VZJ sign up for etocs
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 9 April 2007
Published in Vadose Zone J 6:233-243 (2007)
DOI: 10.2136/vzj2006.0083
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Totsche, K. U.
Right arrow Articles by Kögel-Knabner, I.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Totsche, K. U.
Right arrow Articles by Kögel-Knabner, I.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Totsche, K. U.
Right arrow Articles by Kögel-Knabner, I.
Related Collections
Right arrow Lysimeter/Rhizosphere Studies
Right arrow Organic Compounds
Right arrow Soil Pollution

ORIGINAL RESEARCH

Single Event–Driven Export of Polycyclic Aromatic Hydrocarbons and Suspended Matter from Coal Tar–Contaminated Soil

Kai Uwe Totschea,*, Steffen Janna and Ingrid Kögel-Knabnerb

a Inst. für Geowissenschaften, Friedrich-Schiller-Univ. Jena, 07749, Jena, Germany
b Lehrstuhl für Bodenkunde, Dep. für Ökologie, Wissenschaftszentrum Weihenstephan, Technische Univ. München, D-85350 Freising-Weihenstephan, Germany
* Corresponding author (kai.totsche{at}uni-jena.de).

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


Received 26 June 2006.



   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
Mobile colloidal and suspended matter is likely to affect the mobility of polycyclic aromatic hydrocarbons (PAHs) in the unsaturated soil zone at contaminated sites. We studied the release of mobile particles and dissolved organic matter as a function of variable climatic boundary conditions, and their effect on the export of PAHs at a coal tar–contaminated site using zero-tension lysimeters. Seepage water samples were analyzed for dissolved organic carbon (DOC), pH, electrical conductivity, turbidity, and particles larger than 0.7 µm. The 16 Environmental Protection Agency PAHs were analyzed in the filtrate <0.7 µm and in the particle fraction. Our results show that extended no-flow periods that are followed by high-intensity rain events, such as thunderstorms, promote the mobilization of particles in the size 0.7 to 200 µm. Mobilization is enforced by extended drying during summer. High particle concentrations are also associated with freezing and thawing cycles followed by either rain or snowmelt events. The export of PAHs is strongly connected to the release of particles in the 0.7- to 200-µm size fraction. During the 2-yr monitoring period, up to 0.418 µg kg–1 PAHs were mobilized in the filtrate (<0.7 µm) while the eightfold mass, 3.36 µg kg–1, was exported with the retentate (0.7–200 µm). Equilibrium dissolution of PAHs and transport in the dissolved phase seem to be of minor importance for the materials studied. Extreme singular-release events occurred in January 2003 and January 2004, when up to 55 µg L–1 PAHs per one single seepage event were observed within the retentate. Freezing and thawing cycles affect the PAH source materials, that is, the remnants of the nonaqueous phase liquid (NAPL). High mechanical strain during freezing results in the formation of particles. At the onset of the thawing and following rain or snowmelt events, PAHs associated with these particles are then exported from the lysimeter.

Abbreviations: DOC, dissolved organic carbon • EC, electrical conductivity • FAU, Formazine Attenuation Units • NAPL, nonaqueous phase liquid • OC, organic carbon • PAH, polycyclic aromatic hydrocarbons.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
The pollution of soil and groundwater with PAHs at former industrial sites is an important environmental issue. The PAHs feature low aqueous solubility (Mackay and Shiu, 1977) and high recalcitrance against degradation (Alexander, 1999; Cerniglia, 1993; Ghoshal et al., 1996; Weigand et al., 2002). They are often found at sites contaminated with NAPLs. The release of PAHs from NAPL has been shown to follow a dissolution process according to Raoult's law (Lane and Loehr, 1992; Lee et al., 1992; Mukherji et al., 1997; Eberhardt and Grathwohl, 2002). Raoult's law describes the aqueous concentration of a single PAH compound at equilibrium as the product of the compound's mole fraction in the NAPL phase and its aqueous solubility. However, growing evidence indicates that, particularly for aged NAPLs, the release is controlled by nonequilibrium conditions and affected by colloids and even larger mobile particles (Villholth, 1999; MacKay and Gschwend, 2001; Totsche et al., 2006; Totsche and Kögel-Knabner. 2004). Nonequilibrium release may arise from the formation of high-viscous boundary layers as a result of aging, that is, the depletion in soluble and volatile compounds, the biochemical transformation and polymerization of NAPLs at the NAPL–air or NAPL–water interface. Rate limitations to the mass transfer of NAPL-borne compounds are the consequence (Luthy et al., 1993; Nelson et al., 1996; Totsche et al., 2003; Ghoshal et al., 2004). Thus, Raoult's law may not be applicable for predicting aqueous concentrations of total PAHs (Mahjoub et al., 2000).

Mobile particles may act as carriers for PAHs and other strongly sorbing solutes and thus facilitate their transport (Chiou et al., 1986; Grolimund et al., 1996; Ryan et al., 1998; Villholth, 1999; MacKay and Gschwend, 2001; Kim and Corapcioglu, 2002; de Jonge et al., 2004). In a recent column outflow study using NAPL-contaminated soil materials, Totsche et al. (2006) found up to 42% of the total exported PAHs in the particle-size fraction 0.7 to 200 µm. The PAHs in the filtrate <0.7 µm were mobilized under rate-limited conditions and showed a similar release behavior to DOC. The authors concluded that the dominant processes are the release of PAH-bearing NAPL fragments or droplets with the first flush—that is, the rewetting and flow initiation at the onset of the experiment or after longer no-flow conditions—the hydraulic mobilization after flow-interrupts, and to a lesser extent, the rate-limited release.

