Environ. Sci. Technol. 2008, 42, 5911–5916
Accumulation Parameters and Seasonal Trends for PCBs in Temperate and Boreal Forest Plant Species LUCA NIZZETTO,† CRISTINA PASTORE,† XIANG LIU,‡ PAOLO CAMPORINI,† DANIELA STROPPIANA,§ BEN HERBERT,| MIRCO BOSCHETTI,§ GAN ZHANG,‡ PIETRO A. BRIVIO,§ KEVIN C. JONES,| A N D A N T O N I O D I G U A R D O * ,† Department of Chemical and Environmental Sciences University of Insubria, Via Valleggio 11, Como CO, Italy, State Key Laboratory of Geochemistry, Guangzhou Institute of Geochemistry, The Chinese Academy of Sciences, Guangzhou, China, IREA-CNR, Institute for Electromagnetic Sensing of the Environment, National Research Council, Via Bassini 15, 20133 Milano, Italy, and Centre of Chemicals Management, Lancaster Environment Centre, Lancaster University, Lancaster, LA1 4YQ, United Kingdom
Received January 22, 2008. Revised manuscript received May 17, 2008. Accepted May 21, 2008.
The concentration of polychlorinated biphenyls (PCBs) in the air and vegetation was measured periodically in two alpine forests, during the growing season. Foliage samples from nine plant species typical of the temperate and boreal environment were collected and analyzed. Leaf concentrations of tri- and tetraCBs showed fast response times with changing temperature and gas-phase concentrations, suggesting that a partitioning equilibrium is approached relatively rapidly (few days) in the field. Heavier compounds showed kinetically limited accumulation trends, not reaching equilibrium during the growing season. Results were used to estimate the bioconcentration factors or equilibrium plant/air partition coefficient (KPA) for each species. Values of log KPA (calculated on a mass/volume basis) ranged between 0.78 and 1.96 and were correlated to the log KOA. Uptake trends of the higher chlorinated compounds showed intraspecific differences which were partially explained by the specific leaf area (SLA).
Introduction A number of studies have focused on the accumulation of persistent organic pollutants (POPs) to plant foliage (e.g., refs 1–5). Vegetation accumulation represents the first step for POPs to enter the terrestrial food web and constitutes a key factor in determining both wildlife and human exposure to pollutants. Uptake from the atmosphere rather than the soil has been shown to be the most important pathway for compounds with KOA > 6 and KAW > -6 (6–8), which covers most of the known POPs. Due to their chemical composition * Corresponding author tel.: +39-031-2386480; fax: +39-0312386449; e-mail:
[email protected]. † University of Insubria. ‡ The Chinese Academy of Sciences. § IREA-CNR. | Lancaster Environment Centre, Lancaster University. 10.1021/es800217m CCC: $40.75
Published on Web 07/02/2008
2008 American Chemical Society
and large surface area, leaves efficiently intercept atmospheric organic pollutants both in the gas phase and associated to particles. There is still uncertainty over the dynamics of uptake (9), with high variability in both the plant-air equilibrium partition coefficients (KPA) (10–13) and the time needed to reach such an equilibrium (14, 20). Studies of the dynamics of POP accumulation to the principal forest species in field conditions are still limited (21), and data on bioconcentration factors are available only for a few species. The present paper reports results for a field study designed to assess the accumulation behavior of PCBs in selected dominant species of the temperate and boreal environment. Changes in leaf concentrations with time were monitored during the growth season, while also determining the air concentrations and a number of ecological and physical parameters. The main aims of the study were to evaluate the accumulation behavior of each species and to determine uptake parameters and KPA values.
