminants (e.g., Moran et al. 2006). Concentrations

1232

reported for the US have mostly been below 1 μg L−1 . Impact of chlorinated water from mains and sewers has been suggested as a major source (Bishop et al. 1998). The international health based drinking water limit for chloroform is 300 μg L−1 (World Health Organisation 2006), the limit in the Drinking Water Directive of the European Union is 100 μg L−1 (sum of all trihalomethanes) (European Council of Ministers 1998), but levels of concern as low as 1.5 μg L−1 have been suggested based upon a lifetime cancer risk of 10−5 (Mills et al. 1998). The Danish drinking water limit is 10 μg L−1 for chloroform from natural sources and 1 μg L−1 from contamination (Miljøministeriet 2007). The Danish limit for chloroform in drinking water without an identified natural source is lower than with a natural source, because chloroform is then considered an indicator of groundwater contamination and must therefore comply with the limit of 1 μg L−1 for chlorinated contaminants. Since the first reports of naturally occurring chloroorganic compounds in groundwater (Grøn 1995), it has been demonstrated that chloroform can be found in groundwater at concentrations up to 1.6 μg L−1 without identified sources of contamination suggesting natural forest soil processes as the origin (Laturnus et al. 2000). In field studies, it was shown that spruce forest soils could release chloroform to soil air and to the atmosphere with lower release rates from beech forest and grassland soils (Haselmann et al. 2000a). A biogenic production of chloroform in the soils was supported by a seasonal variation in the releases with high emissions in humid and mild periods with high microbiological activity in the soil (spring and autumn, respectively) and low releases in dry cold or warm periods with low microbiological activity (winter and summer, respectively) (Haselmann et al. 2002). In a recent review (McCulloch 2003), a coupling between soil microbial processes and chloroform in soil pore water and groundwater was suggested based upon the occurrences in the different compartments. Laboratory soil incubation studies further supported the hypothesis of a natural origin of chloroform (Haselmann et al. 2000b). The production of chloroform in forest soils was linked to soils with organic or humus layers

Environ Monit Assess (2012) 184:1231–1241

(Hoekstra et al. 2001), and to the occurrence of inorganic chloride to be incorporated into released chloroform (Hoekstra et al. 1998). The authors established in situ tracer experiments using chloride-37 and showed that after the addition of chloride-37 to soil, the formation of CH37 Cl3 was observed. Similar results were reported recently by Matucha et al. (2010). The application of radioisotope tracers has to be considered as a sophisticated method to identify formation mechanisms of naturally produced chloroform. For example, by adding a Na37 Cl solution to plants from temperate forest ecosystems it was found that the plants emitted CH37 Cl3 , and, thus, clearly indicating the use of chloride as part in the formation of chloroform (Laturnus and Matucha 2008). Additionally, stable isotopes can be used to obtain detailed information on the origin of chloroform. By determining the concentrations of, for example, stable carbon isotopes and calculating the carbon isotope ratios, it is possible to identify whether, for example, chloroform is naturally produced or of anthropogenic origin (Hunkeler and Aravena 2000; Auer et al. 2006; Jendrzejewski et al. 2001; Jochmann et al. 2006). For the natural formation of chloroform in soil, a mechanism based upon chloroperoxidase enzymes possibly released from soil microorganisms has been suggested (Asplund et al. 1993; Laturnus et al. 1995). The chloroperoxidase mechanism is based on the formation of chlorinated organic matter, such as chlorinated humic substances, during degradation of lignin and soil organic matter (Öberg et al. 1997; Niedan et al. 2000; Reina et al. 2004). Studies of chlorinated organic matter in soils suggested an increase in net production of chlorinated organic matter with decreasing pH (Öberg et al. 1996). Furthermore, the occurrence of higher chlorinated organic matter contents in soils was correlated to a higher chloride content (Johansson et al. 2003), and a decrease in the contents of soil chlorinated organic matter was described after addition of inorganic nitrogen (Johansson et al. 2001). However, it should be recognized that the interactions of different processes, such as production, leaching, and degradation of soil chlorinated organic matter are complicated and their causality not fully elucidated (Laturnus et al. 2005). Still, to the degree

