Behaviour of four soil-active herbicides in a boreal podzol

Forest Ecology and Management, 31 ( 1990 ) 125-152

125

Elsevier Science Publishers B.V., Amsterdam - - Printed in The Netherlands

B e h a v i o u r of Four S o i l - A c t i v e H e r b i c i d e s in a Boreal Podzol SHELDON H E L B E R T 1

Department of Geography, Memorial University of Newfoundland, St. John's, New[oundland, A 1B 3X9 (Canada) (Accepted 20 December 1988)

ABSTRACT Helbert, S.,1990. Behaviouroffoursoil-active herbicidesin a borealpodzol. For. Ecol. Manage., 31:125-152. There is inadequate information concerning herbicide persistence in soilsof the boreal forest. A fieldstudy of herbicide behaviour in an orthic humo-ferric podzol began in the spring of 1983 of northeast central Newfoundland on a south-facing slope near G a m b o Pond. Dicamba, hexazinone, tebuthiuron and picloram were applied on four test plots within a clear-cut site. Under stringent controls the LF, Ae, Bf, Bc and C soilhorizons were sampled up to 486 days after application. Residue analyses of samples from the soil horizons show the distribution of herbicides through the profileover time, and indicate each herbicide'spropensity to leach and persist.Total a.i. (rag) values are corrected for recoveries, soil moisture, bulk density and are expressed for a given volume of soil I m 2 and 0.x m thick. Results show that allherbicides leach to the Bf horizon, occasionally leach to the B C horizon and rarely leach to the C horizon. More than 75% of the residues detected per sample period were found to be from the organic surface layer. The results of LF horizon analysis indicated the following: 78, 83, 84 and 93% were the average quantities detected per sample period for dicamba, tebuthiuron, picloram and hexazinone, respectively.Ninety percent of dicamba, hexazinone and picloram disappeared in lessthan 486 days, whereas the point estimate for tebuthiuron fallsoutside the range of predictable values greater than 486 days. Initial application rates varied and are reflectedin the residue concentrations detected.

INTRODUCTION

At a time when the use of herbicides in the natural environment isof growing importance in silviculture,agriculture and maintenance of rights-of-way, scientists, government regulatory agencies and the general public are becoming increasingly interested in the adequacy of data regarding residual biological

activity (Fletcher, 1960; House et al., 1967; Kimmins, 1975; Brown, 1978; Hance, 1Present address: Environment and Lands, Govt. of Newfoundland, Grand Falls, Nfld. A2A lW9, Canada.

0378-1127/90/$03.50

© 1990 Elsevier Science Publishers B.V.

126

s. HERBERT

1980; Khan, 1980; Smith, 1982; Eschenroeder et al., 1983; Nigg et al., 1983; Moyeux et al., 1984; Helbert, 1986; Neary et al., 1986; Thornton, 1986). This study examines two aspects of herbicide behaviour of principal concern in reforestation and habitat development in a boreal forest podzol. The first aspect is the unbiased mean distribution of residues down through the soil profile to the lower limit of sampling, and the second is their measured quantities as distributed through time. These aspects respectively address the degree of herbicide leaching which occurs during sample intervals and the quantities of herbicides which remain unaltered and potentially active in the soil. Each of four herbicide products (Tordon-101, DyCleer-24, VelparL and Spike-80W) 1 were applied on four separate field plots in central Newfoundland. Soil was then sampled from each of the field plots which were specifically analyzed for dicamba (3,6-dichloro-o-anisic acid), hexazinone (3-cyclohexyl6-(dimethylamino)-l-methyl-l,3,5-triazine 2,4(1H,3H)-dione) tebuthiuron (N- [5- (1,1-dimethylethyl)-l,3,4-thiadizol-2-yl]N,N,-dimethylurea), and picloram (4-amino-3,5,6-trichloropicolinic acid). METHODS AND PROCEDURES

Site characteristics Location. The site is located on the south-facing slope of Gambo Pond at an approximate elevation of 66 m a.s.1. (cartographic series: Gambo Pond, Newfoundland, 2D/9, RF = 1:50 000, grid reference 008997). Climate. The region exhibits a continental-type climate moderately to strongly influenced by marine conditions, a 'modified continental' climate, with 11001500 mm of precipitation year-1 (Banfield, 1981 ). Vegetation. The area is the most distinctly boreal part of the island (Damman, 1983). The vegetation found in the area is associated with black-spruce stands (Picea mariana) which dominate much of the area, possibly because of the high frequency of fires (Banfield, 1981, p. 174). More recently, humans have influenced the local floral community by clear-cutting the forest in 1981. Soil. Under the Canadian System of Soil Classification the soil being studied is defined as an Orthic Humo Ferric Podzol, Gambo series (Wells and Heringa, ~Dow Chemical produces the Tordon-101 which contains picloram and 2,4-D in a 1:4 ratio; only picloram was analyzed. Velsicol Chemicals markets DyCleer-24 with a dicamba and 2,4°D formulation ratio of 1 : 2; only dicamba was analyzed. Dupont of Canada produces the hexazinone formulation under the trade name of VelparL, and Elanco markets a formulation of tebuthiuron under the name of Spike°80W.

