| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a AgResearch, Hamilton, New Zealand
b Currently with Environmental & Turf Services, Inc., 11141 Georgia Ave. #208, Wheaton, MD 20902
c HortResearch, Ruakura Research Centre, Hamilton, New Zealand
* Corresponding author (qinglima{at}aol.com).
Received for publication May 1, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: DT50 and DT90, times for 50 and 90% of the initial residues to dissipate, respectively
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Soil properties, as well as pesticide concentration, have been reported to significantly influence pesticide fate and efficacy. In studying atrazine dissipation in laboratory flasks and in the field, Gan et al. (1996) found that atrazine dissipated faster in a clay loam soil (fine loamy, mixed, mesic Typic Haplaquoll) than in a sandy loam soil (sandy, mixed, mesic Typic Hapludoll). They postulated that higher organic carbon content in the clay loam soil (4.09%) than the sandy loam soil (2.49%) increased the microbial populations and activities, thus increasing the degradation rate of atrazine in the clay loam soil. Rahman et al. (1978) and Rahman and Matthews (1979) also reported that more herbicides were required for 80% of weed control in soils with high organic matter content (>19%) than those with relatively low organic matter content (<8%). In examining various soil factors, pesticide concentrations, and soil amendments on degradation of pesticide mixtures, Schoen and Winterlin (1987) reported that pesticide concentration was the single most important factor influencing pesticide persistence in soils.
The persistence of pesticides in soils is often indicated by DT50, the time for 50% of initial residues to dissipate. When dissipation of a pesticide in soils can be adequately described by first-order kinetics, DT50 is equal to the first-order dissipation half-life, which is the time required for 50% dissipation of residual concentration at any given time. However, when the dissipation cannot be adequately described by first-order kinetics, the persistence of the pesticide may not be sufficiently represented by a single half-life or a single DT50 as the dissipation may depend on concentration, or it may have different rates at different times. Then, more complex models and indices are needed to adequately describe pesticide dissipation and persistence (Grover et al., 1997; Hill and Schaalje, 1985; Reyes and Zimdahl, 1989; Wolt, 1997).
Adequacy of model description can be subjective as it depends on the study objectives and error-tolerance levels at which the model fit can be accepted. For example, the coefficient of determination (r2) is often used to determine the goodness-of-fit (GoF) of first-order kinetics to describe pesticide dissipation in soils. One may accept an r2 value of 
0.8 as adequate; another may accept an r2 of 0.9 as adequate. Lack of a consistent standard for GoF has led to adoption of multiple criteria for GoF evaluations (Haan, 1977; Loague and Green, 1991; Nash and Sutchiffe, 1970).
The Hamilton clay loam and the Horotiu sandy loam are two major soils in the Waikato region of New Zealand. Acetochlor has been routinely used in this region to control annual broad-leaved weeds in maize (Zea mays L.). As part of a study to develop strategies for maximizing herbicide use efficacy while minimizing its environmental impacts, we determined the persistence of acetochlor in the field as affected by soil properties and application rates. We also examined the performance of five dissipation models for predicting pesticide persistence in the field soils.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Duplicate soil cores were taken to a depth of 10 cm from each plot using a 7.5-cm-diameter stainless-steel soil sampling tube, on the day of treatment and at 7, 14, 21, and 28 d after the treatment. At 41, 55, 84, 117, 147, and 196 d after the treatment, three soil cores were taken from each plot to a depth of 100 cm using a specially designed Humax soil sampler (Max Hug, Luzern, Switzerland). The sampler consists of a 100-cm-long stainless-steel outer tube with a 25-cm-long inner tube holding a PVC casing in which the soil sample is collected. Sampling to a 100-cm depth is achieved by taking four consecutive 25-cm segments. The soil samples were immediately frozen in their casings and then cut into 5-cm sections. In preparation for analysis the samples were partly thawed and the outside 2 to 3 mm of the sample removed and discarded to minimize cross-contamination. The remaining sample was thawed, bulked with other samples from the same depth collected from the same plot, thoroughly mixed, and passed through a 4-mm sieve. A portion of the sample was used for acetochlor residue analysis.
Laboratory Experiment
The laboratory studies were conducted to help identify major factors affecting acetochlor dissipation and explain acetochlor dissipation observed in the field.
