| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dow AgroSciences, Bldg. 306/A2, 9330 Zionsville Road, Indianapolis, IN 46268
Corresponding author (iwesenbeeck{at}dowagro.com)
Received for publication March 13, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CN, runoff curve number • DEPI, depth of incorporation • K, soil erodibility parameter • P, cropping practice factor • PRZM, Pesticide Root Zone Model • SCS, Soil Conservation Service • SPE, solid-phase extraction • USLELS, slope factor
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Quantifying the runoff potential of diclosulam is important primarily to assess potential effects on sensitive nontarget terrestrial plants through irrigation with surface water. Diclosulam poses no mammalian or avian risks at typical expected environment concentrations.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Runoff Study
The runoff study was conducted on a Tifton loamy sand (fine-loamy, kaolinitic, thermic Plinthic Kandiudult) in Thomas County, GA (approx. 31.5°N, 83.5°W) according to the soil survey map. Actual characterization of the soil showed that the soil consisted of sandy loam from 0 to 15 cm and sandy clay in the deeper soil layers. The soil pH ranged from 4.7 to 6.2 over the soil profile. Soil organic matter ranged from 1.15% in the 0- to 15-cm layer to 0.16% in deeper layers. The simulated rainfall was applied to three approximately 15- x 45-m plots 1 d after the target broadcast application of 48 g ha-1 (combined maximum pre- + post-application rate) of diclosulam. The plots were oriented such that the long dimension was approximately parallel to the 3.5 to 4% slope of the field. Each plot was divided into three equal sections, which were further subdivided into 18 equal grids for the purposes of sampling. Water used to supply the rainfall simulator was obtained from a 1.3-ha pond approximately 45 m downslope from the test plots. The pond was protected by a grass buffer that extended approximately 15 m upslope, and by a diversion ditch downslope from the treated plots that directed runoff water around the irrigation pond.
Plot Preparation
Plots were constructed by excavating a 15-m-long, 1.2-m-wide, and 0.6-m-deep runoff collection area at the downslope end of each plot, and approximately perpendicular to the natural slope of the field. A retaining wall consisting of wooden boards, sealed to the soil using a 0.6-m concrete strip, was built along the upslope edge of the excavated area so that the upper edge of the wall was even with the original soil surface. A galvanized metal gutter was attached to the retaining wall on a 1.5% slope to transport the runoff to the inlet of a large, 60° V-trapezoidal flume fitted with a stilling well and sampling pan used for flow measurement and sampling. After sampling, runoff entered a pit from where it was pumped into the diversion ditch below the plot. Following application of the test material, each plot was hydrologically isolated from the larger field using a grading blade to create soil berms around the top and sides of each plot. The site was left fallow in early 1997 in preparation for the study. Since rainfall was to be simulated 1 d after application of diclosulam to the soil, the seedbed was prepared by discing to a depth of approximately 15 cm along the contour of the field, as well as additional cultivation to remove remaining weeds.
Rainfall Simulator
The rainfall simulator used in this study was developed by PTRL West (Louisville, KY) (Coody et al., 1990) to simulate the drop size distribution and intensity of natural rainfall. The simulator consisted of two laterals equipped with valves and pressure gauges to regulate water input. Water was distributed through pressure regulators (100 kPa output) that were equipped with Nelson Model S-30 irrigation heads and U4 spinner plates (Nelson Irrigation Corp., Walla Walla, WA). The irrigation heads were fitted with Nelson #40 brass nozzles and mounted on PVC risers at a height of approximately 3 m above the soil surface. The irrigation heads were uniformly spaced at 4.5-m intervals along the lateral. The laterals were placed along the edge of the test plot (12.6 m apart). Water for the rainfall simulation was drawn from the irrigation pond via a 7.6-cm-diam. PVC pipe and a 5-cm-diam. flexible hose. The simulator operating pressure was maintained at 150 to 170 kPa using two centrifugal pumps operating in tandem. Rainfall application uniformity was checked using four randomly placed cups in each of the three test plots.
Test Substance Application
Diclosulam was applied to the test plot as a broadcast spray at a target application rate of 48 g a.i. ha-1 in 165 L of spray mix per hectare. The material was applied using a tractor-mounted sprayer in two successive sprays to the plot, each applying one-half of the test material, to improve uniformity. The runoff collection areas, including gutters, concrete, flumes, and sampling electronics were covered with plastic sheeting before each application to minimize contamination. The actual amount of test material applied to the plots was measured using three randomly placed 15-cm-diam. filter pads and three pans (27 x 17.5 cm) filled with soil in each of the three test plots. The filter pads and soil pans were collected immediately after the application and transferred to plastic-coated jars and metal cans, respectively.
