- Email: [email protected]
Influence of climatic and edaphic factors on persistence of glyphosate and 2,4-D in forest soils
ECOTOXICOLOGY
AND
ENVIRONMENTAL
SAFETY
18,230-239
(1989)
Influence of Climatic and Edaphic Factors on Persistence Glyphosate and 2,4-D in Forest Soils N. T. L. TORSTENSSON,* L. N. LUNDGREN,~
of
AND J. STENSTR~M*
*Department of Microbiology and TDepartment of Chemistry, Swedish University of Agricultural Sciences, S- 750 07 Uppsala, Sweden Received October IS.1988 Persistence in soil of the two herbicides glyphosate and 2,4-D was investigated after application for brush control in conifer reforestation areas. Field experiments were carried out at five sites in southern Sweden and six in northern Sweden. Initially, glyphosate disappeared faster in northern soils than in southern soils. This was probably a result of the higher biological activity in the northern soils. However, small amounts of glyphosate were detected in the northern area long after all traces had disappeared in the south, presumably because of the long period during which the soil remained frozen in the northern area and because ofthe slow release of vegetationbound herbicide. One metabolite of glyphosate, aminomethylphosphonic acid (AMPA), persisted longer than glyphosate itself. After 2 years, 8% of the theoretical amount was found in the northern area and, after 1 year, 1% in the southern area. 2,4-D disappeared rapidly from all sites, although minor amounts persisted for several years, probably because of slow release from vegetation-bound residues. 0 1989 Academic press, hc.
INTRODUCTION Herbicides are used in forestry to control unwanted brush vegetation. Various phenoxy acids (2,4-D, 2,4,5-T, and MCPA) have been used in Sweden for this purpose (Barring, 1978). In 1977, spraying with 2,4,5-T was prohibited, and subsequently the herbicide glyphosate has been tested and found valuable for controlling brush (LundHsie, 1985). To judge the risks of possible secondary effects when using herbicides, it is essential to understand their behavior in soil, since most of them will sooner or later reach the soil either directly after spraying or later through downward movements of the residues retained by vegetation. Prediction of the length of time needed for detoxification of herbicides requires detailed knowledge of the specific degradation processes and of the factors influencing their rates. Herbicides can be degraded via photochemical, chemical, and biological reactions. The biological route, in which soil bacteria and fungi play a major role, is the most important source of detoxification (Torstensson, 1980,1987). Research on the degradation of phenoxy acids has been reviewed by Torstensson (1978) and Norris (198 I), and glyphosate studies were reviewed by Torstensson ( 1985). Comprehensive research on the ,degradation of phenoxy acids and glyphosate in agricultural soils has been conducted at many laboratories throughout the world, whereas studies of similar degradation processes in forest soil are scarce. However, certain characteristic differences (e.g., organic matter content, pH) between agricultural and forest soils can predictably lead to differences in the degradation rates of 0 147-65 13/89 $3.00 Copj~@t Q 1989 by Academic Press. Inc. All rights of reproduction in any form reserved.
230
PERSISTENCE
FIG.
OF
1. Location
GLYPHOSATE
of the experimental
AND
2,4-D
231
sites.
herbicides (Torstensson, 1975; Stark, 1982). Consequently, the conclusions based on herbicide persistence tests in agricultural soils are not applicable to forest soils. The investigation presented here was carried out mainly to study the influence of climatic conditions on the persistence of 2,4-D, (2,4-dichlorophenoxy)acetic acid, and glyphosate, IV-(phosphonomethyl)glycine, after application of the herbicides for brush control in a conifer reforestation area. Two areas were chosen for the field experiments, one in the southern part of Sweden and one in the northern part. Study plots within the areas were selected to represent different soil types. MATERIALS
AND
METHODS
Field sites. Field experiments were conducted during 1982-1987 in two climatically different areas of Sweden (Fig. 1). Six sites (l-6) were located close to the Arctic Circle (about 66”3O’ N) and five sites were located (7-l 1) in the south (about 57”30’ N). Some characteristics of the experimental sites are given in Table 1. All sites were clear-cut and replanted with Scats pine (Pinus sylvestris L.). At sites l-6 the ground vegetation consisted of herbs, grasses, and the shrub species Vaccinium myrtillus L., V. vitis idaea L., V. uliginosum L., Empetrum nigrum L., and Calluna vulgaris L. At sites 7- 11 the grasses dominated, while V. vitis idaea L., E. nigrum L., and C. vulgaris L. were less commonly found.
232
TORSTENSSON,
LUNDGREN,
AND STENSTRiiM
TABLE 1 SOILTYPEANDSOMECHARACTERISTICSOFTHEUPPERMOSTO-TO 5-cm LAYER ATTHEVARIOUSEXPERIMENTALSITES
Characteristics of the O- to 5-cm layer
Site
Altitude (n-4
:
115 100
3 4 5
160 170 215
6
295
iit 9
135 190 310
10
300
11
50
Loss on ignition
“Basic” respiration
Soil type
PH
@)
Gl Whr/=wW
Iron podsol pods01 Iron podsol Iron podsol Iron humus pods01 Iron podsol
4.7 5.4
57 25
494 430
5.3 5.0 4.9
21 60
81
495 409 562
4.9
40
527
Iron podsol Brown soil Iron humus podsol Iron humus podsol Iron podsol
5.5 5.0 4.5
33 21 40
267 407 382
4.4
24
340
5.5
11
330
Note. “Basic” respiration accordingto Torstensson andStenstriim ( 1986).