Environ Monit Assess (2012) 184:1231–1241

that these findings are valid for chloroform production, they should be considered in designing chloroform production studies and test methods. A ratio of soil air to ambient air chloroform concentrations well above 1 has been suggested as an indicator of soil chloroform production (Haselmann et al. 2002), and interpretation of the ratios in terms of differences in soil chloroform production has been presented (Haselmann et al. 2000a). A detailed site study supported that measured chloroform soil air to soil ratios were the result of a production in soil (Laturnus et al. 2000). Field and laboratory incubations studies have furthermore been presented for quantification of chloroform production in soil (Haselmann et al. 2000a, b). In the studies, chloroform was released, whereas other chlorinated solvents, such as tetrachloromethane, trichloroethene, and 1,1,1trichloroethane, were not, and, thus, supporting that the studies did in fact provide a quantification of chloroform production in soil rather than release of previously adsorbed compounds. In this paper, we present and compare ambient air and soil air measurements, and field and laboratory incubation studies for their suitability as test methods for the investigation of relative chloroform production in soils. As the regulatory response to and the potential measures against chloroform in groundwater may depend upon whether or not contamination or natural soil processes are the sources, site-specific demonstration of natural chloroform production becomes an important tool in groundwater management, and, thus, reliable and efficient testing methods are needed. For a more detailed discussion of the magnitude of chloroform production in soil, the annual variation of the production and the land use variability in production, reference is made to Laturnus et al. (2002) and Albers et al. (2010).

Materials and methods Site description The study site is located near Viborg in the central part of Jutland, Denmark (N 56◦ 25 , E 09◦ 20 ) with a temperate, coastal climate (average annual precipitation is approximately 750 mm and

1233

average annual air temperature is 7.5◦ C). The main site is a fluvio-glacial sandy deposit from the last glacial with glacial tills to one side with a loamy sandy texture. The parent material is dominated by quartz (approximately 90%) and feldspar (approximately 10%). Heather (Calluna vulgaris (L.) Hull.) dominated the site until the eighteenth century when plantation of coniferous trees began. Today, the plantation is dominated by 30- to 60-year-old Norway spruce (Picea abies (L.) Karst. ‘Acrocona’). A more than 125-year-old beech (Fagus sylvatica L.) forest is present in the western part of the study site. Some parts of the heather have been converted to agricultural land since at least 120 years. Today, soils are podzols or arenosols. The aquifer is in fluvio-glacial sands with few layers of silt and a groundwater table 5 to 10 m below the surface. In the shallowest parts of the aquifer, the groundwater has chloroform concentrations of 0.1 to 2 μg L−1 , and the groundwater is aerobic with an age less than 5 years as shown using chlorofluorocarbon dating. The ratio of stable carbon isotope values in chloroform (δ 13 C) ranged between −13‰ and −27‰ in the shallow groundwater (Laier et al. 2005) and indicated that the origin of the chloroform in groundwater was not industrially produced chloroform, as this would exhibit δ 13 C values of −40‰ to −50‰ (Hunkeler and Aravena 2000; Jendrzejewski et al. 2001; Jochmann et al. 2006). No sources of groundwater contamination were found up gradient of the four sampling sites selected for the study that could impact the groundwater chloroform concentrations at the locations. No major sources of air contamination with chloroform were present in the area. The absence of chloroform contamination at the sampling site is, thus, well established, and accordingly a natural production in soil is the most plausible source of the chloroform found in shallow groundwater. At the site, four sampling locations were selected representing the same geological parent material and original soil development (sandy podzol developed under the heather) but with different present-day soil profiles due to their different vegetation (see Table 1). The site investigations and sampling were done in autumn with ambient air temperatures in the range of 10.9 to

1234

Environ Monit Assess (2012) 184:1231–1241

Table 1 Characteristics of the four sampling sites selected for the study and data (± standard deviation, n = 3) on their soil chemistry Beech forest

Spruce forest

Grassland

Grain field

Norway spruce forest

Hay/pasture since 1987

Soil texture

Mixed young and old forest Fine sand

Fine sand

Fine sand with gravel

Small-grained cereal rotation Fine grained sand with gravel

Profile

Depth [cm]

Horizon

Depth [cm]

Horizon

Depth [cm]

Horizon

Depth [cm]

Horizon

5–1 1–0 – 0–5 5–35 35–100

L F – A B C

9–5 5–2 2–0 0–15 15–45 45–100

L F H A B C

– – – 0–30 30–50 50–105

– – – A B C

– – – 0–28 28–40 40–80

– – – A B C

[%] [%]

5.0 ± 0.16 <0.5 79 ± 44 6.7 ± 1.8 9.4 ± 10 12 ± 4.0 0.38 26.6 ± 2.3 9.9 ± 2.3

[%] [%]

6.2 ± 0.12 5.8 ± 4.5 15 ± 4.7 150 ± 17 <0.5 19 ± 11 0.75 14.9 ± 0.2 2.6 ± 0.1

[%] [%]