B E H A V I O U R OF F O U R SOIL-ACTIVE HERBICIDES IN A B O R E A L P O D Z O L

127

1972; Anonymous, 1978, p. 9B). Podzols occur with more than 50% frequency province-wide (Roberts, 1983, pp. 121, 139), and constitute 15.6% of soils and rockland in Canada (Foth and Schafer, 1980, p. 265). The soil is well drained, of sandy loam texture, having a bulk density range of 170-1380 kg m -3 and low base status, is granular in structure and having a pH range of 3.2-5.5 (CaC12). The common horizon sequence 1 of this soil is: LF, Ae, Bf, BC and C (range of mean profile thickness, 36-72 cm).

Limi~tmns The experiment was designed in conditions typical of much of Canada's boreal forest environment. In addition, the experiment was limited to sampling the soil on a horizon basis. Thus, leaching below the upper part of the C horizon, loss through volatilization and photodegradation at the soil surface, and plant intake are not taken into account.

Field program To be suitable for the experiment, the test site had to meet with the following m i n i m u m criteria: (1) the soil type had to be typical of those found in the boreal environment; (2) the soil unit had to accommodate five plots of 11 × 20 m with a m i n i m u m buffer between plots of 5 m; and (3) the soil density and rockiness should not prevent digging pits with a long-handled shovel and a heavy pick. The herbicides were then applied by a back-pack mist blower at the maximum allowed rates (see Table 1 for total volume applied). Using only water, repeated dress rehearsals were conducted on an off-site plot until the timing was constant and the pattern was consistent. The parallel 1-m swaths were perpendicular to the slope. The muzzle was angled toward the ground and was carried across the swath in an arc-like pattern as the arm swung across the swath. It was later determined that hexazinone was, by error, applied at a rate 11 times higher than recommended.

Sampling program The spatial distribution of sampling points was determined to be a compositesample using a 2-stage systematic grid sample design, with judgement (Cline, 1944, pp. 275-288; Petersen and Calvin, 1965, pp. 54-72; Chesters et al.,1974, pp. 451-453; Armson, 1977, p. 256). The temporal distribution of sampling points was determined according to kinetic half-lifeand persistence data found in literature surveys and by the frequency and intensity of precipitation events afterthe herbicide applications IThe USDA horizonequivalencesare 01, A2, Bir,BC, C.

128

S. H E R B E R T

TABLE 1 Rates of application for selected herbicides Herbicide

Ratio 4

Actual 1 applied

Per-ha equiv.2

MSR 3

200 g a.i2/L 400 g a.i./L

0.80

36

10 L/1000 L water

1:1

Velpar L Hexazinone

25% by wt

1.76

80

6.97 L / h a

11:1

Spike 80W Tebuthiuron

80% by wt

1.36

62

11.0 kg/ha (62 L/ h a)

1:1

60 g a.i./L 240 g a.i./L

0.768

35

35 L/ha in 200-L spray mixture

1:1

Active ingredients DyCleer 24 Dicamba 2,4-D

Tordon 101 Picloram

Formulation

1Amount (L) applied to 0.022 ha in 80 L water. 2In 3636 L water. 3Manufacturers' suggested rates of application. 4Per-ha equiv.: MSR 5Active ingredient. TABLE 2 Soil sample periods (days post-spray) for herbicide residues Herbicide

Dicamba ~ Hexazinone 2 Picloram + tebuthiuron 3

Control 1 Sample periods

-1 -1 - 1

t-1

t-2

t-3

t-4

t-5

t-6

t-7

t-8

t-9

t-10

3 3 5

6 6 11

10 10 32

14 36 64

24 86 310

48 312 483

98 382

312 486

392

486

1Prior to spray, t-0. 221 July 1983 for dicamba and hexazinone. 322 July 1983 for tebuthiuron and picloram.

were made. Complementary to this, the best indication of leaching is represented by a short period of intensive sampling immediately after application before other dissipative processes become dominant (i.e., a period of 1-2 weeks; (Anonymous, 1975, p. 26886) ); Table 2 lists the sample dates. The rectangular plots of 11 × 20 m were subdivided into 1-m 2 grids. The first 1-m from the edge of the plots was excluded from sampling to avoid the problems of boundary effects. The remaining grid squares were divided into three equal transverse

