JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 1 May 2008
Published in J Environ Qual 37:1064-1072 (2008)
DOI: 10.2134/jeq2006.0562
© 2008 American Society of Agronomy, Crop Science Society of America, and 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 Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bayless, E. R.
Right arrow Articles by Lampe, D. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bayless, E. R.
Right arrow Articles by Lampe, D. C.
Agricola
Right arrow Articles by Bayless, E. R.
Right arrow Articles by Lampe, D. C.
Related Collections
Right arrow Best Management Practices
Right arrow Pesticides
Right arrow Vadose Zone Processes and Chemical Transport

Simulated Fate and Transport of Metolachlor in the Unsaturated Zone, Maryland, USA

E. Randall Baylessa,*, Paul D. Capelb, Jack E. Barbashc, Richard M. T. Webbd, Tracy L. Connell Hancocke and David C. Lampef

a USGS, 5957 Lakeside Blvd., Indianapolis, IN 46278
b USGS, 500 Pillsbury Dr., Minneapolis, MN 22807
c USGS, 934 Broadway, Tacoma, WA 98402
d USGS, MS 412, Denver Federal Center, Denver, CO 80225-0046
e USGS, 1730 East Parham Rd., Richmond, VA 23228
f USGS, 5957 Lakeside Blvd., Indianapolis, IN 46278. Commercial names are provided for purposes of identification only and should not be interpreted as an endorsement by the U.S. Geological Survey

* Corresponding author (ebayless{at}usgs.gov).

Received for publication December 28, 2006.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
An unsaturated-zone transport model was used to examine the transport and fate of metolachlor applied to an agricultural site in Maryland, USA. The study site was instrumented to collect data on soil-water content, soil-water potential, ground water levels, major ions, pesticides, and nutrients from the unsaturated zone during 2002–2004. The data set was enhanced with site-specific information describing weather, soils, and agricultural practices. The Root Zone Water Quality Model was used to simulate physical, chemical, and biological processes occurring in the unsaturated zone. Model calibration to bromide tracer concentrations indicated flow occurred through the soil matix. Simulated recharge rates were within the measured range of values. The pesticide transport model was calibrated to the intensive data collection period (2002–2004), and the calibrated model was then used to simulate the period 1984 through 2004 to examine the impact of sustained agricultural management practices on the concentrations of metolachlor and its degradates at the study site. Simulation results indicated that metolachlor degrades rapidly in the root zone but that the degradates are transported to depth in measurable quantities. Simulations indicated that degradate transport is strongly related to the duration of sustained use of metolachlor and the extent of biodegradation.

Abbreviations: MESA, metolachlor ethanesulfonic acid • MOXA, metolachlor oxanilic acid • RZWQM, root zone water quality model • PRZM, pesticide root zone model


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
THE PURPOSE of this study was to examine the transport and fate of metolachlor (2-chloro-N-(2-ethyl-6-methylphenyl)-N-(2-methoxy-1-methylethyl) acetamide) and the metolachlor degradation compounds metolachlor ethanesulfonic acid (2-[(2-ethyl-6-methylphenyl)(2-methoxy-1-methylethyl)amino]-2-oxoethanesulfonic acid) (herein referred to as MESA) and metolachlor oxanilic acid (2-[(2-ethyl-6-methylphenyl)(2-methoxy-1-methylethyl)amino]-2-oxoacetic acid) (herein referred to as MOXA) in the unsaturated zone beneath an agricultural site in Maryland, USA. Field measurements and numerical models can be a potent tool for identifying factors affecting pesticide attenuation. In this study, surface meteorology, agricultural management, and data from an instrumented, unsaturated profile were combined with modeling capabilities of the Root Zone Water Quality Model (RZWQM) (Ahuja et al., 2000) to simulate the transport and fate of metolachlor and its degradates at the site.

The unsaturated zone is an important environmental compartment for understanding the degradation and attenuation of pesticides (Bayless, 2001). Ground water is susceptible to contamination by pesticides that leach through the unsaturated zone, which may be a primary pathway for pesticides to enter aquifers (Flury, 1996). Environmental agencies and the agricultural industry collaboratively seek to minimize pesticide application without compromising crop production. This dilemma creates a need to understand the magnitude and interaction of complex physical, chemical, and biological processes in the unsaturated zone that vary through space and time (Barbash, 2005).