Experimental Section Sampling. Sampling was performed in an alpine valley (Lys Valley, Aosta, Italy), on the slopes of Mont Mars (2650 m), in two sites located at 1100 and 1400 above sea level (5055133 N, 411661 E and 5055160 N, 412732 E, respectively, UTM, ED 50). Vegetation Samples. Single species leaf samples were collected from the dominant tree species. At 1100 m, the following species were collected: white ash (Fraxinus excelsior), chestnut (Castanea sativa), lime (Tilia cordata), hazelnut (Corylus avellana), mountain ash (Sorbus aucuparia), and maple (Acer pseudoplatanus); at 1400 m the following were collected: beech (Fagus sylvatica), larch (Larix decidua), spruce (Picea abies), and mountain ash. All of the sampled species (except the spruce) completely lose their foliage in the autumn. For spruce (the only evergreen species), new needles of the current year were sampled during their development. Sampling started on April 18, 2005 and occurred weekly until the end of June. After that, only one collection was made each month in July, September, and October. The last collection occurred on October 17 during litter fall. In the first three sampling events, only the understory species at 1100 m (namely, hazelnut, maple, and mountain ash) had produced leaves, while bud burst for the upper canopy species (white ash, chestnut, and lime) occurred at the beginning of May. At the 1400 m site, leaf collection started approximately 10 days later, given the temperature-driven delay in the bud burst. The maple and hazelnut samples collected at the first date included buds or bud protective structures, which were manually separated or naturally lost by the following sampling dates. Leaves were collected randomly in forest plots of about 2500 m2 using latex gloves at a height ranging between 1.5 and 6 m from the soil. Samples were kept in solvent-rinsed glass jars and frozen at -20 °C, until extraction. A total of 158 samples were collected and analyzed. Air Samples. A total of 13 air samples were collected using a high-volume sampler located in a clearing at 1400 m near the forest site. A previous study showed that the concentrations measured here represent the exposure conditions for both sites (22). Sampling started on the 16th of April and ended on the fourth of July. Samples were collected continuously over intervals of 5-7 days (yielding 500-700 m3 samples). The particle phase was trapped on a glass fiber filter while gas-phase POPs were adsorbed on two polyurethane foam (PUF) plugs. VOL. 42, NO. 16, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5911
FIGURE 1. Vegetation parameters and temperatures. (A) Points represent LAI (leaf area index), while lines are air temperatures. (B) Mean SLA (specific leaf area) in the canopy (calculated as the average of all of the species of that forest); error bars represent standard deviations. Dotted rectangles include the time window adopted for the derivation of KPA. Analysis. A detailed description of extraction, cleanup, and instrumental analysis for vegetation and air samples is reported elsewhere (13, 22). Analyses were performed for PCBs (tri-PCB 18, 28/31; tetra-PCB 52, 44, 49, 64, 74, 70; pentaPCB 104, 101, 99, 97, 87, 110, 118, 105; hexa-PCB 136, 151, 149, 132/153, 138, 156; hepta-PCB 188, 187, 183, 174, 177, 180; octa-PCB 201, 203, 194). Quality Assurance/Quality Control. Laboratory blanks were included as one in every five samples. Method detection limits (MDL) for the vegetation samples were determined as 3 times the laboratory blank values (considering an average extracted amount of leaves of 5 g) and varied from 9 to 31 pg g-1 dw. Field blanks for atmospheric samples were taken at a rate of one every six samples. In this case, the MDL was calculated as 3 times the level of the field blanks and ranged from 0.8 to 3 pg m-3 (considering an average sampled volume of 500 m3), depending on the compound. No blank correction was performed. PCB 40 and 128 were added to each sample to monitor the recovery of the method, which was 75 ( 11% and 84 ( 7% (n ) 188), respectively. All samples were corrected individually for recovery. An internally validated leaf sample was analyzed routinely in order to monitor for laboratory performance. After recovery correction, results for five PCBs belonging to different isomer groups were 95 ( 21% of the certified values (n ) 3). Analysis of replicates was performed on a subset of samples (n ) 10). The relative standard error for replicate analyses never exceeded 7%. Recoveries for air samples were monitored by spiking PUFs with a recovery standard of 13C12-labeled PCB congeners (13C12 PCB 28, 52, 101, 138, 153, 180, 209) and ranged between 91 and 105%. No recovery correction was applied to air samples. Vegetation Survey and Temperatures. The leaf area index (LAI) (m2 of leaf/m2 of ground surface) was measured weekly at each sampling site using a LAI 2000 sensor (Li-Cor, Lincoln, NE). Each measurement represented the average LAI on a 100-m-long transect diagonally crossing the selected plot. Specific leaf area (SLA, cm2 g-1 dw) was measured weekly on the basis of a subset of samples of each individual species. The set was composed of leaves covering all of the size classes present at that stage. The measurement method is reported elsewhere (23). Air temperature was recorded hourly at each forest site throughout the campaign using data loggers.
