soil& Tillage Research ELSFMER
Soil & Tillage
Research
39 (1996)
241-257
Tillage and application effects on herbicide leaching and runoff D.W. Watts a, J.K. Hall b’* a Coustul b Depurtment
Plains
Soil,
Water,
ofA~ronomy,
I16
and Agric.
Plunt
Research
Sci. und
Center,
Ind.
Bldg.,
16802.
261 I W. Lucus
Penn.syluaniu USA
State
St., Florence. University,
SC 29501, Uniurrsity
USA Park.
PA
Accepted 1I June 1996
Abstract Herbicides are key products in sustaining agricultural production and, to minimize agro-environmental concerns regarding their use, continued assessment of their behavior under different management practices is required. Leaching and runoff losses of four herbicides applied preplantincorporated (PPI) were evaluated in two tillage systems over a 3-year period (1989- 1991). Scant leaching during the droughty 1991 growing season limited treatment evaluations to 2 years. Herbicides were applied at recommended rates (1.7 and 2.2 kg active ingredient (a.i.> ha-‘) to conventional tillage (CT) and mulch tillage (MT) corn (&a mays L.) fields on Hagerstown silty clay loam (fine, mixed, mesic Typic Hapludalf). Tillage treatments were defined as moldboard plow-disk-harrow (CT) and single-disking (MT). During this study, CT followed 5 years of corn production in a comparable CT system on this site and, similarly, MT followed a 5-year no-tillage (NT) system. Herbicides were applied preemergence (PRE) to CT and NT in the .5-year study and preplant-incorporated (PPI) in this study. Herbicide mobility in subsurface drainage was evaluated from herbicide mass transported to pan lysimeters installed 1.2 m deep. Surface drainage losses of these chemicals were determined from residues in runoff collected with automated sampling and recording equipment. Leachate volumes were greater from MT than CT in 1989 and 1990 and exceeded all seasonal losses during the previous 5 years under NT management. Comparisons of total seasonal leachate discharged to pan lysimeters within and among studies and herbicide mass leached showed that timing of leachate-inducing precipitation relative to herbicide application was the key factor in regulating herbicide translocation. Herbicide mass transported through the root zone averaged from less than 0.1% to 0.9% of applied rates in CT and from 1.4% to 5.1% in MT. Leachate-availability of herbicide residues and extent of herbicide longevity in this soil under MT conditions were similar to previous findings under NT management. Despite these behavioral
* Corresponding author. 0167- 1987/96/$15.00 Pff
SO 167.1987(96)0
Copyright 0 1996 Elsevier Science B.V. All rights reserved. 1058.6
242
D.W. Wutts. J.K. Hall/Soil
& Tillqr
Research
39 (1996)
241-257
similarities for herbicides among tillages, herbicide mass discharged per unit of percolate was most often lower for MT compared with NT, particularly in early growing seasons of comparable precipitation. Thus, the PPI treatment in MT appeared to reduce leaching of these chemicals compared with PRE application in NT. Runoff losses of PPI herbicides ranged from 0.35% to 0.77% of applied rates in CT and from 0.13% to 0.28% in MT. Losses of PRE-applied herbicides from NT averaged less than 0.1% of applied rates; maximum yearly losses ranged from 0.06% to 0.18%. Thus, the character of the disked, minimally tilled surface provided a level of impedance to runoff that was greater than achieved with the tilled surface on this 3 to 5% slope, but less than previously obtained with an untilled, mulch-covered surface. Keywords:
Leaching; Pan lysimeters; Runoff; Herbicides; Conventional tillage; Mulch tillage; No-Wage; Zra
muys L.
1. Introduction One basic thrust of agronomic research is to define best managementpractices for pesticide application in crop production systems that will maintain a high level of chemical efficacy without accentuating non-point source pollution of non-target areas. Since agricultural sustainability will continue to require the safe and proper use of pesticides, the effects of tillage system environments and pesticide application techniques on pesticide behavior and fate require further investigation. The main pathways of non-point sourcepollution, surface and subsurfacetransport of pesticidesare not mutually exclusive and abatementof pesticide concentration and mass in one pathway by some management scheme often exacerbates or provides little attenuation of pesticide residues in the other. Moreover, determining the transport protess that has the greatest impact on water quality is difficult to assesssince widely diverse opinions exist on which processmay be more detrimental to water resources. Our research (Hall et al., 1991; Hall and Mumma, 1994) clearly demonstrated that environmental impact ‘trade-offs’ exist with different tillage practices. An untilled, mulched surface markedly reduced runoff of preemergenceand postemergenceapplied herbicides, but its undisturbed topsoil matrix promoted more leaching of these chemicals, presumably by preferential macropore flow, as considerable evidence indicates (Edwards et al., 1989; Andreini and Steenhuis, 1990; Shipitalo et al., 1990). Conversely, the disrupted macroporematrix in a tilled surface limited herbicide leaching but the lack of mulch predisposedthe surface to a greater loss of these chemicals in runoff water. Greater chemical loss in runoff water from tilled surfaces has been documented by Wauchope (1978). The indigenous
character
of these tillage
practices
in our studies (Hall
et al., 1991;
Hall and Mumma, 1994) was consistentover 5 years for herbicide compoundsrepresenting different herbicide classes,namely, chloro-s-triazines, a substituted amide, and a benzoic acid derivative. Quality and quantity assessments of herbicide mobility varied between the two transport processestemporally and between and within herbicide classes.However, in a broad sense,each of these tillage environments influenced these processesuniformly on an annual basis in the well-drained soil. The net effect was greater losses from the untilled
soil due to greater
percolate
discharge
of chemicals.