In preparation for this study, Nolan et al. (2005) evaluated the capabilities of multiple unsaturated-zone transport models to simulate data sets describing agricultural management practices from various regions of the USA. Of particular importance was the flexibility to simulate integrated matrix and preferential flow, biological systems, multiple pesticide degradation compounds, and various agricultural management practices. Extensive public-domain documentation of the mathematical basis of the model and published application to a variety of complex problems was also required. Nolan et al. (2005) determined that the RZWQM (Ahuja et al., 2000) simulated pesticide transport in various agricultural settings with the smallest simulation errors among the transport models tested. Based on that study, the RZWQM was selected to simulate processes affecting metolachlor and degradate attenuation at the Maryland study site.

The RZWQM was developed in the 1980s by the U.S Department of Agriculture-Agricultural Research Service (http://gpsr.ars.usda.gov/products/rzwqm.htm). The RZWQM drew on existing simulation codes that included CREAMS (chemicals, runoff, and erosion from agricultural management systems) (Knisel, 1980), GLEAMS (groundwater loading effects of agricultural management systems) (Leonard et al., 1987), NRTM (nitrogen tillage residue management) (Shaffer and Larson, 1987), Opus (Smith, 1992), and PRZM (the pesticide root zone model) (Carsel et al., 1985) and incorporated routines to simulate tile drainage, preferential flow, and nutrient and pesticide transformation processes (Ahuja et al., 2000). The RZWQM was originally published in 1992, was re-released in 1998 with a Microsoft Windows–compatible interface, and has received continuous updates and improvements. The code was extensively tested at USDA Management Systems Evaluation Areas in the Midwestern USA (Farahani et al., 1999; Ghidey et al., 1999; Jaynes and Miller, 1999; Landa et al., 1999; Martin and Watts, 1999; Wu et al., 1999) and at several other sites throughout the USA (Malone et al., 2005).

The present modeling effort focused on the pesticide metolachlor because of its widespread use on corn (Zea mays L.) and soybean (Glycine max) crops. Metolachlor (C15H22ClNO2; CAS 51218–45–2) was registered in 1977 as a pre-emergent herbicide for weed control for a variety of crops (FAO/WHO, 1998) and is marketed under several names, including Dual, Pennant, Pimagram, and Bicep. Metolachlor is biodegraded in soil (Barbash et al., 1999; Saxena et al., 1987), photodegrades in water (USEPA, 1980), is sorbed to soil organic matter, and may volatilize from the soil surface (Rice et al., 2002; Burkhard and Guth, 1981). A primary degradation pathway is transformation by aerobic and anaerobic microorganisms (Capel et al., 2008; Barbash et al., 1999). The resolved enantiomer of metolachlor, S-metolachlor, has become more widely used since product registration in 1997.

For the protection of human health, the U.S. Geological Survey (USGS) has computed a Health-Based Screening Level of 70 µg L–1 for metolachlor in water resources. Health-Based Screening Levels are developed in collaboration with the U.S. Environmental Protection Agency (USEPA) and others (USGS, 2006). For aquatic life, Gilliom et al. (2006) used USEPA data to compute benchmark values of 1950 and 780 µg L–1 for acute and chronic exposures of fish, respectively, to the herbicide and 12,550 µg L–1 for acute exposures for invertebrates. For the protection of all life stages of the most sensitive species of freshwater aquatic plants or animals, the Canadian Council of Ministers of the Environment has established an interim water-quality guideline of 7.8 µg L–1 (Environment Canada, 2004; Canadian Council of Ministers of the Environment, 1999).

Metolachlor and its degradates, MESA and MOXA, are widely distributed in the ecosphere (Barbash et al., 1999). During the 1990s, more than 50.8 megagrams of metolachlor were annually carried by the Mississippi River to the Gulf of Mexico (Pereira and Hostettler, 1993), and metolachlor was detected in about 1% of 6 million domestic wells across the USA in 1992 at concentrations ranging from 0.1 to 1.0 µg L–1 (Holden et al., 1992). Metolachlor ethanesulfonic acid and MOXA are more abundant than metolachlor in environmental water samples (Kalkhoff and Thurman, 1999). Metolachlor ethanesulfonic acid reportedly occurs in 2 to 100 times more samples than MOXA (Kalkhoff et al., 1998), and where both are present, concentrations of MESA are as much as 5 times greater than that of MOXA (Eckhardt et al., 1999).

