Published online 27 October 2006
Published in Soil Sci Soc Am J 70:1991-1997 (2006)
DOI: 10.2136/sssaj2006.0091
© 2006 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
Soil Physics
Sorption of a Hydrophilic Pesticide
Effects of Soil Water Content
T. E. Ochsnera,*,
B. M. Stephensb,
W. C. Koskinena and
R. S. Kookanac
a USDA-ARS, Soil and Water Management Research Unit, St. Paul, MN 55108
b Dep. of Soil, Water, and Climate, Univ. of Minnesota, St. Paul, MN 55108
c CSIRO Land and Water, PMB 2, Glen Osmond, SA 5064, Australia
* Corresponding author (ochsner{at}umn.edu)
 |
ABSTRACT
|
|---|
Transport of pesticide to groundwater is governed in part by sorption of the pesticide to soil particles. Sorption may be dependent on soil moisture conditions, but limited data are available from which to elucidate the effect. Our objective was to determine the effect of soil water content on the sorption coefficient of a hydrophilic pesticide. Sorption of dicamba (3,6-dichloro-2-methoxybenzoic acid) was measured in three soils, each at two initial water contents. At low water contents (
0.05 kg kg–1), sorption coefficients were similar for all three soils, ranging from 0.01 L kg–1 for the loamy sand to 0.07 L kg–1 for the silty clay loam. At higher water contents (0.19–0.24 kg kg–1), the sorption coefficient for the loamy sand was unchanged, for the silt loam it was doubled, and for the silty clay loam it was increased almost sixfold. Multiple regression analysis revealed a strong linear relationship between the sorption coefficient and the product of soil water content and organic C content (r2 = 0.86). The number of dicamba sorption sites probably increases with soil organic C content, while the accessibility of these sites appears to increase with soil water content. This may be caused by the decreasing hydrophobicity of soil organic matter with increasing water content. The effects of water content on pesticide sorption require further research and may ultimately have implications for the methods used to determine sorption and for managing pesticide application.
|
INTRODUCTION
|
|---|
IF PESTICIDES LEACH into the groundwater, they have the potential to negatively impact human health and the environment. The leaching risk of a pesticide in soil is characterized primarily by the sorption coefficient (Kd). The sorption coefficient, also called the distribution coefficient, quantifies the distribution of a pesticide between the soil solid phase and the soil solution. Lower sorption coefficients correspond with greater potential for leaching. To correctly assess water quality risks, researchers must understand the factors that influence pesticide sorption coefficients.
Currently, the effect of soil water content on pesticide sorption is not well understood. This is a serious gap because pesticides are often applied on or near the soil surface where the water content varies dramatically. Results from several studies have indicated that sorption of some pesticides increases with soil water content. This phenomenon has been demonstrated for the herbicides atrazine [6-chloro-N-ethyl-N'-(1-methylethyl)-1,3,5-triazine-2,4-diamine], linuron [N'-(3,4-dichlorophenyl)-N-methoxy-N-methylurea], metsulfuron methyl (methyl 2-[[[[(4-methoxy-6-methyl-1,3,5-triazin-2-yl)amino]carbonyl]amino]sulfonyl]benzoate), and the fungicide triadimefon [1-(4-chlorophenoxy)-3,3-dimethyl-1-(1H-1,2,4-triazol-1-yl)-2-butanone] (Berglöf et al., 2000a, 2000b, 2003; Rochette and Koskinen, 1996). Additionally, Roy et al. (2000) reported significant water content effects on the sorption of five other fungicides, with the direction and magnitude of the effects depending on the properties of the chemicals and the sorption mechanisms. They found that sorption of the more hydrophilic compounds increased at higher water contents. Some preliminary data have also suggested a positive water content effect on dicamba sorption (Koskinen et al., 2006).