Particle mobilization can be triggered by a change of pH, speciation, or ionic strength, or a change of the hydrodynamic forces due to variations of the flow velocity or to transient flow conditions (McDowell-Boyer 1992, Ryan and Gschwend 1994, Ryan and Elimelech 1996, Bunn et al., 2002, Lenhart and Saiers, 2002). Such changes occur as a result of the variability of the climatic boundary conditions (precipitation, temperature) and the heterogeneity of the soils (McCarthy and McKay, 2004). However, little is known about the release of PAHs from the unsaturated zone at NAPL-contaminated sites under natural conditions (Saison et al., 2004). In particular, the effect of climatic forcing on particle release and on the fate of PAHs under field conditions remains unclear.

Our objective was to study the release of PAHs, DOC, and colloids and particles as a function of variable climatic boundary conditions. Zero-tension lysimeters were installed at an abandoned industrial site and packed with coal tar–contaminated soil material. Previous column experiments with the same soil material (Totsche et al., 2006) showed that particles up to 200 µm might be mobile within the coarse gravelly soil material. Therefore, the seepage water was filtered at 0.7 µm to distinguish between PAHs associated with large colloids and suspended particles (>0.7 µm) and PAHs in dissolved form or associated with small colloids (<0.7 µm). The applicability of Raoult's law was tested by comparing calculated and observed PAH concentrations in the filtrate (<0.7 µm).


   Materials and Methods
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results and Discussion
 Conclusions
 REFERENCES
 
Site Description and Soil Materials
The study was conducted at an abandoned industrial site within the city of Munich, Germany. The site had been used for tar distillation, fuel production, and mineral-oil processing for about 110 yr until it was put out of action in 1984. The parent materials are quaternary glacial and postglacial gravelly sediments of limestone and dolomite (80%), with minor amounts of igneous and metamorphic rocks (Baumann et al., 2002). The percentage of fine material (<2 mm) is generally <20% (w/w). Typical soil types are Calcaric Regosols and Calcaric Luvisols. Because of construction activities, the natural soil profiles have been destroyed. The unsaturated soil zone is contaminated by aged remnants of the fuel production, mineral and tar oils, in the NAPLs designated below. Although the site was partly remediated, high amounts (45 g kg–1 dry soil) of residual NAPLs are still left, heterogeneously and discontinuously distributed over the site at various depths. Major NAPL compounds are PAH and petroleum-derived hydrocarbons.

At present, the site is prepared for building. This includes excavation, filling, leveling, and compaction. Such activities will result in the disruption of the integrity of the NAPLs. Concomitantly, new NAPL surfaces are formed and exposed while others are coated with mineral soil material. Thus, it is expected that the construction activities will have a severe effect on the release of PAH from disturbed residual NAPL.

Properties of the Residual NAPLs
The NAPLs are heterogeneously distributed in the soil matrix in the form of residual rigid fragments closely attached to soil aggregates. To analyze the properties and mean composition of the NAPLs, 10 kg of the NAPL-contaminated soil were exhaustively extracted with hexane. Chemical analysis (Petrolab, Speyer, Germany) revealed aromatic compounds as major components (0.65kg kg–1 NAPL). The 16 Environmental Protection Agency priority pollutants (PAHs) amounted to 0.20 kg kg–1, with a dominance of the three-ring compounds (Table 1). The remaining part of the NAPLs, approximately 0.35 kg per kg–1 pure NAPL, was characterized as asphaltenes and paraffines. The density of the NAPLs was slightly larger than 1, and the residual water content less than 0.023 kg kg–1 NAPL. Molecular weight was determined to be 295 g mol–1. Based on these results, the NAPL is characterized as highly aromatized and viscous (almost rigid) tar oil.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Content of the individual polycyclic aromatic hydrocarbons in the coal tar.

 
Physical and Chemical Characterization of the Soil Material
Particle-size distribution was determined by dry (>2 mm) and wet sieving (sand fractions) and a sedimentation method with X-ray attenuation measurement (silt and clay fractions) (Micromeritics GmbH, Mönchengladbach, Germany). Total carbon content of the air-dried, homogenized, and ground soil material was quantified with an element analyzer (VarioEL; Elementar Analysensysteme, Hanau, Germany); inorganic and organic carbon content was calculated from the difference before and after combustion of the organic material at 560°C. The pH values were determined in deionized water and in 0.01 M CaCl2 solution using a soil-to-solution ratio of 1:2.5 (Avery and Bascomb, 1974). Extractable iron and manganese were determined by oxalate extraction and dithionite-citrate-bicarbonate extraction according to Schwertmann (1964) and Mehra and Jackson (1960), respectively, and quantified by inductively coupled plasma optical emission spectroscopy (Vista Pro CCD Simultaneous ICP-OES; Varian GmbH, Darmstadt, Germany). Bulk densities were determined with an excavation method as described by Blake and Hartge (1986). The physical and chemical properties of selected locations of the soil are given in Tables 2 and 3.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Depth distribution of selected soil properties at two locations (texture classes according to the German classification system).

 

View this table:
[in this window]
[in a new window]

  		TABLE 3. Composition of the lysimeters. (All lysimeters were covered by 0.05-m-thick vegetation collected at the site.)