A limited number of studies have examined metolachlor attenuation in the unsaturated zone. Studies have examined data derived from laboratory experiments (Jebellie et al., 2004; Sanyal and Kulshrestha, 2003; Wietersen et al., 1993) and field sampling efforts (Zacharias et al., 1999; Hippe and Hall, 1996; Smith and Parrish, 1993; Trevisan et al., 1993; Mullaney et al., 1991). Results of these studies have shown that metolachlor concentrations generally taper off to values less than the method reporting limit between 0.1 and 0.5 m and are immeasurable in the saturated zone. Numerical models, including AgriFlux (Novak et al., 2003), GLEAMS (Zacharias et al., 1999), LEACHM (Trevisan et al., 1993), PRZM (Trevisan et al., 1993), PRZM2 (Jebellie et al., 2004), MACRO (Balderacchi et al., 2002), and Opus (Zacharias et al., 1999), have been used in conjunction with field and laboratory data to examine the capabilities of models to simulate metolachlor leaching and to identify factors affecting attenuation. Most models have demonstrated the capability to simulate metolachlor concentrations near the surface but have been less accurate at forecasting distributions below the root zone. Inadequate algorithms to compute transport by preferential flow, as well as the occurrence of pesticide transformations, have generally been identified as the causes of poor correlation between measured and simulated concentrations. The fate and transport of MESA and MOXA have not been widely described.

The RZWQM was applied to field data to identify appropriate model parameters and to examine relevant processes affecting the transport and fate of metolachlor in the unsaturated zone. Simulations examined processes occurring over two time periods: (i) a "3-yr" period from March 2002 through October 2004, during which field measurements were made and agricultural management information were available, and (ii) a "20-yr" period from January 1984 through October 2004 that appended hypothetical agricultural management and probabilistic weather data to the front of the 3-yr simulation. The 3-yr simulations were used for model calibration and to obtain model parameter values to produce reasonable simulations of the measured data. The 20-yr simulations were used to evaluate the potential impact of historical agricultural management practices on (i) the accumulation and storage of pesticide degradation compounds in the unsaturated zone and (ii) the transport of those chemicals to the water table.


   Methods
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Methods
Results and Discussion
 Conclusions
 REFERENCES
 
Study Design
The present investigation was nested within a larger set of studies conducted by the USGS National Water-Quality Assessment Program intended to help understand the behavior and fate of agricultural chemicals in all compartments of the ecosphere at sites, representing important cropping practices within the USA. At the Maryland study site, a single vertical profile through the unsaturated zone was sampled, instrumented, and monitored to provide the data needed to calibrate the numerical model. The model simulations produced information about pesticide fate and chemical loading to other environmental compartments.

Data Collection and Analysis
A study site was selected near Kennedyville, Maryland, where corn and soybeans are normally cropped in alternate years. The climate is coastal temperate with hot, humid summers and cold, damp winters. The study site has a surface slope of 1.5%, a relatively shallow water table (approximately 5 m), is not tiled or irrigated, and soil textures range from loam to sandy loam. The study site represents many sites throughout the USA where row-crop agriculture is the predominant practice. A detailed description of the study site, the installation of monitoring equipment, and the methods used to collect and analyze field data can be found in Capel et al. (2008).

The study site was instrumented to continuously measure soil-water potential, soil-water content, water-table elevation, and weather-related variables and occasionally to collect samples for water-quality analyses. Instruments included one piezometer (screened 7.32–8.23 m below land surface), five heat-attenuation probes (0.39, 1.08, 2.26, 4.09, and 5.49 m depth), four time-domain reflectometers (0.49, 0.79, 1.13, and 1.37 m), two pan lysimeters (0.52 and 1.22 m depth), and five pressure-vacuum lysimeters (0.58, 1.27, 2.44, 4.27, and 5.45 m depth). Water samples could not be extracted from the deepest pan lysimeter and the two deepest pressure-vacuum lysimeters, and sampling was discontinued after initial attempts were unsuccessful. An argillic horizon positioned directly above the deeper pan lysimeter may have limited the amount of recharge reaching the deeper instruments (William Guertal, U.S. Geological Survey, written commun., 7 Nov. 2006). Measured weather variables included humidity, precipitation, solar radiation, wind speed and direction, and air temperature. Water samples were collected quarterly from March 2004 through March 2005 and after three rainfall events in April, June, and July 2004. Gaps in weather data resulting from instrument failure were filled with measurements from nearby research stations (Table 1 ). Weather data for the period 1984 through 2001 were generated using the CLIGEN stochastic model (Zhang and Garbrecht, 2003) using 44 yr of record from the Baltimore, MD, airport. Water samples were analyzed for concentrations of major ions, pesticides, pesticide-degradation compounds, and nutrients (Capel et al., 2008). On many occasions, one or more lysimeters did not produce enough water to permit analysis for any or all constituents. Core samples were collected from the vicinity of each lysimeter cup and the well screen during their installation, and these samples were analyzed for particle-size distribution and concentrations of pesticides, inorganic constituents, and organic matter; bulk density was computed from these data (Table 2 ). A potassium bromide solution (15 mg m–2) was applied to a 30-m2 area of land surface directly above the unsaturated-zone instruments on 12 May 2004 to produce conservative-ion breakthrough data for calibrating the advection-dispersion component of the solute transport model.