5.W.
Wutts. J.K. Hull/Soil
& Tilkuge Resew&
39 (19961241-257
243
In continuing studies on this same site, we chose an alternate approach to tillage and herbicide application practices in an attempt to identify a management scheme that harbored the best characteristics of conventional tillage (CT) and the conservation tillage (CnT) system, no-tillage (NT), in ameliorating herbicide transport. Several researchers (Baker and Laflen, 1979; Hall et al., 1983) showed that blending of herbicides within the topsoil effectively reduced losses of these chemicals in runoff water. However, little attention has been given to the effect of herbicide incorporation within the topsoil surface on transport of these chemicals by leaching. Consequently, the previous 5 year NT system was converted to mulch tillage (MT) and the CT area was left intact. Coupled with tillage rotation, herbicides were soil surface incorporated before planting, permitting the objective evaluation of this management practice on surface and subsurface losses of herbicides in the two tillage systems. It was postulated that these changes would have several effects: firstly, the MT system may function more like CT since macropore continuity would be disrupted in early season, thereby altering prominent pathways for rapid leaching of herbicide residues from the surface when they are most concentrated; secondly, random distribution of preplant-incorporated herbicides through a greater topsoil volume compared with a concentrated zone on a NT surface may lead to reduced herbicide load in leachates; lastly, the MT area would possess a surface ‘roughness’, delineated by the randomness of corn stover residue incorporation, that could provide a level of impedance to runoff flow not markedly different from that achieved with the NT surface. Consequently, the expected net effect of rotating tillage from NT to MT with herbicide incorporation would be reduced leaching load of the chemicals without excessively increasing their losses in runoff.
2. Materials
and methods
This research was conducted from 1989 through 1991 at the Russell E. Larson Agricultural Research Center, Rock Springs, Pennsylvania, on a Hagerstown silty clay loam (fine, mixed, mesic Typic Hapludalf). The study site (0.52 ha) was separated equally into two tillage treatments that had been maintained as distinct leachate and runoff sampling areas since 1984. Each area contained three excavated and framed pits that individually housed three pan lysimeters at a depth of 1.2 m to collect root zone percolates. Areas were segregated from each other and surrounded by earthen-dikes to entrap and facilitate collection and measurement of runoff water using an HS-flume and automated recording and sampling equipment positioned at the natural drainage outlet of each area. Land slope, determined from a topographic survey, ranged from 3 to 5% within each area. A comprehensive discussion of the site, including soil description, equipment placement and installation, and previous soil, crop and herbicide management practices and sampling procedures are cited elsewhere (Hall et al., 1989, 1991). Tillage practices maintained on these separate areas from 1984 through 1988 were CT and NT. The CT area was moldboard plowed, disked and spring-tooth harrowed. The NT surface was disturbed only by the fluted-coulter on the NT-planter. Four herbicides were preemergence (PRE) applied to corn on each tillage area at recommended rates. These chemicals included simazine (6-chloro-N,N’-diethyl-1,3,5triazine-
244
D.W. Wutts, J.K. Hull/Soil
& Tillage Research
39 11996) 241-257
2,4-diamine); atrazine (6-chloro-N-ethyl-N’-( 1-methylethyl)- 1,3,5-triazine-2,4diamine); cyanazine (2-[[4-chloro-6-(ethylamino)1,3,5-triazin-2-yl]amino]-2-methylpropanenitrile) and metolachlor (2-chloro-N-(2-ethyl-6-methylphenyl)-N-(2-me~oxy-lmethylethyl) acetamide). Application rates were 1.7 kg active ingredient (a.i.) ha-’ (simazine, atrazine) and 2.2 kg a.i. ha-’ (cyanazine, metolachlor). In addition, dicamba (3,6-dichloro-2-methoxybenzoic acid) was postemergence applied at 0.56 kg a.i. ha-’ and the insecticide, carbofuran (2,3-dihydro-2,2-dimethyl-7-benzofuranyl methylcarbamate), was row-applied (1.0 kg a.i. ha- ’ ) at planting. In 1989, the NT area was converted to a MT system. Mulch tillage consisted of disking approximately 15 cm deep. Percent corn stover residue remaining on the surface after disking was not determined, however, cover averaged 42% in line transect measurements after similar MT treatment of the same area in current studies. The CT area was managed as previously described. The same pesticides and rates were applied with the exception that the insecticide, terbufos (S-[[(l ,I -dimethylethyl) thio]methyl]O,O-diethyl] phosphorodithioate), replaced carbofuran at the same rate. Prior to herbicide application, the pit and flume areas were sealed with plastic to prevent herbicide contamination of the collection vessels by direct application or drift. The four herbicides previously PRE applied were preplant-incorporated (PPI) in both tillage areas with a single pass of a disk set approximately 8 cm deep. Herbicides were applied on 25 May 1989 and 24 May 1990. Both areas were fertilized and limed according to annual soil test recommendations. Corn was planted to achieve a population of 66000 plants ha-‘. Areas between pit frames and runoff flumes that could not be accessedby normal farm equipment were fertilized and planted by hand at the same rates to insure that uniform soil fertility conditions and plant populationsexisted throughout the