Results Vegetation Parameters. Figure 1 summarizes the results for the vegetation parameters. 5912
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 16, 2008
LAI. LAI increased in mid-April at the 1100 m site and ∼10 days later at the 1400 m site. The maximum LAI was about 5 at both locations (Figure 1). Full leaf development was reached around the end of June at each altitude. SLA. SLA generally decreased with time. This is in agreement with observations reported elsewhere (23), and it depends on the dynamics of leaf development. Figure 1 summarizes results reporting the trend of the mean canopy SLAC. Table SI 1 reports the average SLA measured at different times. Air Concentrations. ∑PCB air concentration, CA (pg m-3) (∑ of congeners: 28/31, 52, 101, 118, 138, 153 and 180), averaged 35 pg m-3 with tri- and tetra-CBs dominating the profile (Table SI 2). Levels were consistent with measurements performed 2 years before in the same areas (22) and are in the range of data reported for rural or remote areas in Europe (24, 25), reflecting background air concentrations. The fraction of PCBs associated with the particle phase was negligible. Leaf Concentrations. A summary of plant concentrations CP (pg g-1 dw) is reported in Table SI 3. Levels were consistent with previously reported data for the same area (13). The highest concentrations were recorded in the last sampling events, at the end of the growth season. Essentially, the data set could be divided into low-molecular-weight (LMW) PCBs, which approached equilibrium rapidly during the growing season, and high-molecular-weight (HMW) PCBs, for which the uptake was kinetically constrained.
Discussion Uptake Trends of LMW Compounds. The concentration versus time plot for LMW PCBs gave scattered data varying weekly within a factor of ∼2 (see PCB 28/31 in Figure 2). Such behavior is in agreement with theoretical predictions. McLachlan (26) predicted in his interpretative framework for the assessment of the uptake of semivolatile organic compounds to vegetation that, for compounds such as triand tetra-CBs, the response time of plant concentrations to changes in air concentrations is “short”. The plant-air system will tend to approach an operational equilibrium reflecting an average concentration over a time period characteristic for the response of the system. For LMW compounds (having lower KOA values), the operational equilibrium might be defined as an average gaseous concentration over the past few days. Mackay et al. (27) also suggested that the time to approach equilibrium for compounds with log KOA < 7 is on the order of 2-3 days. Both experimental and modeling assessments (28, 29) showed that the leaf cuticle can represent
FIGURE 2. Plant concentrations with time in the 1100 m (a) and 1400 m (b) forests. the preferential accumulation compartment. The occurrence of accumulation in the inner foliar structures was also demonstrated experimentally for phenanthrene (30). However, transfers into the inner portion of the leaf are expected to be kinetically limited by the low diffusivity of the compounds through leaf waxes and cuticles or through leaf stomata. Therefore, the gaseous exchange occurring at the air-leaf interface and the storage capacity of leaf cuticular waxes and cutin are expected to be key parameters controlling the accumulation of POPs in plant foliage. In field conditions, the exchange of LMW PCBs between air and leaves appears to be very dynamic, and is complicated by the continuous variation in air concentration and air temperature. For example, even though no significant correlation was found between CP and T, temperature is expected to play a key role in controlling the exchange of LMW PCBs. In fact, the equilibrium plant-air partition coefficient (KPA) is an inverse function of T (31). Considering the enthalpy of phase transfer (∆HPA) reported for PCB 28 and PCB 52 in Lolium multiflorum (82.4 and 86.6 kJ mol-1, respectively) (31), the observed mean weekly change in T through the study would be expected to produce a variation in the KPA value of a factor of 2-3, which can drive changes in CP of the same amplitude as those due to the observed changes in CA. During the last month of the sampling campaign (between mid-September and the end of October), a net decrease of about 10 °C was observed in daily temperatures (∼10 °C). Considering the enthalpies of volatilization for the plant-air system reported by Ko¨mp and McLachlan (31), the increased leaf concentrations of LMW PCBs at the end of October compared to the September values (by a factor of 3-5) can be accounted for by declining T. This behavior was not observed for HMW compounds, presumably because of the much slower “response time” of these compounds (26). Uptake Trends of HMW Compounds. In contrast to LMW congeners, concentration versus time plots for HMW compounds showed increasing trends for all of the species. Figure 2 illustrates results for PCB 138 in different plants. Only hazelnut and mountain ash collected at the first three sampling events had a different behavior. In this case,