The causes of the observed water content effects on pesticide sorption are uncertain. Berglöf et al. (2000a) hypothesized that, at higher water contents, more soil solution is in contact with a larger surface area of soil particles, thereby facilitating greater access of the pesticide to sorption sites. Roy et al. (2000) proposed two possible mechanisms for the water content effect. First, higher water content may facilitate pesticide diffusion to sorption sites. Second, water content may affect the structure of humic substances, leading to more hydrophilic surfaces at higher water contents and thus greater sorption potential for hydrophilic pesticides.
The scarcity of studies on the effect of water content on pesticide sorption is due in part to a lack of appropriate experimental methods. Batch methods, which are frequently used to measure sorption coefficients, use soil/solution ratios that are much lower (i.e., much higher water contents) than those that occur in the field. These methods are generally unsuitable for examining the effect of water content on sorption. Roy et al. (2000) used a direct soil solution sampling method to examine the influence of soil water content on sorption. They used water contents of 0.26 and 0.47 kg kg–1, which are on the higher end of the range of water contents typically observed near the soil surface, and they indicated the need for new methods of measuring sorption at lower water contents. One possible method for measuring sorption at low water contents is the supercritical fluid extraction (SFE) technique. Several SFE studies have provided evidence that pesticide sorption increases with soil water content (e.g., Berglöf et al., 2003), but interpretation of the data is complicated by the fact that the efficiency of the extraction technique may decrease at higher water contents, resulting in an apparent increase in sorption.
The unsaturated transient flow method is another possible method for studying sorption at low water contents (Ahmad et al., 2005; Katou et al., 2001). In this technique, soil spiked with pesticide is packed into a horizontal column, which is then attached to a water source. During infiltration, the chemical is redistributed by piston-like displacement of the antecedent solution (Smiles and Philip, 1978). The sorption coefficient is determined from the distributions of water and pesticide along the column at the end of the experiment. The initial water content of the column can be varied to study the effects of water content on sorption. The transient flow method avoids the possible effect of water content on extraction efficiency, which complicates interpretation of the results from the SFE method. The transient flow method also facilitates sorption measurements at lower water contents than can be studied using the direct sampling method of Roy et al. (2000). For these reasons, we chose the transient flow method for this study.
In our experiments, we used dicamba, a pre- and postemergent broadleaf herbicide sold under the trade name Banvel. During 2002, approximately 11% of U.S. corn (Zea mays L.) was treated with dicamba at a rate of 0.22 kg a.i. ha–1, yielding a total of 512 Mg of active ingredient applied (National Agricultural Statistics Service, 2005). Dicamba is also used on residential lawns. Being a polar, weakly sorbing chemical, dicamba is prone to leaching; however, the existing data on dicamba portend little ecological impact. Caux et al. (1993) reviewed data from studies in Canada and reported that only 8% of surface water and 2% of groundwater samples contained detectable levels of dicamba. They also reported that the half-life of dicamba in soil would be <12 wk under typical conditions in Canada. We chose dicamba not for any environmental risk it poses but rather as a representative of hydrophilic pesticides in general. Our objective was to determine the effect of soil water content on the sorption coefficient of a hydrophilic pesticide.
|
MATERIALS AND METHODS
|
|---|
Dicamba (99% purity) was obtained from Chem Service (West Chester, PA). Uniformly ring-labeled 14C-dicamba (106 MBq mmol–1, radiochemical purity >99%) was purchased from Pathfinder Laboratories (St. Louis, MO). Dicamba has a solubility of 6.5 g L–1 (25°C) and a pKa of 1.87 (Tomlin, 1994).
We studied sorption of dicamba in Tifton loamy sand (fine-loamy, kaolinitic, thermic Plinthic Kandiudult), Waukegan silt loam (sandy-skeletal, mixed, superactive, mesic Typic Hapludoll), and Drummer silty clay loam (fine-silty, mixed, superactive, mesic Typic Endoaquoll). The soils were air dried, ground, and sieved to 2 mm before the experiments. Some basic properties of the soils are listed in Table 1. Values for organic C contents and soil pH were determined by Cecchi et al. (2004), particle size distributions were measured by the hydrometer method (Gee and Or, 2002), and specific surface area was measured by ethylene glycol monoethyl ether retention (Pennell, 2002).