researchsite. Rain was recorded on site from April I through October 31 of each year. During the remaining months, data were obtained from a meteorological station located approximately 1.5 km from the researchsite. Pan lysimeter leachateswere collected throughout the entire year whenever rain or snow-melt was of sufficient magnitude to produce a leaching event. Leachate volumes were recorded and subsampled.Herbicide massleached was calculated from herbicide concentrations in water and correspondingpercolate volume. These data were usedto calculate mean area1leaching dischargeand percentagelossof herbicides within the corn root zone of each tillage area. Runoff water was sampledafter each erosion event from the time of herbicide application through September30. Herbicide concentrations in these samplesand total recorded runoff volume were used to calculate total herbicide losses during the season. Water samples were stored at - 14°C before extraction and analysis. Although resultsand discussionwill focus solely on the PPI chemicals,a method was developed that permitted the simultaneousextraction of all six pesticides(Watts et al., 1994). This method utilizes two types of solid phaseextraction columns, a reverse phase and an anion exchange column connected in seriesto extract all six compoundsfrom a single sample.After elution of the reverse phasecolumn, a gas chromatographequipped with a N-P detector was used to analyze for the chloro-s-triazines (simazine, atrazine, cyanazine) and metolachlor. Detectable limits were 3 kg 1-l for the triazines and 6 mg 1-l for metolachlor. Grand mean percent recoveries for water sampleswere atrazine
D. W. Wafts, J.K. Hull/Soil
& Tillage
Research
(94.7%), simazine (97.8%), metolachlor (98.2%) residues detected in leachates and runoff water Compilation of means and mean separation between r-test) were determined on leaching data using the Cary, NC, and Minitab, Inc., State College, PA, Consulting Center of the University.
39 ( 1996) N-257
245
and cyanazine (94.1%). Herbicide were not corrected for recoveries. tillage systems (pooled, two sample programs of the SAS Institute, Inc., as recommended by the Statistical
3. Results and discussion Although environmental factors vary annually and can influence degradation, persistence and mobility of herbicides within and from soils, periodic reference and association is made between results achieved during this research and the previous 5year study. Particular focus will be given to the CnT practices and comparisons will be made, in general, among tillage systems over comparable sampling periods and within specific time intervals. Soil spatial variability and the limited number of observations between lysimeter-sampling points placed limitations on the degree of statistical significance achieved for mean comparisons among tillage areas. For most means, the confidence intervals were less than 73% (P < 0.27). These statistical results were disappointing since many comparisonsamong tillage systemsin the 5-year study were significant at confidence intervals of 90% or greater. Nonetheless, substantial ‘differences’ noted among tillage systemswill be discussed. 3.1. Precipitation
and leaching
summary
Total rain recorded in 1989 (98.3 cm) was comparable to the 30-year average (97.9 cm) for this area. In 1990, precipitation totals exceeded the 30-year average by 25 cm, but 1991 precipitation was 20 cm lessthan the norm. Herbicides are most susceptibleto leaching or runoff transport during the early growing seasonfrom herbicide application through July (Wauchope, 1978; Triplett et al., 1978; Ha11et al., 1983, 1991). Within this early period, total rainfall was 39.2 cm in 1989 and 23.6 cm in 1990. These amounts accountedfor approximately 58% and 29% of the total precipitation recorded from May through December (Figs. 1 and 2). An additional 38.6 cm or 48% of the total precipitation for this 8-month period occurred during August through October 1990, which exceeded the 30-year average for this locale. In contrast, the 1991 seasonwas very dry and was not a good year for herbicide transport studies. From herbicide application in May through September, 30.5 cm of rain was measured.Three leachate samples were collected during this interval. Most of the leachates collected were dischargedprior to herbicide application or from October through December. Comparatively few herbicide residueswere detected in leachates, concentrations were minimal, total losseswere infinitesimal and no runoff events occurred. Consequently, only results for 1989 and 1990 will be presented. In 1989, 15 leaching events yielded 122 pan lysimeter samples;in 1990, 22 events produced 191 samples.In general, leachate discharge was substantially greater under MT than CT conditions during both years (Figs. 1 and 2). Also, leachate dischargefrom both tillage systemsclosely followed the rainfall distribution during May through July
D.W. Watts, J.K. Hall/Soil
246
& Tillage Research
39 (1996) 241-257
7
1 LEACHATE
- 25
- 20
- 15
c 5 !L
- 10
gj V
0
0
Jun
May
Jul
Sep
W
Ott
Nov
Dee
1989 Fig. I. Total monthly leachate volume and precipitation recorded after herbicide application in 1989 and the 30-year average monthly precipitation for this region. Where no leachate is delineated, none was collected.
and October through 1990 compared with related to antecedent 1990 (14.3 cm) than
November during each year. Greater leachate discharge in May May 1989 after the same amount of precipitation may have been soil moisture since total precipitation for the month was greater in in 1989 (11.7 cm).