 
The soil is composed of calcareous gravels with generally low fractions of sand, silt, and clay. Embedded sand and silt lenses were found (location 1, Table 2). The bulk density reached values up to 2.5 g cm–3 as a consequence of compaction during the construction activities. The carbonate content reached values up to 750 g kg–1, and the content of organic carbon (OC) was between 0.9 and 2.4 g kg–1. Nuclear magnetic resonance studies revealed that the organic matter consisted mainly of the residual and aged tar oil. Natural organic matter was negligible. The pH of H2O and CaCl2 showed values up to 9.4 and 8.0, respectively, due to the high carbonate content. The contents of dithionite soluble iron and manganese reached up to 123 mg kg–1 and 3 mg kg–1, respectively.

Lysimeter Study
The effect of climatic forcing on the release of PAHs, particles, and DOC was studied with commercially available zero-tension lysimeters (emc GmbH, Erfurt, Germany, Fig. 1 ). The monitoring was performed from September 2002 to October 2004, with a three-month interruption in spring and summer 2003 due to vandalism. Three lysimeters differing in the level of contamination, denoted here as Lys1, Lys2, and Lys3, were used. Lys1 reflected the background level for the site, while Lys2 and Lys3 were packed with soil material originating from areas still contaminated with fragments of the aged NAPL. The lysimeter size was 0.5 m by 0.5 m with a total height of 0.35 m (volume = 0.06 m3), with the exception of Lys3, which had a total height of 0.85 m so we could study the effect of prolonged residence time due to a longer flow path.


Figure 1
View larger version (40K):
[in this window]
[in a new window]

  		FIG. 1. Depiction of the lysimeter set-up in the field.

 
All lysimeters were made of stainless steel and were equipped with a stainless steel screen at the bottom (mesh size: 200 µm; emc GmbH) to allow export of particles up to 200 µm. The lysimeters were packed with soil material in a way that resembles the backfilling of the site during construction: excavated soil material with remnants of the NAPL was packed in the lysimeters and covered by an undisturbed, 2- to 5-cm-thick layer of vegetated topsoil collected at the same site. The procedure resulted in homogeneously packed lysimeters with bulk densities similar to those found in the surrounding soil (Table 1). Lys3 was equipped with two soil moisture probes (ECH2O; Decagon Devices, Inc., Pullman, WA) at a depth of 0.2 and 0.5 m. Soil temperatures were determined at depths of 0.15, 0.40, and 0.80 m and at ground level. Because of the repeated vandalism at the field site, these parameters were recorded only during selected periods in winter 2003–2004.

Precipitation was monitored with one rain gauge for each lysimeter. The precipitation was determined and effluent samples were taken immediately after precipitation events high enough to result in a recharge of seepage (event-triggered sampling). The seepage was collected in glass vessels. At each sampling event the vessels were exchanged and transferred to the lab. Thus, all material could be analyzed, including possibly settled matter. In general the period of time between recharge and sample collection was less than 24 h. In an accompanying experiment we found that PAH dissolution from the NAPL took more than 100 d for equilibration at 20°C and more than 150 d at 4°C. Thus, dissolution within the sampling vessels that would result in an increase of dissolved PAH can be neglected.

Analytical Methods
The seepage was analyzed for pH (ion-sensitive electrode, SenTix 41; WTW GmbH, Weilheim, Germany) and for electrical conductivity (TetraCon 625 conductivity cell; WTW GmbH). Turbidity was determined by spectral absorption at 860 nm (Cary 50 UV-Vis Spectrophotometer; Varian GmbH) and expressed as Formazine Attenuation Units (FAU). Before the measurement, samples were shaken horizontally 10 s, allowing 1 min settling. Dissolved organic carbon was measured as nonpurgeable organic carbon using a TOC analyzer (TOC-Analyzer 5050A; Shimadzu, Kyoto, Japan) after filtration <0.45 µm and acidification. Dissolved organic carbon is regarded as a possible carrier for PAHs in the fraction <0.7 µm. The difference between DOC <0.45 µm and the organic carbon <0.7 µm is generally <5% of the total liquid phase organic carbon.

Before PAH extraction, all samples were filtered with fiberglass filters (mesh size 0.7 µm, GF 92; Schleicher and Schuell MicroScience GmbH, Dassel, Germany) to distinguish between dissolved PAHs or PAHs associated with small colloids (filtrate) and those associated with larger colloids and suspended particles (retentate). The PAHs in the filtrate were extracted with a solid-phase extraction technique (Chladek and Marano, 1984). The PAH content of the soil material was determined according to Hartmann (1996). For retentate analysis we used a technique introduced by Kari and Herrmann (1989), modified as follows. Filters with retentate were deep frozen and subsequently freeze-dried. The freeze-dried retentate was then eluted twice with 40 mL hexane/acetone (SupraSolv; Merck, Darmstadt, Germany) and spiked with an internal standard of deuterated PAHs (Supelco; Bellefonte, PA). The solution was ultrasonified for 15 min by ultrasound and filtered with sodium sulfate to remove the water. The solution was then percolated through a hexane/dichlormethane-saturated silica gel/aluminum oxide column followed by 10 mL hexane and 15 mL of hexane/dichlormethane. The solvents were vaporized by rotation evaporator (Rotavapor R-144; Büchi, Flawi, Switzerland) to a minimum amount of 1000 {propto}L and filtered to eliminate remnants of silica gel using fiberglass filters (Macherey and Nagel, Düren, Germany). Finally, the sample was spiked with the recovery standard perylene D-12 (Supelco).