Weather Station
A weather station was installed at the site to measure rainfall, air temperature, soil temperature at 2.5 and 10 cm, relative humidity, solar radiation, and wind speed and direction. Rainfall was measured at 5-min intervals.
Runoff Monitoring
Runoff was measured using both time-paced and flow-proportional samples collected at the downslope edge of each test plot. Flow through the flume was continuously measured using a pressure transducer flow meter installed in the stilling well. Runoff flow rate was calculated by converting the water level in the stilling well using a factory equation stored in the flow meter. Water level measurements were stored in the flow meter's internal memory and downloaded to a laptop computer after each runoff event. Flow-proportional samples were taken using an ISCO (Lincoln, NE) 3700 pump sampler from a sampling pan mounted to the exit of the flume. A 1-L sampling was triggered by the flow meter after each 198 L of water passed through the flume. The sample was delivered directly to a stainless steel 200-L drum.
Time-proportional samples were taken from the same sampling pan using a second ISCO 3700 pump sampler. Aliquots were collected at 2-min intervals and delivered into glass jars located in the base of each sampler. Five consecutive aliquots were collected in a single sampling jar to represent 10 min of runoff.
Soil Sample Collection
Duplicate soil cores (0–35 cm) were collected from each of the three sampling sections within each plot just prior to the runoff event for water content analysis. Soil samples were also collected from three randomly selected locations within each of the three subplots to determine the levels of diclosulam at four times during the study: pre-application, post-application, pre-runoff, and post-runoff. Prior to the rainfall event, a single soil core was taken from each of the subplots to determine the diclosulam profile with depth. After the application of diclosulam but prior to the rainfall event, a single core (0–30 cm) was removed from each of the three subplots within each plot. After the simulated rainfall event, two 0- to 90-cm cores were collected from each of the three subplots within each plot. The cores were frozen and shipped to Dow Agro-Sciences where they were cut into 15-cm increments. Soil water content was also determined for these samples.
All soil and water samples, except for soil characterization cores, were placed into a field freezer or refrigerated storage at the site immediately after collection. Samples were transferred from the field freezer or refrigerator to the appropriate field laboratory cooling chamber and then shipped to Dow Agro-Sciences for analysis. All samples were analyzed within approximately 3 mo of sampling.
Analytical Methods
Soil and sediment samples were analyzed for residues of diclosulam as described by Batzer et al. (1997). Residues of diclosulam were extracted from soil or sediment using acidified acetone. An aqueous magnesium acetate solution was added to the extract to precipitate extracted soil matrix. The supernatant was transferred, the acetone evaporated under nitrogen, and the residue isolated using C18 solid-phase extraction (SPE). The eluant from the C18 SPE was evaporated to dryness and the residue reconstituted in ethyl acetate. The residue was further isolated using neutral alumina SPE. The eluant from the alumina SPE was evaporateed to dryness and the analytes derivatized with iodoethane to form ethyl derivatives. The analytes were partitioned between an aqueous sodium chloride solution and toluene containing the internal standard. The concentration of N-ethyl diclosulam was determined in the final solution by gas chromatography with mass selective detection (GC–MSD) with a limit of detection (LOD) of 0.3 ng g-1 and a limit of quantitation (LOQ) of 1.0 ng g-1.
Diclosulam residues in water were analyzed as described by Blakeslee et al. (1998). Aliquots of water were acidified and the residues isolated on a C18 SPE column. The eluant from the column was evaporated to dryness under nitrogen and the analytes derivatized with iodoethane to form ethyl derivatives. The analytes were partitioned between an aqueous sodium chloride solution and toluene containing the internal standard. The concentration of N-ethyl diclosulam was determined in the final solution by gas chromatography with mass selective detection (GC–MSD) with a limit of detection of 0.03 ng mL-1 and a limit of quantification of 0.10 ng mL-1.
