Environ. Sci. Technol. 2003, 37, 4325-4329
Assessing Forest Fire as a Potential PCDD/F Source in Queensland, Australia J O E L L E A . P R A N G E , * ,†,§ CAROLINE GAUS,† ROLAND WEBER,| O L A F P A¨ P K E , ‡ A N D J O C H E N F . M U ¨ LLER† National Research Centre for Environmental Toxicology (EnTox), University of Queensland, 39 Kessels Road, Coopers Plains, Queensland, Australia 4108, School of Public Health, Griffith University, Logan Campus, University Drive Meadowbrook, Queensland, Australia 4131, Universita¨t Tu ¨ bingen, Institute for Organic Chemistry, Auf der Morgenstelle 18, 72076 Tu ¨ bingen, Germany, and Ergo Forschungsgesellschaft mbH, Geierstrasse 1, 22305 Hamburg, Germany
Forest fires are suggested as a potential and significant source of polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs), even though no studies to date provide sufficient evidence to confirm forest fires as a source of PCDD/Fs. Recent investigations in Queensland, Australia have identified a widespread contamination of PCDDs (in particular OCDD) in soils and sediments in the coastal region from an unknown source of PCDD/Fs. Queensland is predominately rural; it has few known anthropogenic sources of PCDD/Fs, whereas forest fires are a frequent occurrence. This study was conducted to assess forest fires as a potential source of the unknown PCDD/F contamination in Queensland. A combustion experiment was designed to assess the overall mass of PCDD/Fs before and after a simulated forest fire. The results from this study did not identify an increase in ∑PCDD/Fs or OCDD after the combustion process. However, specific non-2,3,7,8 substituted lower chlorinated PCDD/Fs were elevated after the combustion process, suggesting formation from a precursor. The results from this study indicate that forest fires are unlikely to be the source of the unknown PCDD contamination in Queensland, rather they are a key mechanism for the redistribution of PCDD/Fs from existing sources and precursors.
Introduction Combustion processes such as municipal waste incineration are suggested to be a major contributor to the ubiquitous contamination of polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/Fs) in the contemporary environment (1, 2). A number of studies suggest forest fires as a potential and significant source of PCDD/Fs (1, 3-5). However, the evidence to date regarding forest fires as a source of PCDD/Fs remains ambiguous. Only a few studies have successfully measured PCDD/Fs in the atmosphere during a forest fire (6, 7). The authors * Corresponding author phone: +61 7 3000 9198; fax: +61 7 3274 9003; e-mail:
[email protected]. † University of Queensland. § Griffith University. | Universita ¨ t Tu ¨ bingen. ‡ Ergo Forschungsgesellschaft mbH. 10.1021/es0343454 CCC: $25.00 Published on Web 08/22/2003
2003 American Chemical Society
from these studies concluded that the data were insufficient to confirm forest fires as a source of PCDD/Fs (6). A number of other studies have analyzed soils and vegetation before and after fires to determine the influence of fire on the environmental PCDD/F contamination (8-11). Only one of these studies suggests that the increase in PCDD/Fs was due to the forest fire (8). On the other hand, a range of combustion studies has been conducted to determine the emission of PCDD/Fs from natural wood and/or biomass (4, 12, 13). In comparison to waste woods the combustion of natural woods results in lower PCDD/F emissions (12). The original PCDD/F contamination of the materials combusted has been determined in only one of these studies, and in this case only one replicate sample was analyzed for PCDD/Fs (4). The authors from that study detected a 10-fold increase in the emission of PCDD/Fs (on a TEQ basis) than in the unburnt biomass and concluded that the results provided preliminary evidence that PCDD/Fs are formed during forest fires (4). In Queensland, Australia a widespread distribution of elevated concentrations of PCDD/Fs has been identified in topsoils and sediments extending along the more than 2000 km coastline, with relatively few industrial or known PCDD/F sources (14, 15). The contamination shows a similar PCDD/F congener profile in most samples, dominated by PCDDs, with OCDD contributing 90-99% ∑PCDD/Fs, while PCDFs are low and often below the limit of quantification. To date, the specific PCDD/F congener and isomer distributions detected in soils and sediments from the Queensland coastal region could not be associated with known PCDD/F industrial or other anthropogenic activities, and the source of this contamination remains unidentified. A preliminary investigation in Queensland identified elevated concentrations of PCDD/Fs in the atmosphere during a “prescribed burn” (an induced forest fire, used as a management tool, to reduce biomass and severity of wildfires) in comparison to the control sampling of atmosphere during a period of nonburning (7). The PCDD/F profiles detected in the prescribed burn were unlike typical combustion profiles, with a high D:F ratio and the elevation of specific non-2,3,7,8-substituted PCDD/F isomers, indicating that PCDD/F formation may be occurring via precursors, during the prescribed burn. Atmospheric measurements, however, did not provide sufficient information to determine the source of PCDD/Fs in the fire nor could they be used to assess the role of forest fire on the PCDD/F contamination in Queensland. The present study was carried out to determine whether forest fires are a source of PCDD/Fs or primarily a mechanism for the redistribution of existing PCDD/Fs in the Queensland coastal environment. This study was conducted using a combustion chamber to simulate the conditions of a typical Queensland forest fire. The distributions of PCDD/Fs in the pre- and postcombustion components were quantified to assess the role of forest fires on the PCDD/F contamination in Queensland.