BEHAVIOUR OF FOUR SOIL-ACTIVE HERBICIDES IN A BOREAL PODZOL

129

sections. The long axes of the plots were oriented downslope; thus the first section where the sampling began was the bottom one-third of a plot. The composite was made up of nine units, three units from each horizon from each section; the weight of the composite was approximately 2 kg. For each section a 1-m 2 pit was dug exposing the five horizons. Once the digging of the three pits was complete, then three samples representing the upper, middle and lower portions of the C horizon of the first section were composited in a bucket lined with a disposable plastic bag. Following this, the 4th, 5th and 6th samples from the second pit, and then the 7th, 8th and 9th samples from the third pit were composited for the gross sample. This was done for each successive overlying horizon. In order to minimize cross-contamination, each plot had five permanently marked buckets, one for each horizon. The conically shaped steel trowels used for extracting samples were thoroughly cleaned between sampling horizons. The depths of each of the nine samples were recorded, and the pits were filled in immediately after the sampling was completed. In both digging and filling the pits, care was taken not to contaminate the ground surface with excavated soil and not to trample over the plots. New pits were dug for each sample date and they were located at least 1 m apart; the distances to subsequent pits were incremented by 1-m grids until no obstacles to digging (tree stumps, rock outcrops, skid tracks, old logs) were encountered. When the end of the row was reached, the digging was moved to the second row up or down from the initial row and the procedure continued until the experiment was complete (486-day period).

Residue analysis Dicamba. The Ontario Research Foundation (Sheridan Park, Mississauga, Ontario) did the analysis of dicamba, but not 2,4-D, for Velsicol Chemicals of Canada, Ltd. They planned to use H P L C / U V technique for the dicamba analysis, but severe interferences were encountered mainly due to the very high organic-matter content of many of the samples. A gas chromatographic (GC) methods coupled with an ether/acid extraction was again found unsuitable because of gross interferences. They finally adopted a GC method involving an alkaline extraction. Hexazinone. The Atlantic Pesticide Laboratory at the KentviUe Agricultural Center (Nova Scotia) did the analysis of hexazinone for Dupont of Canada. The procedure adopted for the determination of hexazinone in soil was a method by R.F. Holt of D u P o n t (1981). Tebuthiuron. Elanco Chemicals did residue analyses in their own laboratories in Indianapolis, Indiana. The analytical procedures for tebuthiuron were not made available.

4oo -

I

2,

i

t

/

after day 9B residues ore only detected in the LF horizon

Samplingperiod(dayspostspray)

I i !

I I

2000,

0 t"-

p.

30,

0 LF residues

o C residues

0 BC residues

t> Bf residues

<~ Ae residues

I,

2,

3'

o

10,

20'

50, 40,

100'

~q

E .~

v

200,

300,

500, 400,

E • ~

1O00,

x o4

6

3000,

>~

5DO0, 4000,

.

. 200

. 300

. $

~-

4OO

~ ~

\

+ TO~Wrmmidu4min s ~ profile

Logend ¢I Roqrlsl~O¢~ l~ne . data pomt~L~ 5(30

. - •

Sampling period (days postsproy)

. 100

•

"i, ~

day3

. s 4 e 1 7 . ~og(¥O

r** = -.976, RMS = 0.0..323

,og(Xi) = 2 . e l s 6 -

,,~,

.~

"'*",~

Dicambo Persistence and Regression Lines

E

Soil horizons

of sail layer for 100 days

D i s L r i b u t i o n In A P o d z o l ,

Fig, 1. Dicamba residues in a podzol.

i5

0

E

5,

10,

20 =

"O

~o,

E

40 -

"*6

o O~

50=

l oo.

E

._'

200 =

300 -

(~ X 04

E

500,

I000,

2000 •

4000 3000-

Dicamba

total mg a.i./volume

0

4-

10E

40

5o

2oo

3oo

400

tooo

0

,~o

~-

~o

- "*-~L'.

,;o

\ \

SQmpling period (dQys postsprey)

4°

. . . . .

Fig. 2. Hexazinone residues in a podzol.