The PAHs were analyzed using a gas chromatograph coupled to a mass selective detector (GC 8000, MD 800; Fisons Instruments, Manchester, UK), equipped with a DB 5 MS column (internal diameter, 0.25 mm; film thickness, 0.25 µm; J. and W. Scientific, Folsom, CA). We used splitless injection with an injector temperature of 280°C and the following oven temperature program: 1 min 85°C, 85–160°C (15°C min–1), 160–300°C (5°C min–1), 15 min 300°C. The PAHs were quantified with a mixture of seven deuterated PAHs (PAH surrogate cocktail; Cambridge Isotope Laboratories, Inc., Andover, MA). We quantified the PAH recoveries of the internal standard by adding an external standard (perylene D-12, Supelco).

Raoult's Law
Observed PAH concentrations in the fraction <0.7 µm were compared with the PAH concentrations calculated using Raoult's law. The equilibrium concentration cw (mg L–1) of a solute in the aqueous phase is controlled by its mole fraction in the NAPL and its aqueous solubility:

Formula
where ct (mg g–1) is its tar-oil concentration, MWt (g mol–1) is the molecular weight of the tar phase, and Sw (mol L–1) is the compound's molecular subcooled liquid solubility. To calculate the equilibrium concentrations, we used the measured single PAH concentrations in the coal tar (Table 2 and 3), measured molecular weight of 295 g mol–1 for the coal tar, and subcooled liquid solubilities of the single PAH calculated after Peters et al. (1997) using aqueous solubilities from Mackay and Shiu (1977) and Walters and Luthy (1984). Additionally, we calculated equilibrium concentrations assuming minimum and maximum molecular weights of 230 and 780 g mol–1, respectively, for the coal tar (Lee et al., 1992) to cover a wider span of coal-tar types.

Statistics
To minimize the effect of extreme observations on the correlation statistics, we used Spearman's rank correlation to analyze the relationships between the parameters. This distribution-free analog of correlation analysis can be used to compare two independent random variables, each at several levels. Spearman's rank correlation works on ranked data rather than directly on the data itself. For each correlation calculated, the F-values were determined and the according significance levels of 0.1, 0.05, and 0.005 are presented.


   Results and Discussion
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
Conclusions
 REFERENCES
 
Temperature, Precipitation, Seepage Water, and Soil Moisture Conditions
The mean annual air temperature at the site is 8.9°C (1991–2000). The monthly mean air temperatures during the monitoring period reached their maximum value in August 2003 with 22.4°C and their minimum value in February 2003 with –4.4°C. The mean annual precipitation is 815 mm (1991–2000). The precipitation amount of the entire sampling period (2 yr) reached 1784 mm. Mean pH and electrical conductivity (EC) of precipitation was about 6.3 and 45 µS cm–1, respectively. In general, DOC was below the detection limit. The vandalism at the site interrupted monitoring between July and September 2003, affecting the rain gauges and the logger that was used to record the temperature and probes. We concluded that the lysimeters were not affected because the vegetation that grew on the them was not disturbed or destroyed. Thus, we are confident that the vandalism had no effect on the results.

Figure 2 shows the temperature and soil moisture conditions. Soil temperatures indicate major soil freezing events at the end of December 2003 to a depth of about 0.3 m. The soil moisture content was generally low because of the low water-holding capacity of the coarse material. The very low water content observed during summer 2004 resulted from an extremely hot and dry period.


Figure 2
View larger version (29K):
[in this window]
[in a new window]

  		FIG. 2. (a) Temperatures at ground level and at three depths below ground level (bgl), and (b) soil moisture conditions at two depths within Lys3.

 
Peak values of the precipitation and seepage water were 70 and 60 L m–2, respectively (Fig. 3 ). In general the seepage reacted quickly to the precipitation, with only small temporal offset. This was to be expected as evapotranspiration was low and the coarse material has a low water-holding capacity. Deviations between the course of precipitation and seepage water resulted from (i) freezing or snow fall in December 2002, January 2003, February 2003, December 2003, and January 2004; (ii) the corresponding snowmelts in December 2002, March 2003, and February 2004; and (iii) increased evapotranspiration in summer 2004.


Figure 3
View larger version (20K):
[in this window]
[in a new window]

  		FIG. 3. Mean precipitation and mean amount of seepage water during the 2-yr monitoring period. Single events connected with no-flow and drying periods are indicated by S; those connected with freezing–thawing are indicated by W.

 
Course of pH
The pH values of the seepage water of all lysimeters were generally lower than the equilibrium pH (H2O) of the material (Table 1) and varied from 7.1 to 8.4 (Fig. 4a ). The pH decreased after strong precipitation events and increased during dry periods. The lowest pH was observed in summer 2004.


Figure 4
View larger version (37K):
[in this window]
[in a new window]

  		FIG. 4. Course of (a) pH, (b) electrical conductivity, (c) turbidity, and (d) colloid and particle (0.7–200µm) concentration of the three lysimeters. Single events connected with no-flow and drying periods are indicated by S; those connected with freezing–thawing are indicated by W. The numbers in the subplots indicate the maximum values.

 
The pH is controlled by dissolution–precipitation of soil carbonates and release of organic matter. While the former results in a pH increase, the latter would eventually decrease the pH due to the deprotonation of acidic functional groups. From previous column experiments with the same material, we know that the dissolution of carbonates and the release of organic matter are strongly rate-limited (Totsche et al., 2006). The residence time of the soil solution, which is controlled by the intensity and duration of the precipitation, is thus another crucial factor that controls the seepage pH.