The water retention curves for the three soils were determined using Tempe cells, pressure plates, and a WP4-T dewpoint potentiameter (Decagon Devices, Pullman, WA). The Tempe cells were used from saturation to –65 kPa, the pressure plate was used for the –100-, –200-, and –500-kPa readings, and the WP4-T was used below –1000 kPa. Triplicate samples were tested for each soil at each pressure. The resulting data were fitted using the van Genuchten (1980) soil water retention model:
 | [1] |
where 
is the mass water content (kg kg–1), h is the matric potential, r is the residual water content, s is saturated water content, 
and n are the empirical fitting coefficients, and m is equal to 1 – 1/n. We set r equal to the air-dry moisture content for each soil. The water retention parameters and the associated coefficient of determination (R2) for each soil are listed in Table 2.
The water retention curves were then used to select the initial conditions for the experiments. First, experiments were conducted with initial water contents for each soil of 0.05 kg kg–1 but differing matric potentials. For the loamy sand this water content corresponded to a matric potential around –100 kPa on the measured retention curve. Therefore, experiments were then conducted with the silt loam and silty clay loam at initial water contents associated with matric potentials around –100 kPa on the measured water retention curves. For the silty clay loam, this water content was 0.23 kg kg–1. A similar initial water content (0.24 kg kg–1) was thus chosen for the final experiments with the loamy sand. Details of the initial conditions are listed in Table 3. The initial matric potentials in Table 3 are the matric potentials associated with the specified water contents on the initial drainage curve of the water content–matric potential relationship (Dane and Hopmans, 2002).
View this table:
[in this window]
[in a new window]
|
Table 3. Dicamba sorption coefficients (Kd) determined by the unsaturated transient flow method for three soils each at two different initial water contents and matric potentials. Shown are the mean (with standard deviation in parentheses) of three replicates.
|
|
Soils were moistened to the desired initial water content using a 5 mM CaCl2 solution containing 14C-labeled dicamba; the final dicamba concentration was 1 mg kg–1. The solution was applied to the soil using a spray bottle for lower initial water contents and dropwise (using a syringe needle) for higher initial water contents. While applying the solution, the soil was thoroughly mixed. After wetting, the soil was refrigerated for 24 h at 4°C. Previous research with similar soils showed that sorption equilibrium for dicamba was attained in <24 h and that no dicamba degradation occurred for incubation times <2 d (Menasseri et al., 2003).
The soil was then packed into an acrylic column consisting of 20 square sections (0.9 cm thick), each with a 2.5-cm-diameter hole in the center for the soil. After packing, the column was positioned horizontally, the inlet was connected to a Marriotte bottle permitting infiltration, and the other end was kept open to the atmosphere. The bottle supplied 5 mM CaCl2 solution (without dicamba) to the column inlet at a constant pressure throughout each run. Between runs, the inlet pressure was adjusted by raising or lowering the bottle. The inlet pressures were between –10 and 10 cm of water and were selected to provide slow flow rates and acceptable run times. Run times varied between 0.5 and 4 h, and flow rates varied between 0.7 and 3.4 cm h–1 depending on soil texture, initial water content, and inlet pressure.
After the wetting front reached three quarters of the way through the column, infiltration was halted and the individual sections were immediately separated. The total volume of water that infiltrated the column was recorded. Triplicate columns were evaluated at each initial water content for each soil, except for the silt loam at an initial water content of 0.19 kg kg–1 for which two columns were evaluated. The bulk densities (
b) for the packed columns were 1.22 ± 0.04 Mg m–3 for the loamy sand, 1.03 ± 0.06 Mg m–3 for the silt loam, and 0.99 ± 0.09 Mg m–3 for the silty clay loam. These are the same bulk densities to which the Tempe cells were packed for determining the water retention curves.
