Reducing herbicide losses from tile-outlet terraces.
In an effort to reduce atrazine runoff to surface water, the U.S. Environmental Protection Agency (EPA) approved atrazine label revisions proposed by Novartis in 1993. Label revisions included (1) a reduction in the maximum rate that could be applied annually from 3.0 to 2.5 lb./acre (3.4 kg/ha to 2.8 kg/ha) with additional rate restrictions to 1.6 lb./acre (1.8 kg/ha) if the treated soil was highly erodible land with less than a 30% residue cover; (2) a no-spray setback of 66 ft (20 m) from treated areas to points where runoff discharged directly to a perennial or intermittent stream or river; and (3) various no-spray setback distances from wells, reservoirs lakes, and ponds. In 1994 DuPont and EPA agreed to adopt similar label changes for cyanazine.
EPA indicated in an interpretive statement that standpipes or surface inlet riser pipes in tile-outlet terrace systems are considered a discharge point requiring a setback if the outlet of the subsurface drainage system is within 66 ft (20 m) of a stream or river. The setback could be implemented either at the riser pipe or at the outlet point. Generally, the expense of redesign and reconstruction at the outlet point is prohibitive. Therefore, producers use a setback at each riser pipe in their tile-outlet terrace system to meet the label requirements. Lack of weed control in the setback area presents a yield and weed seed problem for producers. These areas may be sprayed with an alternative herbicide, but this creates additional management and expense. The use of practices other than the no-spray setback that are equally effective at reducing herbicide runoff would save producers time and money.
In 1993, Iowa State University (ISU) completed a field study to evaluate the effectiveness of practices to reduce the runoff of cyanazine and atrazine from tile-outlet terraces (Mickelson et al. 1994). They evaluated three herbicide application treatments on disk-tilled silty clay loam soil. The treatments were (1) surface applied herbicide, no-setback; (2) surface applied herbicide, setback; and (3) surface applied, herbicide incorporation, no-setback. They found that the setback did not reduce herbicide outflow concentrations or losses compared to the no-setback beyond the reduction expected as a result of treating less area with herbicides. Also, incorporation generally reduced concentration of herbicides in the outflow compared to the surface applied treatments.
The ISU study only evaluated the affect of incorporation and setback. Other practices, such as reduced tillage, or no-till, were not examined. Fawcett et al. (1994) reviewed research from 1967 to 1991 that included simulated and natural rainfall studies. He concluded that conservation tillage slows water runoff, increases infiltration, and reduces herbicide runoff. Reduction ranged from 42% to 70% for reduced tillage systems compared to moldboard plow systems. However, he cautioned heavy rainfalls, soon after application, and soils with a low infiltration rate may reduce the effectiveness of conservation tillage in reducing herbicide runoff.
The study reported herein was conducted in 1994 by the University of Nebraska-Lincoln, in cooperation with DuPont, Novartis, Fawcett Consulting, and Missouri Valley Agri-Services. The objectives of this study were to evaluate the effectiveness of various tillage/herbicide management practices in reducing cyanazine and atrazine runoff concentration and total mass loss, and to evaluate the effectiveness of the required no-spray setback at reducing herbicide runoff loss from tile-outlet terrace outflow.
Material and methods
Tile-outlet terraces are designed as surface impoundments to collect water and convey it from the field through subsurface tile lines (Laflen et al. 1978). Various tile-outlet terrace designs are used, ranging from broad-based terraces, where the entire terrace can be farmed, to grass-back-slope terraces where one or more sides are not farmed [ILLUSTRATION FOR FIGURE 1 OMITTED]. Riser pipes are located next to the terrace bank, or in the case of broad-based terraces, several crop rows away from the terrace crest.
Three locations were used for this study with different terrace designs at each site. In Bedford, Iowa, grass-back-slope terraces were farmed nearly to their crest. In Oregon, Missouri, broad-based terraces were farmed on both terrace slopes. In Tekamah, Nebraska, narrow-based terraces were not farmed. The tile riser pipes in Nebraska were adjacent to the terrace, with no crop rows between the riser and the terrace. In Iowa and Missouri, the riser pipes were 20 to 40 ft (6 to 12 m) from the terrace crest with crop rows on each side. The no-spray setback area ranged from 0.20 to 0.32 acres (0.081 to 0.13 ha), depending on the location of the riser pipe.