relatively high concentrations were observed at the first collection, then decreased to a minimum in the following two events. This behavior may be related to the following points: • The presence of a protective structure coating the buds through the winter (and hence being exposed during this time at low temperatures) could have contributed to determining the high initial values. In the following samplings, these structures were progressively lost or became negligible compared to leaf biomass. • During the first sampling events, hazelnut and mountain ash leaves were still very small (0.7-2-cm-long), thin, and light. This is also evidenced by the relatively elevated SLA value at this stage. In these conditions, they could have become saturated relatively rapidly. Subsequent leaf “thickening” during growth would have then “diluted” the CP. These points could explain the differences observed for these species compared with the other broadleaved species, which showed a completely different mechanism of leaf development. After this early phase (starting from mid-May), plant leaves showed increasing uptake (Figure 2). By the last sampling event, just before leaf fall, the plants appeared to have not reached partitioning equilibrium with the air, for the HMW congeners. The data could be described by a single- or twocompartment kinetic uptake model using eqs 1 and 2, respectively: CP ) A(1 - e-kt) + y0
(1)
CP ) a(1 - e-bt) + c(1 - edt) + y0
(2)
Table SI 4 shows results obtained from the regression analysis. Equation 2 provided a significantly better description of trends for mountain ash, beech, and larch. For the other species, the two-phase uptake trend was less evident. Mountain ash showed similar behavior in both sites and reached higher concentrations more rapidly than the other species. Figure 3a shows three examples which summarize the trends observed for PCB 153 in all of the species. In the first VOL. 42, NO. 16, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5913
FIGURE 3. (a) Comparison between the concentration vs time plot for PCB 153 in mountain ash (white points), maple (gray points), and white ash (black points) collected at the 1100 m site. (b) Concentrations/SLA. Trend lines were determined by regressing the points with the following functions: y ) A(1 exp(a · t)) dotted and gray lines; y ) A(1 - exp(a · t)) + B(1 exp(b · t) black line. phase (hereafter called phase 1; initial ∼3 weeks), concentrations increased relatively quickly; then, a slower and curvilinear uptake trend (phase 2) was maintained until the end of the season. Two-phase uptake was previously observed under laboratory conditions and was explained by assuming that leaves may be constituted by two compartments: the leaf surface and the internal reservoir (19, 32). Phase 1 involves the fast exchange of POPs between the air and the leaf surface, while phase 2 is controlled by diffusion from the plant surface into the internal reservoir. A main difference between the present results and those obtained under controlled conditions is the length of phase 1, which lasted only a few hours in the laboratory observations (performed using other species) (19). The considerably longer duration of phase 1 observed here in some species is probably related to the leaf development strategy. For example, leaves with an initial enhanced surface development and continuous formation of new leaves during the canopy development period likely allocate more resources for the production of surface structures which constitute an important fraction of the total leaf mass. Mountain ash, beech, and larch, for example, showed this leaf development strategy, and for them, leaf “thickening” only occurred in a second stage. Samples collected during the canopy development period for these species contained a significant fraction of freshly synthesized leaves, constituting new surfaces available for fast net accumulation. Other species tended to elongate buds at the beginning of the leaf development and displaced new leaves “all at once”. In this case, there was not a continuous replacement of “fresh” surfaces during the growth as described above. Pseudo Plant Air Uptake Rates. From eq 1, the kinetic parameter k represents an estimate of the pseudo overall uptake rates k (d-1) for the transfer of HMW PCBs from the air to the leaves. Therefore k was estimated for the slow uptake phase in each species. Note that the values obtained represent an estimate of the “operational” rate constant, only valid for the observed conditions, describing the transport of PCBs 5914