At the end of every run, 1.45 g of soil from each column section was saved for oxidation and the remaining soil was weighed, oven dried at 105°C for 24 h, and reweighed to determine the mass water content. To plot the water content distributions, the column bulk density was used to convert the mass water content of each section to volumetric water content (v). Triplicate 0.33-g soil samples from each column section were oxidized using a sample oxidizer (Model 307, Packard Instrument Co., Downers Grove, IL). The volatilized CO2 was captured in 20-mL scintillation vials containing Carbosorb E and Permafluor E+ (Packard Instrument Co.). Efficiency of oxidation and trapping (97%) was routinely determined by combusting known aliquots of 14C-glucose solution applied to cellulose powder.
The amount of 14CO2 obtained from each column section was quantified using a liquid scintillation counter (Model 1500 Tri-Carb, Packard Instrument Co.). Samples were counted until the counts per minute (CPM) counting error was <2% of the CPM value. The measured radioactivity in CPM was converted to disintegrations per minute (DPM) using the external standard ratio method with quench standards to correct for quenching. The DPM values were converted to micrograms of dicamba using the specific activity of the radio-labeled dicamba. Appropriate corrections were included so that the resulting concentrations were on a per unit mass of dry soil basis. In two of the experiments, we analyzed soil that was left out of the column. The resulting concentrations were within 2% of the concentrations measured in the sections near the outlet of those columns after the experiment, indicating negligible sorption of dicamba to the column.
Sorption coefficient calculations were basically as described by Ahmad et al. (2005). In brief, water flowing into the horizontal columns displaced the antecedent solution, creating a plane of separation. The pesticide concentration beyond this plane consisted of the pesticide initially present in that region plus the pesticide from the antecedent solution behind the plane. The location of the plane of separation at the end of an experiment was designated as x*. For positions beyond x*, the total pesticide content (sorbed and in solution) per unit mass of dry soil, M (mg kg–1), was plotted against the solution volume per unit mass of dry soil, v/b (m3 Mg–1). This produced a linear relationship of the form
 | [2] |
where Q is the sorbed concentration (mg kg–1) and C is the solution concentration (g m–3). Linear regression was used to determine Q and C. The sorption coefficient is simply the ratio Q/C.
In theory, x* is determined based on the volume of water that infiltrated the column and the final water content distribution within the column. In practice, x* was determined by finding the maximum in the pesticide concentration distribution and setting x* just behind the location of that maximum. This produced better regression statistics and more consistent sorption coefficients than determining x* based on the water content distribution. Similarly, Katou et al. (2001) found it necessary to adjust x* to avoid the effects of solute dispersion near the theoretical location of the plane of separation.
At an initial water content of 0.24 kg kg–1, the loamy sand was near saturation, and the unsaturated transient flow method was not feasible. Therefore, we implemented an alternate method. The soil was prepared as before, but in this case the column was positioned vertically rather than horizontally. A 0.22-µm nylon transfer membrane (part no. 1213410, Osmonics, Minnetonka, MN) was placed at the bottom of the column, a rubber gasket fit to surround the membrane, and a wire mesh placed between the membrane and the outlet. A constant suction of 50 kPa was then applied to the outlet. Runs were terminated when a predetermined amount of solution had been sucked into an attached graduated cylinder. Calculations were similar to those for the horizontal columns.
|
RESULTS
|
|---|
Figure 1
shows typical water content and pesticide concentration profiles for the unsaturated transient flow method. The solid vertical lines represent the location of the plane of separation, x*. The plane of separation lagged behind the wetting front as both antecedent water and solute were pushed ahead by the infiltrating water. Both inlet pressure and initial water content influenced the water content and pesticide distributions (data not shown). Lower inlet pressures resulted in wetting fronts and concentration peaks that were less sharp than those for higher inlet pressures. The wetting fronts and concentration peaks also became more diffuse as the initial water content increased. These effects are consistent with theoretical predictions for horizontal infiltration (Philip, 1957, 1958).