Chemicals and application. In the field experiments plots of 5 X 5 m were used with three replicates and a 2-m-wide free zone between the plots. The herbicide formulation used for glyphosate was Roundup (glyphosate-isopropyl-ammonium, 360 g a.i./liter, Monsanto) and that for 2,4-D was Gullviks 2,4-D ester (2,4-dichlorophenoxyacetic-butoxyethyl ester, 500 g a.i./liter, Gullviks), both at a dose of 2 kg a.i./ ha. A Sonett knapsack sprayer was used to apply the herbicides. The herbicide was dissolved in 2 liters water per plot and sprayed on the experimental plots in two directions at right angles to each other to ensure that the spray was distributed as evenly as possible. Spraying was according to the normal schedule used for coniferous reforestation: on lo- 12 August 1982 at sites 1-6 and on 1S- 19 August 1983 at sites 7- 11. Soil samples. Soil samples taken after the spraying were used for analysis of herbicide residues and measurements of respiration, pH, and ignition loss. Because of the low mobility of 2,4-D and glyphosate under soil conditions similar to those found here (Stark, 1982; Torstensson, 1985), all samples were taken from the 0- to 5-cm layer. The samples for residual analysis were collected using a sampler with a surface area of 1.5 cm*, apart from samples for analysis of glyphosate from sites l-6 which were taken using a sampler with a surface area of 10 cm*. The latter sampler was also used for all other types of samples. Five samples were taken per plot; thus 15 samples were collected for each herbicide. Samples used for measurements of respiration, pH, ignition loss, and those serving as background values for residue analysis were taken from the control plots, 3 X 5 samples per site on each sampling occasion. Samples of dead birch leaves and branches were collected for residue analysis of 2,4-D from four of the experimental sites ( 1, 5, 8, and 10). Immediately following sampling, all sam-
PERSISTENCE
OF GLYPHOSATE
AND 2,4-D
233
ples were frozen at -20°C. The material was transported to Uppsala from the experimental sites in thermoboxes and then stored at -20°C until the analysis. The first sampling was done on the day of spraying, and two or three samples were taken during the late summer and autumn. The last sample was collected as late as possible during the autumn but before the first ground frost (sites l-6 in late September and sites 7- 11 in mid-November). The first sampling in the following spring was done in late May at sites l-6 and in early May at sites 7- 11. Analytical procedures. Analyses of glyphosate and its metabolite, aminomethylphosphonic acid (AMPA), were done according to Lundgren (1986). Frozen samples were combined and analyzed together as follows. Samples from sites l-6 were dried at 100°C for 15 hr, milled to a powder, and carefully mixed. To a small part of each sample (3.5 g) was added a 2.5 ppm aqueous N-(phosphonomethyl)+alanin solution (3.0 ml, internal standard). After 15 hr at 10°C the samples were extracted and analyzed according to Lundgren (1986). Correction factors for drying of the samples, subsequent extraction, working up, and HPLC procedures were determined for the various soils by analyzing dried and powdered control samples from sites l-6 spiked to a concentration of 2.86 pg g-’ with glyphosate and AMPA as described above. Control soils were also analyzed to determine the background values. The drying process for samples from sites l-6 was found to be rather laborious. For samples from sites 7-l 1, the amount of each sample was reduced to fit for extraction of whole samples. Samples from sites 7- 11 were placed in IOOO-ml Erlenmeyer flasks, 20 ppm internal standard solution ( 10 ml) was added, and the flasks were placed at 10°C for 15 hr. Subsequently, aqueous 0.1 M triethylamine solution (300 ml) was added and the flasks were shaken for 1 hr at room temperature. A small volume (30 ml) of each extract was centrifuged at 23008 for 10 min, and the concentrations of glyphosate and AMPA in the supematant were determined according to Lundgren ( 1986). Correction factors for extraction, working up, and HPLC procedures were determined for the various soils by analyzing control soils from sites 7-l 1 (65 g fresh wt) spiked with 100 pg glyphosate and AMPA. Control soils were also analyzed to provide background values. The results are expressed as micrograms of glyphosate or AMPA per soil core, 1.5 cm2 X 5 cm. The limit of detection was 0.05 pg per core for both glyphosate and AMPA. Analysis of 2,4-D was according to Stark (1982). Frozen samples were thawed and analyzed separately. The soil was transferred to test tubes and 4 ml of acetone acidified with phosphoric acid was added. The tubes were then shaken for 1 min and 6 ml chloroform was added. Again the tubes were shaken (1 min X 3 at hourly intervals) and then filtered. The internal standard, 2,4,5-trichlorophenoxyacetic acid, was added to the filtrate and then evaporated to dryness. The residue was redissolved in 2 ml borontrifluoride in methanol and heated in a water bath at 60°C for 1 hr. Two milliliters of saturated NaCl solution was added and shaken for 1 min, followed by the addition of 2 ml toluene. The solution was shaken again for 1 min, and the toluene phase retained. 2,4-D was determined on a Carlo Erba gas chromatograph with an OV 101 (3%) column and EC detector. Temperatures used were injector, 225°C; column, 180°C; and detector, 250°C. Recovery varied between 68 and 78%. The results are expressed as micrograms of 2,4-D per soil core, 1.5 cm2 X 5 cm. The limit of detection was 0.05 pg per core. The vegetation residues were milled before the analysis of 2,4-D with the same method as that used for the soil samples. The “basic” respiration rate was measured according to Torstensson and StenStrom (1986). Soil cores were transferred to an incubation vessel designed to fit a
Glyphosate AMPA 2,4-D
Glyphosate AMPA 2,4-D
Glyphosate AMPA 2,4-D
735 (2 ye=@
1080 (3 ye=-@
1463 (4 YeaM
Note. For 2,4-D x + SE, n = 15. a &soil core, 1.5 cm* X 5 cm. b Not determined.