In general precipitation pH was weakly acidic and not buffered. Its infiltration will result in the dilution of the resident soil solution. The extent of the dilution, however, depends on the intensity and duration of the precipitation. Low intensity will cause only little dilution of the resident soil solution, with almost no seepage. Such conditions allow more time for carbonate dissolution and, consequently, an increase in the pH. During times of reduced microbial activity (e.g., during cold seasons), effects from the dissolution of the carbonates will outbalance those connected with the formation of dissolved or colloidal phase organic matter, resulting in higher pH. During times of enhanced microbial activity, on the other hand, we expect a decrease in the pH. We observed the lower pH values in the summers of 2003 and 2004. This effect was more distinct for Lys1 and Lys2, both of which had a higher topsoil/subsoil ratio. Thus, the travel distance for organic matter released in the OC-rich topsoil layer was smaller than the one in Lys3, as the carbonates attenuated or even outbalanced the pH decrease caused by the DOC.

In contrast, lasting and high-intensity rainfall and snowmelt events caused a marked dilution of the soil solution, increased flow velocities, and therefore reduced contact time. This results in a decrease of the pH in the seepage water, which is then controlled more by the precipitation. The significant negative correlation of pH with precipitation supports this interpretation (Table 4).


View this table:
[in this window]
[in a new window]

  		TABLE 4. Spearman's rank correlation coefficients calculated for selected bulk liquid parameters, polycyclic aromatic hydrocarbons (PAHs) in both fractions and colloids and particles (0.7–200 µm). Level of significance as obtained from the F-value of the respective rank correlation.

 
Course of Dissolved Organic Carbon
Figure 5 shows the course of DOC. In general, DOC export of the three lysimeters differed in the concentration level but not in temporal behavior. We observed extreme release of DOC at the beginning of the monitoring in October 2002 and after no-flow periods in June and October 2003. The high DOC concentration leveled off rapidly and reached values of 3 mg L–1 (Lys1, Lys2) and 5 mg L–1 (Lys3). During summer periods, DOC concentrations showed values up to 18 mg L–1, while winter values are generally below 6 mg L–1. For 2004 we observed a gradual increase of the DOC effluent levels for all three lysimeters, with marked maxima in June and September.


Figure 5
View larger version (25K):
[in this window]
[in a new window]

  		FIG. 5. Dissolved organic carbon (DOC) of the three lysimeters. Single events connected with no-flow and drying periods are indicated by S; those connected with freezing–thawing are indicated by W.

 
Sources for DOC in the seepage included the release of NAPL constituents such as alkanes or PAH, the (micro)biological formation of organic materials, and the desorption of organic matter from the solid phase. Laboratory column studies with pristine soils (Weigand and Totsche, 1998; Münch et al., 2002), contaminated soils (e.g., Wehrer and Totsche, 2005, 2007), and in particular for the material studied here (Totsche et al., 2006) demonstrated that the release of DOC is rate-limited. Thus, the mean residence time or the travel distance is of paramount importance for the concentration of DOC in the seepage, but also for the chemical composition of DOC.

The extreme release of DOC ("first flush") observed at the onset of the monitoring in October 2002 and once again in October 2003 has been reported in particular for column experiments (Weigand and Totsche, 1998; Reemtsma et al., 1999; Münch et al., 2002; Laegdsmand et al., 2004). Although the maximum values of the initial OC export from the three lysimeters cannot be compared directly because of the difference in the materials, first-flush behavior is frequently observed both for disturbed and undisturbed soil materials.

The initial release of our experiments agrees with lysimeter studies comparing the composition of seepage water after various periods of drought (Guggenberger et al., 1998), as well as batch experiments using forest topsoil and subsoil materials (Kaiser and Zech, 1998). First flush of DOC is attributed to the export of readily available OC that is released during no-flow conditions. We assume that different sources of OC are responsible for the export of DOC and such for the first flush. Release of OC from the topsoil material is mainly the result of biological activity. Upon infiltration, plant-derived organic matter degradation products (e.g., lignin and lignocellulose fragments), microbial metabolites (e.g., polysaccharides), or microbial debris accumulated within the soil material are exported (Christ and David, 1996; Guggenberger et al., 1998; Kaiser and Zech, 1998). Moreover, drying of soil material is known to produce water-soluble organic matter due to the lysis of microbial cells (Christ and David, 1994). This organic material dissolves rapidly upon rewetting (Kaiser and Zech, 1998) and is washed out with the seepage water. Peak values of DOC in the seepage water after dry periods have been observed by Lamersdorf et al. (1998) as well. In January 2004 only, high amounts of infiltrating rain water led to a high dilution and thus to decreased DOC concentrations. This behavior is explained on the one hand by the effect of freezing–thawing events that led to a disruption of microbial tissues and an increase of the DOC values and on the other by the inverse relationship of DOC concentrations and the water flux (Kalbitz et al., 2000).

Release from the subsoil material is attributed to the mobilization of OC, which was imported from topsoil. Because of the high content of carbonates and the longer residence time, the subsoil material has a higher Ca2+ activity and higher ionic strength in the pore water than the topsoil material. Under such conditions, dissolved and colloidal phase OC imported from the topsoil is immobilized due to flocculation, resulting in a relative accumulation (Münch et al., 2002). Upon infiltration of rain, which is free of or low in DOC and has low ionic strength and low Ca2+ activity, these mechanisms are reversed: Precipitated colloidal OC is redispersed and released, resulting in the first flush and the lasting export of large amounts of DOC. A major source for DOC and such for the first flush is the residual NAPL. The pretreatment of the soil will cause disintegration of the NAPL, thus contributing to the formation of NAPL fragments, which are mobilized on infiltration. Wehrer and Totsche (2005) and Totsche et al. (2006) reported extreme initial release of DOC in column experiments with contaminated materials, which compares well with the first flush observed for the lysimeters in this study.