Four herbicide management practices were replicated three times on 12 individual tile-outlet terraces at each location. The practices were (1) herbicide applied after tillage, no setback-"tilled;" (2) herbicide applied after tillage with setback-"setback;" (3) herbicide applied on no-till, no setback-"no-till;" and 4) spray applied with tillage incorporation, no setback-"incorporated." The average drainage area to the tile terraces was 1.5, 2.32, and 2.47 acres (0.61, 0.94, and 1.0 ha) in Iowa, Missouri and Nebraska respectively. Drainage areas ranged in size from 0.82 to 2.7 acres (0.33 to 1.1 ha) in Iowa, from 0.35 to 2.1 ha in Missouri and from 0.38 to 2.2 ha in Nebraska.
At all locations, plots had been under the same tillage system for at least four years. In Iowa and Nebraska the tilled, incorporation, and setback treatments were created with a one-pass tillage on no-till plots, while in Missouri these treatments were in disk-till plots. Spring anhydrous ammonia injection was the only soil disturbance on no-till treatments. A pre-plant disk was the only tillage and incorporation tool used at the Nebraska locations, with the exception of soil disturbance near the terraces for new construction and maintenance. A John Deere Mulch-Master was used for tillage and herbicide incorporation in Iowa and Missouri.
Each location had a previous soybean or corn crop. Iowa was in corn following corn, and both Missouri and Nebraska were in corn following soybeans. In Iowa, surface residue cover ranged from 30% to 45% in no-till, and 20% to 35% following tillage. In Missouri, crop residue was less than 15% for the tilled treatments and greater than 50% for no-till. The Iowa soil was Nira clay loam (fine-silty, [TABULAR DATA FOR TABLE 1 OMITTED] typic hapludolls) on 2% to 6% slopes. The Missouri soil was Monona silt loam (fine-silty, typic hapludolls) on 11% to 17% slopes. The Nebraska soil was Burchard clay loam (fine-loamy, typic argiudolls) on 11% to 17% slopes.
Herbicides were applied by custom application with commercial spray equipment. In Iowa, atrazine 4L, with 1.61 lb/acre (1.8 kg/ha) active ingredient (a.i.) atrazine and Bladex 4L with 2.41 lb/acre (2.7 kg/ha) a.i. cyanazine, was applied on May 5. In Missouri, Aatrex 90 DF, with 1.61 lb/acre (1.8 kg/ha) a.i. atrazine and Bladex 90 DF with 2.41 lb/acre (2.7 kg/ha) a.i. cyanazine, was applied on April 27. In Nebraska Atrazine 4L, with 0.88 lb/acre (0.99 kg/ha) a.i. atrazine and 2.5 lb/acre (2.8 kg/ha) a.i. cyanazine, was applied on May 11. An alternative herbicide, with no cyanazine or atrazine active ingredient, was sprayed in the setback areas for weed control.
A runoff event was defined as water ponding at the riser pipes to a depth greater than 1.0 ft (0.30 m) above the soil surface in all 12 terraces. An orifice flow restriction was placed in each riser pipe to increase ponding time and allow additional time for sampling. Occasionally, small runoff events ponded water at one or more terraces to a depth less than 1.0 ft (0.3 m). These runoff events were not sampled.
Mickelson et al. (1994) found that because of thorough mixing of water ponded at tile-terrace outlets, herbicide concentration changed very little over time. Therefore, one water sample was used to characterize the herbicide concentration for the ponded volume. However, in Missouri a second sample was obtained when ponding occurred for more than two hours. Samples were collected in 0.13 gallon (500 ml) amber bottles and refrigerated until they were shipped in coolers to Morse Laboratory, Sacramento, California, for chemical analysis. Samples were analyzed by immuno assay techniques.
Pressure transducers with data loggers were installed at the riser pipes in two randomly selected no-till plots and two randomly selected tilled plots at each location. The elevation of each transducer was surveyed so that the depth of water above the transducer could be related directly to the total depth of water. A topographic survey was completed to determine a depth-volume relationship for the terrace. This relationship was used to relate depth of ponded water to volume of water and to compute the total runoff volume for each event. Because the treatments at each site were contiguous, relatively small drainage areas (Figure 1) with the same soil and tillage treatments, the measured runoff volume per acre was assumed representative of the runoff volume from unmeasured terraces with that same treatment.
