A Regional Atmospheric Fate and Transport Model ... - ACS Publications

National Exposure Research Laboratory, MD-E243-03,. Research Triangle ... system is adapted to simulate the regional transport and fate of atrazine, o...
0 downloads 0 Views 869KB Size
Environ. Sci. Technol. 2002, 36, 4091-4098

A Regional Atmospheric Fate and Transport Model for Atrazine. 1. Development and Implementation ELLEN J. COOTER* AND WILLIAM T. HUTZELL U.S. Environmental Protection Agency, National Exposure Research Laboratory, MD-E243-03, Research Triangle Park, North Carolina 27711

The Community Multiscale Air Quality (CMAQ) modeling system is adapted to simulate the regional transport and fate of atrazine, one of the most widely used herbicides in the United States. Model chemistry and deposition are modified, and a gas-to-particle partitioning algorithm is added to accommodate semivolatile behavior. The partitioning algorithm depends on humidity, temperature, and particulate matter concentration and composition. Results indicate that gaseous atrazine will usually dominate warm season atmospheric concentrations, but particulate form can surpass gas forms when atmospheric humidity is high (>70%) and less-acidic (pH > 2.5) aqueous aerosol component is present. Implementation of the modified CMAQ for atrazine is illustrated, and, within the limits of our current understanding, preliminary transport and fate patterns appear to be reasonable. This research represents one of the first attempts to include a gas-to-particulate matter partitioning mechanism in an Eulerian grid-model.

Introduction Atrazine is one of the most widely applied herbicides in the United States (1). It is used to control broadleaf and grass weeds in corn and sorghum crops and has been applied both as a preemergent and postemergent agent in the United States and Canada for more than 30 years (2). General population exposure to this relatively nonreactive, semivolatile chemical, is most likely to occur through ingestion of contaminated drinking water derived from domestic wells or public water supplies providing minimal levels of treatment (3, 4). There is evidence that atrazine may disrupt endocrine function in mammals (5), and it has been related to observed declines in submerged aquatic vegetation in the Chesapeake Bay (6). Atrazine is commonly detected in soil, air, groundwater, and surface waters adjacent to agricultural fields as well as remote locations (6-13). Previous regional atmospheric transport assessments for atrazine have used range-oftransport (e.g. ref 14) and source-receptor (e.g. ref 15) modeling approaches. Range-of-transport studies are often multimedia screening level assessments in which the maximum linear distance traveled is modeled as a function of advection (16), dispersion (17), and degradation half-life in the atmosphere and adjacent surface media, e.g., soil, water, * Corresponding author phone: (919)541-1334; fax: (919)541-1379; e-mail: [email protected]. NOAA Air Resources Laboratory, Atmospheric Sciences Modeling Division on assignment to the U.S. Environmental Protection Agency, National Exposure Research Laboratory, MD-E243-01, Research Triangle Park, NC 27711. 10.1021/es011371y Not subject to U.S. Copyright. Publ. 2002 Am. Chem. Soc. Published on Web 08/22/2002

and vegetation. More sophisticated Lagrangian sourcereceptor analysis follows a pollutant “puff” along an atmospheric trajectory whose path is determined by meteorological conditions (18). The approach is computationally fast and is reliable for nonreactive pollutants when boundary conditions along the trajectory can be assumed to be constant. Eulerian grid models add yet another level of complexity regarding process detail and computational intensity. They solve numerical expressions describing vertical and horizontal transport, chemical, and emission processes by dividing the model domain into a large number of cells. Cells interact through advective and diffusive transport, i.e, dynamic boundary conditions. Internally, cells evolve pollutant concentrations based on chemistry, deposition (both wet and dry), and other transformation or removal processes (18). Although few Eulerian air quality models include the physicochemical process detail necessary to simulate semivolatiles in the atmosphere, the approach is well-suited to investigate this behavior as it relates to urban and regional photochemistry and particulate matter. In this paper, a partitioning algorithm is proposed, and its dependence on chemical and environmental factors is discussed. The algorithm, along with other modifications for semivolatiles, is implemented in the U.S. Environmental Protection Agency’s (EPAs) Community Multiscale Air Quality (CMAQ) model to explore the atmospheric fate of atrazine. Preliminary results are presented and interpreted.