However, the maxima observed in June 2003, June 2004, and September 2004 within the temporal course of DOC cannot be explained by the processes connected with the first-flush phenomena. More likely, they mirror the fact that the release of DOC is rate-limited and that microbial activity results in an increased production of mobilizable organic substances, especially during the warm season. The development of the biological soil activity with increasing mean soil temperatures explains the general trend of increasing DOC levels during spring 2003 and spring and summer 2004 (with summer 2003 data missing due to vandalism). The declining temperatures in combination with more lasting rain events caused the decreasing DOC concentrations in fall 2004. The maxima in the DOC concentration are generated if extended no-flow periods are followed by a rain event large enough to produce seepage. During the no-flow period, mobilizable organic materials accumulate as a consequence of desorption–dissolution from the contaminant source and non–steady-state biological formation of DOC.

Rate-limited release of DOC and material-dependent export levels have already been found for various natural and contaminated soils (Münch et al., 2002; Wehrer and Totsche, 2005; Weigand and Totsche, 1998). The different DOC levels of the three lysimeters were additionally affected by the different levels of contamination of the lysimeters. In addition to the natural organic matter, the NAPL phase served as a source of DOC mainly in Lys2 and Lys3. Despite its lower contamination, Lys2 showed higher DOC values than Lys3 during spring and summer 2004. This might be explained by the smaller height of Lys2, which promoted the desiccation of the soil monolith and thus a larger production of water-soluble organic matter.

Course of Electrical Conductivity, Turbidity, and Colloids and Particles >0.7 µm
The EC of all lysimeters showed increasing values during spring and summer periods and decreasing values in autumn and winter (Fig. 4b). Sharp declines were observed after snowmelt in March 2003 and after high amounts of precipitation in January and July 2004. The EC values of the lysimeters were between 200 and 400 µS cm–1. High EC values were observed at the start of the experiment, decreasing in the following months.

As with pH, EC is mainly affected by the dissolution of carbonates and the release of organic substances. While the dissolution of carbonates will result in increased EC, the release of organic matter will not inevitably result in an increase. It depends on the presence and amount of ionic functional groups and the protonation state. In general the EC follows the course of the DOC (strong positive and significant correlation of DOC and EC; Table 4). During summer months the EC of all lysimeters showed increased values as a result of increased DOC concentrations. The seepage water of Lys3 showed a generally higher EC than Lys1 and Lys2. The higher amount of organic matter of Lys3 results may be one reason. However, rate-limited dissolution of minerals, such as the carbonates, might also result in a much higher EC level. In column experiments conducted with the same soil material, the dissolution of the carbonates was found to be rate-limited (Totsche et al., 2006). Thus, the higher EC can be explained by the longer travel distance, which matches more closely the saturation length, that is, the distance the solution needs to travel through the porous media to completely achieve the thermodynamic dissolution equilibrium of the carbonates.

The concentrations of colloids and particles in the size 0.7 to 200 µm showed levels of 15 mg L–1 for Lys1 and Lys2 and about 3 mg L–1 for Lys3 (Fig. 4d). Increased particle concentrations were observed after strong precipitations. This observation is in agreement with results obtained for macroporous soil that has been subjected to simulated rain events (Ryan et al., 1998). As indicated by the turbidity data, the snowmelt in March 2003 and the thunderstorm in July 2004 resulted in extremely high exports of particles from Lys1. The turbidity of the seepage of Lys1 and Lys2 was generally about 20 FAU. Lys3 showed lower values of turbidity (Fig. 4c). Increased turbidity values occurred after high amounts of precipitation.

The colloids and particles were mobilized predominantly by hydrodynamic forces due to the infiltration of the seepage water and the propagation of the water–air interface. This conclusion is supported by the strong and significant positive correlation of the retentate mass with the precipitation, and the even stronger positive correlation with the seepage water (Table 4). The first flush, that is, the mobilization of colloids and particles that were generated during prolonged no-flow periods, adds to the overall export. We found a higher amount of mobilized particles the longer the precipitation-free period lasted.

Export of PAHs: The Role of Particles and Single-Release Events
Figure 6 shows the course of the PAH concentration in the retentate and in the filtrate. The amount of PAH exported from Lys1 was very low in the filtrate (<0.5 µg per seepage event) as well as in the retentate (<3 µg per seepage event). For Lys2, the PAH exported with the filtrate and the retentate were generally below 1 µg and 10 µg per seepage event, respectively. The mean amount of PAH released from Lys3 was about 3 µg per seepage event in the filtrate and 23 µg in the retentate per seepage event. The overall PAH export was dominated by the colloids and particles in the size fraction 0.7 to 200 µm. For the whole monitoring period, 0.418 µg kg–1 (135 µg per lysimeter) were mobilized from Lys3 in the filtrate (<0.7 µm) while the eightfold mass, 3.36 µg kg–1 (1085 µg per lysimeters), was exported in the retentate. Similar distributions were observed for the Lys1 and Lys2, although the total PAH export was much lower due to the lower contamination.