Experimental Design Study Site. The combustion experiments undertaken in this study were conducted outdoors in the Doongal region of the Wongi State forest, located approximately 45 km inland from the coast of Queensland. The Wongi State forest is a dry eucalypt woodland of two predominate species: Spotted gum (Corymbia citriodora) and Broad-leaf Red Iron Bark (Eucalyptus fibrosa). The combustion experiment was conducted in situ to avoid the alteration of natural environmental VOL. 37, NO. 19, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4325
conditions (e.g. fuel and moisture content), to reduce contamination of the fuel and soil during transport, and to conduct the experiment where the atmospheric PCDD/F concentrations are known to be low. Two combustion experiments were designed for the purpose of this study, with conditions simulated to represent a typical prescribed burn (i.e. litter density of 15-18 tonnes/hectare and maximum temperature range of 350-400 °C). Exp #1 involved the combustion of leaf litter alone, and exp #2 involved the combustion of leaf litter on a bed of ambient soil. The ignition of leaf litter was performed using a gas lighter; the material was ignited in three locations to ensure adequate combustion of the leaf material. The combustion experiments were carried out in a combustion chamber specifically designed for the purpose of this study. The chamber consisted of a stainless steel grate and/or tray where the material was combusted and combined with a stainless steel hood and chimney, which channeled the smoke emitted during the burning process to a filter (GFF, Schleicher and Schuell No. 9). A glass condenser trap was used to cool the emission samples before collection on a PUF/XAD-2 resin trap. Experiment #1 - Assessment of Material Burnt. The initial experiment in this study was conducted to assess the origin of the PCDD/Fs by the combustion of leaf litter material only. For the purpose of this experiment, the fuel consisted of leaf litter (comprised of fallen and partially degraded eucalyptus leaf, twigs, and bark) collected from the forest floor of the Wongi State forest. The collection of leaf litter involved subsamples collected with stainless steel equipment, within an area of the Wongi State forest, unburnt for 9 years. The subsamples were homogenized to allow a representative sample of the region. The leaf litter was combusted on stainless steel mesh trays (mess size 5 mm), in 1 kg allotments (a total of 3 kg was combusted). After the combustion of each 1 kg of litter the litter ash was collected from the tray (and any which had fallen through to the bottom grate). The leaf ash was weighed and homogenized, and subsamples were taken for PCDD/F analysis. The cumulative smoke was collected on the filter/condenser/XAD traps and subsequently analyzed for PCDD/Fs. Experiment #2 - Forest Fire Simulation. The forest fire simulation experiment was conducted to assess the emission of PCDD/Fs during a “typical” forest fire. For this combustion experiment 25 kg of ambient soil was collected within an area unburnt for 9 years. The soil was homogenized using stainless steel sampling equipment, and a subsample was collected for PCDD/F analysis. The soil was placed into a stainless steel tray (80 cm × 80 cm) at a depth of 10 cm. The fuel consisted of a subsample of the leaf litter collected for exp #1. A total of 3 kg of the litter material was combusted in the chamber on top of the soil. The bulk of leaf ash was collected from the surface of the soil after the combustion of each 1 kg avoiding the collection of the surface soil. Following the combustion of the 3 kg, the soil was differentiated into soil-ash (the top 5 mm soil and remaining ash), burnt soil (top 2 cm of soil distinguished by a darkened color due to the burning process), and the underlying soil (remaining soil fraction in the tray). All fractions (leaf litter, leaf ash, and soils) were weighed, homogenized, and subsampled for PCDD/F analysis. The smoke was collected on the filter/condenser/XAD traps and subsequently analyzed for PCDD/Fs. Temperature. The temperature during the combustion process was measured with an infrared heat gun and a thermocouple, at various locations in the combustion chamber to determine the heat output at various stages in the system. Temperatures were determined in the leaf litter material during combustion (350-450 °C) but not at the direct point of leaf ignition at the soil surface (250-350 °C), at 2 cm 4326
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 19, 2003
below soil surface (110-150 °C), and the PUF/XAD-2 resin cartridge (22-25 °C). The temperature at the PUF/XAD-2 resin cartridge did not exceed ambient temperatures (23-25 °C); therefore, it was assumed that PCDD/F formation occurred within the combustion system and not after sample collection. Analysis. Samples were analyzed for all 2,3,7,8-substituted PCDD/Fs and ∑tetra-hepta homologues at ERGO Forschungsgesellschaft mbH, in Hamburg Germany, using a certified method (16). Non 2,3,7,8-substituted PCDD/F isomers were calculated using a ratio of the signal height of the individual isomers to the 13C labeled standards. Isomer verification was determined by the analysis on both a DB5 and SP23-31 column. Unburnt soil, unburnt leaf, leaf ash, and burnt soil were collected and analyzed in replicates. The mean normalized difference between replicates was 15%. Details of the analytical methods have been reported elsewhere (14).