6.4 ± 0.17 3.8 ± 3.6 11 ± 0.35 160 ± 28 8.3 ± 10 10 ± 3.1 0.75 11.5 ± 0.9 1.9 ± 0.4

[%] [%]

Vegetation

pHa Fluoridea Chloridea Nitratea Phosphatea Sulfatea Soil densityb Water contentc TOC

5.2 ± 0.26 2.8 ± 3.8 15 ± 3.7 38 ± 58 <0.5 17 ± 2.0 0.56 32.9 ± 2.9 5.6 ± 0.7

L litter layer, H humus layer, B precipitation layer, F fragmentation layer, A leached layer, C original matter, dw dry weight kg−1 dw] b [g cm−3 ], estimated from the weight of the soil used to fill 1/3 of a 120-mL glass flask c [%] of fresh weight soil a [mg

16.5◦ C and soil temperatures in the range of 10.0 to 13.2◦ C. Selected data on the soil chemistry of the four locations are given in Table 1. Soil, soil air, and air sampling During field sampling, four different test methods were established: sampling of ambient air for concentration measurements, sampling of soil air for concentration measurements, sampling of soil air for field production studies, and sampling of soil for laboratory incubation studies and soil characterization. At each of the four sampling locations, three sampling spots were selected within an area of 10 × 10 m yielding three replicates for each test method. To conduct the laboratory incubation experiments and as there was no possibility to analyze the soil directly in the field for example for soil humidity, samples of the four different soil types were collected from 0 to 10 cm, homogenized, and stored in 3-L glass jars with 1/3 headspace at a temperature of 20◦ C in the dark

until used. Analyses of the soil for a variety of physico-chemical parameters were done in the laboratory 5 days after sampling. The results are presented in Table 1. Ambient air and soil air samples were collected on adsorbent tubes made of glass (L 150 mm, ID 4 mm, OD 6 mm) filled with 200 mg HayeSep R adsorbent (mesh 80/100, Supelco). Prior to sampling, the sampling tubes were pre-cleaned with ultra-pure helium (temperature program: 5 min at 22◦ C, 15 min at 200◦ C, 10 min at 22◦ C). After pre-cleaning, the sampling tubes were sealed at both ends with Swagelok© fittings and polytetrafluoroethylene (PTFE) ferrules, wrapped into aluminum foil and stored in a transport container. Ambient air samples were collected by connecting the adsorbent tube to a flow regulator followed by a membrane pump. One liter of ambient air was collected against the wind direction 50 cm above ground at a flow rate of 100 mL min−1 . For soil air, a special stainless steel cylinder with a volume of 3 L (depth 10 cm, ID 20 cm) was used enabling

Environ Monit Assess (2012) 184:1231–1241

soil air sampling of up to 1 L from a defined soil layer without co-sampling the surrounding layers. The adsorbent tube was connected to the cylinder and 600 mL soil air was collected at a flow of 100 mL min−1 . All adsorbent tubes were analyzed immediately after returning to the laboratory from the field sampling. For more details of the methods, see Laturnus et al. (2000). Field incubation studies To measure the chloroform production in soil, a method was used already established and described previously by Haselmann et al. (2000a). Samples for soil chloroform production were taken directly after the collection of soil air using the same stainless steel cylinder at the same sampling spot without having removed the cylinder from the soil prior taking samples for soil chloroform production. While the steel cylinder was still in the soil, the soil air within the cylinder was replaced with 2 L of synthetic air within 20 min. To monitor a possible release of chloroform into soil, the cylinder was left in the ground for 2.5 h. To subsequently collect the trapped soil air with produced chloroform, an adsorbent tube was connected to the cylinder and 600 mL of soil air was collected with a pump at a sampling rate of 100 mL min−1 . The ambient temperature during field incubation varied between 10.9 and 16.5◦ C. Soil humidity was determined first in the laboratory. Three replicate incubations were done at each sampling site, i.e., one for each sampling spot. Laboratory incubation studies The laboratory incubation studies included release of chloroform from untreated soil and from soil spiked with sodium chloride. The soils from the three sampling spots at each location were mixed to one composite sample per location, but not sieved. All soil samples were treated in the same way. Three untreated and three chloridespiked replicates were prepared for each sampling location using the composite soil samples. Incubation studies were done in two series after 14 and 21 days of storage. Furthermore, a 61-day pre-incubation study was done to investigate how