Figure 6
View larger version (42K):
[in this window]
[in a new window]

  		FIG. 6. Course of the polycyclic aromatic hydrocarbons in the two fractions <0.7 µm and >0.7 µm.

 
Only a minor amount of the PAHs was exported with the filtrate fraction <0.7 µm. If we assume that this fraction is predominantly composed of dissolved PAHs, their concentration should be controlled by a dissolution process according to Raoult's law (Lane and Loehr, 1992; Lee et al., 1992; Mukherji et al., 1997; Eberhardt and Grathwohl, 2002; Wehrer and Totsche, 2006). By calculating the aqueous equilibrium concentrations of the single PAH based on their (subcooled) solubility and mole fraction in the NAPL, a comparison with the measured PAH concentrations should reveal the applicability of Raoult's law and thus the prevalence of a dissolution process as the governing release mechanism (Fig. 7 ). Our predictions of the aqueous concentration of PAHs in the filtrate failed. Only benzo[b]fluoranthene and benzo[k]fluoranthene were in the range of their calculated aqueous equilibrium concentrations. We attribute the smaller PAH concentrations to the fact that the time for equilibration was too short between the different seepage water events. Mean precipitation-free periods of time were approximately 2 wk. A long-term kinetic experiment performed with the tar-oil material from the same site revealed that equilibration took more than 100 d at 20°C and more than 150 d at 4°C. Applying Raoult's law after these long-term equilibration times gave very good agreement of measured and predicted concentrations. Thus, the driving process for the release of PAHs from the NAPL is indeed a dissolution process according to Raoult's law; however, it is strongly rate limited.


Figure 7
View larger version (17K):
[in this window]
[in a new window]

  		FIG. 7. Comparison of the mean aqueous equilibrium concentrations calculated after Raoult's law and mean concentrations of the single polycyclic aromatic hydrocarbon (PAH) in the filtrate of Lys3. The molecular weights of the coal tar for the three variants are 295, 230, and 780 g mol–1. Naphthalene (Nap), acenaphthylene (Acy), acenaphthene (Ace), fluorene (Flr), phenanthrene (Phe), anthracene (Ant), fluoranthene (FlA), pyrene (Pyr), benz[a]anthracene (BaA), chrysene (Chr), benzo[b]fluoranthene (BbFlA), benzo[k]fluoranthene (BkFlA), benzo[a]pyrene (BaP), indeno[1,2,3-c,d]pyrene (IcdP), dibenz[a,h]anthracene (DahA), benzo[g,h,i]perylene (BghiP).

 
Table 5 shows the total mobilized amounts of PAHs in both fractions, DOC, and colloids and particles (0.7–200 µm) released from the lysimeters during the monitoring period. The total mass of DOC released during the monitoring period increased in Lys1, Lys2, and Lys3. The amount of mobilized colloids and particles increased from Lys3 to Lys2 to Lys1.


View this table:
[in this window]
[in a new window]

  		TABLE 5. Total mobilized amounts of dissolved organic carbon (DOC), colloids and particles (0.7–200 µm), and polycyclic aromatic hydrocarbons (PAHs) in both fractions eluted from the lysimeters during the 2-yr monitoring period and the corresponding percentages mobilized via single-release events.

 
Remarkably, single-release events are obvious for PAHs and colloids and particles in the size fraction 0.7 to 200 µm for all lysimeters. Up to 85% of the total mass of PAHs were released from Lys3 with single events (Table 4). Single events are thereby defined as follows: If the difference between two consecutive release events for particles or PAHs >0.7 µm is larger than two times the mean amount exported for the whole monitoring period, this event is called a single event. Such events were observed in November 2002, December 2002–January 2003, March 2003, October 2003, January 2004, July 2004, and August 2004. In summer and autumn (October 2003, July 2004, August 2004), these events were provoked by intensive precipitation, such as thunderstorms, after prolonged dry periods. During winter months (November 2002, December 2002–January 2003, March 2003, January 2004) they occurred after soil freezing followed by thawing and a high amount of snowmelt or rain water infiltrating the soil material.

All single-release events were characterized by high values of turbidity and high concentrations of colloids and particles in the size 0.7 to 200 µm (Fig. 4c and 4d). In summer this was accompanied by high concentrations of DOC for all lysimeters and increased values of PAHs released in the retentate of Lys3 (Fig. 6). Most striking were the extremely high exports of PAHs in the retentate of Lys3 as a result of single-release events in winter (Fig. 8b ).


Figure 8
View larger version (33K):
[in this window]
[in a new window]

  		FIG. 8. (a) Accumulated amount of dissolved organic carbon (DOC), colloids and particles (0.7–200 µm), and (b) accumulated amount of polycyclic aromatic hydrocarbons (PAH) in both fractions exported from Lys3. Single events connected with no-flow and drying periods are indicated by S; those connected with freezing–thawing are indicated by W.