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

  		Table 1. Sources of weather data.

 

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

  		Table 2. Soil and substrate properties.

 
Information on agricultural management practices was provided by the producer operating at the site (Table 3 ). The customary corn–soybean crop rotation was disrupted during the period of this study by an unusually wet April 2004. Consequently, soybeans were planted in 2003 and 2004; corn was planted in 2002 and 2005. Agricultural management practices simulated by the model included tillage, crop planting, applications of inorganic fertilizers and pesticides, and harvest. Dates and details of agricultural practices were simulated to agree with information provided by the producer. Chemical properties of pesticides were acquired from the RZWQM database, updated where possible with more recently published values, or estimated from chemical structures when values were otherwise unavailable (Table 4 ) (Capel et al., 2008).


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

  		Table 3. Agricultural management activities.

 

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

  		Table 4. Model parameters for pesticides.

 
Numerical Model
To the extent possible, the numerical model was parameterized with measured data or values computed from measured data. The upper boundary of the one-dimensional RZWQM model was land surface. Infiltration was simulated using the Green-Ampt equation (Green and Ampt, 1911). The simulated soil column extended 10 m below land surface. The soil column was discretized into six layers based on distinct changes in particle-size data or visual appearance of a core collected at the site. The physical soil properties of the uppermost layer were identical to layer 2, but it was added (as layer 1) to allow simulation of different microbiology, residue, and macropore properties. The six layers were rediscretized internally by RZWQM into 208 computational layers with thicknesses ranging 1 to 5 cm, including approximately 110 unsaturated computational layers (depending on the depth of the fluctuating water table). The six layers were parameterized with measured particle-size and organic-matter data. Soil-water content–soil-water potential–unsaturated hydraulic conductivity functions were modeled using a slightly modified, nonhysteretic form of the Brooks-Corey relations internally parameterized with values suggested by Rawls et al. (1982) for the specified particle-size distribution. The hydraulic conductivity of near-surface layers was automatically modified by RZWQM after tillage and subsequent reconsolidation. Inclusion of the surface crust–macropore flow capability of RZWQM was tested but was not necessary to simulate the water velocities indicated by the bromide tracer test. Initial values of soil-water potential were estimated for the six layers; values ranged from 0 at the base of the profile to –50 cm of water at land surface. Initial values of the soil-water content and unsaturated hydraulic conductivity were internally computed by RZWQM from the initial values of soil-water potential. The initial water-table elevation was set to 5 m. The basal boundary condition was a constant flux rate of 3 x 10–5 cm s–1 and was intended to simulate leakage from the saturated zone. Initial values of temperature were set to 15°C for all layers. The warm-up period before pesticide application and the iterative solving procedures allowed temperature and soil-water content to equilibrate to system conditions before pesticide application and transport. Water movement was computed using an iterative finite-difference solution of the Richard's equation. The nonreactive component of transport was computed by solving the advection-dispersion equation. A more detailed description of the methods used to parameterize and calibrate the model can be found in Capel et al. (2008).