Incubation Study
An aqueous solution (1.0 mL) of the formulated product (Roustabout) was fortified in the Horotiu soil (equivalent to 50.0 g dry wt.) in flasks (250 mL) to make it equivalent to 5.0 kg a.i. ha–1. The flasks were maintained at 10, 22, and 30°C at 60% of maximum water holding capacity (MWHC) and at 40 and 80% of MWHC at 22°C. Water was added once a week to bring the flasks up to the predetermined weight. Two flasks were taken from each temperature regime at designated times over a 112-d period for analysis. An incubation study was not conducted for the Hamilton soil.
Equilibrium Adsorption Measurements
A modification of the standard procedure of the Organisation for Economic Co-Operation and Development (1990) was used for determining acetochlor equilibrium adsorption constant in each soil. Aqueous solution (4.0 mL) of Roustabout (equivalent to 5.0 kg a.i. ha–1) was added to the moist soil (equivalent to 50.0 g dry wt.) in flasks (250 mL). The flasks were sealed with parafilm and stored at 4°C in dark for 24 h. Soil in a set of four flasks was extracted, each with 100.0 mL of methanol and water (70/30, v/v), for 1 h on an orbital shaker (230 rpm) to obtain the "total" residues. Soil in another set of four flasks was extracted, each with 100.0 mL of 0.2 M CaCl2 solution, to obtain the "aqueous" residues. The adsorbed residues were assumed to be "total" residues minus "aqueous" residues. The equilibrium adsorption coefficient was calculated as the ratio of adsorbed concentration to aqueous concentration.
Chemical Extraction and Analysis
All soil samples for acetochlor analyses (equivalent to 50.0 g dry wt.) were shaken with 100.0 mL of methanol and water solution (70:30, v/v) in flasks for 1 h and then allowed to stand for 1 h. An aliquot of the supernatant (10.0 mL) was extracted three times with dichloromethane (7.0, 3.5, 3.5 mL) and the combined extracts were slowly evaporated under N2 until dry. The extracts were redissolved in methanol (0.5 mL) and then made up to 1.0 mL with purified water. The redissolved extracts were analyzed by high performance liquid chromatography (HPLC) with a UV detector at 230 nm. The column was Prodigy (150 x 4.6 mm; SGE Int., Victoria, Australia) packed with 5-µm ODS(3) held at 35°C. The mobile phase was 56:44 methanol to water (v/v) at a flow rate of 1.0 mL min–1. This analytical procedure gave a recovery of 89 ± 14% for a range of spiked concentrations from 200 to 2000 µg kg–1. With a 50-µL injection volume, the analytical detection limit was 40 µg kg–1. The mean recovery was used to correct acetochlor concentrations in the field.
Measurements of Soil Properties
Fresh soil was taken from the topsoil horizon (0–15 cm) of each field for measuring soil water content, soil bulk density (
), soil particle size distribution (Day, 1965), and soil organic carbon (OC) content (Nelson and Sommers, 1982). The sand, clay, and OC contents and were respectively 29.0%, 31.0%, 4.6%, and 1.03 Mg m–3 for the Hamilton clay loam, and 58.0%, 17.0%, 8.7%, and 0.87 Mg m–3 for the Horotiu sandy loam. The soil pH (in 0.01 M of CaCl2 solution) was 5.4 for the Horotiu soil and 5.6 for the Hamilton soil.
Pesticide Dissipation Models
Model 1: Simple First-Order Kinetics
For this model, the equation is:
 | [1] |
Model 2: Quadratic Model
For this model, the equation is a combination of the first- and second-order equations:
 | [2] |
Model 3: First-Order Double-Exponential Model
For this model, the equation is:
 | [3] |
Model 4: First-Order Discontinuous Biphasic Model
For this model, the equation is:
 | [4] |
Dissipation occurs at different rates (k1 and k2) within two time domains, t1 and t2, each following first-order kinetics (Model 1). The rapid and slow dissipation phases (t1 and t2) were determined according to the slopes of the dissipation curves, and then the rate constants were obtained by separately fitting Model 1 to data for respective time domains. There might be an abrupt discontinuity in the dissipation rate at the phase boundary. Usually both DT50 and DT90 are used to indicate the persistence of pesticides.