1235

longer exposure of the soil microorganisms to added chloride would impact the soil chloroform production, i.e., whether or not adaptation of soil microorganisms to added chloride occurred. The soil incubation studies were done in 120 mL glass flask pre-cleaned with acetone: hexane (1:1) and stored at 70◦ C until incubation start. Fifteen to 30 g of untreated soil were filled in a glass flask to which 1–2 mL of clean water (Milli-Q system from Millipore) were added. The amount of soil in each flask varied with the density of the soils (see Table 1), as a headspace inside the incubation flask of around 2/3 of its volume was aimed at. The amount of water added to each flask was adjusted to the water content of the untreated soil to obtain the same water content in all flasks. The soil humidity of the four different soil types varied between 11.5% and 32.9% (see Table 1). Immediately after the soil was filled into the glass flasks and the water was added, the flasks were sealed with an aluminum cap with PTFE/butyl septum. The glass flasks were incubated for 2.5 h in the dark at a temperature of 20◦ C followed by immediate analysis of the headspace above the soil for chloroform. The chloride-spiked incubation studies were done in the same way as described in the paragraph above. To attain identical soil humidity as in the incubation studies of untreated soils, 1 to 2 mL of a 750 mg L−1 sodium chloride solution were added to the soil samples. The amount of chloride added to the soil was 50 mg kg−1 dry weight soil and corresponded to one to four times the chloride concentrations determined in the different soil types (see Table 1). The pre-incubation study was done similarly to the incubation studies described above. However, prior to sealing the flasks with aluminum caps for an incubation period of 2.5 h, the flasks were stored unsealed in the dark at a temperature of 20◦ C for 2 months allowing free exchange of air within the flask with the surrounding air. Prior to the laboratory pre-incubation study, the amount of water to be added to the soil to avoid desiccation of the soil sample during the long pre-incubation periods was determined in a water evaporation test (Table 2). In this desiccation test, samples of each soil type were added to 120-mL glass flasks according to the experimental set-up

1236

Environ Monit Assess (2012) 184:1231–1241

Table 2 Water evaporation test at storage temperature on different soils investigated in this study Days

Beech forest

Spruce forest

Grassland

Grain filed

Water loss during incubation time [%] 0 2 5 9 13 20

0 9.1 17.6 28.3 39.0 61.2

0 8.9 18.9 27.7 37.7 64.4

0 11.4 28.8 35.0 50.7 91.3

0 13.3 29.7 50.5 84.4 100

Storage and incubation temperature was at 20◦ C

of the laboratory incubation study. The incubation flasks were stored without a lock in the dark at 20◦ C and the remaining water content was determined after different time intervals (Table 2). From the results of the experiment, the amount of water necessary to reduce humidity fluctuation and to avoid desiccation of the soil was set to 1.5 mL tap water added to the soil once a week.

5 min, heating rate 5◦ C min−1 , 200◦ C isothermal for 5 min (total analysis time = 42 min). Identification and quantification was done by external calibration with chloroform (p.a.) dissolved in methanol (p.a.). For soil samples, three replicates of soil from each location were analyzed separately. Water content was determined after drying at 105◦ C overnight or to constant weight. Total organic carbon content (g TOC kg−1 ) was determined on ball-milled subsamples using a LECO CNSanalyzer with IR-detector (LECO Corporation, St. Joseph, USA). The pH was determined in 0.01 M CaCl2 with a soil to liquid ratio of 1 to 2.5. Fluoride, chloride, nitrate, phosphate, and sulfate were determined with a DIONEX ionchromatograph (Dionex Corporation, Sunnyvale, USA) after extraction in Milli-Q water applying a soil to liquid ratio of 1 to 10. The results were normalized to soil dry weight. Quality control

Analysis All air samples were analyzed by thermodesorption cryofocusing gas chromatography with electron capture detection (GC-ECD) as described in details in Laturnus et al. (2000). In short, the analyses were performed as follows. Prior to thermodesorption, the adsorption tubes were flushed for 5 min with ultra-purified helium (pre-cleaned by OMI-3 gas cleaner, Supelco) to remove water from the adsorption tubes. The chloroform was desorbed from the adsorption tubes at 200◦ C for 10 min with a helium flow of 80 mL min−1 and preconcentrated on a cold-trap submerged into liquid nitrogen. Transfer onto the capillary column was done by rapidly heating the cold-trap with boiling water. Headspace analysis of the soil incubation samples was done by purging the headspace in the bottles over the soil with ultra-purified helium for 15 min and analysis by cryofocusing GC-ECD as described above. For detailed description, see Haselmann et al. (2000b) and Borch et al. (2003). Gas chromatographic separation was done on CP-PoraBond Q column (Chrompack, L 25 m, ID 0.32 mm, film thickness 5 μm) using the following temperature program: 40◦ C isothermal for