"T

o×

N

r-

c" O

(I)

"23

m I1)

..2'

¢xl

x

c5

E

2000

~

i

x

0

"-"

cq

E F

Ae

~C

~I,

BC

<3 Bf

I>

o LF

.RM$ = 2 3 9 6 3

~e-

-I x

~

Sampling period (days postspray)

n = 7, exclu~le$ day 3

r~ = -.975,

\

Legend _ _ -ft. rm= u k a ~ i Ii =e;4~14=11

H,.,,z~.on,,Po~t,,n,:,~o°da~,,~s.~o,,Un"~

Soil horizons

7(]D, 1~0'

3000,

HexQzinone Distribution In A Podzol, totul mg o.i./volume of soil layer for 500 days

O bl O

O

N

g

£

oo

0

0

4a

C

"~

,

" " - -

4o

~o

~ " ~ - ' - "-------" "-~---"

,

~

:~

Ae

~C

~BC

b

o LF

~

~

800'

0

900 '

I000'

2o00,

3000.

x

Iog(YT)

~a

z6o

do~

1I

~o

&32

4~o

"~.

Legend

r~

Lines

Somplinq period (days postspray)

n = 4, a x c l u d e l

"~x.

and Regression

,-.,.=:9~f..~s=~,!.:%~

\

Persistence

X; = 3 0 3 4 . 1 - 8 6 4 . 6 6 ,

t

Tebuthiuron

Soil horizons

--


,"

/

8

E

3 10

v

x

d

q'~

" - .
~

,

Sampling period (days postspray)

,~

Fig. 3. Tebuthiuron residues in a podzol.

I--

~o

100

20O

~0O

400

1000

2000

g2 -s

.Q

v

E "5

Cn

d

•--

F

X

d

E

~-~

38OO

total mg a . i . / v o l u m e of soil layer for 5 0 0 days

Tebuthiuron Distribution In A Podzol,

=

Sampling period (days postspray)

0

.S ,

2

1o

~"

~'1

Fig. 4. Picloram residues in a podzol.

o_ t3

-6

o

O

E

(D

v (/)

o~ E

d

F

X

E

•

40,

10, 8,

I> Ae

o LF

i

,"V'-.. ~

Soil horizons

o I---

cb "o

E

20,

60,

E

d

SO,

C)

lO0,

200

day .32

a Regrem~n I~ cloto po~nta

lOW pro#de

.m Legend

Sampling period (days postspray)

~..,excludes

~;(xD = 3 ~ 3 o - ~ . 2 4 o o ~ ,og(Y~) r,, =_-.980. RMS= 0.0..399

Picloram Persistence and Regression Lines

Odx

E

Picloram Distribution In A Podzol, total rng a i./volume of soil layer for 500 days

0 N 0

0

o~

o

C

0

C

134

s. H E R B E R T

Picloram. Dow Chemicals of Canada did residue analyses for picloram but not 2,4-D in their own laboratories in Midland, Michigan. The methods followed were those of Dow's, ACR 73.3 developed by Bjerke (1973) and ACR 73.3 S.2, a supplement to Bjerke's method developed by Johnson (1977) Numerical program The data as received from the laboratories were corrected where necessary for recoveries, soil moisture, bulk density and horizon thickness, thereby transforming unadjusted sample p p m values (y) into adjusted total mg a.i. (Ya) for a given volume of soil values. The general form of the equation used for correcting the unadjusted results presented in column (4) to the adjusted values presented in column (5) (Tables 4-7 ) is as follows: ya = [y X ( 1 0 0 / > fr )/0d] × (Pb × D X 1002 ) X 10- 3

(1 )

where: [y × (100/fr)/0d I = residue ". where: y is residue values (ppm); (100/fr), factor for recovery rate; 0d, soil dryness (unit fraction ); and: Pd × D × 1002, dry-weight of soil volume (g). where: Pa is bulk density of horizons (g cm-a; (DX 1002, term which expresses residue values for 1-m 2 sample area, with specified horizon D cm deep; D, horizon depth (cm); 1002, 1 m 2 sample area; and: 10-3, factor to convert #g to mg and m to cm. Preliminary graphical regressions of the transformed data were prepared by summing the residue values of each horizon for each sample point giving a total for a soil profile i m 2 × 0.x m thick. This total is independent of location within the profile, which alleviates any between-horizon variation and graphically depicts herbicide persistence when viewed over time. These time-dependent graphs then indicated an approximate descriptive model for the herbicides' disappearance. Persistence data presented on the second graph of each of Figs. 1-4 suggests a negative linear least-squares model for hexazinone and tebuthiuron, and a negative-exponential model for dicamba and picloram. These models were t h e m employed to produce descriptive regression equations (Figs. 1-4) which in turn were used to estimate herbicide disappearance times (td; see Table 3). RESULTS The two graphs per figure given in Figs. 1-4 present the results of the herbicide analysis (total mg of a.i. ) as residues adjusted for recoveries, soil moisture, mean bulk densities, mean horizon thickness, and expressed for a given volume of soil (1 m 2 × 0.x m). The first graph of each figure shows remaining aValuesexpressedon a dry-weightbasis (#g g- 1).