3.2. Herbicide mass in root zone leachates In general, herbicide mass leached during the 1989 and 1990 collection seasons was coincident with leachate discharge and was greater from MT than CT management
200
7, t
D
150
-
-j 40
1 LEACHATE
it-
I
I
I
1
MT
30
RAINFALL
A E!. 20
3
0 Jun
Jut
*w
Sep
Ott
Nov
Dee
1990 Fig. 2. Total monthly leachate volume and precipitation recorded after herbicide application in 1990 and the 30-year average monthly precipitation for this region. Where no leachate is delineated, none was collected.
D.W. Watts, J.K. Hall/Soil
& Tillqe
70-
0-l 2 -0 w 5 z A 2 2 w ; a 5 I 3 5 l-
CT 0 m
I MT m m m B
600 40-
39 (1996)
247
241-257
Simazine Atrazine Cyanazine Metotachlor
GO50-
Research
Simazine Atrazine Cyanarine Metolachlor
30-
20lo-
o-
I
May
Jun
Jul
Aug
Sep
Ott
Nov
Dee
1989 Fig. 3. Total herbicide mass leached after herbicide herbicide residue is delineated. none was detected.
application
in 1989. Where
leachate was collected
and no
(Figs. 3 and 4). The most notable variance to this trend occurred in May 1989 where herbicide residueswere only detected under CT conditions in the small leachatevolume collected. Mean area1lossesof the three chloro-s-triazines were comparablein tilled soil during 1989, ranging from 1530 to 1570 pg mV2, which representedlossesof 0.69% to 0.93% of rates applied (data not presented). Triazine losseswere more variable under MT conditions (6813 to 8503 +g m -2; 3.0% to 5.1%). Theseherbicides also exhibited a higher leaching potential than metolachlor, which was displaced at levels of 0.37% and 2.46% of the rates applied to CT and MT, respectively. Leaching trends noted in 1989 for all herbicides continued in 1990. Herbicide mass displacedwas less,particularly in CT, where losseswere reduced by approximately 79% to 94% for all herbicides comparedwith massleachedin 1989. Likewise, herbicide mass leached from MT in 1990 was 8% to 44% less than 1989 leaching losses. Only cyanazine leachedin amountsslightly more than loads measuredin 1989. Although total leachate discharge in 1990 exceeded 1989 levels and substantially so under MT conditions (Figs. 1 and 2), the dominant early seasonrainfall in 1989 and substantial leachate discharge in June 1989 contributed to the greater herbicide loads with both tillage practices during this year. Herbicides were generally more ‘leachate-available’ under MT than CT conditions regardlessof year. The most leachate-availablechemicals in this soil were also the most persistent compounds, atrazine and simazine. Their availability and longevity were expressedby their dominance in early seasonleachatesand also by their late season
248
D.W. Watts. J.K. Hall/Soil 900
70
60
1
9 u' .?iii
600-
4 3
500 -
r" a, ?$ B & I i5 i5 I-
241-257
65
800 h
& TiNclge Research 39 (1996)
700-
55
40030 300-
25 20
200-
15 100-j
10 5 0
May
Jun
Jul
Sep
act
Nov
Dee
L
1990 Fig. 4. Total herbicide mass leached after herbicide application in 1990. Where leachate was collected and no herbicide residue is delineated, none was detected.
discharge in both tillage systems (Figs. 3 and 4). Residues of these chemicals in leachateswere dominant in MT, especially during Septemberthrough December 1990 when rainfall for the most part exceeded the long-term average for this region. During this period also, l- to 2-p.g residuesof the less persistentcyanazine were also detected in leachatesfrom both tillages, despite considerabledischarge of this compound from MT during the first 3 months. These samecharacteristics were noted for atrazine and simazine under NT management(Hall et al., 1991) and appearto be a consistentfeature of CnT systems. Considerable percolation within the MT root zone was contrary to theoretical considerationsthat a minimal tilling of the NT surface would disrupt macroporesand limit water translocation through this disturbed matrix, particularly in early season.On the other hand, several researchers(Quisenberry and Phillips, 1976; Phillips et al., 1989) identified macropore flow beneath tilled matrices and showed that water under negative pressurecan enter simulated macroporesafter a continuous water film is establishedon macroporewalls. Consequently, the substantialpercolate volumes collected in 1989 (198 1) and 1990 (454 1) from MT, which exceeded the volumes (5 to 114 1) obtained in previous seasons(1984 through 1988) of NT management(Hall et al., 1991), may be largely attributed to macropore flow. Phillips et al. (1989) also reasonedthat soluble compoundsin water entering a tilled surface may equilibrate with this entire soil matrix, thus, soluble compounds in water entering macropores beneath a tilled layer may