All samples, measurements and incubations were done in triplicate to enable estimation of the precision (repeatability). Precision was estimated as standard deviation and presented with the data as pertinent. For air and soil air measurements, blank adsorption tubes were transported, handled, and analyzed as the field samples except for the exposure to air to provide the true field blanks used for estimation of the field detection limit from the field blank standard deviation as described in Haselmann et al. (2002). Laboratory detection limits were estimated from the laboratory blank standard deviation. The instrument detection limit for chloroform was calculated as three times the baseline noise. An analytical recovery of 98% was determined by desorption of chloroform added to an adsorbent tube. Data below the detection limit were generally not included in the results presented. However, if required for calculation or presentation purposes, data point set to 1/2 times the detection limit was used. Results and discussion To evaluate the potential for naturally produced chloroform in soil as a source of chloroform in

0.019 ± 0.0027 0.193 ± 0.041 <0.007 <0.007 Values ± standard deviation (n = 3). Measurements below the detection limits were omitted from the calculations

<0.017 0.57 ± 0.034 0.032 ± 0.003 <0.017 0.033 ± 0.017 0.64 ± 0.074 0.044 ± 0.016 0.024 0.13 ± 0.054 0.62 ± 0.024 0.072 ± 0.035 0.041 ± 0.011 7.4 ± 2.3 10.9 ± 2.7 7.5 ± 1.7 6.6 ± 3.2 Beech forest Spruce forest Grassland Grain field

0.921 ± 0.921 1.5 ± 0.84 1.2 ± 0.75 0.75 ± 0.25

11.7 ± 4.5 31.8 ± 8.4 15.9 ± 7.6 15.1 ± 5.1

[pmol g−1 2.5 h−1 ] [pmol g−1 2.5 h−1 ]

With chloride added [pmol L−1 ] [pmol L−1 ]

[pmol L−1 2.5 h−1 ]

Laboratory production 14 days 21 days [pmol L−1 2.5 h−1 ] Without chloride added Field production Soil air concentration Ambient air concentration Vegetation

Table 3 Soil air concentrations, ambient air concentrations and production rates of chloroform determined in field and laboratory studies

groundwater, an approach could be to determine the relative soil chloroform production, i.e., to measure the production of soils in the suspected source area relative to the (presumably) lower production of soils on comparable geology but outside the suspected source area. The most appropriate method for estimation of the relative chloroform production in soil would be the method with the largest difference between the chloroform releases measured for high and for low production soils. In this study, three methods to determine a natural production of chloroform have been compared: (a) soil air to ambient air ratio, (b) measurement of a direct production in soil, and (c) measurement of the emission of chloroform from soil incubated in the laboratory. To determine the most effective method for chloroform production studies a comparison was done for every method between soil with lowest chloroform production (grain field) and soil with highest chloroform production (spruce forest soil; Table 3). The percentage differences between the two soil sites for every method are given in Fig. 1. The ratio between soil air concentration and ambient air concentration has previously been used to evaluate and compare chloroform production at different soil sites (Haselmann et al. 2000a, b). The background for this is that higher chloroform concentrations can occur in soil air than in ambient air only if a release is taking place in the soil. Comparison of the soil air to ambient air ratios for grain field soil and spruce forest soil revealed only a small difference between both soil sites (Fig. 1, Table 3). The soil air to ambient air ratios were similar at both soil sites, i.e., the differences between high and low production soils was poor. It should be mentioned though, that the soil air to ambient air ratio showed better differences, when soil air was sampled at 50 cm depth in contrast to the 0 to10 cm sampling depth used here (data not shown). The reason probably was due to a high upwards diffusion of produced chloroform from the top soil into the atmosphere, which leads to low concentrations of chloroform in the upper centimeters of the soil even in periods of high biological production, such as spring and autumn (Grøn 1995; Laturnus et al. 2000). The differences between field sites for the field soil production method increased to around 50%

1237 61 days [pmol g−1 2.5 h−1 ]

Environ Monit Assess (2012) 184:1231–1241

1238

Environ Monit Assess (2012) 184:1231–1241

100

difference [%]