Glyphosate AMPA 2,4-D
243 (1 yead
( 1St spring)
Glyphosate AMPA 2,4-D
283
co.05 co.05 <0.05
0.1 + 0.1
<0.05 0.3
0.3 0.9 0.6 + 0.2
0.8 1.7 0.3 f 0.1
0.9 1.4 3.5 f 0.6
3.4’r;6.9
2.4
9.1 x0.05 9.7 & 1.0
Glyphosate AMPA 2,4-D
Glyphosate AMPA 2,4-D
V-g)
1
Substance
47 @ePt)
0
Time (days)
AMOUNTS OF GLYPHOSATE, AMPA,
co.05
0.1
-co.05
0.1 0.9 <0.05
0.6 2.5 0.2kO.l
it: Nbb
0.7 1.4 3.3 f 0.4
2.2 2.0 1.3f 0.2
12.8 co.05 9.2 zt 1.5
2
<0.05 -co.05 <0.05
-co.05 0.2 0.1 f 0.1
0.6’;r90.2
0.3
0.2Yo. 1
0.6
1.4 1.5 4.9 f 0.7
2.0 0.9 0.5 It 0.1
co.05 9.1 f 1.3
10.0
3
Site
x0.05 0.7 -co.05
0.2 3.1 -co.05
0.9 5.3 0.4kO.l
2.6 3.1 0.9 f 0.4
3.4 3.5 2.4 + 0.5
4.2 3.3 1.5z!z0.3
12.5 co.05 9.8 + 1.0
4
AND 2,4-D” FOUND AT THE NORTHERLY EXPERIMENTAL
TABLE 2
0.2
<0.05 0.2 co.05
0.2 1.7 0.1 + 0.1
0.;:
1.0
E 1.4io.7
0.8 2.0 3.0 + 0.4
1.9 1.0 1.1 f 0.3
-co.05 10.6+ 1.1
11.6
5
SITES
<0.05
0.1
<0.05
<0.05 0.6 x0.05
0.2 1.2 0.5kO.l
1.1 kO.3
1.9 3.2 1.7 zk0.5
13.0 <0.05 10.0+0.8
6
3 2 g 3 H:
p
“8
3 g ;;I z M
E
PERSISTENCE
OF GLYPHOSATE TABLE
AMOUNTSOFGLYPHOSATE,
235
AND 2,4-D
3
AMPA, AND~,~-D*FOUNDATTHESOUTHERLY EXPERIMENTALSITES Site
Time (days)
Substance
7
8
9
10
11
(Ad
Glyphosate AMPA 2,4-D
13.5 <0.05 13.0 + 1.5
19.7 co.05 12.3 2 1.6
17.7 <0.05 13.2 + 1.2
16.3 to.05 12.0 f 1.5
15.0 co.05 14.8 + 0.6
(SW)
Glyphosate AMPA 2,4-D
8.1 1.3 2.0 Ik 0.5
6.8 1.1 2.1 + 0.6
WV)
Glyphosate AMPA 2,4-D
1.9 0.4 1.2k0.3
1.3 1.9 1.3 -t 0.2
265 ( 1 st spring)
Glyphosate AMPA 2,4-D
0.9 1.5 0.6 f 0.3
380 (1 yea)
Glyphosate AMPA 2,4-D
0.2 0.4 0.6 -+ 0.2
732
Glyphosate AMPA 2,4-D
0
53
87
(2 YeaN
-co.05 0.1 <0.05
5.0 co.05 ND’
5.3 0.3 1.3 kO.3
i4; 1.9iO.l
1.9 0.4 0.8 t- 0.2
2.5 0.6 0.7 kO.1
2.2 0.7 0.5 kO.1
1.3 1.9 0.8 -t 0.2
0.4 0.4 0.5 -t 0.2
0.2 0.6 0.5 + 0.1
1.2 0.5 0.3 50.1
0.2 0.4 0.3 +- 0.2
co.05 0.3 0.1 f 0.1
-co.05 <0.05 0.2kO.l
0.6 0.4 -co.05
x0.05 0.2 co.05
x0.05 co.05 <0.05
-co.05 <0.05 -co.05
<0.05 0.1 -co.05
Note. For 2,4-D x f SE, n = 15. ’ &soil core, 1.5 cm* X 5 cm. ’ Not determined.
Gilson respirometer. They were then equilibrated for 5 days at 20°C. Next, the rate of respiration (R) was determined. The results are expressed as microliters of O2 used per hour and per soil core, 10 cm* X 5 cm. Soil pH was determined in water (soil: water = 1:2). RESULTS
AND
DISCUSSION
The amounts of glyphosate, AMPA, and 2,4-D found in the soil samples at various times after their application are given in Tables 2 and 3. Glyphosate reached low levels faster at sites l-6 than at sites 7- 11 during the weeks after application. On the last sampling occasion before winter at sites l-6,47 days after application, the mean amount of glyphosate per sample was 2.4 pg compared with 6.5 pg at sites 7-l 1 after 53 days. Not until 87 days after application did the mean amount of residue reach the same low level at sites 7- 11. This might be an effect of higher amounts of glyphosate reaching the soils at sites 7-l 1 (about 50% of applied) than at sites l-6 (about 40%) as a consequence of the lower frequency of cover plants on the ground. However, there might also be other explanations. Glyphosate is not directly utilized as an energy substrate by soil microorganisms (Sprankle et al., 1975; Nomura and Hilton, 1977; Torstensson and Aamisepp, 1977)
236
TORSTENSSON,
0.14 r
LUNDGREN,
AND STENSTRijM
a= O.OiB93+O.ooO18 R rr0.9821
0.068 200
400 R (plOph-'1
600
FIG. 2. Correlation between rate of respiration (I?) and capability to decompose glyphosate (a) in field experiments 1- 11.
and undergoes cometabolic decomposition (Torstensson, 1980); i.e., the energy derived does not support microbial growth. Microbial production of enzymes necessary for breaking down glyphosate is not induced by the herbicide itself but by other sub strates for the microorganisms. Thus, one would expect cometabolic processes to depend on the general microbial activity of the soil, as has been found for glyphosate (Lijnsjii et al., 1980; Torstensson and Stenstrom, 1986). The data on the decomposition of glyphosate from application to the last sampling in the first autumn can be linearized by the formula c=co-kft, where c is the actual concentration of glyphosate at time t, co is the initial concentration, and k is the rate constant (Torstensson and Stenstriim, 1986). However, the dose reaching the soil in the different field experiments varies due to the vegetation cover (Tables 2 and 3), and the rate constant k is directly proportional to co when co is low (Stenstrom, 1989). Thus, k= acoanda
= k/co,
where the quotient a is a co-independent rate constant. Therefore, a, but not k, can be used to compare, for example, the decomposition of a pesticide in different soils. A plot of a against the “basic” respiration rate of the different soils (Table 1) shows a high correlation between them (Fig. 2). This correlation between the biological activity of soils and the capacity to degrade glyphosate has been found previously for a number of forest and agricultural soils (Torstensson and Stenstriim, 1986). It is also noteworthy that the northern soils ( 1-6) have a higher capacity to degrade glyphosate than the southern ones (7- 11, Fig. 2). During winter, with ground frost for about 7 months at sites l-6 and 4 months at sites 7- 11, the reduction in glyphosate level was slight (Tables 2 and 3). Moreover, during the following summer, the disappearance was comparatively slow at all sites, and even 2 years after application 0.2- 1.O pg glyphosate per sample was found at sites
PERSISTENCE
OF
GLYPHOSATE
AND
237
2,4-D
TABLE 4 AMOUNTS
OF
2,4-D FOUND IN SAMPLESOF BIRCH RESIDUES” Site
Time I year
Vegetation residues Leaves Branches
2 years
Leaves
3 years
Branches Branches
1
5
8
10
20 60
11 35
’ mg/kg dry wt. ’ Not determined.
l-6 (Table 2). The slow disappearance of the minor residues of glyphosate was not thought to be caused by insufficient capacity of the soils to degrade the herbicide but by adsorbed glyphosate being slowly released from binding sites in vegetation residues (Coupland, 1985). The metabolite AMPA from glyphosate is known to accumulate in soils and has been previously reported to decompose more slowly than the parent compound (Comes et al., 1976; Nomura and Hilton, 1977; Rueppel et al., 1977). In this investigation, the accumulation was most pronounced in the northern soils (Table 2). At site 4, soil samples contained as much as 5.3 pg AMPA per sample 2 years after the application of glyphosate, which is 18% of the theoretical maximum. On average, soils from sites l-6 contained 2.3 pg AMPA per sample after 2 years (8%) while soils from sites 7- 11 contained only 0.3 pg per sample after 1 year ( 1%). After application of glyphosate and 2,4-D in August, there was a clear difference between their patterns of disappearance until late September. The 2,4-D disappeared rapidly from all sites, and about 7 weeks after application only 0.5-2.1 pg 2,4-D per sample was found on sites 2-l 1, i.e., about 2-7% of the amount originally applied. Larger amounts were found only in samples from site 1 (3.4 pg per sample). This indicates that 2,4-D is rapidly utilized as an energy substrate for metabolic decomposition by soil microorganisms, as discussed by Torstensson ( 1978, 1980) and Norris ( 198 1). The important rate-determining edaphic and climatic factors, water, pH, and temperature, do not seem to have been limiting to any appreciable extent during the period. Instead, the rate was probably determined by the supply of the energy substrate 2,4-D to those soil microorganisms able to utilize it. However, the small amounts of 2,4-D remaining after the winter period (Table 2 and 3) were persistent, as demonstrated by the fact that 0.1-0.7 pg per sample was found in soils from sites l-6 2 years after application. This is probably due to the persistence of 2,4-D in vegetation residues (Table 4; Norris, 198 1; Torstensson, 1983). The length of time during which small amounts of both glyphosate and 2,4-D were detected in soil samples differed between sites l-6 and 7- 11. Certain amounts of the herbicides bound in vegetation residues would be protected from decomposition until they were either desorbed or decomposed together with the vegetation residues. During the first winter, desorption of 2,4-D apparently occurred at sites l-6 (Table
238
TORSTENSSON,
LUNDGREN,
AND STENSTRGM
2). Desorption of 2,4-D has also been observed in a forest area unintentionally exposed to an overdose (Torstensson, 1983). However, decomposition of vegetationbound herbicides together with vegetation residues seems to be a probable explanation for the long persistence ofsmall amounts of 2,4-D and probably also glyphosate as found in this investigation. The divergence between the two groups of sites can then easily be explained on the basis of different vegetation-residue decomposition rates as a consequence of different climatic conditions. CONCLUSIONS (i) After application of glyphosate or 2,4-D in mid-August to reforestation areas, most of both herbicides dissipated rapidly during the autumn. However, minor amounts of both herbicides were retained by vegetation residues and then released during a period of up to 1 year in the southerly areas and up to 3 years in the northerly areas. (ii) There was a high correlation (Y = 0.98) between the rate of respiration of the soils and the normalized co-independent rate constant for dissipation of the cometabolically decomposed glyphosate. (iii) The soils of experimental sites in northern areas generally had a higher rate of respiration and, thus, according to (ii) above, a higher rate of dissipation of glyphosate during the first autumn. ACKNOWLEDGMENTS We thank 0. Forsberg and L. Funke for technical assistance. This work was supported by a grant from the Swedish Council for Forestry and Agricultural Research.
REFERENCES B;~RRING, U. S. M. (1978). The use of phenoxy herbicides in Swedish forestry: Amounts, types, and mode of application. Ecol. Bull. 27,219-230. COMES, R. D., BRUNS, V. F., AND KELLY, A. D. (1976). Residues and persistence ofglyphosate in irrigation water. Weed Sci. 24,47-50. COUPLAND, D. (1985). Metabolism of glyphosate in plants. In The Herbicide Glyphosute (E. Grossbard, and D. Atkinson, Eds.) pp. 25-34. Butterworths, London. L~~NsJG, H., STARK, J., TORSTENSSON, L., AND WESSBN, B. (1980). Glyphosate: Decomposition and LUNDGREN, L. N. (1986). A new method for the determination ofglyphosate and (aminomethyl)phosphonit acid residues in soils. J. Agric. Food Chem. 34,535-538. LUNDHBIE, K. (1985). Efficiency of glyphosate in forest plantations. In The Herbicide Glyphosute (E. Grossbard and D. Atkinson, Eds.) pp. 328-338. Butterworths, London. NOMURA, N. S., AND HILTON, H. W. (1977). The adsorption and degradation of glyphosate in five Hawaiian sugarcane soils. Weed Res. 17, 113- 12 1. NORRIS, L. A. (198 1). The movement, persistence, and fate of the phenoxy herbicides and TCDD in the forest. Residue Rev. 80,65-l 35. RUEPPEL, M. L., BRIGHTWELL, B. B., SCHAEFER,J., AND MARVEL, J. T. (1977). Metabolism and degradation of glyphosate in soil and water. J. Agric. Food Chem. 25,5 17-528. SPRANKLE, P., MEGGI’IT, W. F., AND PENNER, D. (1975). Adsorption, mobility and microbial degradation of glyphosate in the soil. Weed Sci. 23,229-234. STARK, J. (1982). Persistence of herbicides in forest soils. Dissertation, Swedish University of Agricultural Sciences, Department of Microbiology, Uppsala, Report 15. STENSTR~~M,J. (1989). Quantitative assessment of herbicide decomposition at different initial concentrations. Toxic. Ass. 4,53-70.
PERSISTENCE
OF
GLYPHOSATE
AND
2,4-D
239
TORSTENSSON,N. T. L. (1975). Degradation of 2,4-D and MCPA in soils of low pH. Environ. Qual. Suf Suppl. 3,262-265. TORSTENSSON, L. (1978). Effects of phenoxyacetic acid herbicides on soil organisms. Ecol. Bull. 27,263284.
TORSTENSSON, L. (1980). Role of microorganisms in decomposition. In Interactions Between Herbicides and the Soil (R. J. Hance, Ed.) pp. 159-l 78. Academic Press, London. TORSTENSSON,L. ( 1983). Undersijkning av 2,4-D:s persistens och effekter pa markorganismer i skogsmark efter oavsiktlig overdosering. Swedish Environmental Protection Board PM 17 16. (In Swedish). TORSTENSSON, L. (1985). Behaviour of glyphosate in soils and its degradation. In The Herbicide Glyphosate (E. Grossbard and D. Atkinson, Eds.), pp. 137- 150. Butterworths, London. TORSTENSSON, N. T. L. (1987). Microbial decomposition of herbicides in soil. In Herbicides: Progress in Pesticide Biochemistry and Toxicology (D. H. Hutson and T. R. Roberts, Eds.), Vol. 6, pp. 249-270. Wiley, New York. TORSTENSSON, N. T. L., AND AAMISEPP, A. (1977). Detoxification of glyphosate in soil. Weed Res. 17, 209-2 12. TORSTENSSON, L., AND STENSTR~M, J. (1986). “Basic” respiration as a tool for prediction of pesticide persistence in soil. Toxic. Ass. 1,57-72.
AND
ENVIRONMENTAL
SAFETY
18,230-239
(1989)
Influence of Climatic and Edaphic Factors on Persistence Glyphosate and 2,4-D in Forest Soils N. T. L. TORSTENSSON,* L. N. LUNDGREN,~
of
AND J. STENSTR~M*
*Department of Microbiology and TDepartment of Chemistry, Swedish University of Agricultural Sciences, S- 750 07 Uppsala, Sweden Received October IS.1988 Persistence in soil of the two herbicides glyphosate and 2,4-D was investigated after application for brush control in conifer reforestation areas. Field experiments were carried out at five sites in southern Sweden and six in northern Sweden. Initially, glyphosate disappeared faster in northern soils than in southern soils. This was probably a result of the higher biological activity in the northern soils. However, small amounts of glyphosate were detected in the northern area long after all traces had disappeared in the south, presumably because of the long period during which the soil remained frozen in the northern area and because ofthe slow release of vegetationbound herbicide. One metabolite of glyphosate, aminomethylphosphonic acid (AMPA), persisted longer than glyphosate itself. After 2 years, 8% of the theoretical amount was found in the northern area and, after 1 year, 1% in the southern area. 2,4-D disappeared rapidly from all sites, although minor amounts persisted for several years, probably because of slow release from vegetation-bound residues. 0 1989 Academic press, hc.
INTRODUCTION Herbicides are used in forestry to control unwanted brush vegetation. Various phenoxy acids (2,4-D, 2,4,5-T, and MCPA) have been used in Sweden for this purpose (Barring, 1978). In 1977, spraying with 2,4,5-T was prohibited, and subsequently the herbicide glyphosate has been tested and found valuable for controlling brush (LundHsie, 1985). To judge the risks of possible secondary effects when using herbicides, it is essential to understand their behavior in soil, since most of them will sooner or later reach the soil either directly after spraying or later through downward movements of the residues retained by vegetation. Prediction of the length of time needed for detoxification of herbicides requires detailed knowledge of the specific degradation processes and of the factors influencing their rates. Herbicides can be degraded via photochemical, chemical, and biological reactions. The biological route, in which soil bacteria and fungi play a major role, is the most important source of detoxification (Torstensson, 1980,1987). Research on the degradation of phenoxy acids has been reviewed by Torstensson (1978) and Norris (198 I), and glyphosate studies were reviewed by Torstensson ( 1985). Comprehensive research on the ,degradation of phenoxy acids and glyphosate in agricultural soils has been conducted at many laboratories throughout the world, whereas studies of similar degradation processes in forest soil are scarce. However, certain characteristic differences (e.g., organic matter content, pH) between agricultural and forest soils can predictably lead to differences in the degradation rates of 0 147-65 13/89 $3.00 Copj~@t Q 1989 by Academic Press. Inc. All rights of reproduction in any form reserved.
230
PERSISTENCE
FIG.
OF
1. Location
GLYPHOSATE
of the experimental
AND
2,4-D
231
sites.
herbicides (Torstensson, 1975; Stark, 1982). Consequently, the conclusions based on herbicide persistence tests in agricultural soils are not applicable to forest soils. The investigation presented here was carried out mainly to study the influence of climatic conditions on the persistence of 2,4-D, (2,4-dichlorophenoxy)acetic acid, and glyphosate, IV-(phosphonomethyl)glycine, after application of the herbicides for brush control in a conifer reforestation area. Two areas were chosen for the field experiments, one in the southern part of Sweden and one in the northern part. Study plots within the areas were selected to represent different soil types. MATERIALS
AND
METHODS
Field sites. Field experiments were conducted during 1982-1987 in two climatically different areas of Sweden (Fig. 1). Six sites (l-6) were located close to the Arctic Circle (about 66”3O’ N) and five sites were located (7-l 1) in the south (about 57”30’ N). Some characteristics of the experimental sites are given in Table 1. All sites were clear-cut and replanted with Scats pine (Pinus sylvestris L.). At sites l-6 the ground vegetation consisted of herbs, grasses, and the shrub species Vaccinium myrtillus L., V. vitis idaea L., V. uliginosum L., Empetrum nigrum L., and Calluna vulgaris L. At sites 7- 11 the grasses dominated, while V. vitis idaea L., E. nigrum L., and C. vulgaris L. were less commonly found.
232
TORSTENSSON,
LUNDGREN,
AND STENSTRiiM
TABLE 1 SOILTYPEANDSOMECHARACTERISTICSOFTHEUPPERMOSTO-TO 5-cm LAYER ATTHEVARIOUSEXPERIMENTALSITES
Characteristics of the O- to 5-cm layer
Site
Altitude (n-4
:
115 100
3 4 5
160 170 215
6
295
iit 9
135 190 310
10
300
11
50
Loss on ignition
“Basic” respiration
Soil type
PH
@)
Gl Whr/=wW
Iron podsol pods01 Iron podsol Iron podsol Iron humus pods01 Iron podsol
4.7 5.4
57 25
494 430
5.3 5.0 4.9
21 60
81
495 409 562
4.9
40
527
Iron podsol Brown soil Iron humus podsol Iron humus podsol Iron podsol
5.5 5.0 4.5
33 21 40
267 407 382
4.4
24
340
5.5
11
330
Note. “Basic” respiration accordingto Torstensson andStenstriim ( 1986).
Chemicals and application. In the field experiments plots of 5 X 5 m were used with three replicates and a 2-m-wide free zone between the plots. The herbicide formulation used for glyphosate was Roundup (glyphosate-isopropyl-ammonium, 360 g a.i./liter, Monsanto) and that for 2,4-D was Gullviks 2,4-D ester (2,4-dichlorophenoxyacetic-butoxyethyl ester, 500 g a.i./liter, Gullviks), both at a dose of 2 kg a.i./ ha. A Sonett knapsack sprayer was used to apply the herbicides. The herbicide was dissolved in 2 liters water per plot and sprayed on the experimental plots in two directions at right angles to each other to ensure that the spray was distributed as evenly as possible. Spraying was according to the normal schedule used for coniferous reforestation: on lo- 12 August 1982 at sites 1-6 and on 1S- 19 August 1983 at sites 7- 11. Soil samples. Soil samples taken after the spraying were used for analysis of herbicide residues and measurements of respiration, pH, and ignition loss. Because of the low mobility of 2,4-D and glyphosate under soil conditions similar to those found here (Stark, 1982; Torstensson, 1985), all samples were taken from the 0- to 5-cm layer. The samples for residual analysis were collected using a sampler with a surface area of 1.5 cm*, apart from samples for analysis of glyphosate from sites l-6 which were taken using a sampler with a surface area of 10 cm*. The latter sampler was also used for all other types of samples. Five samples were taken per plot; thus 15 samples were collected for each herbicide. Samples used for measurements of respiration, pH, ignition loss, and those serving as background values for residue analysis were taken from the control plots, 3 X 5 samples per site on each sampling occasion. Samples of dead birch leaves and branches were collected for residue analysis of 2,4-D from four of the experimental sites ( 1, 5, 8, and 10). Immediately following sampling, all sam-
PERSISTENCE
OF GLYPHOSATE
AND 2,4-D
233
ples were frozen at -20°C. The material was transported to Uppsala from the experimental sites in thermoboxes and then stored at -20°C until the analysis. The first sampling was done on the day of spraying, and two or three samples were taken during the late summer and autumn. The last sample was collected as late as possible during the autumn but before the first ground frost (sites l-6 in late September and sites 7- 11 in mid-November). The first sampling in the following spring was done in late May at sites l-6 and in early May at sites 7- 11. Analytical procedures. Analyses of glyphosate and its metabolite, aminomethylphosphonic acid (AMPA), were done according to Lundgren (1986). Frozen samples were combined and analyzed together as follows. Samples from sites l-6 were dried at 100°C for 15 hr, milled to a powder, and carefully mixed. To a small part of each sample (3.5 g) was added a 2.5 ppm aqueous N-(phosphonomethyl)+alanin solution (3.0 ml, internal standard). After 15 hr at 10°C the samples were extracted and analyzed according to Lundgren (1986). Correction factors for drying of the samples, subsequent extraction, working up, and HPLC procedures were determined for the various soils by analyzing dried and powdered control samples from sites l-6 spiked to a concentration of 2.86 pg g-’ with glyphosate and AMPA as described above. Control soils were also analyzed to determine the background values. The drying process for samples from sites l-6 was found to be rather laborious. For samples from sites 7-l 1, the amount of each sample was reduced to fit for extraction of whole samples. Samples from sites 7- 11 were placed in IOOO-ml Erlenmeyer flasks, 20 ppm internal standard solution ( 10 ml) was added, and the flasks were placed at 10°C for 15 hr. Subsequently, aqueous 0.1 M triethylamine solution (300 ml) was added and the flasks were shaken for 1 hr at room temperature. A small volume (30 ml) of each extract was centrifuged at 23008 for 10 min, and the concentrations of glyphosate and AMPA in the supematant were determined according to Lundgren ( 1986). Correction factors for extraction, working up, and HPLC procedures were determined for the various soils by analyzing control soils from sites 7-l 1 (65 g fresh wt) spiked with 100 pg glyphosate and AMPA. Control soils were also analyzed to provide background values. The results are expressed as micrograms of glyphosate or AMPA per soil core, 1.5 cm2 X 5 cm. The limit of detection was 0.05 pg per core for both glyphosate and AMPA. Analysis of 2,4-D was according to Stark (1982). Frozen samples were thawed and analyzed separately. The soil was transferred to test tubes and 4 ml of acetone acidified with phosphoric acid was added. The tubes were then shaken for 1 min and 6 ml chloroform was added. Again the tubes were shaken (1 min X 3 at hourly intervals) and then filtered. The internal standard, 2,4,5-trichlorophenoxyacetic acid, was added to the filtrate and then evaporated to dryness. The residue was redissolved in 2 ml borontrifluoride in methanol and heated in a water bath at 60°C for 1 hr. Two milliliters of saturated NaCl solution was added and shaken for 1 min, followed by the addition of 2 ml toluene. The solution was shaken again for 1 min, and the toluene phase retained. 2,4-D was determined on a Carlo Erba gas chromatograph with an OV 101 (3%) column and EC detector. Temperatures used were injector, 225°C; column, 180°C; and detector, 250°C. Recovery varied between 68 and 78%. The results are expressed as micrograms of 2,4-D per soil core, 1.5 cm2 X 5 cm. The limit of detection was 0.05 pg per core. The vegetation residues were milled before the analysis of 2,4-D with the same method as that used for the soil samples. The “basic” respiration rate was measured according to Torstensson and StenStrom (1986). Soil cores were transferred to an incubation vessel designed to fit a
Glyphosate AMPA 2,4-D
Glyphosate AMPA 2,4-D
Glyphosate AMPA 2,4-D
735 (2 ye=@
1080 (3 ye=-@
1463 (4 YeaM
Note. For 2,4-D x + SE, n = 15. a &soil core, 1.5 cm* X 5 cm. b Not determined.
Glyphosate AMPA 2,4-D
243 (1 yead
( 1St spring)
Glyphosate AMPA 2,4-D
283
co.05 co.05 <0.05
0.1 + 0.1
<0.05 0.3
0.3 0.9 0.6 + 0.2
0.8 1.7 0.3 f 0.1
0.9 1.4 3.5 f 0.6
3.4’r;6.9
2.4
9.1 x0.05 9.7 & 1.0
Glyphosate AMPA 2,4-D
Glyphosate AMPA 2,4-D
V-g)
1
Substance
47 @ePt)
0
Time (days)
AMOUNTS OF GLYPHOSATE, AMPA,
co.05
0.1
-co.05
0.1 0.9 <0.05
0.6 2.5 0.2kO.l
it: Nbb
0.7 1.4 3.3 f 0.4
2.2 2.0 1.3f 0.2
12.8 co.05 9.2 zt 1.5
2
<0.05 -co.05 <0.05
-co.05 0.2 0.1 f 0.1
0.6’;r90.2
0.3
0.2Yo. 1
0.6
1.4 1.5 4.9 f 0.7
2.0 0.9 0.5 It 0.1
co.05 9.1 f 1.3
10.0
3
Site
x0.05 0.7 -co.05
0.2 3.1 -co.05
0.9 5.3 0.4kO.l
2.6 3.1 0.9 f 0.4
3.4 3.5 2.4 + 0.5
4.2 3.3 1.5z!z0.3
12.5 co.05 9.8 + 1.0
4
AND 2,4-D” FOUND AT THE NORTHERLY EXPERIMENTAL
TABLE 2
0.2
<0.05 0.2 co.05
0.2 1.7 0.1 + 0.1
0.;:
1.0
E 1.4io.7
0.8 2.0 3.0 + 0.4
1.9 1.0 1.1 f 0.3
-co.05 10.6+ 1.1
11.6
5
SITES
<0.05
0.1
<0.05
<0.05 0.6 x0.05
0.2 1.2 0.5kO.l
1.1 kO.3
1.9 3.2 1.7 zk0.5
13.0 <0.05 10.0+0.8
6
3 2 g 3 H:
p
“8
3 g ;;I z M
E
PERSISTENCE
OF GLYPHOSATE TABLE
AMOUNTSOFGLYPHOSATE,
235
AND 2,4-D
3
AMPA, AND~,~-D*FOUNDATTHESOUTHERLY EXPERIMENTALSITES Site
Time (days)
Substance
7
8
9
10
11
(Ad
Glyphosate AMPA 2,4-D
13.5 <0.05 13.0 + 1.5
19.7 co.05 12.3 2 1.6
17.7 <0.05 13.2 + 1.2
16.3 to.05 12.0 f 1.5
15.0 co.05 14.8 + 0.6
(SW)
Glyphosate AMPA 2,4-D
8.1 1.3 2.0 Ik 0.5
6.8 1.1 2.1 + 0.6
WV)
Glyphosate AMPA 2,4-D
1.9 0.4 1.2k0.3
1.3 1.9 1.3 -t 0.2
265 ( 1 st spring)
Glyphosate AMPA 2,4-D
0.9 1.5 0.6 f 0.3
380 (1 yea)
Glyphosate AMPA 2,4-D
0.2 0.4 0.6 -+ 0.2
732
Glyphosate AMPA 2,4-D
0
53
87
(2 YeaN
-co.05 0.1 <0.05
5.0 co.05 ND’
5.3 0.3 1.3 kO.3
i4; 1.9iO.l
1.9 0.4 0.8 t- 0.2
2.5 0.6 0.7 kO.1
2.2 0.7 0.5 kO.1
1.3 1.9 0.8 -t 0.2
0.4 0.4 0.5 -t 0.2
0.2 0.6 0.5 + 0.1
1.2 0.5 0.3 50.1
0.2 0.4 0.3 +- 0.2
co.05 0.3 0.1 f 0.1
-co.05 <0.05 0.2kO.l
0.6 0.4 -co.05
x0.05 0.2 co.05
x0.05 co.05 <0.05
-co.05 <0.05 -co.05
<0.05 0.1 -co.05
Note. For 2,4-D x f SE, n = 15. ’ &soil core, 1.5 cm* X 5 cm. ’ Not determined.
Gilson respirometer. They were then equilibrated for 5 days at 20°C. Next, the rate of respiration (R) was determined. The results are expressed as microliters of O2 used per hour and per soil core, 10 cm* X 5 cm. Soil pH was determined in water (soil: water = 1:2). RESULTS
AND
DISCUSSION
The amounts of glyphosate, AMPA, and 2,4-D found in the soil samples at various times after their application are given in Tables 2 and 3. Glyphosate reached low levels faster at sites l-6 than at sites 7- 11 during the weeks after application. On the last sampling occasion before winter at sites l-6,47 days after application, the mean amount of glyphosate per sample was 2.4 pg compared with 6.5 pg at sites 7-l 1 after 53 days. Not until 87 days after application did the mean amount of residue reach the same low level at sites 7- 11. This might be an effect of higher amounts of glyphosate reaching the soils at sites 7-l 1 (about 50% of applied) than at sites l-6 (about 40%) as a consequence of the lower frequency of cover plants on the ground. However, there might also be other explanations. Glyphosate is not directly utilized as an energy substrate by soil microorganisms (Sprankle et al., 1975; Nomura and Hilton, 1977; Torstensson and Aamisepp, 1977)
236
TORSTENSSON,
0.14 r
LUNDGREN,
AND STENSTRijM
a= O.OiB93+O.ooO18 R rr0.9821
0.068 200
400 R (plOph-'1
600
FIG. 2. Correlation between rate of respiration (I?) and capability to decompose glyphosate (a) in field experiments 1- 11.
and undergoes cometabolic decomposition (Torstensson, 1980); i.e., the energy derived does not support microbial growth. Microbial production of enzymes necessary for breaking down glyphosate is not induced by the herbicide itself but by other sub strates for the microorganisms. Thus, one would expect cometabolic processes to depend on the general microbial activity of the soil, as has been found for glyphosate (Lijnsjii et al., 1980; Torstensson and Stenstrom, 1986). The data on the decomposition of glyphosate from application to the last sampling in the first autumn can be linearized by the formula c=co-kft, where c is the actual concentration of glyphosate at time t, co is the initial concentration, and k is the rate constant (Torstensson and Stenstriim, 1986). However, the dose reaching the soil in the different field experiments varies due to the vegetation cover (Tables 2 and 3), and the rate constant k is directly proportional to co when co is low (Stenstrom, 1989). Thus, k= acoanda
= k/co,
where the quotient a is a co-independent rate constant. Therefore, a, but not k, can be used to compare, for example, the decomposition of a pesticide in different soils. A plot of a against the “basic” respiration rate of the different soils (Table 1) shows a high correlation between them (Fig. 2). This correlation between the biological activity of soils and the capacity to degrade glyphosate has been found previously for a number of forest and agricultural soils (Torstensson and Stenstriim, 1986). It is also noteworthy that the northern soils ( 1-6) have a higher capacity to degrade glyphosate than the southern ones (7- 11, Fig. 2). During winter, with ground frost for about 7 months at sites l-6 and 4 months at sites 7- 11, the reduction in glyphosate level was slight (Tables 2 and 3). Moreover, during the following summer, the disappearance was comparatively slow at all sites, and even 2 years after application 0.2- 1.O pg glyphosate per sample was found at sites
PERSISTENCE
OF
GLYPHOSATE
AND
237
2,4-D
TABLE 4 AMOUNTS
OF
2,4-D FOUND IN SAMPLESOF BIRCH RESIDUES” Site
Time I year
Vegetation residues Leaves Branches
2 years
Leaves
3 years
Branches Branches
1
5
8
10
20 60
11 35
’ mg/kg dry wt. ’ Not determined.