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 1. Water content and pesticide distributions (sum of pesticide sorbed and in solution) after infiltration for loamy sand, silt loam, and silty clay loam at initial water contents of 0.05, 0.19, and 0.06 kg kg–1, respectively. The solid line denotes the location of the plane of separation.
|
|
Figure 2
shows typical data and regression lines for total pesticide content plotted vs. solution volume for positions ahead of the plane of separation. At low initial water contents, little antecedent solution was present to be pushed ahead of the infiltrating solution, so the plane of separation was not far behind the wetting front. The result is a cluster of data points at a low solution volume and relatively few points with higher solution volumes. At higher initial water contents, the distance between the plane of separation and the wetting front was greater. This led to a more uniform spread of data points between the maximum and minimum solution volumes and potentially led to better regression statistics. Counteracting this advantage at higher initial water contents is the fact that displacement of the antecedent solution can become nonuniform, possibly due to air entrapment.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 2. Total pesticide concentration vs. solution volume for loamy sand, silt loam, and silty clay loam at initial water contents of 0.05, 0.19, and 0.06 kg kg–1, respectively. Only data from positions ahead of the plane of separation are included. Kd = sorption coefficient.
|
|
The mean sorption coefficients from three replicates for each soil at two different initial water contents are listed in Table 3. Across all three soils, the sorption coefficients range from 0.00 to 0.40 L kg–1. These relatively low values indicate that the chemical is quite mobile in soil, as expected given dicamba's high solubility. Previous studies have reported sorption coefficients for dicamba from 0.01 to 0.52 L kg–1 (Johnson and Sims, 1998; Menasseri et al., 2003; Oliveira et al., 2001), a range that is consistent with our results.
The data in Table 3 indicate that the sorption coefficient of dicamba generally increased with increasing initial water content. The sensitivity of the sorption coefficient to the initial water content depended on the soil. The loamy sand showed no change in sorption as initial water content increased, the silt loam exhibited a doubling of sorption (from 0.04 to 0.08 L kg–1), and the silty clay loam showed the greatest change (from 0.07 to 0.40 L kg–1).
Statistical Analysis
A multiple regression model was constructed to identify variables with statistically significant effects on the sorption of dicamba in these soils. Three primary variables were considered: soil organic C content, initial water content, and initial matric potential. Cross-product terms of these three variables were also included in the model to identify significant interactions. The partial slopes for the three primary variables and the three cross-product terms were determined by least squares fitting of the model to the sorption coefficients from 17 soil columns. The results are presented in Table 4. This model resulted in an R2 of 0.90 and was significant at the p < 0.001 level; however, only one of the partial slopes, the one for the cross-product of organic C content and initial water content, was significant based on the 95% confidence intervals.
View this table:
[in this window]
[in a new window]
|
Table 4. Parameter estimates, 95% confidence intervals, coefficients of determination (R2 for the full model or r2 for the reduced model), and F statistics for the full and reduced regression models. The independent variables are organic C content (OC), initial water content (g), matric potential ( m), and their cross-products. The dependent variable is the dicamba sorption coefficient (L kg–1).
|
|
In light of this finding, a reduced model was tested. This reduced model was a simple linear regression model consisting of only the organic C content–initial water content cross-product and an intercept. The reduced model gave an r2 of 0.86 and was significant at the p < 0.001 level (Table 4). An F test failed to reject the null hypothesis that the variables omitted from the reduced model had no predictive value (F = 0.75 < Fcrit = 3.33 at p < 0.05). This somewhat surprising result implies that organic C content alone was not a significant predictor for the observed variability in sorption coefficients; however, the product of organic C content and initial water content was strongly and linearly related to sorption. The F test also indicated that no significant relationship existed between sorption and matric potential or between sorption and the product of organic C content and matric potential.