The 3-yr model was iteratively calibrated to the system hydrology, nitrogen cycle, and pesticide distribution; this procedure followed an approach similar that described in Hanson et al. (1999) but focused on pesticide and degradate concentrations rather than plant production as a calibration target. The 3-yr model was calibrated to bromide concentrations as an indicator of transport by advection and dispersion. Simulated and measured soil-water contents were compared, and the simulated water flux to the saturated zone was compared with a range of recharge rates computed or this site (Nolan et al., 2007). Parameters affecting the distribution and partitioning of organic matter and microorganisms were then adjusted to improve the correlation between the simulated and measured distribution of dissolved nitrate in the unsaturated zone. In the RZWQM, microorganism populations do not significantly affect pesticide dissipation; however, they degrade crop residue and create organic matter in the soil that affects pesticides and degradate sorption. Model-sensitive pesticide properties (Koc, soil half-life) were then adjusted within the reported range of values from the literature and computed from chemical-structure properties (Capel et al., 2008) to fit measured and simulated pesticide concentrations. The mean error (Ahuja et al., 2000), mean absolute error, and modified coefficient of efficiency (Legates and McCabe, 1999) were statistical indicators used to identify the best pesticide-model parameterization.

The 3-yr model calibration focused on matching simulation results to observations at 0.5 m where weather, soil hydraulics and chemistry, microbiology, and agricultural management practices were expected to have the most profound effects on the hydrologic system and chemical concentrations would be most variable. Bromide and pesticide concentrations also occurred most commonly in measurable concentrations at the 0.5 m depth.

The 20-yr simulations were not used for model calibration because of uncertainty in the accuracy of hypothetical agricultural management practices and the probabilistic weather data. The 20-yr simulations used a 2-yr corn–soybean rotation believed to represent historical agricultural management in the region, except for the plantings of soybean in the 2003 and 2004 growing seasons. The switch from metolachlor to S-metolachlor and the accompanying change in application rates were simulated beginning in 2001. Hydrologic and chemical parameters developed during the 3-yr calibration simulations were used in the 20-yr simulations.


   Results and Discussion
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Methods
 Results and Discussion
Conclusions
 REFERENCES
 
Soil Water Content and Ground Water Recharge
The input values of hydraulic and hydrologic variables were adjusted until the model accurately simulated the concentrations of bromide tracer measured at a depth of 0.5 m (Fig. 1 ). Adjustments to hydraulic variables computed from grain-size analyses were tested but did not improve calibration. Preferential flow capabilities of RZWQM also were not required to fit the measured data. Although the RZWQM model seemed to acceptably simulate bromide concentrations at 0.5 m, breakthrough occurred late in the study period, and it was not possible to assess the model's ability to reproduce the entire breakthrough curve. Bromide detections below 0.5 m were few and infrequent. Of 30 samples analyzed for bromide, concentrations of bromide in 21 were at or below the method reporting limit (0.05 mg L–1); six of the samples with measurable concentrations were collected at 0.5 m.


Figure 1
View larger version (8K):
[in this window]
[in a new window]

  		Fig. 1. Measured and simulated bromide concentrations at 0.5 m depth. Squares indicate measured bromide concentrations; triangles indicate bromide concentrations less than the minimum reporting limit.

 
The primary indicator of acceptable calibration of the hydraulic component of the model was the agreement of measured and simulated bromide concentrations. The absolute values and trends in simulated and measured soil-water contents were similar (Fig. 2 ). The mean error in soil water contents was 0.02 and was greater near land surface and diminished with depth.


Figure 2
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 2. Measured and simulated soil-water contents at 1.1 m depth.

 
In agreement with measured data, the simulated depth of the water table ranged from 5.0 to 5.2 m below land surface during most of the simulation period. The simulated daily flux of water from the unsaturated zone to the water table ranged from 0.005 to 0.430 cm d–1, being greatest from December through July and least from August through November. The simulated mean annual water flux from the unsaturated zone to ground water for the period of study was 0.119 cm d–1. Flux to the water table was highest during years when soybeans were cropped. The simulated water-flux rates compared well with the range of recharge values computed by Nolan et al. (2007) for the study area (0.003–0.210 cm d–1).