D.W. Wutts. J.K. F~all/Soil
& Tillage
Reseurch
39 (1996)
241-257
249
provide a greater solute concentration than in water flowing in macropores open to the soil surface, wherein solute would originate from the soil surface or from macropore surfaces. Based on the leaching patterns observed on this site in 1989 and 1990, it appeared that this concept aligns more with herbicide solute transport in a minimally tilled surface than in a moldboard-plowed, disked and harrowed surface. Moreover, movement of water and herbicide solute into untilled soil surfaces also maintained higher subsurface loading of herbicides than in tilled soil (Hall et al., 1991). Percolate volume and herbicide load from the CT system varied yearly with seasonal distribution and frequency of precipitation but absolute chemical mass leached was always less than the mass transported from MT management (Figs. 3 and 4) and from NT management (Hall et al., 1991). Depth and manner of tilling did not vary during the last 7 years of CT management. Thus, matrix effects on herbicide solute ‘retention’ and transport within a CT surface compared with a MT surface appeared to be different. 3.3. Area1 losses among tillage practices Direct rotation of NT to MT eliminated the opportunity to make yearly comparisons of herbicide mobility between these different CnT systems. In retrospect, having a NT system of the same age for comparison with the converted NT system would have been ideal since this area would have served as a control in evaluating the hypothesized effects of tillage disruption and chemical incorporation on herbicide mobility. Soil properties and characteristics in continuous, long-term tillage practices have been evaluated but the effects of tillage rotation on crop production and the time required to achieve a ‘fixed’ set of physico-chemical characteristics within the new system is less understood. Some farmers practicing NT corn production will intuitively rotate this system to CT for a l-year to 2-year period, reasoning that this change improves the physical nature of the soil and aids in weed and insect control. Since our study area contained only two confined tillage systems, rotation choices to evaluate the hypotheses were limited. Converting from 5 years of NT to MT and continuing with the 5-year-old CT system in place seemed to be the logical choice. Conversion of the CT area to NT may not have provided the same matrix effects typified by the ‘aged’, 5-year NT system (Dick and Daniel, 1987; Edwards et al., 1988). In the absence of an intact NT system, atrazine and cyanazine leaching in MT was compared with that achieved on the same site in the previous 5 years. Although tillage rotation had little effect on altering percolation through the MT solum to 1.2 m compared with a NT system, prescribed differences in application management appeared to affect the magnitude of leaching losses within CnT and CT management when compared among years. Rainfall from application in May through July was comparable for 1984, 1985 and 1990 (26.0, 26.7,23.6 cm; Figs. 2 and 5). Likewise, comparable totals were recorded for 1986 and 1989 (34.6 and 39.2 cm; Figs. 1 and 5). Similarities notwithstanding, total recorded leachate for this period was considerably less for NT than MT conditions, regardless of year (Table 1). Thus, based upon leachate discharge during the early period, herbicide mass leached in 1989 and 1990 should have surpassed loads recorded during the 5 years of NT management. Atrazine mass leached in 1986 during the early period exceeded loads recorded in
measured
0.3 0.3 0.6 1.7 35.3
0.9 1.2 2.1 6.7 31.3
0.3 1.8 2.1 I I.3 18.6
1986 0.1 0. I I.4 7.1
I987
1.4 1.4 9.8 14.3
-
I988 I .o 16.3 3.8 21.2 24. I 87.9
1989 0.7 8.2 8.9 51.7 17.2
of each year.
0.8 0.6 8.4 9.8 33.8 28.9
1990
of each year
I984
-b
1985
December
1984
through NT/MT
application
CT
in CT and CnT from herbicide
a Values reported for May represent the period from herbicide application through 3 I May b No leachate recorded during this time. ’ From herbicide application through 3 I December. * Percent of total yearly losses collected from herbicide application through 3 I July.
May a (1) June (1) July (I) 3-month totals (1) Yearly ’ totals (I) Early d season losses (%)
Table I Total leachate
11.3 6.4 17.7 57.7 30.7
1985
7.3 6.1 2.0 15.3 32.8 46.6
1986
0.6 I.1 1.6 4.0 40.0
I987
14.2 14.2 I 13.6 12.5
1988
0.2 131.1 15.8 147.0 197.8 74.3
1989
35.9 2.0 41.2 79.2 453.6 17.5
1990
D.W. Watts.
J.K. Hall/Soil
& Tillage Research
39 (1996)
251
241-257
20
g = 1 .G 2
15
10
5
:
0
-- 1984 -Fig. 5. Total 5-year study.
monthly
precipitation
-- 1985 --
recorded
-- 1986 --
after herbicide
-- 1987 --
application
through
- 1988December
in the previous
1989 and 1990; also, losses in 1985 exceeded those obtained in 1989 (Table 2). During these 4 years, 83 to 95% of the total atrazine losses occurred during the early growing season. Considering that rainfall was below normal in early 1990 (Fig. 2) and only 18% of the total leachate discharge was recorded early (Table I), atrazine yield (50 p,g) after July in 374 1 of percolate was minuscule (Table 2). In comparison, an additional 40 1 in 1985 and 18 1 in 1986 yielded 96 and 53 p.g, respectively. In 1989, 74% of the total leachate was measured in early season and an additional 51 1 discharged through December yielded 112 p,g of atrazine. Given that early season leaching magnitude would affect leaching losses later, the ratio of atrazine mass transported to measured root zone leachate for this period was greater in 1985 and 1986 than 1989 and 1990 (Tables 1 and 2). With few exceptions, greater mass to leachate ratios were also obtained in the other seasons of NT management compared with MT conditions. In summary, where rainfall was comparable during the early period but percolate transmission varied between MT and NT conditions, albeit among studies, atrazine mass transported was not predictable. If gravitational water in macropores at the surface or beneath the disked layer was principally involved in herbicide solute transport, then it may be concluded from a comparison of CnT systems that the PPI treatment reduced atrazine mass leached. These suggestions were strengthened by results for cyanazine mobility. The higher water-solubihty of cyanazine compared with atrazine was expressed by the greater cyanazine load discharged during the early seasons of 1985 through 1990 (Tables 2 and 3). However, the leachate-availability of cyanazine, as expressed by the transported mass to leachate ratio, was greater in NT during 1985 through 1988 than in MT during 1989 and 1990 (Tables 1 and 3), supporting the premise that area1 losses of this chemical were reduced by the PPI treatment. Area1 losses of atrazine and cyanazine in CT were also influenced, in part, by the PPI treatment. Leached quantities of these herbicides in 1989 were greater than in any other
of early
1.3 0.5 I .8 2.3 78.3
58.3 51.7 110.0 110.2 99.8
1985
atrazine
51.8 36.8 88.6 97.5 90.9
1986
0.1 0.7 14.3
0.1 -
1987
losses in leachate
34.9 141.8 1.0 177.7 190.6 93.2
1989 5.7 49.3 55.0 109.9 50.0
1984
NT/MT
473.7 167.7 641.4 737.0 87.0
1985
g PRE application
of each year.
0.8 0.5 19.2 20.5 32.3 63.5
1990
CT and CnT followin
13.0 13.0 24.4 53.3
1988
from
a Values reported for May represent the period from herbicide application through 3 I May b No leachate recorded during this time. ’ From herbicide application through 3 I December. * Percent of total yearly losses collected from herbicide application through 3 I July.
-”
1984
CT
season and yearly
May a ( wg) June ( pg) July ( pg) 3-month totals ( /.~g) Yearly totals ’ ( pg) Early d season losses (%I
Table 2 Comparison
686.9 177.8 47.2 914.6 967. I 94.6
1986
I.4 107.8 109.2 115.6 94.5
1987
in 1984 through
107.9 107.9 180.3 59.8
1988
0.0 525.1 8.5 533.6 645.8 82.6
1989
748.2 3.1 65.2 816.5 866.8 94.2
1990
1988 and PPI in 1989 and 1990
s ?J $ 9 ;:3% 2 g s
,” -F g R. Y z
$
? ? 2 ?
of early
51.3 29.6 80.9 80.9 100.0 0.5 0.5 100.0
0.5 4.8 4.8 4.8 100.0
68.0 120.9 2.0 190.9 190.9 loo.0
1.8 8.2 10.0 14.6 68.5
of each year.
0.0 0.0 9.0 9.0 12.4 72.6
1990
a Values reported for May represent the period from herbicide application through 3 I May b No leachate recorded during this time. ’ From herbicide application through 3 1 December. d Percent of total yearly losses collected from herbicide application through 3 I July.
I .6 0.1 I .7 I .7 100.0
80.3 45.5 125.8 126.2 99.7
I989
-
I988
-b
1987
I984
1986 410.9 59.2 470. I 471.0 99.8
I985
PRE application
1985
from CT and CnT following
I984
losses in leachate NT/MT
cyanazine
CT
season and yearly
May a ( pg) June f pg) July ( /vg) 3-month totals (pg) Yearly ’ totals ( pg) Early d season losses (8)
Table 3 Comparison
1458.1 187.2 41.5 1686.8 1693.8 99.6
1986
156.9 157.2 99.8
0.7 156.2 -
I987
in 1984 through
191.9 191.9 285.7 67.2
1988
0.0 740.7 8.4 749. I 799.8 93.7
1989
824.6 2.9 18.3 845.8 846.9 99.9
1990
1988 and PPI in 1989 and 1990
i fs
$ B 5 % 2 3 B 2
2 2 R Y z s 2
\
G
3 ze
i
254
D.W. Wafts, J.K. Hu#/Soii
& Tiilugr
Research
39 fI996124/-257
season (Tables 2 and 3). However, ratios of herbicide mass transported to percolate discharge in early season were greater from PRE-applied chemicals in 1985 and 1986 compared with PPI chemicals in 1989 and 1990. Ratios were also higher in 1984, 1987 and 1988 compared with 1990. Thus, blending of atrazine and cyanazine into the topsoil by disking appeared to reduce their leachability in CT management. Collectively, these results demonstrated that a NT system rapidly transmits herbicides through the root zone during early season rainfall as other work demonstrated (Isensee et al., 1990; Shipitalo et al., 1990; Hall et al., 1991). Herbicide solute within the micropore network of this system may subsequently diffuse and merge with macropore water and move preferentially and/or move by diffusion and displacement through the micropore matrix (Germann et al., 1984; Shipitalo et al.. 19901, accounting for herbicide load discharged in late season. In previous studies (Hall and Mumma, 1994) on this Hagerstown silty clay loam, dicamba residues were detected in PL leachates from CT and NT systems as late as 6 months after application during several seasons. Entrapment and translocation within the micropore matrix of this well-drained soil were postulated as being critical factors involved in the atypical residence time and late season ‘breakthrough’ of this highly soluble, anionic, low persistence herbicide. Additionally, the hydrophobic organic matter that accumulates on the NT surface and within earthworm burrows can aid in the preferential movement of water and dissolved solutes under dry soil conditions with radial capillary movement