Model 5: First-Order Two-Compartment Model
Hill and Schaalje (1985) proposed the following two-compartment model for describing the fast and slow dissipation of pesticides:
 | [5] |
 | [6] |
As for Models 3 and 4, an analytical solution cannot be simply derived from Eq. [6] for DT50; a numerical method or a fitting procedure has to be used to obtain DT50, as detailed below. This two-compartment model is similar to that proposed by Hamaker and Goring (1976).
Model 1 describes a log-linear decay of pesticides with time, whereas Models 2 through 5 describe log-nonlinear, fast and slow dissipation of pesticides with time. Model parameters were obtained from the best-fit models using Microsoft Excel-Solver (Microsoft, 2000), an add-ins analysis ToolPak of Microsoft Excel. The Solver uses the Generalized Reduced Gradient (GRG2) nonlinear optimization code and returns the best-fit parameters by minimizing the sum of squares of the residuals between measured and fitted values. An equal weight of 1.0 was assigned to every measured concentration in the optimization irrespective of its value. A log-transformation of measured and fitted concentrations was used to calculate the sum of squares of the residuals. Also, parameters were made squared in the optimization to ensure the minimizing routine in the right direction. All optimized model parameters were obtained using the Solver for consistency, even through the parameters for Model 1 can be easily obtained by linear regression of the log-transformed concentrations. The optimized parameters were C0 and k1 for Model 1; C0, k1, and k2 for Model 2; C10, C20, k1, and k2 for Models 3 and 4; and C0, k1, k2, and kr for Model 5. The initial guess for C0 was the measured initial concentration and those for C10 and C20 were half of the C0 value at each application rate. The initial guess for all rate constants was the first-order rate constant obtained by linear regression of log-transformed concentrations at each application rate. This optimization scheme returned stable numerical values for a range of initial guesses for each set of the optimized parameters. The DT50 and DT90 values were then obtained numerically from the best-fit models.
The effects of application rates and soil properties on dissipation of acetochlor in the field were examined by comparing acetochlor dissipation rates (DT50 and DT90 values) at two application rates in two types of soils. In these comparisons, the same time interval was used for all application rate–soil combinations. Visual graphical comparison, r2, and root mean square error (Loague and Green, 1991) were used to evaluate the goodness-of-fit. All statistical analyses were performed at the 0.05 significance level.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
Models 2 through 5 predicted essentially the same DT50 values at the low application rate (Table 1), 9.4 ± 0.8 d. This compares with a DT50 value of 20.3 d predicted by Model 1. Likewise, Models 2 through 5 predicted a mean DT90 value of 56.1 ± 0.4 d as compared with a DT90 value of 67.4 d predicted by Model 1 (Table 1). The question arises as to which DT50 and DT90 values should be reported. Statistically those obtained from Models 2 through 5 should be used because of smaller RMSE values (Table 1). The initial rapid and later slow dissipation nature simulated by Models 2 through 5 suggests that a single DT50 or DT90 value may not be able to represent the overall dissipation pattern. Therefore, both DT50 and DT90 values should be reported. At the high application rate, all five models predicted nearly identical DT50 (17.9 ± 1.9 d) values and DT90 (63.0 ± 0.9 d) values (Table 1).
Tests with artificially generated datasets using Model 1 demonstrate that all five models would give essentially the same DT50 value (and DT90 value) if Model 1 fits the data adequately, as judged by both r2 and RMSE. The implication of this observation is that the DT50 and DT90 values derived from Models 2 through 5 can serve as an additional condition to verify the adequacy of Model 1 description. For example, Model 1 predicted statistically the same DT50 value (and DT90 value) and RMSE as Models 2 through 5 at the high application rate (Table 1). Therefore, Model 1 adequately described acetochlor dissipation. On the other hand, Model 1 predicted a significantly different DT50 value (and DT90 value) and RMSE from Models 2 through 5 at the low application rate (Table 1). Thus, Model 1 did not adequately describe acetochlor dissipation as compared with Models 2 through 5, and Models 2 through 5 should be used instead.