For these soils, strong linear relationships exist between organic C content, clay content, and specific surface (Pearson correlation coefficients > 0.98). As a result, only one of these three variables could be included in the multiple regression model to avoid collinearity. The multiple regression was repeated with first clay content and then specific surface substituted for organic C content. The results were largely unaffected. Only the interaction of the chosen soil property and the initial water content was significant. Organic C content resulted in better regression statistics than clay content or specific surface (data not shown), and it has been previously identified as being strongly related to dicamba sorption (Oliveira et al., 2001), so it was retained for the reduced model.
|
DISCUSSION
|
|---|
Interpretation
Two requirements for solute sorption are that sorption sites exist and that they are accessible to the solute. We hypothesize that the number of dicamba sorption sites increases with the organic C content and that the accessibility to those sorption sites increases with water content. Previous studies with the batch method have shown that the dicamba sorption coefficient correlates significantly with soil organic C (Oliveira et al., 2001), so it is reasonable to hypothesize that the number of sorption sites for dicamba is positively correlated with organic C content. But that is only half of the story. The wettability of soil organic matter is a function of soil water content. Soil organic matter tends to be hydrophobic at low soil water contents and transitions to a more hydrophilic state at higher water contents (Ellerbrock et al., 2005; Michel et al., 2001). Thus at low water contents the soil solution will have limited interaction with sorption sites in the soil organic matter. As water content increases, the organic matter becomes more hydrophilic and the accessibility of the sorption sites increases. This is similar to the hypothesis proposed by Roy et al. (2000) to explain the effect of water content on the sorption of the herbicide diuron [(2,4-dichlorophenoxy)acetic acid].
This interpretation is consistent with the data from these experiments. The silty clay loam had the greatest organic C content and presumably the largest number of sorption sites. Increasing the initial water content from 0.06 to 0.23 kg kg–1 greatly increased the accessibility of sorption sites, as indicated by the marked increase in the sorption coefficient. In contrast, the loamy sand had low organic C content and presumably few sorption sites. In this case, sorption was limited by the number of sorption sites, not the accessibility of the sites; therefore, increasing the initial water content did not increase the sorption coefficient. The silt loam soil probably possessed an intermediate number of sorption sites, and the sensitivity of sorption to initial water content was also intermediate.
Is there evidence in the literature to support the effect of water content shown in our experiments? Johnson and Sims (1998) measured dicamba sorption coefficients for six soils using the batch method and using the soil thin-layer chromatography method in which the pesticide is applied as a spot onto air-dry soil. Across all six soils, the average sorption coefficients were 0.25 L kg–1 for the batch method and 0.03 L kg–1 for the thin-layer chromatography method. These data provide some additional evidence that dicamba sorption increases with water content.
Comments on the Methods Used
The unsaturated transient flow method is useful for quantifying the sorption of mobile agrochemicals at water contents representative of field conditions. Many of the other methods used to evaluate sorption are not able to accurately determine sorption at low water contents. This is a serious limitation when sorption depends on water content. However, the unsaturated transient flow method also has some limitations. One problem is determining sorption at high initial water contents. At initial water contents corresponding to roughly 50% saturation or greater, it is difficult to attain an accurate measurement of sorption. The water flow becomes less uniform, the r2 values for the linear regression decline, and the estimates of the sorption coefficients become less certain.