80

60

40

20

0 soil air ratio

field production studies

laboratory incubation studies

Fig. 1 Percentage differences between the results determined from spruce forest soil (highest production) and grain field soil (lowest production). Error bars indicate ± standard deviation for n = 3

while the laboratory incubation showed the highest difference (Fig. 1, Table 3). Evidently, the laboratory incubation studies provided the best difference between high and low production soils. The difference for the field production method may be improved by extending the incubation period, as this is known to increase the production up to a factor of 3 after 38 h (Haselmann et al. 2000a), but for routine use this may not be practical enough. The precision estimated as relative standard deviation was in the range of 20% to 75% for the soil air to ambient air method, and 25% to 50% for the field soil production method. The precision includes both the precision of the measurement methods and the significant soil heterogeneity derived variability (Haselmann et al. 2000a). In general, the precision of the laboratory studies was better than for the field method as would be expected considering that the laboratory incubation studies were done with composite samples, and, thus, were less impacted by soil heterogeneity. So far, the approach through laboratory incubation seems to be the most effective one. However, for practical use of the methods, the allowable and optimal storage period for soil samples in the laboratory incubation studies needed be determined too. Figure 2a shows the difference between production in spruce forest soil and grain field soil after 14, 21, and 61 days of pre-incubation of the soil samples. Although the 61 days pre-incubation

period showed the highest differences, the actual chloroform production was decreasing and the precision is deteriorating with pre-incubation length (Fig. 2b). For the grain field soil, the chloroform production is approaching (21 days) or gets below the test detection limit (61 days). Thus, a maximum pre-incubation period of 14 days is suggested, as this provided a reasonable combination of difference and precision in this study. In contrast, Albers et al. (2010) described an increase with longer incubations periods (>30 days). However, they assumed that a low temperature during storage of soils samples prior the incubation lead to a lag phase were microbes needed to build up sufficient biomass first before the formation of chloroform occurred. For further experimental set-ups, the authors suggested to collect soil in productive periods of the year and store the soil samples at temperatures close to the soil temperature at time of the sampling. The effects of preincubation periods shorter than 14 days were not studied here, and, thus, cannot be evaluated. Furthermore, the decrease in chloroform production with pre-incubation length may be due to different factors, such as change in soil chemistry, in nutrient availability and/or in microbial community, and can in an irreproducible way lead to a bias of the measured chloroform production. Therefore, reference is made to methods dedicated to quantification of soil chloroform production (Haselmann et al. 2000a, b; Albers et al. 2010), if emission estimates rather than relative production evaluation is the purpose of a study. In tests such as the OECD test for ready biodegradability of contaminants in water, nutrients are added to prevent limitation of microorganism growth and contaminant degradation (OECD 1992). Similarly, addition of, e.g., chloride or nitrogen to laboratory incubations can be considered, as these have been suggested to be important for the processes controlling the presence of chlorinated organic compounds in soils (Johansson et al. 2001, 2003). From Fig. 3, the effect of adding chloride to one to four times the original soil water chloride concentration appeared to be a reduction in the production of chloroform rather than an increase, although the differences were not statistically significant. It can be argued that the chloride concentration in the

Environ Monit Assess (2012) 184:1231–1241

a 97 96

difference [%]

Fig. 2 Percentage differences between chloroform production estimate for spruce forest and grain field obtained with three different pre-incubation periods (a) and the corresponding chloroform production with test standard deviation (b). Error bars indicate ± standard deviation (n = 3)

1239

95 94 93 92 91 14 days

21 days

61 days

pre-incubation period

release rate [pmol g -1 2.5hr -1)]

b

0.8 spruce forest grain field

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

10

20

30

40

50

60

70

pre-incubation period [days]

spruce forest soil was already excessive and, thus, excluded an effect of further chloride addition. However, the same lack of chloride effect was found for the other vegetation types with five to seven times less inherent chloride (see Table 1). The reason for the high chloride concentration in spruce forest soil is probably a high dry deposition of sea-derived chloride that is efficiently captured

0.8

release rate [pmol g -1 2.5hr -1]

Fig. 3 Chloroform production for the four vegetation types with and without added chloride with test standard deviation. Bars indicate ± standard deviation (n = 3)

by the spruce forest as described for a similar area by Hansen and Postma (1992). The results of chloride addition to the soil incubations, thus, did not support to include chloride additions into a test method for determination of a natural chloroform production. In this study, the highest chloroform production was found in spruce forest soil with the lowest