 
Dry periods followed by thunderstorms in summer and autumn and freezing–thawing cycles in winter were most effective in the production of colloids and particles (0.7–200 µm). High mechanical strain caused by shrinking and swelling due to wetting–drying cycles in summer, as well as the freezing of the soil material in winter, favored the formation of particles. Temperature and water-content measurements confirmed the formation of soil ice in winter 2004 and the extreme drying down to a depth of approximately 30 cm, where the volumetric water content dropped below 2% (Fig. 2). Particles and colloids produced under such extreme environmental conditions are mobilized and exported with the following precipitation or snowmelt, which is also demonstrated by high and significant correlation between the amount of seepage and the amount of colloids and particles (Table 5). In addition, decreased pH and EC values favor the release of colloids and particles (Fig. 4a and 4b). Colloid and particle mobilization as a result of decreased pH and EC or increased flow velocity or rain intensities have been reported by others (Bunn et al., 2002; Kaplan et al., 1993; McDowell-Boyer, 1992; Rousseau et al., 2004; Ryan and Elimelech, 1996; Ryan et al., 1998; Saiers and Lenhart, 2003). The interplay of wetting–drying cycles followed by intensive rain and freezing–thawing cycles followed by snowmelt events is an important trigger for the formation and redistribution of colloids and particles and such for the PAHs in the unsaturated soil zone.

The cumulative amounts of DOC, colloids and particles (0.7–200 µm), and PAHs within the two fractions exported from Lys3 are presented in Fig. 8. The interplay of dry periods and thunderstorms during summer months resulted in a release of DOC and colloids and particles up to 69 and 80 mg per seepage event, respectively, whereas up to 97 mg DOC and up to 266 mg colloids and particles per seepage event were exported from Lys3 during winter events.

For the export of PAHs in the filtrate, only slightly increased values were observed at the single-release events. No strong relation was found for DOC and PAHs <0.7 µm (Table 5). This was not expected because column experiments suggested an effect of DOC on the release of PAHs (Totsche et al., 2006). It must be noted, however, that the column experiments were run under saturated flow conditions, and consequently, the effect of propagating air–water interface was disregarded. Moreover, the changes of flow velocity (two different flow velocities, multiple flow interrupts) imposed to the column were smaller than those under thunderstorm flow conditions. The mechanical strain and, in particular, the formation and propagation of the air–water interfaces as a consequence of wetting–drying (and freezing–thawing) are much more effective in the mobilization and export of colloids and particles than the flow velocity variations of the column experiments.

Another notable observation was the large export of PAHs in the retentate from Lys3 at single events in winter (Fig. 8b). Two single seepage events (January 2003 and January 2004) together resulted in an export of 600 µg PAHs in the fraction 0.7 to 200 µm. We believe that the freezing and thawing cycles seem to effect the PAH source materials, that is, the remnants of the NAPL phase. With decreasing temperature, the NAPL not only lost fluidity but also plasticity. The high mechanical strain exerted by the freezing water will thus likely result in the formation NAPL fragments, which will be mobilized at the onset of thawing and following rain or snowmelt events, and so the PAHs associated with these fragments.

The importance of the freezing of the source material was demonstrated at the release event in March 2003. Peak values of turbidity and particle concentration were found for Lys1 and Lys2 as the result of a snowmelt (Fig. 4c and 4d). Again, this was caused by increased hydrodynamic forces due to the high amount of melt water infiltrating the soil. However, this event resulted in an increased release of neither particles in Lys3 nor of PAHs in any lysimeter. The freezing before this event seems not to have progressed deep enough to cause a release of particle-associated PAH or NAPL fragments.


   Conclusions
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
REFERENCES
 
We showed that the export of PAHs from the studied contaminated soil was dominated by the release via the fraction of colloids and particles size 0.7 to 200 µm. In contrast to column experiments conducted with the same material (Totsche et al., 2006), no significant correlation of the PAH export was found with the release of DOC. The comparison of the concentrations of PAHs in the size fraction <0.7 µm with predictions based on Raoult's law reveal an apparent overestimation of the concentrations. This, however, is explained by the insufficient time for equilibration of the aqueous phase with the NAPL phase at the prevailing conditions in the lysimeters.

The occurrence of single- and extreme-release events resulted from freezing–thawing cycles or prolonged periods of no-flow during dry summer periods, when potentially mobilizable particles are formed. These conditions when followed by high-intensity rain events, snowmelts, or thunderstorms lead to high loads of the seepage water with particles, DOC, and PAHs. Because such conditions are difficult to mimic with laboratory column experiments, the relevance of this is vastly ignored.

Our results emphasize the role of the large colloids and particles and single events for release, transport, and redistribution of PAHs under natural climatic conditions. Our study demonstrates that the variability of climatic and soil parameters, such as the intensity, amount, and duration of precipitation, soil moisture, and soil temperature are important factors for the mobilization of PAH, DOC, and colloidal particles. Without a profound understanding of the interaction of these factors and the effect of their spatiotemporal variability on unsaturated flow conditions, an estimation of the release and mobility of PAHs at contaminated sites is not possible.


   ACKNOWLEDGMENTS
 
We wish to thank Bärbel Angres for conducting the PAH analysis and Claudia Guggenberger for assistance in conducting the experiment. This work was financially supported by the Bayerisches Staatsministerium für Umwelt, Gesundheit und Verbraucherschutz (StMUGV) and by the Deutsche Forschungsgemeinschaft (Contract No. To 184/5-2).


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
P. Sharma, H. M. Abdou, and M. Flury
Effect of the Lower Boundary Condition and Flotation on Colloid Mobilization in Unsaturated Sandy Sediments
Vadose Zone J., August 1, 2008; 7(3): 930 - 940.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Totsche, K. U.
Right arrow Articles by Kögel-Knabner, I.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Totsche, K. U.
Right arrow Articles by Kögel-Knabner, I.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Totsche, K. U.
Right arrow Articles by Kögel-Knabner, I.
Related Collections
Right arrow Lysimeter/Rhizosphere Studies
Right arrow Organic Compounds
Right arrow Soil Pollution

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome