Environ. Sci. Technol. 2004, 38, 6044-6050
Application of Controlled Mesocosms for Understanding Mercury Air-Soil-Plant Exchange M . S . G U S T I N , * ,† J . A . E R I C K S E N , † D. E. SCHORRAN,‡ D. W. JOHNSON,† S. E. LINDBERG,§ AND J. S. COLEMAN‡ Department of Natural Resources and Environmental Sciences, University of Nevada, Reno, Nevada 89557, Desert Research Institute, Reno, Nevada 89512, and Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831
Whole system elemental mercury (Hg0) flux was measured for ∼1.5 years using two large gas exchange mesocosms containing ∼100 two-year old aspen trees (Populus tremuloides) planted in soil with elevated mercury concentrations (12.3 µg/g). We hypothesized that during leafout, whole mesocosm Hg0 flux would increase due to movement of Hg0 in the transpiration stream from the soil to the air. This hypothesis was not supported; plants were found to assimilate Hg0 from the contaminated air, and whole system Hg0 emissions were reduced as plants leafed-out due to shading of the soil. Surface disturbance, watering, and increases in soil moisture, light, and temperature were all found to increase whole system Hg0 flux, with light being a more significant factor. Although surface soils were maintained at 15-20% moisture, daily watering caused pulses of Hg0 to be released from the soil throughout the experiment. Data developed in this experiment suggested that those processes acting on the soil surface are the primary influence on Hg emissions and that the presence of vegetation, which shields soil surfaces from incident light, reduces Hg emissions from enriched soils.
Introduction Most investigations of the effect of vegetation on mercury (Hg) biogeochemical cycling have been done using single plant experiments (1-5) or in field settings (6-9). These studies have shown that plant species and air and soil Hg concentrations influence vegetation-air Hg exchange and that plants may act as either a source or sink for atmospheric Hg. Frescholtz et al. (10) demonstrated that uptake of atmospheric elemental Hg (Hg0) (the dominant form of Hg in the atmosphere) by foliage is an important process. They showed that foliage Hg concentrations were controlled predominantly by air Hg0 concentrations and to a lesser extent by soil Hg concentrations. Investigations using a single plant gas exchange system demonstrated that Hg0 emission from and deposition to foliage are dynamic processes and that the net flux changes as air concentrations change (4, 5). Litterfall in forested ecosystems is a significant source of Hg to the forest floor (7, 9, 11, 12); however, it is not clear whether the Hg in foliage is derived from the local soils or the atmospheric pool. If litterfall is a new source of Hg to * Corresponding author e-mail:
[email protected]. † University of Nevada. ‡ Desert Research Institute. § Oak Ridge National Laboratory. 6044
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 22, 2004
ecosystems (cf. refs 10 and 13), estimates of atmospheric Hg deposition might have to be revised upward substantially, and controlling this input of Hg to ecosystems will be difficult. Understanding the relative importance of vegetation as a source or sink for atmospheric Hg is also critical for balancing the global Hg budget. Our ability to understand ecosystem level responses to environmental contaminants is limited by the technology we have to study the processes. With a single plant system, we cannot be sure that the processes we are monitoring are similar to those that would occur at a larger scale or how representative individual plant fluxes are relative to fluxes occurring in a forest stand. For example, Ericksen and Gustin (4) found that the range in Hg0 deposition associated with individual potted aspen was greater than that measured from trees growing in a stand under similar conditions. Other studies have found that single plant studies fall short when attempting to predict system level responses in modeling and scaling exercises (14). At the forest stand level, in situ data are often confounded by uncontrolled environmental variables. Studies of Hg at the controlled mesocosm scale are nonexistent because of the lack of large facilities for doing such work. This study investigated the role of plants in the biogeochemical cycling of Hg using two internationally unique gas exchange mesocosms termed Ecologically Controlled Enclosed Lysimeter Laboratories (EcoCELLs) at Desert Research Institute, Reno, NV. This project was the first use of this facility for the study of contaminant fluxes. The original working hypothesis was that whole cell Hg0 flux would increase as the plants leafed-out due to the movement of Hg from the soil to the air by plants via transpiration. In addition to monitoring whole cell flux, soil Hg0 flux was monitored separately using small dynamic field chambers (15), Hg0 and CO2 in soil gas were measured (16), and detailed analyses of Hg uptake by plants growing in the EcoCELLs were performed (13). Ancillary experiments with the same plant species were done using multiple plant exposure chambers (10) and with a single plant gas exchange system (4, 5). This paper describes the system-level fluxes and the biogeochemical cycling of Hg within the mesocosms.