The vertical column outflow method was used to overcome some of the problems associated with higher initial water contents. This method was applied with the loamy sand at an initial water content of 0.24 kg kg–1. Since no solution infiltrates the column during the experiment with this method, uniform water flow is not essential. The only requirement is that a sufficient water content gradient is generated to facilitate the regression analysis. With this method, the water content and pesticide distributions displayed peaks about 13 cm from the outlet (Fig. 3
). Linear regression on the data before the peak in the pesticide distribution permitted determination of the sorption coefficient by Eq. [2] (Fig. 4
). The r2 values were 0.33, 0.73, and 0.84 for three replicate columns of the loamy sand at an initial water content of 0.24 kg kg–1. The resulting sorption coefficients were essentially zero (Table 3), in agreement with those measured at an initial water content of 0.05 kg kg–1. The vertical column outflow method yielded reasonable sorption coefficients for this experiment, but a thorough test of the method is beyond the scope of this research.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3. Water content and pesticide distributions (sum of pesticide sorbed and in solution) after vertical outflow for loamy sand at an initial water content of 0.24 kg kg–1. Data to the left of the solid line were used to estimate the sorption coefficient.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4. Total pesticide concentration vs. solution volume for loamy sand at an initial water content of 0.24 kg kg–1. Only data from positions to the left of the solid in Fig. 3 are included. Kd = sorption coefficient.
|
|
|
CONCLUSIONS
|
|---|
In summary, sorption of the hydrophilic pesticide dicamba increased as soil water content increased; however, the effect of water content was not uniform across different soil textures. In the silty clay loam, the influence of water content on sorption was dramatic: increasing the initial water content from 0.06 to 0.23 kg kg–1 increased the sorption coefficient almost sixfold. The increase in sorption was less for the silt loam, and there was no increase in sorption for the loamy sand as water content increased. The greater the soil organic C content, the greater was the sensitivity of sorption to water content.
Two interesting questions arise from consideration of this data. First, do analytical methods that use low soil/solution ratios (i.e., unrealistically high soil water contents) overestimate sorption of hydrophilic pesticides in soils with appreciable organic C content? The data from this study suggest an affirmative answer, but clearly more research is needed. Second, in soils with appreciable organic C content, can leaching of hydrophilic pesticides be reduced by applying them when the soil surface is moist? The fate and transport of these compounds is influenced by the interplay of many processes besides sorption, so these simple laboratory results cannot be extrapolated to predict what would occur in the field. However, these results do suggest some intriguing possibilities for in situ testing of management practices to reduce pesticide leaching and protect groundwater resources.
|
ACKNOWLEDGMENTS
|
|---|
We thank Todd Schumacher, USDA-ARS, St. Paul, MN, for his assistance with the experiments. Dr. Rai Kookana appreciates the interactions on this topic with Dr. Hidetaka Katou, National Institute for Agro-Environmental Sciences, Tsukuba, Japan, and Dr. Riaz Ahmad, Chiba University, Japan.
Received for publication February 23, 2006.
|
REFERENCES
|
|---|
- Ahmad, R., H. Katou, and R.S. Kookana. 2005. Measuring sorption of hydrophilic organic compounds in soils by an unsaturated transient flow method. J. Environ. Qual. 34:1045–1054.[Abstract/Free Full Text]
- Berglöf, T., W.C. Koskinen, J. Brücher, and H. Kylin. 2000a. Linuron sorption–desorption in field-moist soils. J. Agric. Food Chem. 48:3718–3721.[CrossRef][ISI][Medline]
- Berglöf, T., W.C. Koskinen, M.J. Duffy, K.A. Norberg, and H. Kylin. 2003. Metsulfuron methyl sorption–desorption in field-moist soils. J. Agric. Food Chem. 51:3598–3603.[CrossRef][ISI][Medline]
- Berglöf, T., W.C. Koskinen, H. Kylin, and T.B. Moorman. 2000b. Characterization of triadimefon sorption using supercritical fluid extraction and accelerated solvent extraction techniques. Pest Manage. Sci. 56:927–931.[CrossRef]
- Caux, P.Y., R.A. Kent, M. Tache, C. Grande, G.T. Fan, and D.D. MacDonald. 1993. Environmental fate and effects of dicamba: A Canadian perspective. Rev. Environ. Contam. Toxicol. 133:1–58.[Medline]
- Cecchi, A.M., W.C. Koskinen, and H.H. Cheng. 2004. Sorption–desorption of phenolic acids as affected by soil properties. Biol. Fertil. Soils 39:235–242.[CrossRef]
- Dane, J.H., and J.W. Hopmans. 2002. Water Retention and Storage: Laboratory, Introduction. p. 675–680. In J.H. Dane and G.C. Topp (ed.) Methods of soil analysis. Part 4. SSSA Book Ser. 5. SSSA, Madison, WI.