The date, rain depth, runoff volume per ha and runoff depth for each sampled event is provided in Table 1. Runoff data for at least one no-till plot and one tilled plot was obtained for each location. Rainfall depths varied from 0.47 to 2.75 in (12 to 70 mm). Generally, greater than 1.18 in (30 mm) of rainfall was needed to produce a runoff event; however, in Missouri on June 11, a runoff event occurred with 0.51 in (13 mm) of rainfall in a very short period of time on soil already near field capacity.
Of note in Table 1 is the similarity of the runoff depth for the setback and no-till plots in Iowa, and the incorporated and no-till plots in Nebraska. At each site a single tillage pass into no-till was used to create the tilled treatment. However, based on similar runoff amounts it was concluded that a single tillage pass did not significantly alter the surface hydrology at either site. This point will be further emphasized in following sections.
Comparison of mean herbicide concentrations. The herbicide runoff concentrations varied between the three locations because of variable application rates and different rainfall dates and quantities. This variability between locations was not a concern because the primary use of the data is to compare differences in treatments at a location, not between locations. Three runoff events were sampled in Iowa and Missouri, but only one runoff event occurred in Nebraska. The first sampled event occurred at 28 and 39 days after the herbicide application in Iowa and Missouri, respectively. The single Nebraska event occurred 54 days after the application.
Figure 2 is a plot of the Iowa data showing the trend of decreasing cyanazine concentration in runoff with time after application. A similar pattern occurred for atrazine (data not presented). Figure 3 shows that cyanazine concentrations with time at Missouri deviated slightly from the Iowa trend. This is likely a result of the [TABULAR DATA FOR TABLE 2 OMITTED] short time between Events 1 and 2 in combination with the short duration, high intensity rainfall that produced Event 2. High rainfall intensity generally causes greater loss of herbicides (Baker and Johnson 1979; Wauchope 1978).
In Iowa and Missouri concentrations of both atrazine and cyanazine were greatest from the tilled and no-till plots, where 100% of the contributing area was sprayed (Table 2). Lower concentrations resulted from the incorporated and setback treatments where herbicides were less available for runoff either because of incorporation below the soil surface or a smaller sprayed area.
An analysis of variance (ANOVA) paired comparison evaluation of herbicide concentration was done for each runoff event (Table 2). The cyanazine concentration data from Missouri show significant statistical differences between tilled and incorporation for the first two runoff events (P = 0.05) and between tilled and no-till for all three events. For all three events, there is no statistical difference between the tilled, setback and corrected setback. The corrected setback concentration is the setback concentration increased by the area ratio to account for the contributing drainage area that was not sprayed (Table 2).
In Iowa, the cyanazine data shows no significant difference between tilled and no-till. This similarity in concentration, and the similarity in runoff volume per ha between tilled (setback) and no-till treatments in Iowa (Table 1) leads to the confusion that a one-pass tillage with the Mulch-Master did not significantly alter the surface hydrology compared to the no-till soil. Thus, at Iowa, the filled treatments did not represent an actual tillage difference.
The Iowa cyanazine data shows significant differences between tilled and incorporated and between tilled and setback for runoff Event 1. However, when the setback is corrected for the area not sprayed, there is no statistical difference. For Events 2 and 3 there is no significant difference between tilled, setback, and corrected setback.
In general the atrazine data was less informative. Differences were found between tilled and no-till at Missouri and between tilled and setback for event 1 at Iowa; however, for the other events there are no statistical differences found with the atrazine data. Similarly, no statistical differences were found in the Nebraska data with cyanazine or atrazine.
In Table 2 the average percent change in herbicide concentration compared to tilled is shown for each treatment and for the corrected setback value. This was computed as the arithmetic mean of the percent of reduction for each runoff event. The percent of change of values show several things. First, incorporation reduced atrazine concentrations from 14% to 23% and cyanazine concentrations from 18% to 37% compared to the tilled treatment. It is important to note that the tilled plots in Missouri represent actual long-term disk [TABULAR DATA FOR TABLE 3 OMITTED] tillage, not the first-year conversion from no-till, as was the case in Iowa. However, at both Iowa and Missouri, in either no-till or disk-till, incorporation with the Mulch-Master reduced cyanazine concentrations in runoff between 27% and 37%. Atrazine concentrations were also reduced with the exception of the third runoff event in Missouri. The third event in Missouri had an increase for incorporation over tilled of 0.8 ug/L, which represented a 53% increase, resulting in the average increase of 3% shown in Table 2. For Event 1 and Event 2 the atrazine concentration from incorporated treatments was reduced 22% and 28%, respectively.