0.7

no chloride chloride

0.6 0.5 0.4 0.3 0.2 0.1 0 beech forest

spruce forest

grassland

grainfield

1240

pH, nitrate concentration and density, and highest chloride concentration and water content (Table 1; Fig. 3). The second highest chloroform production was found in beech forest soil with the second lowest pH, density, and nitrate. The lowest chloroform productions were observed for grassland and grain field soils with similar—and highest—pH, density, and nitrate concentrations. The data obtained for all soils were not sufficient to provide statistical support for a general conclusion regarding the relation between soil pH, chloride concentrations, nitrate concentrations, density, water content, and chloroform production. However, the results showed that nitrate spiking apparently is not an essential part of the incubation methods. Conversely, removal of soil nitrate or adjustment of soil pH is hardly practical for a test of production of chloroform in soil. In contrast, high water content and low soil density apparently support the formation of chloroform possibly due to a better and faster accessibility of microbial substrates in this type of soil for microorganisms considered to be involved in the formation of chloroform (Laturnus et al. 2005). Different to changes in nitrate concentrations and pH, however, the variation of these factors can easily be adjusted in laboratory set-ups. A final point of discussion is the fact that the “true” chloroform production of soils is not known, and references or standards for validation of the field measurements and laboratory incubation studies are not yet available. All methods hitherto introduced do have their limitations. For the measurement of soil air to ambient air concentrations, the results can not be directly related to a soil volume and a production period primarily due to diffusion. For the field soil production method, the release of chloroform adsorbed previously to the soil and diffusion out of the bottom of the measuring cylinder can impact the results. Finally, the manipulation of the soil and inherent microorganisms as done in laboratory incubation studies will impact the production results. Additionally, soil production varies with season of the year (i.e., temperature and humidity) (Haselmann et al. 2002) and with the contents of decaying organic matter (e.g., position in a forest) (Haselmann et al. 2000a; Hoekstra et al. 2001). Until macrocosm studies on chloroform production in soils under

Environ Monit Assess (2012) 184:1231–1241

fully controlled conditions have been performed, reliability upon relative measurements is the only choice for an estimation of the production of chloroform in soils in, for example, groundwater resource management or atmospheric chlorine budgeting. Acknowledgements Funding from the Danish Environmental Protection Agency (DEPA) and Energy Viborg (EV) is gratefully acknowledged. The authors further acknowledge comments by Søren M. Kristiansen, Troels Laier, Daniel Hunkeler. The help and support from Susanne Rasmussen (DEPA), Palle Fløe Holm (EV) and Ulrik Vestergaard (EV) have been essential for the success of the studies.

References Albers, C. N., Laier, T., & Jacobsen, O. S. (2010). Formation and leaching of chloroform from coniferous forest soils. Applied Geochemistry, 25, 1525–1535. Asplund, G., Christiansen, J. V., & Grimvall, A. (1993). A chloroperoxidase-like catalyst in soil: Detection and characterization of some properties. Soil Biology and Biochemistry 25, 41–46. Auer, N. R., Manzke, B. U., & Schulz-Bull, D. E. (2006). Development of a purge and trap continuous flow system for the stable carbon isotope analysis of volatile halogenated organic compounds in water. Journal of Chromatography A, 1131, 24–36. Bishop, P. K., Misstear, B. D., White, M., & Harding, N. J. (1998). Impacts of sewers on groundwater quality. Water and Environment Journal, 12, 216–223. Borch, T., Ambus, P., Laturnus, F., Svensmark, B., & Grøn, C. (2003). Biodegradation of chlorinated solvents in a water unsaturated soil. Chemosphere, 51, 143–152. European Council of Ministers (1998). Council Directive 98/83/EC of 3 November 1998 on the quality of water intended for human consumption. Grøn, C. (1995). AOX in Groundwater. NaturallyProduced Organohalogens. In Selected and edited proceedings of the f irst international conference on naturally produced organohalogens (pp. 49–64), 14–17 September 1993, Delft, Holland. Hansen, B. K., & Postma, D. (1992). Groundwater acidification in the unsaturated zone of a sand deposit: Processes and modeling. In Proceedings of the 7th international symposium on water-rock interaction (pp. 389–392), 13–18 July 1992, Utah, USA. Haselmann, K. F., Ketola, R. A., Laturnus, F., Lauritsen, F. R., & Grøn, C. (2000a). Occurrence and Formation of Chloroform at Danish Forest Sites. Atmospheric Environment, 34, 187–193. Haselmann, K. F., Laturnus, F., Svensmark, B., & Grøn, C. (2000b). Formation of chloroform in spruce forest soil—results from laboratory incubation studies. Chemosphere, 41, 1769–1774.