l-6 (Table 2). The slow disappearance of the minor residues of glyphosate was not thought to be caused by insufficient capacity of the soils to degrade the herbicide but by adsorbed glyphosate being slowly released from binding sites in vegetation residues (Coupland, 1985). The metabolite AMPA from glyphosate is known to accumulate in soils and has been previously reported to decompose more slowly than the parent compound (Comes et al., 1976; Nomura and Hilton, 1977; Rueppel et al., 1977). In this investigation, the accumulation was most pronounced in the northern soils (Table 2). At site 4, soil samples contained as much as 5.3 pg AMPA per sample 2 years after the application of glyphosate, which is 18% of the theoretical maximum. On average, soils from sites l-6 contained 2.3 pg AMPA per sample after 2 years (8%) while soils from sites 7- 11 contained only 0.3 pg per sample after 1 year ( 1%). After application of glyphosate and 2,4-D in August, there was a clear difference between their patterns of disappearance until late September. The 2,4-D disappeared rapidly from all sites, and about 7 weeks after application only 0.5-2.1 pg 2,4-D per sample was found on sites 2-l 1, i.e., about 2-7% of the amount originally applied. Larger amounts were found only in samples from site 1 (3.4 pg per sample). This indicates that 2,4-D is rapidly utilized as an energy substrate for metabolic decomposition by soil microorganisms, as discussed by Torstensson ( 1978, 1980) and Norris ( 198 1). The important rate-determining edaphic and climatic factors, water, pH, and temperature, do not seem to have been limiting to any appreciable extent during the period. Instead, the rate was probably determined by the supply of the energy substrate 2,4-D to those soil microorganisms able to utilize it. However, the small amounts of 2,4-D remaining after the winter period (Table 2 and 3) were persistent, as demonstrated by the fact that 0.1-0.7 pg per sample was found in soils from sites l-6 2 years after application. This is probably due to the persistence of 2,4-D in vegetation residues (Table 4; Norris, 198 1; Torstensson, 1983). The length of time during which small amounts of both glyphosate and 2,4-D were detected in soil samples differed between sites l-6 and 7- 11. Certain amounts of the herbicides bound in vegetation residues would be protected from decomposition until they were either desorbed or decomposed together with the vegetation residues. During the first winter, desorption of 2,4-D apparently occurred at sites l-6 (Table
238
TORSTENSSON,
LUNDGREN,
AND STENSTRGM
2). Desorption of 2,4-D has also been observed in a forest area unintentionally exposed to an overdose (Torstensson, 1983). However, decomposition of vegetationbound herbicides together with vegetation residues seems to be a probable explanation for the long persistence ofsmall amounts of 2,4-D and probably also glyphosate as found in this investigation. The divergence between the two groups of sites can then easily be explained on the basis of different vegetation-residue decomposition rates as a consequence of different climatic conditions. CONCLUSIONS (i) After application of glyphosate or 2,4-D in mid-August to reforestation areas, most of both herbicides dissipated rapidly during the autumn. However, minor amounts of both herbicides were retained by vegetation residues and then released during a period of up to 1 year in the southerly areas and up to 3 years in the northerly areas. (ii) There was a high correlation (Y = 0.98) between the rate of respiration of the soils and the normalized co-independent rate constant for dissipation of the cometabolically decomposed glyphosate. (iii) The soils of experimental sites in northern areas generally had a higher rate of respiration and, thus, according to (ii) above, a higher rate of dissipation of glyphosate during the first autumn. ACKNOWLEDGMENTS We thank 0. Forsberg and L. Funke for technical assistance. This work was supported by a grant from the Swedish Council for Forestry and Agricultural Research.
REFERENCES B;~RRING, U. S. M. (1978). The use of phenoxy herbicides in Swedish forestry: Amounts, types, and mode of application. Ecol. Bull. 27,219-230. COMES, R. D., BRUNS, V. F., AND KELLY, A. D. (1976). Residues and persistence ofglyphosate in irrigation water. Weed Sci. 24,47-50. COUPLAND, D. (1985). Metabolism of glyphosate in plants. In The Herbicide Glyphosute (E. Grossbard, and D. Atkinson, Eds.) pp. 25-34. Butterworths, London. L~~NsJG, H., STARK, J., TORSTENSSON, L., AND WESSBN, B. (1980). Glyphosate: Decomposition and LUNDGREN, L. N. (1986). A new method for the determination ofglyphosate and (aminomethyl)phosphonit acid residues in soils. J. Agric. Food Chem. 34,535-538. LUNDHBIE, K. (1985). Efficiency of glyphosate in forest plantations. In The Herbicide Glyphosute (E. Grossbard and D. Atkinson, Eds.) pp. 328-338. Butterworths, London. NOMURA, N. S., AND HILTON, H. W. (1977). The adsorption and degradation of glyphosate in five Hawaiian sugarcane soils. Weed Res. 17, 113- 12 1. NORRIS, L. A. (198 1). The movement, persistence, and fate of the phenoxy herbicides and TCDD in the forest. Residue Rev. 80,65-l 35. RUEPPEL, M. L., BRIGHTWELL, B. B., SCHAEFER,J., AND MARVEL, J. T. (1977). Metabolism and degradation of glyphosate in soil and water. J. Agric. Food Chem. 25,5 17-528. SPRANKLE, P., MEGGI’IT, W. F., AND PENNER, D. (1975). Adsorption, mobility and microbial degradation of glyphosate in the soil. Weed Sci. 23,229-234. STARK, J. (1982). Persistence of herbicides in forest soils. Dissertation, Swedish University of Agricultural Sciences, Department of Microbiology, Uppsala, Report 15. STENSTR~~M,J. (1989). Quantitative assessment of herbicide decomposition at different initial concentrations. Toxic. Ass. 4,53-70.
PERSISTENCE
OF
GLYPHOSATE
AND
2,4-D
239
TORSTENSSON,N. T. L. (1975). Degradation of 2,4-D and MCPA in soils of low pH. Environ. Qual. Suf Suppl. 3,262-265. TORSTENSSON, L. (1978). Effects of phenoxyacetic acid herbicides on soil organisms. Ecol. Bull. 27,263284.
TORSTENSSON, L. (1980). Role of microorganisms in decomposition. In Interactions Between Herbicides and the Soil (R. J. Hance, Ed.) pp. 159-l 78. Academic Press, London. TORSTENSSON,L. ( 1983). Undersijkning av 2,4-D:s persistens och effekter pa markorganismer i skogsmark efter oavsiktlig overdosering. Swedish Environmental Protection Board PM 17 16. (In Swedish). TORSTENSSON, L. (1985). Behaviour of glyphosate in soils and its degradation. In The Herbicide Glyphosate (E. Grossbard and D. Atkinson, Eds.), pp. 137- 150. Butterworths, London. TORSTENSSON, N. T. L. (1987). Microbial decomposition of herbicides in soil. In Herbicides: Progress in Pesticide Biochemistry and Toxicology (D. H. Hutson and T. R. Roberts, Eds.), Vol. 6, pp. 249-270. Wiley, New York. TORSTENSSON, N. T. L., AND AAMISEPP, A. (1977). Detoxification of glyphosate in soil. Weed Res. 17, 209-2 12. TORSTENSSON, L., AND STENSTR~M, J. (1986). “Basic” respiration as a tool for prediction of pesticide persistence in soil. Toxic. Ass. 1,57-72.