Results and Discussion Forest Fires as a Source of PCDD/Fs in Queensland, Australia. To elucidate the source and distribution of PCDD/ Fs in the smoke in this combustion study the absolute quantity of PCDD/Fs in all components (i.e. smoke, leaf, ash, and soils) of the combustion system were evaluated for combustion exp #1 and #2, before and after the combustion event. The mass of the PCDD/Fs in each component was calculated from the concentration of the individual PCDD or PCDF congeners and homologue groups in the sample (ng kg-1) × weight (kg) (see Supporting Information for calculations). The total mass of PCDD/Fs of the post- to precombustion was compared for individual PCDD/F congeners and homologue groups to evaluate the effect of combustion process on overall PCDD/F distribution. For the ∑tetra-hepta PCDD/F homologues and OCDD/F congeners, the results suggest that in exp #1 and #2, no formation occurred during the combustion process (Table 1). On an individual congener basis, the mass of OCDD did not increase postcombustion in either experiment. In fact, in exp #2 the mass of OCDD in the soils, smoke, and ash postcombustion could not account for OCDD in the unburnt soil. The results from these combustion experiments suggest that forest fires are not the source of the elevated OCDD detected in Queensland soils. However, in exp #2 there was an increase of the TCDD (2-fold), PnCDD (1.6-fold), and TCDF (1.4-fold) homologue groups, after the combustion process (Table 1), in particular in the burnt soil. The increase of lower chlorinated PCDD/Fs in exp #2 suggests that fire may play a role in the formation of the lower chlorinated PCDD/Fs, in particular with respect to processes occurring in the soil. To investigate the possible formation of the lower chlorinated PCDD/Fs the unburnt, burnt, and smoke components in exp #2 were investigated. Formation of the Lower Chlorinated PCDD/Fs. The concentration of ∑PCDD/Fs detected in the unburnt leaf litter material utilized for this combustion experiment was 28((3) pg g-1 dw. The homologue profile in the unburnt leaf was dominated by PCDDs (D:F ratio 5.2), in particular OCDD (which contributed to 53% of ∑PCDD/Fs). After the combustion of the leaf material in exp #2, the subsequent leaf ash contained elevated concentrations of ∑PCDD/Fs (180 pg g-1 dw), with a higher contribution of OCDD to ∑PCDD/ Fs (97%), see Figure 1. In this experiment, the leaf ash was collected from the surface of the soil, and although care was taken not to collect the soil, the higher concentration of ∑PCDD/Fs, particularly OCDD in the leaf ash, may be attributed to the soil component. The concentration and distributions of PCDD/Fs in the soil were altered with combustion process (Figure 1). The
TABLE 1. Total Mass of the PCDD/F Homologues (ng) in Each Component, Pre- and Postcombustion for Experiments #1 and #2 postcombustion precombustion
TCDD PnCDD HxCDD HpCDD OCDD TCDF PnCDF HxCDF HpCDF OCDF ∑PCDD ∑PCDF ∑PCDD/F
experiment #1
unburnt leaf
exp #1 total
unburnt soil
exp #2 total
leaf ash
4.3 2.9 7.7 8.3 40 4.9 2 2.8 2