Model Development The U.S. EPA has developed the CMAQ modeling system to investigate scientific and regulatory questions related to acid deposition, tropospheric ozone, and particulate material over scales ranging from urban to intracontinental (19). The CMAQ contains modules representing advection, eddy diffusion, air chemistry, aerosol physics, in-cloud, and precipitation processes. Advection and diffusion satisfy mass conservation and include removal by dry deposition (20, 21). Chemical production and loss terms are determined by a modified version (22, 23) of the Carbon Bond IV mechanism (24), a chemical kinetic mechanism for urban and regional photochemistry. The CMAQ module for aerosol physics determines concentrations of trimodal size distributed particulate material of diameter less than 10 µm (25). Each particulate mode can be an internal mixture of several compounds. Aitken and accumulation modes include particles of diameter 2.5 µm or less (PM2.5), that are either emitted or are produced by gas-to-particulate matter conversion processes. A thermodynamic mechanism (26), based on temperature and relative humidity, determines concentrations of water, sulfate, ammonium, and nitrate in the fine modes. Organic and elemental carbon are also included in these modes. The remaining mode represents coarse material having diameters between 2.5 and 10 µm and includes dry inorganic material emitted by natural and anthropogenic sources. In-cloud and precipitation processes simulate aqueous chemistry and wet deposition by cloud droplets (27-29). CMAQ adaptations required to address the behavior of atrazine in the atmosphere include the addition of atrazine gas chemistry, an algorithm to describe the mass exchange of atrazine between its gas and particulate forms, and deposition of each atrazine form from the atmosphere. One gas-phase atrazine form and three particulate-sorbed forms, one for each particulate mode, are introduced. It is assumed that only the gas form undergoes chemical degradation through a reaction with hydroxyl (OH) radicals (30). Structural VOL. 36, NO. 19, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4091

TABLE 1. Chemical Properties of Atrazine chemical and physical data for atrazinea principle use

broadleaf and grass herbicide photosynthetic inhibition 1912-24-9 C8H14ClN5 215.68 249.1 (cm3 mol-1)

mode of action CAS registration number molecular formula molecular weight molar vol (LaBas approximation)b density melting point boiling pointc water solubility solid vapor pressure (@ 293K) subcooled liquid vapor pressure (@298) Henry’s Law constantd acid dissociation constant (pKa) octanol-water partition coeff (Log Kow) soil adsorption (Koc)e OH reaction rate in air f

1.187 (g cm-3) 446 (K) 544 (K) 33 (mg L-1) 4.0 × 10-5 (Pa) 1.75 × 10-3 (Pa) 2.48 × 10-4 (Pa m3 mol-1) 1.7 2.4 100 (mL 2.73 × 10-11 (cm 3 molecules-1 s-1) 0.015 (per day) 0.951 (per day)

a Unless otherwise noted, (63) @ 298 K. b Reference 64. c Estimated by a structural analysis calculator, the SPARC online calculator, http:// ibmlc2.chem.uga.edu/sparc (accessed June 2002). d Reference 42. Estimated based on solid vapor pressure and solubility @ 298 K. e Reference 42. Representative value based on a range of lab and field experiments across soil type, organic content, and pH. f References 31 and 65. g Reference 30.