of water out of earthworm burrows and into the surrounding micropore matrix occurring with prolonged exposure to water (Edwards et al., 1989). Stehouwer et al. (19931 showed that sorption of atrazine was increased on the organic C-enriched linings of macropore burrows created by the earthworm, Lumbricus terrestris L., compared with sorption in the bulk soil. Thus, retention and release of residual herbicides in these large biopores can regulate the leachability and distribution of these chemicals throughout the entire pore network. On the other hand, random orientation and anchoring of crop residues (corn stover) in a CnT surface such as MT provide micro-depressional topography that pools surface water reducing runoff potential, yet also provide ‘conduits’ for water entry and flow between the organic litter and aggregated topsoil (Onstad and Voorhees, 1987). A micro-saturated soil zone at the interface between a minimally tilled and untilled matrix in the upper regions of the soil profile may provide for water entry into macropores beneath this zone despite having a water-unsaturated pore ‘column’ above this zone. Likewise, water under negative pressure may enter macropores in response to a water film on macropore walls (Phillips et al., 1989). These conditions may permit as much or more drainage of water through macropores beneath minimally tilled layers as through the same sized pores connected to the soil surface. Consequently, if these characteristics prevailed under MT management, accounting for the large amount of percolation in this system, reduced chemical mobility compared with NT was a function of the PPI treatment. 3.4. Herbicides
in runoff water
Seven runoff events were recorded during the 1989 growing season; no runoff occurred in 1990. On a seasonal basis, total runoff volume was 336 and 229 m3 ha-’
D.W. Watts. J.K. Hall/Soil Table 4 Mean area1 and percent
losses of herbicides
& Tillugr
in runoff
Research
39 (1996)
Simazine Atrazine Cyanazine Metolachlor
9.2 a 12.9 8.7 7.8
255
water
1989 CT(g
241-257
1985 to 1988 ha-‘)
MT (g ha- ‘)
CT (%o)
MT (%Io)
CT f%)
3.7 4. I 3.8 3.0
0.55 b 0.77 0.39 0.35
0.22 0.28 0.17 0.13
0.29 (0.62) 0.18 (0.36) 0. I 1 (0.30) 0.1 I (0.25)
a Calculated from water volume and mean herbicide concentrations for each erosion number of events within each growing season. b Calculated from area1 losses of herbicides and herbicide rates applied. ’ Mean (maximum) percent runoff losses in the previous study.
NT (o/o) ’
0.03 0.06 0.02 0.02 event
(0.18) (0.10) (0.06) (0.06)
and the total
under CT and MT management, respectively. The largest single runoff event, 25 days after herbicide application, produced 99 m3 ha-’ from the CT area and 66 m3 ha-’ from the MT area. Therefore, MT management reduced total seasonal runoff and maximum event losses by one-third compared with CT. In the previous study (Hall et al., 19911, NT reduced seasonal runoff by 54 to 78% compared with CT during 1986 through .1988. Consequently, in this single season, the partially mulch-covered, rough surface of MT was not as effective in reducing runoff as the mulch, untilled surface. Total area1 and percent herbicide losses in runoff were higher under CT than MT conditions (Table 4). Maximum percent losses were 0.77% for atrazine under CT and 0.28% for atrazine under MT. Although percentage losses represented very low herbicide amounts, one may conclude that the PPI treatment did not effectively alter runoff losses of these chemicals from CT compared with maximum quantities transported in runoff from PRE-applied chemicals in the previous study. Other work in Pennsylvania showed that runoff losses of PPI atrazine in CT were reduced compared with PRE-application, however, the incorporation technique differed (Hall et al., 1983). On the other hand, since maximum herbicide losses from CnT systems were not widely divergent and were generated from 80 m3 ha-’ (NT) and 229 m3 ha- ’ (MT) of runoff, it may be concluded that herbicide incorporation in MT compensated for the reduction in surface mulch cover in this system compared with maximum cover under NT management.
4. Summary
and conclusions
Minimum tillage of a previously untilled Hagerstown silty clay loam surface after 5 years of corn production had little effect on reducing the volume of percolate through the solum as postulated. Leachate-availability of herbicides was greater under MT than CT conditions, consequently, greater mass yield was detected yearly, in general, from this CnT system. Herbicide load in percolate was greater for the chloro-s-triazines (atrazine, simazine and cyanazine) than metolachlor and was more dependent on the amount of leachate-inducing precipitation within the first 2 to 3 months after application than in total seasonal precipitation. The characteristic herbicide availability and extent of
256
D. W. Watts, J.K. Hall/Soil
L Tillage
Research
39 (1996) 241-257
herbicide longevity found under MT conditions were similar to herbicide behavior observed previously under NT conditions on this soil. Despite similarities in herbicide behavior among the CnT practices, herbicide load discharged to pan lysimeters per unit of percolate was most often lower for MT compared with NT, particularly when early season herbicide loads were compared among years of similar rainfall. As a consequence, it was concluded that incorporating sprayed herbicides preplant within the MT system appeared to reduce the transport of these chemicals in root zone leachates compared with preemergence application in NT. Additionally, the partially mulched character of the MT surface provided a level of impedance to runoff flow that was greater than achieved in an adjacent CT system, but probably less than expected from a NT system based upon an intuitive evaluation from previous results in Pennsylvania and elsewhere. Since erosional losses of herbicide mass principally occur in the aqueous phase, any management scheme that reduces surface drainage can provide a significant element in conservation and retention of applied herbicides. The disked, MT system with PPI herbicides exhibited this capability.
Acknowledgements
We gratefully acknowledge the support of this research by the PA Department of Agriculture, Harrisburg, PA (Contract: ME 449020) and the assistance of Mary Kay Amistadi, John Bathgate and Edward Bogus for their expertise and help in sample collection, processing and analysis and preparation of graphs for this report.
References Andreini, M.S. and Steenhuis, T.S., 1990. Preferential paths of flow under conventional and conservation tillage. Gecdenna, 46: 85 102. Baker, J.L. and Laflen, J.M., 1979. Runoff losses of surface-applied herbicides as affected by wheel tracks and incorporation. J. Environ. Qual., 8: 602-607. Dick, W.A. and Daniel, T.C., 1987. Soil chemical and biological properties as affected by conservation tillage: environmental implications. In: T.J. Logan, J.M. Davidson, J.L. Baker and M.R. Overcash (Editors), Effects of Conservation Tillage on Groundwater Quality - Nitrates and Pesticides. Lewis Pub]., Chelsea, MI, pp. 125-147. Edwards, W.M., Shipitalo, M.J. and Norton, L.D., 1988. Contribution of macroporosity to infiltration into a continuous corn no-tilied watershed: Implications for contaminant movement. J. Contam. Hydrol., 3: 193-205. Edwards, W.M., Shipitalo, M.J., Owens, L.B. and Norton, L.D., 1989. Water and nitrate movement in earthworm burrows within long-term no-till cornfields. J. Soil Water Conserv., 44: 240-243. Germann, P.F., Edwards, W.M. and Owens, L.B., 1984. Profiles of bromide and increased soil moisture after infiltration into soils with macropores. Soil Sci. Sot. Am. J., 48: 237-244. Hall, J.K. and Mumma, R.O., 1994. Dicamba mobility in conventionally tilled and non-tilled soil. Soil Tillage Res., 30: 3-17. Hall, J.K., Hartwig, N.L. and Hoffman, L.D., 1983. Application mode and alternate cropping effects on atrazine losses from a hillside. J. Environ. Qual., 12: 336-340. Hall, J.K., Murray, M.R. and Hartwig, N.L., 1989. Herbicide leaching and distribution in tilled and untilled soil. J. Environ. Qual., 18: 439-445.
D.W. WUNS, J.K. Hull/Soil
& Tillage Research
39 (1996)
241-257
257
Hall, J.K., Mumma, R.O. and Watts, D.W.. 1991. Leaching and runoff losses of herbicides in a tilled and untilled field. Agric. Ecosyst. Environ., 37: 303-314. Isensee, A.R., Nash, R.G. and Helling, C.S., 1990. Effect of conventional vs. no-tillage on pesticide leaching to shallow groundwater. J. Environ. Qual., 19: 434-440. Onstad, C.A. and Voorhees, W.B., 1987. Hydrologic soil parameters affected by tillage. In: T.J. Logan, J.M. Davidson, J.L. Baker and M.R. Overcash (Editors), Effects of Conservation Tillage on Groundwater Quality - Nitrates and Pesticides. Lewis Publ., Chelsea, MI, pp. 9% 124. Phillips, R.E., Quisenberry, V.L., Zeleznik, J.M. and Dunn, G.H., 1989. Mechanism of water entry into simulated macropores. Soil Sci. Sot. Am. J., 53: 1629-1635. Quisenberry, V.L. and Phillips, R.E., 1976. Percolation of surface-applied water in the field. Soil Sci. Sot. Am. J., 40: 484-489. * Shipitalo, M.J., Edwards, W.M., Dick, W.A. and Owens, L.B., 1990. Initial storm effects on macropore transport of surface-applied chemicals in no-till soil. Soil Sci. Sot. Am. J., 54: lS30- 1536. Stehouwer, R.C., Dick, W.A. and Traina, S.J., 1993. Characteristics of earthworm burrow lining affecting atrazine sorption. J. Environ. Qual.. 22: 181-185. Triplett, G.B.. Jr., Conner. B.J. and Edwards, W.M., 1978. Transport of atrazine and simazine in runoff from conventional and no-tillage corn. J. Environ. Qual., 7: 77-83. Watts. D.W., Bogus, E.R., Hall, J.K. and Mumma, R.O., 1994. Simultaneous extraction of six pesticides using a dual column extraction procedure. J. Environ. Qual., 23: 383-386. Wauchope, R.D., 1978. The pesticide content of surface water draining from agricultural fields - A review. J. Environ Qual., 7: 459-472.