Environ Monit Assess (2012) 184:1231–1241 Haselmann, K. F., Laturnus, F., & Grøn, C. (2002). Formation of chloroform in soil—A year-round study at a Danish spruce forest site. Water, Air, and Soil Pollution, 139, 35–41. Hoekstra, E. J., de Leer, E. W. B., & Brinkmann, U. A. T. (1998). Natural formation of chloroform and brominated trihalomethanes in soil. Environmental Science & Technology 32, 3724–3729. Hoekstra, E. J., Duyzer, J. H., de Leer, E. W. B., & Brinkmann, U. A. T. (2001). Chloroform— concentration gradients in soil air and atmospheric air, and emission fluxes from soil. Atmospheric Environment, 35, 61–70. Hunkeler, D., & Aravena, R. (2000). Determination of compound-specific carbon isotope ratios of chlorinated methanes, ethanes, and ehtenes in aqueous samples. Environmental Science & Technology, 34, 2839– 2844. Jendrzejewski, N., Eggenkamp, H. G. M., & Coleman, M. L. (2001). Characterisation of chlorinated hydrocarbons from chlorine and carbon isotopic compositions: Scope of application to environmental problems. Applied Geochemistry, 16, 1021–1031. Jochmann, M. A., Blessing, M., Haderlein, S. B., & Schmidt, T. C. (2006). A new approach to determine method detection limits for compound-specific isotope analysis of volatile organic compounds. Rapid Communication in Mass Spectrometry, 20, 3639–3648. Johansson, E., Ebenå, G., Sandén, P., Svensson, T., & Öberg, G. (2001). Organic and inorganic chlorine in Swedish spruce forest soil: Influence of nitrogen. Geoderma, 101, 1–13. Johansson, E., Sandén, P., & Öberg, G. (2003). Spatial patterns of organic chlorine and chloride in Swedish forest soil. Chemosphere, 52, 391–397. Laier, T., Jacobsen, O. S., Thomsen, O., Grøn, C., Hunkeler, D., & Laturnus, F. (2005). Chloroform production in spruce forest soils—A potential problem for groundwater use in drinking water supply in Denmark. Geophysical Research Abstracts, 7, 05884. Laturnus, F., Mehrtens, G., & Grøn, C. (1995). Haloperoxidase-like activity in spruce forest soil—A source of volatile halogenated organic compounds? Chemosphere, 31, 3709–3719. Laturnus, F., Lauritsen, F. R., & Grøn, C. (2000). Chloroform in a pristine aquifer system: Toward and evidence of biogenic origin. Water Ressources Research, 36, 2999–3009. Laturnus, F., Haselmann, K. F., Borch, T., & Grøn, C. (2002). Terrestrial natural sources of trichlorometh-

1241 ane (chloroform, CHCl3)—An overview. Biogeochemistry, 60, 121–139. Laturnus, F., Fahimi, I., Gyndler, M., Hartmann, A., Heal, M. R., Matucha, M., et al. (2005) Natural formation and degradation of chloroacetic acids and volatile organohalogens in forest soil—Challenges to understanding. Environmental Science and Pollution Research, 12, 233–244. Laturnus, F., & Matucha, M. (2008) Chloride—A precursor in the formation of volatile organochlorines by forest plants? Journal of Environmental Radioactivity, 99, 119–125. Matucha M, Clarke N, Lachmanova Z, Forczek S T, Fuksova K, & Gryndler M (2010) Biogeochemical cycles of chlorine in the coniferous forest ecosystem— Practical implications. Plant Soil Environ, 56, 357–367. McCulloch, A. (2003). Chloroform in the environment: Occurrence, sources, sinks and effects. Chemosphere, 50, 1291–1308. Miljøministeriet (2007). Bekendtgørelse om vandkvalitet og tilsyn med vandforsyning. Bekendtgørelse 1664. Mills, W. B., Lew, C. S., & Loh, J. Y. (1998). Predictions of potential human health and ecological risks from power plant discharges of total residual chlorine and chloroform into rivers. Environmental Science & Technology, 32, 2162–2171. Moran, M. J., Hamilton, P. A., & Zogorski, J. S. (2006). Volatile organic compounds in the Nation’s ground water and drinking-water supply wells—A summary (p. 6). U. S. Geological Survey Fact Sheet 2006–3048. Niedan, V., Pavasars, I., & Öberg, G. (2000). Chloroperoxidase-mediated chlorination of aromatic groups in fulvic acid. Chemosphere, 41, 779–785. Öberg, G., Börjesson, I., & Samuelsson, B. (1996). Net change in organically bound halogens in relation to soil pH. Water, Air, and Soil Pollution, 89, 351–361. Öberg, G., Brunberg, H., & Hjelm, O. (1997). Production of organically bound chlorine during degradation of birch wood by common white-rot fungi. Soil Biology and Biochemistry, 29, 191–197. OECD (1992). Test no. 301: Ready biodegradability. In OECD guidelines for the testing of chemicals, Section 3: Degradation and accumulation. OECD iLibrary. doi:10.1787/2074577x. Reina, R. G., Leri, A. C., & Myneni, S. C. B. (2004). Cl K-edge X-ray spectrocopic investigation of enzymatic formation of organochlorines in weathering plant material. Environmental Science & Technology, 38, 783–789. World Health Organisation (2006). Guidelines for drinking-water quality (3rd ed.).