Secondly, in Missouri and Nebraska, herbicide concentrations in runoff from no-till were as much as 4 to 5 times greater than from tilled treatments, but in Iowa, concentrations from no-till were at most 1.3 times greater. This again suggests there is little difference in tillage treatments in Iowa, and that the single pass with a Mulch-Master did not create a "tilled" treatment different from the no-till treatment.
Thirdly, in Missouri and Iowa, the setback treatment reduced the atrazine and cyanazine concentrations between 11% and 27%. However, the corrected setback changes were -8% and +7%, suggesting that the actual reduction is in large part a reduction in proportion to the area not sprayed.
Nebraska results are confounding because the concentrations from the setback treatments were greater than the tilled, even though less surface area was sprayed. The longer time after application to the first runoff event and data from only one runoff event limit the reliability of conclusions from the Nebraska data. Several rainfall events that caused runoff water to collect in some terraces, but insufficient water for sampling, may have contributed to these inconsistent results.
Comparison of herbicide total mass loss. The total mass of herbicide lost in runoff is a better indicator of treatment differences than is concentration because the mass lost also integrates differences in total runoff. The mass loss from each event was summed for each treatment and divided by the total area in that treatment (Table 3).
In Iowa and Missouri, the total mass loss per unit area was generally highest for the tilled soil, with 100% of the area sprayed. In Iowa, reductions in total mass loss from incorporation and setback compared to tilled support the same conclusions as did concentration, that is, incorporation reduced total herbicide losses by approximately 25% (22% to 25%), and the setback was effective only in proportion to the area not sprayed. Also, no-till showed mass loss changes of -8% and +11% compared to tilled, suggesting little treatment differences.
In Missouri mass loss reduction from incorporation were 21% to 37%. However, the no-till treatment showed total mass reductions of 94% and 91%, compared to tilled for cyanazine and atrazine, respectively, despite a 3- to 5-fold difference in concentrations. This reduction is attributable to the fact that no measurable runoff water reached the riser pipes in the no-till treatments for two of the three runoff events. No-till soils in Missouri produced a total of only 0.67 in (17 mm) of runoff compared to 2.40 in (61 mm) for the tilled soils, a reduction of 72%. This reduction in total runoff volume dramatically illustrates the potential benefit of no-till as a best management practice for reducing herbicide runoff at that location.
However, the Nebraska data do not appear to support this conclusion. Mass loss increases from no-till of 100% and 220%, compared to tilled, occurred in Nebraska. It is important to note that the Nebraska site had only one runoff event whereas the Missouri data show losses from three runoff events. Using only a single Nebraska event to evaluate total mass loss creates a bias against no-till that can also be shown by looking at a single event from the Missouri data.
On a three-storm basis the Missouri data shows a total herbicide mass loss reduction for no-till of 91% and 94% (Table 3). However, on a single-storm event basis (Event 3) there is an increase in mass loss from no-till of 185% and 90% for atrazine and cyanazine respectively, compared to tilled, similar to the percent increases for the Nebraska single-storm data (100% and 220%). For a single storm event, no-till showed an increase in runoff mass, but over the course of a number of storm events no-till showed a reduction compared to tilled. The reduction seen over the three measured events in Missouri illustrates the long-term advantage of no-till when total runoff amounts are reduced.
Measurable runoff from no-till plots in Missouri did not occur until 84 days after the herbicide application, whereas for tilled treatments it occurred 39 days after the application. An unusually low rainfall year could have biased the runoff results. However, ten years (1985 to 1994) of twenty-four hour rainfall records for the Missouri location, excluding the flood year of 1993, indicate this was not the case. Average June and July rainfall during this period was 4.02 and 4.69 in (102 and 119 mm) respectively, with an average of 1.5 rainfall events per month greater than 1.0 in (25 mm). In 1994, the year of this study, there was 107 and 117 mm rainfall and 1.0 and 2.0 events greater than 1 in (25 mm) in June and July respectively.