Metolachlor
Simulation results can be sensitive to the values of many input parameters and are most sensitive to pesticide half-life and Koc values. Model parameterization complicates studies of the fate of metolachlor because values of model-sensitive parameters for the degradation compounds are not widely available. Three half-lives were used to describe transformation of the pesticide metolachlor-foliar contact, soil surface, and soil profile (Table 4). The half-life of metolachlor in soil reportedly ranges from 15 to 190 d (Accinelli et al., 2001; Aga and Thurman, 2001; Dinelli et al., 2000; Walker and Brown, 1985), with the shorter periods generally determined through field studies (Sanyal et al., 2000). For this study, a soil profile half-life of 10 d resulted in simulated concentrations of metolachlor that were less than the method reporting limit and produced the best agreement between simulated and measured degradate concentrations; metolachlor did not occur in measurable concentrations in field samples. The unique foliar contact and soil-surface half-lives were acquired from databases internal to the numerical model SWAT (Soil & Water Assessment Tool; Neitsch et al., 2002). The Koc of metolachlor has been reported to range from 88 to 364 (Jack Barbash, U.S. Geological Survey, written commun., 17 Nov. 2005). A metolachlor Koc value of 182 was used in the model. The suitability of the selected half-lives and Koc for metolachlor could not be judged on the basis of concentrations measured in lysimeters because concentrations were less than the method reporting limit in all samples; the value was consequently evaluated based on (i) simulated metolachlor concentrations not exceeding the method reporting limit at the depths of any lysimeters and (ii) producing MESA and MOXA in quantities that allowed model calibration to measured quantities of those compounds.

Metolachlor ESA and OXA
Values of the half-lives and Koc for MESA and MOXA have not been widely reported. Soil-layer aerobic half-lives for MESA and MOXA that produced good agreement between simulation results and measured data were 70 and 50 d, respectively. The relative rates were estimated from measured abundances in samples collected at a USGS National Water-Quality Assessment site in central Nebraska with a thick (25 m) unsaturated zone (Hancock et al., 2008). The initial formation ratio of 70/30 (MOXA/MESA) was also estimated from the central Nebraska data. The model-calibrated Koc values for MESA and MOXA (13.5 and 17, respectively) that minimized simulation errors were adjusted from initial values (1.6 and 51, respectively) computed from chemical structure properties (Jack Barbash, U.S. Geological Survey, written commun., 17 Nov. 2005). Theoretical parameter values commonly need to be adjusted when they are applied in measured laboratory and field studies (Flury, 1996).

The pesticide-transport model was calibrated to aqueous concentrations of MESA and MOXA collected at 0.5 m in 2004 (Fig. 3 ). It was anticipated that the conditions at 0.5 m would be the least complicated in terms of discriminating the breakthrough of recently applied metolachlor and degradates from the carry-over concentrations of pesticides applied during previous years. Deeper lysimeter data would likely be affected by chemical accumulation and storage as well as variable biotransformation rates associated with depth-dependent changes in temperature, organic matter concentration, and redox conditions. The simulated concentrations of MESA at 0.5 m ranged from 0.3 to 5.1 µg L–1. The mean error and mean absolute error of simulated MESA concentrations at 0.5 m were –0.083 µg L–1 and 0.129, respectively, and indicated that the model did not favor under- or overprediction (Table 5 ). Simulated concentrations of MOXA at 0.5 m ranged from 0.090 to 1.01 µg L–1. The mean error and mean absolute error of simulated MOXA concentrations at 0.5 m were 0.140 and 0.169 µg L–1, respectively, and indicated only a slight bias for underpredicting MOXA concentrations.


Figure 3
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 3. Simulated and measured (A) metolachlor ethanesulfonic acid (MESA) and (B) metolachlor oxanilic acid (MOXA) concentrations at 0.5 m. Diamonds indicate measured degradate concentrations.

 

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

  		Table 5. Errors computed for measured and simulated chemical concentrations.

 
The calibrated model indicated that metolachlor is virtually completely degraded in the uppermost 1 m of the unsaturated zone (Fig. 4A ). The MESA and MOXA concentrations peaked at approximately 1 m (Fig. 4B and 4C). Despite smaller concentrations, MESA was transported deeper into the unsaturated zone compared with MOXA as a result of a longer half-life. Although the model errors were relatively low, the simulated concentrations of MESA and MOXA at 2.4 m were substantially lower than measured values (Fig. 4B and 4C). The discrepancy was noteworthy because the accuracy of estimating degradate flux to ground water hinges on accurate predictions in deeper parts of the unsaturated zone. Plausible explanations for the discrepancy between measured and simulated concentrations of MESA and MOXA at 2.4 m included (i) the 3-yr calibration simulations were not long enough to permit penetration of pesticide degradates to the 2.4-m depth, particularly in the fine-grained, matrix-flow dominated soils at this site; (ii) the corn–soybean–soybean cropping used for the short-term model calibration did not adequately simulate loading of metolachlor if historical applications were made on a consistent 2-yr corn–soybean rotation; and (iii) that biotransformation rates depend on depth-dependent variables that were not well represented by half-lives applied uniformly through the vertical profile (e.g., Federle et al., 1986). The first two concerns were examined by constructing a 20-yr model of the site. The third concern was addressed by applying degradate half-lives in a 20-yr model that simulated biotransformation only in the aerobic-soil zone, thereby reducing the computed biotransformation of degradates below the root zone. This approach was supported by studies that have shown that biotransformation rates can be much faster in the root zone than in deeper parts of the unsaturated zone (Vinther et al., 2001).