BEHAVIOUROF FOURSOIL-ACTIVEHERBICIDES IN A BOREAL PODZOL

135

TABLE 3 50, 75, 90 a n d 95% disappearance times for four selected herbicides

Disappearance time ( d a y s ) ~ Herbicides

All horizons

Name

Appl. rate

Dicamba Hexazinone Tebuthiuron Picloram

7.6 19.5 8.8 2.1

(kg/ha) &

Conc. +

50%

75%

90%

95%

1X 11 X 1X 1X

25 186 * 7

37 318 * 18

61 492 * 55

89 * * 131

6.3 X 15.2 X 8.8 X 0.5 X

9 123 296 16

13 255 * 38

22 430 * 119

32 * * 281

Detected rate ( k g / h a ) $ Dicamba Hexazinone Tebuthiuron Picloram

48.3 27.0 29.4 1.1

Rounded to the nearest day. + Concentration rates are n times the commercially suggested maximum *Days computed to be greater than 500 days; only values within the range of given data are valid. ~Disappearance times based on the initial rate of application. *Disappearance times based on the maximum herbicide detected over all sample periods.

residues in each of the five horizons sampled over time, indicating the compounds' persistence and propensity to leach. The second graph of each figure portrays the total sum of adjusted residues, which gives a clearer picture as to the persistence of the herbicides independent of movement between horizons. In addition, a linear least-squares regression line is presented for comparison with the persistence curve. Associated with this graph is a descriptive regression equation for estimating disappearance times (td), correlation coefficient (r2), residual mean square (RMS), sample size (n) and which outliers were excluded. These regression equations are used to compute the tso, tTs, tgo and t9s which are listed in Table 3. The data for the graphs are presented in Tables 4-7, in which column (4) lists the residue data as received from the laboratories (ppm), column (5) shows the corrected residue figures as total mg a.i. per unit volume of soil, and column (6) gives herbicide weight detected as a percent of the total detected for the associated sample day; these values add up to 100%. Column (7) presents the residues detected as a percentage of the original amount of a.i. in the spray solution applied to the soil surface; these values should add up to less than 100%.

136

s. HERBERT

TABLE 4 Dicamba residues: m e a n horizon thickness, ppm, mg, percent of total mg per sample date, a n d percent of total herbicide applied

Day 3 LF:* AE: BF: BC: ^ C: ^ Totals Day 6 LG:* AE: BF: BC: C: A Totals Day 10 LF:* AE: BF: BC: C: Totals Day 14 LF:* AE: BF: BC: C: Totals Day 24 LF:* AE: BF: BC: C: Totals Day 48 LF:* AE: BF: BC: C: Totals

(1) (2) M e a n horizon

(3)

(4) Residues

(5)

Depth (cm)

Range (cm)

Unadj. (ppm) s

Adjusted (mg) ~

Count (n)

(6) Horiz. % ~

(7) Herbicide % +

6 6 29 0 0 43

! v ! 0 0 3

n.a. n.a. n.a. 0 0 n.a

11.40 0.01 n.d. 0.0 0.0 11.41

670.9 1.8 0.0 0.0 0.0 672.7

99.7 0.3 0.0 0.0 0.0 100.0

81.87 0.22 0.0 0.0 0.0 82.08

8 8 35 12 0 63

! ! ! ! 0 4

n.a. n.a. n.a. n.a. 0 n.a.

35.70 0.36 1.08 0.20 0.0 37.34

3735.6 66.9 968.0 57.1 0.0 4827.6

77.4 1.4 20.0 1.2 0.0 100.0

455.83 8.16 118.12 6.97 0.0 589.08

8 5 30 15 8 66

36 36 36 36 11 155

6.0-21.2 0.0-16.8 14.6-48.4 2.5-20.0 0.0-15.0 n.a

15.40 1.18 0.52 0.05 n.d. 17.15

911.5 129.9 336.8 16.1 0.0 1394.4

65.4 9.3 24.1 1.2 0.0 100.0

111.23 15.85 41.09 1.97 0.0 170.15

8 8 26 16 10 68

38 38 38 38 13 165

7.0-25.0 2.3-20.9 17.3-42.2 10.2-20.0 5.0-17.0 n.a.

3.87 0.16 0.01 n.d. n.d. 4.04

333.7 28.6 5.5 0.0 0.0 367.8

90.7 7.8 1.5 0.0 0.0 100.0

40.72 3.48 0.67 0.0 0.0 44.88

9 3 34 15 11 72

39 39 39 39 13 169

5.1-20.8 0.0-10.6 15.8-44.5 10.5-19.8 4.0-22.0 n.a.