The rainfall data suggests the runoff results in Missouri were not biased by an exceptionally low rainfall year. Despite this some caution should be used when extending the conclusions from this study regarding impacts of no-till. In this study, while significant runoff occurred from tilled plots 39 days after herbicide application, runoff from no-till did not occur until 84 days after application. Had an earlier runoff event occurred, no-till herbicide runoff reductions would likely have been different than what was observed. Also, poorly drained soils, or soils with a buried clay pan, will likely have greater herbicide runoff losses from no-till than the well-drained soils in this study.
Summary and conclusions
This study evaluated the effectiveness of various tillage/herbicide management practices for reducing atrazine and cyanazine runoff from tile-outlet terrace systems. Of particular interest was the effectiveness of the no-spray setback area around tile riser pipes.
Concentrations of both atrazine and cyanazine were generally greatest from the tilled and no-till plots, where 100% of the contributing area was sprayed. Lower concentrations were measured from the incorporated and setback treatments where herbicides were less available for runoff, either because of soil incorporation or a reduced spray area. Conclusions from this study are:
(1) The setback treatment reduced atrazine and cyanazine losses are only in proportion to the area not sprayed. When the losses were corrected to account for the area not sprayed, the setback losses were generally equal to the tilled treatment with 100% of the area sprayed.
(2) At the Iowa site a single tillage with a Mulch-Master into no-till did not create a "tilled" treatment in the first year that was significantly different than no-till. This conclusion is supported by similar runoff volumes and herbicide concentrations from tilled and no-till treatments in Iowa.
(3) Incorporation with a Mulch-Master into no-till or disk-till reduced atrazine and cyanazine runoff concentrations by approximately 25%.
(4) When runoff occurred from no-till plots herbicide concentrations from a single event were three to five times greater than from tilled.
(5) Based on three runoff events in Missouri, herbicide total mass loss was reduced by greater than 90% for no-till compared to tilled. This was primarily because total water runoff was reduced by 72%.
(6) A single runoff event evaluation of no-till losses can bias the conclusions. The evaluation of several runoff events during a growing season suggests that no-till reduces total herbicide loss by reducing total water runoff amounts from well-drained silt loam soils.
Management practices recommended based on this study to reduce herbicide runoff from tile-outlet terraces include (1) herbicide incorporation into disk-till or no-till to reduce herbicide runoff by 25%; and, (2) no-till management to reduce total runoff and total herbicide losses by as much as an order of magnitude compared to tilled.
Baker, J.L., and H.P. Johnson. 1979. The effect of tillage system on pesticides in runoff from small watersheds. Transactions of the ASAE 22(3):554-559.
Fawcett, R.S., B.R. Christensen, and D.P. Tierney. 1994. The impact of conservation tillage on pesticide runoff into surface waters: A review and analysis. Journal of Soil and Water Conservation. 49(2):126-135
Laflen, J.M., H.P. Johnson, and R.O. Hartwig. 1978. Sediment modeling of impoundment terraces. Transactions of the ASAE 21(6):1131-1135.
Mickelson, S.K., J.L. Baker, S.W. Melvin, and R.S. Fawcett. 1994. Effects of soil incorporation and setbacks on herbicide runoff loss from a tile-outlet terraced field. Report to Ciba-Geigy and DuPont. February 23, 1994.
Wauchope, R.D. 1978. The pesticide content of surface water draining from agricultural fields-a review. Journal of Environmental Quality. 7(4):459-472.
T.G. Franti is an assistant professor, biological systems engineering, University of Nebraska-Lincoln; C.J. Peter, DuPont Corporation; D.P. Tierney, Novartis Crop Protection Inc.; R.S. Fawcett, Fawcett Consulting; and S.A. Myers Missouri Valley Agri-Services. This manuscript was submitted in April of 1996 and was accepted in March of 1997.
|Printer friendly Cite/link Email Feedback|
|Author:||Franti, T.G.; Peter, C.J.; Tierney, D.P.; Fawcett, R.S.; Myers, S.A.|
|Publication:||Journal of Soil and Water Conservation|
|Date:||Mar 22, 1998|
|Previous Article:||Effects of soil incorporation and setbacks on herbicide runoff from a tile-outlet terraced field.|
|Next Article:||Survey of integrated pest management training needs among retail store employees in Illinois.|