Pesticide Root Zone Model Runoff Study Data
The PRZM model (Version 3.12) was used with measured agricultural, soil, chemical, and weather inputs to predict the edge-of-field water and sediment runoff volumes and the associated diclosulam runoff observed during the runoff study described above. The PRZM model is a one-dimensional, finite-difference model that simulates hydrology and chemical transport within and below the plant root zone (Carsel et al., 1985). Assumptions implicit in the PRZM model include downward only advective transport of soil water and first-order adsorption and degradation kinetics. The model takes into account mass loss due to degradation in the sorbed, dissolved, and vapor phase, and plant uptake of the dissolved phase. Evapotranspiration and runoff are also simulated. Limitations of the PRZM model include the inability to simulate macropore flow and tile drainage.
The most critical parameters in the PRZM model affecting sediment and water runoff are the runoff curve number (CN), soil erodibility factor (K), crop practice factor (P), and slope factor (USLELS). These parameters were estimated from actual field measurements, or using soil characterization properties and look-up tables in the PRZM2 manual. Chemical parameters for diclosulam were taken from laboratory studies. All simulations were conducted using an average half-life (T1/2) of 56 d and an average Koc of 83.9 L kg-1 (Yoder et al., 2000). For modeling purposes, diclosulam was assumed to be soil-applied with a 2-cm incorporation depth (although diclosulam was not incorporated in the field study to simulate worst-case conditions), with the diclosulam concentration linearly decreasing with depth (CAM = 6, DEPI = 2.0, where CAM is defined as the chemical application method and DEPI is the depth of incorporation). This represents a realistic situation where some of the diclosulam may have moved a few millimeters into the soil during the period shortly after rainfall starts but prior to runoff occurring. Model predictions were compared with actual field measurements of water, sediment, and diclosulam concentrations in water and sediment.
Extrapolation to Prevalent Soil in the Southeastern U.S. Peanut Market
The validated model was first used to simulate the edge-of-field runoff of diclosulam using the physical parameters of the field-study scenario under varying weather conditions by using 328 simulation years of real weather from Dublin, GA (NCDC Summary of the Day CD, Earthinfo, Boulder, CO), with diclosulam application timings distributed evenly between 1, 8, 15, and 22 April and 1, 8, 15, and 22 May, the most likely application window for diclosulam applications in the southeastern USA. The 41-yr weather record from Dublin, GA (1955–1995) had a mean annual rainfall of 1186 mm with average monthly rainfalls of 89 and 84 mm for April and May, respectively. The weather data contained 4 daily rainfall events >76 mm (approximately 1-in-10-yr return), 12 daily rainfall events between 51 and 76 mm, and 60 events between 25 and 51 mm in the period from 1 April to 30 May, when all the diclosulam applications were made. The highest single-day rainfall in the dataset was 142 mm. The 328 edge-of-field runoff amounts (expressed as percent of applied for each entire year) were used to generate a probability of exceedence plot for diclosulam runoff, against which the measured field runoff could be compared.
The PRZM model, using the same chemical inputs, was then used to extrapolate edge-of-field runoff loadings from "typical" Tifton and Bibb soil series, the two most prevalent soils in the southeastern peanut market, representing 305000 and 145000 ha of the southeastern peanut growing region, respectively. The Tifton soil is Hydrologic Group B with 81% sand, 5.5% clay, and 0.29% organic matter, while the Bibb soil (coarse-loamy, siliceous, active, acid, thermic Typic Fluvaquent) is Hydrologic Group D containing 55% sand, 10% clay, and 1.16% organic matter. The Bibb series is the only Hydrologic Group D soil among the top 10 most prevalent soils in the southeastern U.S. peanut market. Average soil properties for the Tifton and Bibb soil were taken from the Pesticide Assessment Tool for Rating Investigations of Transport (PATRIOT) database (USEPA, 1993). Worst-case agronomic practices (fallow conditions, no residue, no contour plowing) and a 10% slope, 100 m long (both more than two times greater than in the runoff study), were used in conjunction with 164 yr of real weather from Dublin, GA in the extrapolation simulations.
The simulated annual edge-of-field runoff loadings were used to generate a probability of exceedence plot for diclosulam runoff for these soils. Diclosulam mass in the sediment and water were added together for a single estimate of edge-of-field diclosulam loading. The model predictions of diclosulam associated with the sediment were several orders of magnitude lower than that observed in the water; therefore, adding the loadings does not introduce significant error. The results of the runoff study also indicated that edge-of-field diclosulam loading due to sediment runoff was several orders of magnitude lower than diclosulam loading associated with water runoff.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Diclosulam Application
Based on the calibrated sprayer delivery and the average concentration of diclosulam measured in the spray mix samples, approximately 95, 89, and 96% of the target rate of diclosulam was applied to Plots 1, 2, and 3, respectively. Average results of the filter papers and soil pans indicated that 92, 80, and 79% of the target rate of diclosulam reached the soil.