Materials and Methods Two environmentally controlled EcoCELLs (7.3 × 5.5 × 4.5 m (l × w × d)) that are totally enclosed, naturally lighted plant growth chambers were used for this study (17). Within each EcoCELL are three lysimeter/rhyizotron weighing containers or pots (2.85 × 1.35 × 1.8 m), which have a soil surface area of 11.8 m2. For this project, each container was filled with ∼5 tons of gravel overlain by a liner and ∼4.5 tons of sandy loam topsoil amended with Hg containing mill tailings resulting in a Hg concentration of 12.3 ( 1.4 µg Hg g-1 (n ) 12). The speciation of Hg in the mill tailings is primarily Hg bound to Ag-Au-Fe sulfides, and based on selective extraction procedures, 0.06), regardless of soil Hg content, indicates that uptake of gaseous Hg0 was the predominant pathway by which Hg accumulated in the foliage (13). Estimated mean daily (24 h) foliar atmospheric Hg0 uptake rates (based on tissue Hg concentrations over time for both years) ranged from ∼6 ng m-2 (of leaf area) h-1 in the first month of plant growth to 3 ng m-2 h-1 as LAI increased to its maximum (13). Foliar uptake, measured with a portable field chamber during the day when LAI reached its maximum in the second year, was ∼3 ng m-2 h-1 (13). EcoCELL Resolution for Measurement of Fluxes. Despite the fact that only one Hg0 flux value for both EcoCELLs was 6048
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 22, 2004
obtained every 40 min, the resolution of the whole system flux response to environmental perturbations was very good. To demonstrate this, data for the first week of July 2001, when only one EcoCELL was planted, are presented in Figure 3. During this time, incident solar radiation declined from the 1st to the 3rd and was variable on the 4th and 5th due to intermittent clouds. Diurnal whole cell CO2 uptake and H2O emission for the planted cell were reduced as light intensity was reduced. Soil surface temperatures in both planted and unplanted cell were reduced with the decline in light, and the unplanted cell exhibited the greater response. Mercury emissions followed a diel pattern that peaked at midday and with watering of the soil at ∼1800 h. The peak flux associated with watering was greater than that for midday in most cases. Midday and watering peak emissions were greater for the unplanted cell (Figure 3). Regression analyses for Hg0 flux versus light and temperature (with nighttime and watering data from 1700 to 0600 removed) for the planted cell yielded r2 values of 0.73 and 0.80, respectively (p < 0.05), and for the unplanted cell, 0.78 and 0.67, respectively (p < 0.05). Gustin et al. (15), using data derived from dynamic
FIGURE 4. Schematic diagram summarizing Hg fluxes within the EcoCELLs at midday in August of the second year. Whole cell flux represents net emission. Downward pointing arrows represent inputs to the system. Numbers are in units of ng m-2 h-1. field flux chamber measurements made in November from the EcoCELL soil surface with plants removed, demonstrated that 75.1% of the variability in Hg flux in could be explained by light. In a stepwise regression analysis, the addition of soil temperature improved this to 77% and air H2O vapor concentration to 80% (p < 0.0001). On the basis of calculated activation energies, they suggested that photoreduction was an important process generating Hg0 in EcoCELL soils. A pulse of Hg0 was emitted with every watering event over the course of the experiment. The release of Hg0 from dry soils with the addition of water has been demonstrated in the field and laboratory and has been attributed to replacement of Hg0 adhered to soil particles by the more polar H2O molecule (31, 32). This data set demonstrates that Hg0 will also be released from enriched soils maintained at 15-30% soil moisture content. The greater Hg0 flux response to watering in the unplanted cell may be due to more Hg0 being available due to photoreduction of Hg (II) at the soil surface during the day. Johnson et al. (16) reported that there were strong vertical gradients of CO2 concentrations in soil gas for this experiment, with higher values at depth as is typical of a diffusion driven process. Elemental Hg0 soil gas concentrations did not exhibit consistent vertical gradients within the soil column. However, strong diel variations in Hg0 concentrations were measured with concentrations being highest at midday. They found that Hg flux calculated using soil gas diffusion equations, Hg0 gas concentrations, soil porosity, and soil moisture was 2 orders of magnitude lower than measured whole cell Hg0 fluxes. This indicates that Hg0 flux is not controlled by diffusion, as is the case for CO2 (16). The diel pattern in soil gas concentration mimicking the diel pattern in surface emissions suggests that they are driven by similar processes (i.e., temperature, light, atmospheric turbulence). Increased solar radiation during the day and subsequent heating of the soil could facilitate Hg desorption from soils and movement up through the soil column. Elemental Hg flux was significantly increased with disturbance of the soil surface at the initiation of both growing seasons, demonstrating that disruption of the soil surface will significantly enhance emissions. Surface processes were the main drivers of the observed temporal variations in flux with light having a more dominant influence than soil surface temperature. Watering caused a significant amount of Hg0 to be emitted from the soil on a daily basis, and changes in soil moisture had a profound affect on emissions. Our original hypothesis was that with leaf-out, whole cell Hg0 flux would increase due to transport of Hg0 from the soil to the atmosphere by way of the plant transpiration stream. In contrast, the response was a decline in whole system flux as a result of soil shading by the leaf canopy (Figure 4). Instead of emission from vegetation, deposition was measured using individual plant gas exchange chambers, and Hg0 accumulated in foliage over time. A mass balance of Hg fluxes in August of the second year (Figure 4) illustrates that the
dominant flux was emission from the soils. Inputs to the system were insignificant relative to outputs. Although plants were taking up Hg0 from the air, the rate of uptake was too low (∼3 ng m-2 h-1) to explain the observed difference in emissions between midday emissions in the planted and unplanted EcoCELLs (∼800 ng m-2 h-1) (Figure 4). Mercury added to the system by way of daily watering was a small input. If all the Hg added in irrigation water each day were released at a constant rate over 24 h, it would contribute