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 16, 2008
from the air to the leaf under the influence of all of the forcing variables under nonstandard conditions. When the “two-phase” uptake was evident (i.e., the mountain ash), k for phase 1 was estimated fitting the concentration with a straight line. The value of k obtained for phase 1 in the mountain ash (0.19 d-1) is similar to that reported by Bacci et al. (35) (0.12-0.21 d-1) obtained in chamber experiments for azalea. The rate constants for phase 2 for all of the species were more than 1 order of magnitude lower, ranging between 0.003 and 0.01 d-1. The k value for white ash in phase 2 (0.003 d-1, also applicable for PCB 180 in the spruce) was the lowest. White ash and spruce were the two species with the lowest SLA (Table SI 1) and relatively long leaf development periods. The SLA can be regarded as a measure of the surface available for POP exchange with air per unit of leaf mass, while the longer leaf development time can enhance the growth dilution effect, resulting in a slower apparent uptake rate. Role of the Specific Leaf Area (SLA) in the Uptake Trends. In order to have a qualitative assessment of the role of SLA in the uptake of POPs by plant foliage, leaf concentrations were divided by the SLA values measured at each sampling event. Note that SLA showed a decreasing trend with time for all of the species (Table SI 1). The performed elaboration therefore modified the shape of the observed mass-based uptake curve. Results representing the amount of PCBs accumulated per unit of leaf surface (or pollutant density; pg cm-2) are reported as an example in Figure 3b and in Figure SI 1. After SLA normalization, much of the variability observed in the uptake trends calculated on a mass basis disappeared. For example, in Figure 3b, the differences observed between white ash and maple can be explained by the differences in SLA. White ash appeared to accumulate PCBs less efficiently (on a mass basis), simply because it had less leaf surface per unit of leaf mass available for exchange. On a surface area basis, mountain ash was the most efficient accumulator. The role of SLA in the gaseous plant air exchange of POPs can be described considering Fick’s first law (and assuming no growth dilution and no significant losses of pollutants from the canopy due to degradation, wash out, and other nonboundary layer processes), as follows (36):
(
CP dCp ) 2SLA × kU CA dt KPA
)
(3)
where kU is the plant air mass transfer coefficient (m d-1) and KPA is picograms per gram of dry leaf/picogram per cubic meter of air. This theoretical framework supports the view that the relatively higher SLA values at the beginning of the leaf development can enhance the accumulation rate of POPs. However, given that under the observed experimental conditions some of the assumptions made here are not verified, a more comprehensive framework is needed to describe the plant air exchange of POPs (36). Estimating the Equilibrium Plant-Air Partition Coefficient, KPA. Between the end of May and the end of July the following conditions occurred: • Plant leaves had already reached their maximum development (see LAI plot in Figure 1). • The SLA had already stabilized around its mean value for fully mature leaves (Figure 1). • All of the species were already in phase 2 of the uptake. • The average daily temperature stabilized around 18 °C and did not show any further increase during the selected period (Figure 1). • Air concentrations ranged only by a factor 2 around the mean value (Table SI 2).
TABLE 1. Estimated KPA (pg g-1 of dry leaf)/(pg m-3 of air) ± SD at 25°C
chestnut w. ash lime hazelnut m. ash (1100) maple beech larch (1400) spruce m. ash (1400)
PCB 28/31
PCB 52
PCB 101
PCB 138
PCB 153
PCB 180
7.9 ( 2.8 6.1 ( 1.4
7.6 ( 2.5
28.8 ( 15.9
11.8 ( 3.8
54.1 ( 33 37 ( 1.5 48.6 ( 4.7
91.7 ( 79.5
9.1 ( 4.9 6.1 ( 1.2 9.9 ( 4 6.3 ( 0.6 8.8 ( 2.8 9.2 ( 3.4
7.9 ( 2.2 10 ( 3.2 10.7 ( 4 10.8 ( 4.7 14.2 ( 4.6
18.5 ( 4.7 22.6 ( 20.6 31 ( 3.8 18.3 ( 1.6 29 ( 1.5 20.3 ( 0.9 15.6 ( 3.3 9.1 ( 3.4 31.7 ( 5.1
TABLE 2. Parameters of the Equation log KPA = y0 + a log KOAa
chestnut w. ash lime hazelnut m. ash maple beech larch spruce
y0
y0 SE
a
a SE
P
R2
-2.83 -3.10
0.69 0.80
0.46 0.49
0.07 0.09
-2.12 -1.94 -3.08 -2.60 -2.99 -1.75
0.51 0.50 0.70 0.62 0.25 0.81
0.38 0.37 0.49 0.45 0.48 0.33
0.05 0.05 0.07 0.07 0.03 0.09