3.14 0.10 0.59 0.09 n.d. 3.92

189.9 6.7 420.1 29.3 0.0 646.0

29.4 1.0 65.0 4.5 0.0 100.0

23.7 0.81 51.26 3.58 0.0 78.82

6 7 33 13 7 66

38 38 38 38 11 163

4.9-19.0 0.8-15.7 13.5-52.3 5.1-18.9 0.0-16.0 n.a.

2.24 0.22 0.23 0.09 0.03 2.81

133.4 36.4 163.6 28.6 5.1 367.2

36.3 9.9 44.6 7.8 1.4 100.0

16.28 4.44 19.97 3.50 0.62 44.81

BEHAVIOUROF FOUR SOIL-ACTIVEHERBICIDES IN A BOREALPODZOL

137

T A B L E 4 ( continued )

Day 312 LF:* AE BF: BC: C: Totals Day 382 LF:* AE; BF: BC: C: Totals Day 486 LF:* AE: BF: BC: A C: A Totals

(1) (2) M e a n horizon

(3)

(4) Residues

(5)

(6) Horiz. % *

(7) Herbicide % +

Depth (cm)

Count (n)

Range (cm)

Unadj. (ppm) s

Adjusted (rag)#"

7 8 18 16 10 59

40 40 40 40 11 171

8.1-17.7 1.0-14.4 13.7-24.7 10.1-20.0 5.0-19.0 n.a.

0.07 n.d. n.d. n.d. n.d. 0.07

4.3 0.0 0.0 0.0 0.0 4.3

100.0 0.0 0.0 0.0 0.0 100.0

0.53 0.0 0.0 0.0 0.0 0.53

9 5 27 15 10 66

33 33 33 33 11 143

7.8-23.9 1.0-9.5 18.7-36.2 10.2-19.9 4.0-18.0 n.a.

0.02 rbna rbna rbna rbna 0.02

1.5 0.0 0.0 0.0 0.0 1.5

100.0 0.0 0.0 0.0 0.0 100.0

0.19 0.0 0.0 0.0 0.0 0.19

7 10 29 0 0 46

41 41 41 0 0 123

7.0-18.1 1.0-30.9 18.1-44.2 0 0 n.a.

0.03 n.d. n.d. 0.0 0.0 0.03

2.8 0.0 0.0 0.0 0.0 2.8

100.0 0.0 0.0 0.0 0.0 100.0

0.34 0.0 0.0 0.0 0.0 0.34

SExpressed on a fresh-weight basis and not corrected for recoveries. #'Estimate of total m g of residues expressed on a dry-weight basis by volume for the specified soil layer of i X 1-m area and o.x-m thick (column 1 ). *Figures in column (6) are expressed as a percent of column (6)'s total showing the herbicide's relative distribution within the profile. +Percentage of total applied: 3,6-dichloro-o-anisic acid applied at the rate of 7.6 kg/ha active ingredient. *LF horizon corrected for a 41% dilution factor. !n = 1, where mean is defined by normal random number generator N(£,a2), using I M S L (International Maths and Stats Library, programme G G N M L ) ; means and std. deviations are as follows: LF, 12,2, 2.1, n=839; Ae, 6.9, 2.6, n--839; BF, 25.3, 4.9, n=839; BC, 14.7, 1.0, n--734; C, 9.8+, 1.9+, n=69. M e a n residue value of duplicate samples. n.d., None detected, <0.01 ppm, limit of detection (0.01 ppm). ^ Horizon not sampled. n.a., not applicable. rbna, Ontario Research Foundation received samples but were not analyzed.

138

s. H E R B E R T

TABLE5 Hexazinone residues: m e a n horizon thickness, ppm, rag, percent of total mg per sample date, a n d percent of total herbicide applied (1) (2) M e a n horizon

(3)

(4) Residues

(5)

Depth (cm)

Range (cm)

Unadj. (ppm) $

Adjusted (mg) ~

Count (n)

(6) Horiz. % *

(7) Herbicide %+

Day 3 LF:* AE: BF: BC: ^ C: ^ Totals

7 7 22 0 0 36

1 ! ! 0 0 3

n.a. n.a. n.a. 0 0 n.a.

73.90 0.08 0.06 0.0 0.0 74.04

1383.2 8.6 18.8 0.0 0.0 1410.6

98.0 0.6 1.3 0.0 0.0 100.0

53.64 0.33 0.73 0.0 0.0 54.70

5 11 28 14 0 58

! !