Figure 4
View larger version (19K):
[in this window]
[in a new window]

  		Fig. 4. Simulated distribution of aqueous and sorbed concentrations of (A) metolachlor, (B) metolachlor ethanesulfonic acid (MESA), and (C) metolachlor oxanilic acid (MOXA) in the unsaturated zone. Units of adsorbed species are µg g–1; aqueous species are mg L–1.

 
Twenty-year simulations were done using a persistent agricultural management pattern that included a 2-yr corn–soybean rotation and metolachlor application during corn-cropped years. As with the 3-yr simulations, the 20-yr simulations predicted that metolachlor would dissipate in the upper 1.0 m of the unsaturated zone (Fig. 5A ). The transport of MESA and MOXA, however, increased to deeper parts of the unsaturated zone with the longer simulation period. This was particularly true for the simulations that limited degradation to the aerobic part of the unsaturated zone; those simulations indicated that degradates would eventually reach the water table (Fig. 5B). Simulations that maintained a uniform degradation rate throughout the entire unsaturated zone indicated that degradation compounds would potentially not reach the water table after 20 yr of biannual application (Fig. 5C). The modified coefficient of efficiency statistic indicated that the 20-yr simulations were less accurate at simulating the measured data than the 3-yr calibration simulations (Table 5).


Figure 5
View larger version (24K):
[in this window]
[in a new window]

  		Fig. 5. Plot of (A) metolachlor after 3, 5, 10, 15, and 20 yr of biannual metolachlor applications, (B) metolachlor ethanesulfonic acid (MESA) with aerobic-soil zone (only) degradation, and (C) MESA with uniform soil-profile degradation.

 
Ground water concentrations of MESA and MOXA in wells screened beneath the unsaturated zone ranged from 1.84 to 16.0 µg L–1 and from 1.48 to 10.0 µg L–1, respectively (Table 6 ). Metolachlor ethanesulfonic acid and MOXA were detected in 50 and 33% of the samples, respectively; these detection rates are less than those in samples from the lysimeters and indicate that conditions in the saturated zone were not static. Although concentrations of MESA and MOXA simulated at 2.4 m with the 20-yr model (1.23–21.4 µg MESA L–1; 0.59–12.7 µg MOXA L–1) were similar to values observed in the saturated zone, concentrations simulated at the approximate depth of the water table (5 m) were generally less than the method reporting limit. This result may indicate that MESA and MOXA concentrations in the saturated zone are attributable only in part to vertical flux of degradates through the unsaturated zone at this particular site.


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

  		Table 6. Summary of water and soil-water quality data.

 
Model Limitations
Model complexity and model limitations can be directly related. The limitations can be traced to the number and availability of values for the input parameters and the quantity and quality of measured data used for model calibration (Dubus et al., 2003; Ahuja and Ma., 2002). The RZWQM is a comprehensive code that requires extensive parameterization. Initialization of crop residue, soil organic matter, microorganism populations, and associated rate coefficients requires parameter values that are not commonly measured (Hanson et al., 1999) but have direct and indirect bearing on the transport and fate of metolachlor. Some parameter values must be derived during model calibration, and the variability of parameter values implies that models will be non-unique (Malone et al., 2005). Other parameterizations would have benefited from additional field measurements in time and space. Guidance for calibrating RZWQM parameters indicates using some data for calibration and the remainder to verify calibration (Hanson et al., 1999); the limited study period and number of field measurements prevented using that approach to improve and evaluate this study. Optimization and parameter estimation routines were not used to identify parameter values that would generate the most numerically precise solution; however, applying those routines to complex agricultural models may not improve calibration (Ahuja and Ma, 2002). Model non-uniqueness and caveats associated with applying multiple theoretical solutions to an actual field site should be considered when evaluating the results of this study.