Measured Runoff
The rainfall simulator provided 77, 80, and 76 mm of water to Plots 1, 2, and 3, respectively, based on the average measurements of the 12 rain gauges placed in each plot. These data indicate that the average simulated rainfall inputs amounted to approximately 96% of a 1-in-5-yr, 2-h rainfall (81 mm) for the study area (Hershfield, 1961). The addition of 4 mm of natural rainfall during the simulation at Plot 2 resulted in it exceeding the 1-in-5-yr 2-h storm for the area by 5%.
Significant runoff volumes were generated in all the plots as a result of the simulated rainfall events (Table 2). Runoff hydrographs for each of the plots are shown in Fig. 1 . Table 2 shows the results of the discrete and composite sampling, as well as the average of the two methods. Mean water yield was 23.1, 22.7, and 29.9 mm, representing 29.9, 28.4, and 39.4% of the applied rainfall, respectively, for Plots 1 through 3. The higher water runoff from Plot 3 could be partially explained by the higher antecedent water content in that plot compared with the other two. A total of 3231.3, 2309.6, and 8336.5 kg ha-1 of sediment was lost from Plots 1 through 3, respectively. The significantly higher sediment loss from Plot 3 is partly explained by the higher water loss from the site. Diclosulam runoff in the water phase averaged 1790.7 mg ha-1, or 4.05% of the amount applied (average of discrete and flow proportional samples for all plots). Diclosulam associated with sediment averaged 53.9 mg ha-1, or 0.12% of the amount applied, for a total average diclosulam loss via water and sediment runoff of 1844.6 mg ha-1, or 4.2% of the applied material (Table 2). Diclosulam runoff losses as measured using the discrete and flow-proportional methods correspond quite closely.
|
|
Similar rainfall events occurring at later times, after degradation and sorption of diclosulam had occurred, would result in proportionally less edge-of-field runoff. Previous studies have shown the increasing sorption of diclosulam to soil over time (Yoder et al., 2000), which would reduce the amount of diclosulam available to the water phase for runoff. Pantone et al. (1992) reported higher herbicide concentrations and loss in surface runoff 1 d after herbicide application than 30 d after application. Additionally, this study was conducted under worst-case agricultural practices (e.g., freshly tilled seedbed, rainfall 1 d after application). Conservation practices such as conservation tillage, contour plowing, and vegetative filter strips will most likely further reduce the edge-of-field losses of this compound.
Pesticide Root Zone Modeling of Runoff Study
The PRZM model, using measured field soil and weather inputs, and average chemical properties predicted the total observed runoff of diclosulam (sediment + dissolved) for each of the three plots within 4% and within 1% for the average of the three plots (Table 3). The model consistently overpredicted water runoff by 58% on average but underpredicted sediment losses by 20% on average. Calibration against these outputs was not attempted. Although the total diclosulam runoff predictions are quite good, the model did not predict the partitioning between the dissolved and sediment phases that was observed in the field. The PRZM model underpredicted diclosulam transported with the sediment by an order of magnitude; however, since >97% of the diclosulam is transported in the aqueous phase and water yield predictions were reasonable, the total diclosulam runoff predictions were very good (within 1%).
|
|
Monte Carlo Extrapolation of Runoff Study Results
The probability of exceedence of percent runoff and instantaneous peak pond concentrations obtained from simulating 328 yr of runoff for the soil and slope conditions at the Georgia study site are shown in Fig. 3
. Results of these simulations suggest that the runoff study, although it simulated a 1-in-5- to 1-in-10-yr return (80th–90th percentile) rainfall event, it resulted in a 99th percentile runoff event because the simulated rainfall event occurred immediately the day following the application of diclosulam. In other words, based on 328 simulated years of Dublin, GA weather, a 1-in-10-yr intensity storm would be expected to occur exactly 1 d after application only once in 100 yr. The 90th percentile runoff event, considering both weather and application date, would result in runoff of 1.76% of the applied amount.
|
|
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. G. Davis, C. C. Truman, S. C. Kim, J. C. Ascough II, and K. Carlson Antibiotic Transport via Runoff and Soil Loss J. Environ. Qual., October 27, 2006; 35(6): 2250 - 2260. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