- Ellerbrock, R.H., H.H. Gerke, J. Bachmann, and M.-O. Goebel. 2005. Composition of organic matter fractions for explaining wettability of three forest soils. Soil Sci. Soc. Am. J. 69:57–66.[Abstract/Free Full Text]
- Gee, G.W., and D. Or. 2002. Particle-size analysis. p. 255–293. In J.H. Dane and G.C. Topp (ed.) Methods of soil analysis. Part 4. SSSA Book Ser. 5. SSSA, Madison, WI.
- Johnson, R.M., and J.T. Sims. 1998. Sorption of atrazine and dicamba in Delaware Coastal Plain soils: A comparison of soil thin layer and batch equilibrium results. Pestic. Sci. 54:91–98.[CrossRef]
- Katou, H., K. Uchimura, and B.E. Clothier. 2001. An unsaturated transient flow method for determining solute adsorption by variable-charge soils. Soil Sci. Soc. Am. J. 65:283–290.[Abstract/Free Full Text]
- Koskinen, W.C., T.E. Ochsner, B.M. Stephens, and R.S. Kookana. 2006. Sorption of isoxaflutole diketonitrile degradate (DKN) and dicamba in unsaturated soil. J. Environ. Sci. Health B (in press).
- Menasseri, S., W.C. Koskinen, and P.Y. Yen. 2003. Sorption of aged dicamba residues in soil. Pest Manage. Sci. 60:276–304.
- Michel, J.-C., L.-M. Riviere, and M.-N. Bellon-Fontaine. 2001. Measurement of the wettability of organic materials in relation to water content by the capillary rise method. Eur. J. Soil Sci. 52:459–467.[CrossRef]
- National Agricultural Statistics Service. 2005. Agricultural chemical use database. Available at pestmanagement.info/nass/ (verified 15 Aug. 2006). NASS, Washington, DC.
- Oliveira, R.S., W.C. Koskinen, and F.A. Ferreira. 2001. Sorption and leaching potential of herbicides on Brazilian soils. Weed Res. 41:97–110.[CrossRef]
- Pennell, K.D. 2002. Specific surface area. p. 295–315. In J.H. Dane and G.C. Topp (ed.) Methods of soil analysis. Part 4. SSSA Book Ser. 5. SSSA, Madison, WI.
- Philip, J.R. 1957. The theory of infiltration: 5. The influence of the initial moisture content. Soil Sci. 84:329–339.
- Philip, J.R. 1958. The theory of infiltration: 6. Effect of water depth over soil. Soil Sci. 85:278–286.
- Rochette, E.A., and W.C. Koskinen. 1996. Supercritical carbon dioxide for determining atrazine sorption by field-moist soils. Soil Sci. Soc. Am. J. 60:453–460.[ISI]
- Roy, C., P. Gaillardon, and F. Montfort. 2000. The effect of soil moisture content on the sorption of five sterol biosynthesis inhibiting fungicides as a function of their physicochemical properties. Pest Manage. Sci. 56:795–803.[CrossRef]
- Smiles, D.E., and J.R. Philip. 1978. Solute transport during absorption of water by soil: Laboratory studies and their practical implications. Soil Sci. Soc. Am. J. 42:537–544.[Abstract/Free Full Text]
- Tomlin, C. 1994. Dicamba. p. 298–299. In C. Tomlin (ed.) The pesticide manual. 10th ed. Br. Crop Protection Counc., Farnham, Surrey, UK.
- van Genuchten, M.Th. 1980. A closed-form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Sci. Soc. Am. J. 44:892–898.[ISI]
TOP
ABSTRACT
INTRODUCTION