The adequacy of Model 1 description at the high application rate can also be seen by comparing the fitted rate constants (Table 2). The fitted k1 for Model 1 was equal to that in the rapid dissipation pool of each of Models 2, 3, and 5 (Table 2). Moreover, Model 3 had the same rate constant in the rapid and slow dissipation pools, which is also equal to k1 of Model 1 (Table 2). Furthermore, the fitted initial concentration by Model 1 was the same as those by Model 2, Model 3 (sum of C10 and C20 because k1 = k2), and Model 5. All these suggest that Model 1 adequately described acetochlor dissipation.
|
Dissipation of Acetochlor in Horotiu Sandy Loam Soil
Fitting Model 1 to the measured concentrations resulted in predicted concentrations on the day of application that were approximately half of the measured concentrations at both application rates (Fig. 2)
. This compares with the maximum difference factors of 1.4 and 1.1 in the Hamilton soil at the low and high application rates, respectively (Fig. 1). Values of r2 and RMSE of the fits and the derived DT50 and DT90 for all models are given in Table 1. Relatively larger RMSE values for Model 1 (Table 1) further indicate that it did not adequately describe acetochlor dissipation in the Horotiu soil as Models 2 through 5.
|
Examinations of the DT50 and DT90 values in Table 1 show that differences exist in these benchmark values among the five models. Model 1 calculated approximately the same DT50 values at the low and high application rates, 13.6 and 13.9 d, respectively. Likewise, the calculated DT90 values were nearly identical (Table 1). However, these values are significantly greater than those calculated by Models 2 through 5 (Table 1). This was also observed for the Hamilton soil. Since Models 2 through 5 represented the field data better than Model 1 (Table 1), we conclude that Model 1 did not adequately describe acetochlor dissipation and that the benchmark values (DT50 and DT90) calculated by Model 1 overestimated acetochlor persistence in the soil.
Models 2 through 5 calculated nearly the same DT50 (5.1 ± 0.9 d) values and DT90 (28.8 ± 0.8 d) values at the low application rate, but significantly different DT50 values at the high application rate (Table 1). However, the mean DT50 (6.6 ± 2.8 d) values and DT90 (30.8 ± 1.7 d) values calculated by Models 2 through 5 at the high application rate were not significantly different from those at the low application rate. Thus, it appears that application rate did not significantly affect the DT50 and DT90 values in the Horotiu soil.
To further explore the influence of C0 fittings on DT50 and DT90 values, we applied the same fitting procedure to Models 1, 2, and 5, but set C0 equal to the measured initial concentration. We then compared the fitted parameters and derived indices (Table 3). Acetochlor dissipation data at the low application rate were used for the comparison because the biggest difference was observed between measured and fitted C0. Models 1, 2, and 5 were used because they are continuous functions and contain C0 explicitly. Analyses of the results in Table 3 revealed the same conclusion as obtained when C0 was considered as a fitting parameter, although the fitted parameter values were different.
|
Comparing Acetochlor Dissipation in Hamilton Clay Loam and Horotiu Sandy Loam
Because Model 1 did not adequately describe acetochlor dissipation in the Horotiu soil, Models 2 and 3 were selected to compare acetochlor persistence in these two soils. The DT50 values calculated by Model 2 at the low application rate were 10.6 and 4.8 d for the Hamilton soil and the Horotiu soil, respectively. Similarly, the calculated DT90 values were 56.6 and 27.8 d for the Hamilton soil and the Horotiu soil, respectively. These results suggest that the persistence of acetochlor was significantly influenced by soil properties.
Examining the soil properties for the topsoil horizons of both soils reveals that the most contrasting difference with respect to acetochlor dissipation is soil organic carbon content, which was 8.7% for the Horotiu soil and 4.6% for the Hamilton soil. Higher organic carbon content in the Horotiu soil would cause more acetochlor adsorption by the soil based on the concept proposed by Chiou et al. (1979). An equilibrium adsorption study conducted to verify this hypothesis showed that acetochlor equilibrium adsorption constant was 5.8 ± 1.8 L kg–1 for the Horotiu soil and 3.4 ± 0.6 L kg–1 for the Hamilton soil. Higher adsorption constant in the Horotiu soil resulted in more acetochlor retained in the soil surface layer. In contrast, acetochlor had a smaller adsorption constant in the Hamilton soil and thus moved relatively deeper into the soil profile.
Because more acetochlor was retained near the soil surface in the Horotiu soil, it had more intense and direct interactions with extreme temperature, solar radiation, and other soil surface dissipation processes such as volatilization and photodegradation, leading to more chemical and biological dissipation. As a result, the persistence of acetochlor in the Horotiu soil was shorter than that in the Hamilton soil. Sheath and Boom (1985) reported that the maximum temperature at a 0.5-cm depth in the Horotiu soil could reach 45 to 50°C in summer and decreased rapidly with depth. The recorded maximum temperature at a 10-cm depth was 25.8°C. Soil temperature at a 0.5-cm depth in the Hamilton soil was not measured. However, the recorded maximum temperature at a 10-cm depth was 23.2°C for the same period of time. Presumably, soil surface temperature in the Hamilton soil was also lower than that in the Horotiu soil.