0 4

n.a. n.a. n.a. n.a. 0 n.a.

202.0 0.28 0.28 0.11 0.0 202.67

2520.5 47.4 111.7 25.0 0.0 2704.7

93.2 1.8 4.1 0.9 0.0 100.0

97.74 1.84 4.33 0.97 0.0 104.88

6 8 25 15 8 62

37 37 37 37 13 161

4.97-17.4 0.0-21.9 15.8-34.2 10.3-19.5 4.0-14.0 n.a.

99.50 0.16 0.08 0.30 n.d. 100.04

1552.0 19.7 28.5 72.9 0.0 1673.1

92.8 1.2 1.7 4.4 0.0 100.0

60.18 0.76 1.11 2.83 0.0 64.88

6 9 26 15 9 65

40 40 40 40 15 175

5.0-15.0 0.0-18.5 13.1-39.6 10.1-19.8 3.0-18.0 n.a.

85.40 0.85 0.39 0.63 0.24 87.51

1465.2 117.8 144.5 153.2 37.2 1918.0

76.4 6.1 7.5 8.0 1.9 100.0

56.82 4.57 5.60 5.94 1.44 74.37

8 9 18 14 14 63

43 43 43 43 15 187

7.1-17.6 0.8-20.9 10.1-25.3 10.0-19.3 8.0-23.0 n.a.

98.20 0.42 0.43 0.31 0.10 99.46

1991.2 58.2 110.3 70.3 24.1 2254.2

88.3 2.6 4.9 3.1 1.1 100.0

77.21 2.26 4.28 2.73 0.94 87.41

5 8 26 15 13 67

37 37 37 37 15 163

4.0-13.9 0.8-19.6 16.1-36.7 10.0-19.8 5.0-20.0 5.0-20.0

20.30 0.12 0.20 0.13 0.10 20.85

298.7 15.5 77.7 33.1 23.5 448.5

66.6 3.5 17.3 7.4 5.2 100.0

11.58 0.60 3.01 1.28 0.91 17.39

Day 6 LF:* AE: BF: BC: C: ^ Totals Day 10 LF:* AE: BF: BC: C: Totals

! !

Day 36 LF:* AE: BF: BC: C: Totals

Day 86 LF:* AE: BF: BC: C: Totals

Day 312 LF:* AE: BF: BC: C: Totals

BEHAVIOUROF FOURSOIL-ACTIVEHERBICIDESIN A BOREALPODZOL

139

TABLE 5 ( continued )

D a y 382 LF:* AE: BF: BC: C: Totals Day 486 LF:* AE: BF: BC: C: Totals

(1) (2) Mean horizon

(3)

(4) Residues

(5)

(6) Horiz. % ~

(7) Herbicide % +

Depth (cm)

Count (n)

Range (cm)

Unadj. (ppm) s

Adjusted (mg) ~

5 8 27 15 12 67

35 35 35 35 14 154

4.4-14.5 0.0-16.7 12.6-47.5 10.0-19.8 5.0-18.0 n.a.

16.30 0.34 0.22 0.16 0.07 17.09

232.8 48.0 96.9 44.5 16.6 438.8

53.1 10.9 22.1 10.1 3.8 100.0

9.03 1.86 3.76 1.73 0.64 17.02

6 5 27 15 9 62

37 37 37 37 14 162

5.9-15.4 0.3-13.5 11.6-37.5 10.7-19.7 3.0-17.0 n.a.

10.20 0.07 n.d. n.d. n.d. 10.27

192.7 5.2 0.0 0.0 0.0 197.9

97.4 2.6 0.0 0.0 0.0 100.0

7.47 0.20 0.0 0.0 0.0 7.68

s Expressed on a dry-weight basis and not corrected for recoveries. ~Estimate of total mg of residues expressed on a dry-weight basis by volume for the specified soil layer of 1 × 1-m area and 0.x-m thick (column 1 ). Figures in column (6) are expressed as a percent of column (6)'s total showing the herbicide's relative distribution within the profile. +Percentage of total applied: 3-cyclohexyl-6- (dimethylamino)-l,3,5-triazine 2,4 (1H,3H)-dione applied at the rate of 19.5 kg/ha of active ingredient. *LF horizon corrected for a 41% dilution factor. ~n= 1, where mean is defined by normal random number generator N(2, a2), using IMSL {International Maths and Stats Library, programme GGNML); means and std. deviations are as follows: LF, 12.2, 2.1, n=839; Ae, 6.9, 2,6, n=839; Bf, 25.3, 4.9, n=839; BC, 14.7, 1.0, n=734; C, 9.8+, 1.9+, n=69. ^ Horizon not sampled. n.d., None detected, -<0.04 ppm, limit of quantification (0.04 ppm). n.a., not applicable. DISCUSSION

Herbicide distribution The three patterns clearly shown on the first graph of each of Figs. 1-4 are: (1) the distinct separation of values of the organic LF horizon from the mineral Ae to C horizons; (2) the pattern which represents the displacement process where values of the lower horizons (Bf, BC and C ) increase as the values of the upper horizons (LF, Ae and Bf) decrease; and (3) the tendency for the values of the Ae horizon to be less than those of the Bf horizon.