analysis (15, 31) provides a temperature independent reaction rate constant. The constant has not been measured, and other estimates (30, 32) suggest the value varies by more than a factor of 5. The product of the OH reaction is an unspecified compound that is assumed not to undergo further chemical reaction. It is assumed, based on speculation reported in Goolsby et al. (33), to have the same volatility and solubility as atrazine (Table 1). One gas and three particulate forms are added to represent the reaction product. Neglect of chemical degradation in the aerosols effectively prolongs the atmospheric lifetime of atrazine. Assumptions regarding reversible gas-to-particle exchange mitigate this effect, because particle reserves quickly replace degraded gas form chemical concentrations. Degradation is assumed to be negligible in cloudwater, because model generated cloudwater and aqueous aerosols have lifetimes that are shorter ( 70%. The effect corresponds to forcing atrazine completely into the Aitken and accumulation modes because each mode contains sufficient neutral water to absorb gaseous atrazine. Figure 1B,C shows that lower temperatures or higher PM10 concentrations cannot reproduce this effect’s size below 70% relative humidity. Atrazine’s relatively high Po and low Kow predict that adsorption and absorption do not move atrazine from the gaseous to particulate forms. Lowering temperature increases particulate fraction (Figure 1B) because the partitioning coefficients vary inversely with temperature. Particulate fractions also increase when PM10 concentrations are higher, mostly distributed in the fine modes, and contain a large fraction of organic material (Figure 1C). These conditions produce large amounts of surface area or organic solvents for sorption to occur. Urban areas exhibit such conditions during moderate to high pollution events but most continental settings do not. Hence, the partitioning algorithm predicts that the gaseous forms of atrazine dominate except where the atmosphere is moist and has large concentrations of NH4+ or organic material in PM2.5. Further sensitivity analysis indicates that chemical property uncertainty can impact particulate fraction prediction as well. For example, calculations show that the particulate fraction changes by a factor of 2-4 when Po and ∆E are allowed to vary (see Supporting Information). The partitioning mechanism clearly depends on the way particulate matter is represented. The model predicts only PM10, but accurate simulation of partitioning behavior may require the inclusion of particulate concentrations greater than 10 µm in diameter. This additional material can supply significant aqueous or organic mass for sorption, and its omission could result in the underprediction of particulate form concentrations, and overprediction of gas forms. Atrazine and its degradation product are removed from the atmosphere through dry and wet deposition. Dry

4094

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 19, 2002

FIGURE 2. Atrazine exchange in the lowest layer of the atmosphere for grid cell with lower left corner at -87.387 longitude 39.967 latitude (north central Indiana) on June 7, 1995 (A) 12 midnight to 1 a.m. and (B) 12 noon to 1 p.m. Note: advection includes vertical as well as horizontal components. deposition is modeled as a one way flux out of the lowest layer of the atmosphere. Particulate forms have dry deposition velocities determined by Brownian diffusion, turbulence, and sedimentation. Mass and diameter of the mode determine the sedimentation velocity (23, 44). The gas form deposition velocity is based on an electrical resistance analogue that includes aerodynamic, Ra, quasi-laminar boundary layer, Rb, and canopy (surface), Rc, resistance (45). Ra is computed using similarity theory with heat flux. Rb depends on landuse specific friction velocity and molecular characteristics of gases. Rc is made up of several components, including bulk stomatal, dry cuticle, wet cuticle, ground and in-canopy aerodynamic resistance (45). Cuticle and ground resistances are based on ozone and sulfur dioxide observations and neglect absorption into organic plant or soil material. Values for time-dependent Rc parameters are determined via landuse and mesoscale meteorological models (46), which provide fractional vegetation cover, leaf area index, fractional leaf area wetness, and leaf stress associated with radiation, rootzone soil moisture, temperature, and humidity. Wet deposition of the gas form occurs through in-cloud and below-cloud scavenging. Both processes are assumed to have the same scavenging efficiency. Wet scavenging of particulate forms is described in Roselle and Binkowski (47) and depends on the particle mode and time required to remove all of the water from a cloud volume. Wet scavenging

FIGURE 3. Predicted volatilization and deposition of atrazine (total during a simulated hour), and hourly atrazine and degradation product concentrations, and planetary boundary layer (PBL) height during the simulation period 0000 GMT June 7-2300 GMT June 8 1995, for a model grid cell with lower left corner at -87.387 longitude 39.967 latitude (north central Indiana). Local time (GMT minus 6 h) is indicated along the top of the chart. of gas forms depends on the Henry’s Law, water content of the cloud, and precipitation rate (47).

CMAQ Implementation for Atrazine CMAQ simulations require estimates of chemical releases to the atmosphere. Emission of atrazine during application is reported to be about 0.5% of applied chemical (48). Postapplication emission includes volatilization of local chemical from the soil surface, wind resuspension of soil particles carrying attached chemical, and revolatilization of extralocal inputs, i.e., revolatilization of deposited chemical that has previously undergone regional or long-range transport. Research suggests that wind resuspension (8, 49) and revolatilization of extra-local chemical could be important sources of atmospheric atrazine, but neither process is well understood or quantified and are not considered further at this time. In contrast, volatilization of atrazine applied to local soils has been widely studied. Its magnitude and duration are determined by application method (8), soil properties and condition (8), farming practice (4, 8, 50), meteorology, and, given its relatively long half-life in soil (Table 1), time, e.g., refs 8 and 50-52. Cumulative losses immediately following application (