A laboratory incubation study designed to examine the effects of temperature and soil moisture on acetochlor dissipation in the Horotiu soil demonstrated that acetochlor degradation was barely affected by soil moisture (Fig. 3a) but strongly dependent on temperature (Fig. 3b), with first-order degradation rate constants (k1) being 0.011, 0.049, and 0.091 d–1 at 60% maximum water holding capacity (MWHC) at 10, 22, and 30°C, respectively. These results support the hypothesis that higher temperature near the soil surface in the Horotiu soil caused more acetochlor dissipation. The low vapor pressure (3.4 x 10–8 mm Hg at 25°C) of acetochlor seems to suggest that volatilization is not a major dissipation pathway. However, photodegradation can occur in the field, especially under prolonged lack of rainfall when acetochlor remained near the soil surface. In particular, the similar molecule structure to metolachlor, a photosensitive herbicide, seems to suggest that photodegradation can be a loss pathway for acetochlor near the soil surface. Photochemical reactions generally do not follow Model 1, but show initial rapid and later slow losses (Miller et al., 1989). This seems to be consistent with the observed acetochlor loss in the Horotiu soil. We do not have measured data to support either hypothesis.
|
In describing field dissipation of trifluralin and ethalfluralin, Grover et al. (1997) and Wolt (1997) hypothesized a mechanism that initially readily available chemical resulted in apparent rapid dissipation while subsequent increased binding to soil caused noticeable reduction in dissipation rate. When comparing acetochlor dissipation in both field soils, we observed that acetochlor dissipated faster in the Horotiu soil than in the Hamilton soil, although more acetochlor was adsorbed in the Horotiu soil. Thus, it appears that adsorption did not significantly inhibit acetochlor dissipation in the field. Otherwise, acetochlor should have dissipated faster in the Hamilton soil than in the Horotiu soil. As discussed previously, rapid dissipation of acetochlor in the Horotiu soil is likely to be temperature-driven chemical and biological degradation in the topsoil layer.
Fitting Model 3 to the measured data showed that the dissipation rate constants in the rapid dissipation pool in the Horotiu soil were greater than those in the Hamilton soil at both application rates (Table 2). Calculations using Model 3 based on data in Table 2 revealed that readily available chemical in the rapid dissipation pool initially represented 55 and 60% of the applied acetochlor in the Hamilton soil and 70 and 96% in the Horotiu soil at the low and high application rates, respectively. These results suggest that the chemical in the rapid dissipation pool is not equivalent to the dissolved chemical, as Wolt (1997) proposed. Although chemical in soil solution is generally thought to be readily available to microorganisms for biodegradation, there is evidence that adsorption can accelerate chemical degradation (Valentine, 1986). Furthermore, measurements of chemical dissipation under field conditions are generally based on total chemical concentration in the soil. As such, the dissipation model is fitted to the total concentration. The resulting best-fit parameters (e.g., C10 and C20 in Table 2) should represent the total concentration. Therefore, it is inappropriate to interpret chemical concentration in the rapid dissipation pool (C10) as dissolved concentration and that in the slow dissipation pool (C20) as adsorbed concentration.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
First-order kinetics is too often fitted to the observed data and the goodness-of-fit is simply determined based on the magnitude of the coefficient of determination (r2). This study has demonstrated that high r2 alone does not guarantee the adequacy of the model description for pesticide dissipation. Adequacy of model description should be evaluated based on multiple criteria, especially when the model is used to derive a critical index such as DT50. It appears that Models 2 and 3 have advantages over Models 4 and 5 in applications, although all five models should be viewed as empirical in nature.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  Q. L. Ma, A. Rahman, T. K. James, P. T. Holland, D. E. McNaughton, K. W. Rojas, and L. R. Ahuja Modeling the Fate of Acetochlor and Terbuthylazine in the Field Using the Root Zone Water Quality Model Soil Sci. Soc. Am. J., September 1, 2004; 68(5): 1491 - 1500. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | |||