Complications related to natural conditions should be considered when evaluating the results of this study. The study design expected corn to be planted in 2004 and treated with metolachlor. In response to the wet weather, however, soybeans were planted, and metolachor was not applied. This unusual cropping pattern may have bearing on the direct application of these results to other sites. In addition, without data from 2004, there is no information to confirm the model results that metolachlor is completely sorbed or degraded during a standard 2-yr rotation. Although the literature values of chemical properties for metolachlor indicate that it should be highly sorbed and rapidly degraded relative to MESA and MOXA, the availability of field-measured data collected immediately after metolachlor application would have strengthened the model calibration.


   Conclusions
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Methods
 Results and Discussion
 Conclusions
REFERENCES
 
Simulations indicated that metolachlor, an herbicide commonly used on corn, was almost completely transformed to MESA and MOXA within 1 m of land surface. A much smaller quantity of metolachlor was sorbed to soil in the root zone. Metolachlor was not transported to the water table. The degradates MESA and MOXA, however, were transported to the water table within a few years of the initial application of metolachlor to the land surface. Simulations indicate that after 20 yr of biannual application of metolachlor, the concentrations of MESA and MOXA increased notably in the uppermost 2 to 4 m of the unsaturated zone and were transported to the water table. The amount of degradate transported to depth in the unsaturated zone was related to the degradation model that was used and the simulated duration of the biannual application. A larger quantity of the degradates is transported to the water table if metolachlor degradation of MESA and MOXA occurs only in the aerobic part of the unsaturated zone.


   ACKNOWLEDGMENTS
 
The authors thank Lajpat R. Ahuja, James C. Ascough II, and Liwang Ma of the U.S. Department of Agriculture/Agricultural Research Service/Northern Plains Area/Great Plains System Research Unit, Ft. Collins, CO.; Kenneth W. Rojas of the U.S. Department of Agriculture/National Research Conservation Service/Information Technology Center, Ft. Collins, CO; and Donald R. Wauchope, University of Georgia Coastal Plain/U.S. Department of Agriculture/Agricultural Research Service, Tifton, GA, for maintaining and assisting with the Root Zone Water Quality Model. The authors also thank Dr. David Legates, University of Delaware, Newark, DE; Dr. C. Susan B. Grimmond, Indiana University, Bloomington, IN; and Dr. Gordon Heisler, U.S. Department of Agriculture Forest Service, Syracuse, NY for sharing their weather data.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
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.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Methods
 Results and Discussion
 Conclusions
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
P. D. Capel, K. A. McCarthy, and J. E. Barbash
National, Holistic, Watershed-Scale Approach to Understand the Sources, Transport, and Fate of Agricultural Chemicals
J. Environ. Qual., May 1, 2008; 37(3): 983 - 993.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
T. C. Hancock, M. W. Sandstrom, J. R. Vogel, R. M.T. Webb, E. R. Bayless, and J. E. Barbash
Pesticide Fate and Transport throughout Unsaturated Zones in Five Agricultural Settings, USA
J. Environ. Qual., May 1, 2008; 37(3): 1086 - 1100.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
G. V. Steele, H. M. Johnson, M. W. Sandstrom, P. D. Capel, and J. E. Barbash
Occurrence and Fate of Pesticides in Four Contrasting Agricultural Settings in the United States
J. Environ. Qual., May 1, 2008; 37(3): 1116 - 1132.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
R. M. T. Webb, M. E. Wieczorek, B. T. Nolan, T. C. Hancock, M. W. Sandstrom, J. E. Barbash, E. R. Bayless, R. W. Healy, and J. Linard
Variations in Pesticide Leaching Related to Land Use, Pesticide Properties, and Unsaturated Zone Thickness
J. Environ. Qual., May 1, 2008; 37(3): 1145 - 1157.
[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 Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bayless, E. R.
Right arrow Articles by Lampe, D. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bayless, E. R.
Right arrow Articles by Lampe, D. C.
Agricola
Right arrow Articles by Bayless, E. R.
Right arrow Articles by Lampe, D. C.
Related Collections
Right arrow Best Management Practices
Right arrow Pesticides
Right arrow Vadose Zone Processes and Chemical Transport

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