140

s. HERBERT

TABLE6 T e b u t h i u r o n residues: m e a n horizon thickness, ppm, mg, percent of total mg per sample date, a n d percent of total herbicide applied

Day 5 LF:* AE: BF: BC: C: ^ Totals Day 11 LF:* AE: BF: BC: C: Totals Day 32 LF:* AE: BF: BC: C: Totals Day 64 LF:* AE: BF: BC: C: Totals Day 310 LF:* AE: BF: BC: C: Totals

(1) (2) M e a n horizon

(3)

(4) Residues

(5)

Depth (cm)

Range (cm)

Unadj. (ppm) $

Adjusted (mg) &

Count (n)

(6) Horiz. % ~

(7) Herbicide % +

8 8 18 15 0 49

! ! ! ! 0 4

n.a. n.a. n.a. n.a. 0 n.a.

40.0 0.41 0.13 0.10 0.0 40.64

2814.5 54.3 42.7 25.5 0.0 2936.9

95.8 1.8 1.4 0.9 0.0 100.0

282.3 5.44 4.28 2.55 0.0 294.30

5 16 25 12 7 65

! ! ! ! ! 5

n.a. n.a. n.a. n.a. n.a. n.a.

56.00 0.23 0.20 0.12 n.d. 56.55

1899.8 57.3 75.6 23.9 0.0 2056.6

92.4 2.8 3.7 1.2 0.0 100.0

190.37 5.75 7.58 2.39 0.0 206.09

7 4 25 15 8 59

36 36 36 35 13 156

2.5-20.3 0.0-14.6 12.6-41.5 10.4-19.6 0.0-15.0 n.a.

60.00 0.67 0.40 0.18 0.10 61.36

1954.1 42.2 149.4 41.8 13.0 2200.6

88.8 1.9 6.8 1.9 0.6 100.0

195.81 4.23 14.97 4.19 1.31 220.52

6 5 21 15 9 56

38 38 38 37 11 161

1.4-20.7 0.0-16.7 7.1-40.4 10.0-20.0 2.0-24.0 n.a.

71.0 0.14 0.11 0.09 n.d. 71.34

2833.7 11.7 35.4 23.5 0.0 2904.3

97.6 0.4 1.2 0.8 0.0 100.0

283.96 1.17 3.54 2.35 0.0 291.03

8 8 19 14 12 61

41 41 41 41 11 173

8.0-21.0 0.0-35.3 8.8-30.3 10.2-19.9 6.0-20.0 n.a.

28.0n.d.~ 0.07n.d. n.d. 28.07

1452.8 0.0 20.1 0.0 0.0 1472.9

98.6 0.0 1.4 0.0 0.0 100.0

145.58 0.0 2.02 0.0 0.0 147.59

BEHAVIOUROF FOURSOIL-ACTIVEHERBICIDESIN A BOREALPODZOL

141

T A B L E 6 ( continued ) (1) (2) Mean horizon

(3)

(4) Residues

(5)

Depth (cm)

Count (n)

Range (cm)

Unadj. (ppm) $

Adjusted (mg) ~

6 6 25 16 10 63

43 43 43 43 11 181

6.9-15.2 0.8-19.4 12.7-36.4 10.1-19.9 3.0-20.0 n.a.

14.000.200.28n.d. n.d. 14.48

746.4 19.1 105.8 0.0 0.0 871.3

(6) Horiz. % #

(7) Herbicide % +

Day 483 LF:* AE: BF: BC: C: Totals

85.7 2.2 12.1 0.0 0.0 100.0

74.79 1.92 10.61 0.0 0.0 87.31

SExpressed on a fresh-weight basis and corrected for recoveries. ~Estimate of total mg of residues expressed on a dry-weight basis by volume for the specified soil layer of 1 X 1-m area and 0.x m thick (column 1 ). ~Figures in column (6) are expressed as a percent of column (6)'s total showing the herbicide's relative distribution within the profile. + Percentage of total applied: N- [ 5 - ( 1,1-dimethylethyl ) - 1,3,4-thiadizol-2-y 1 ] N,N,-dimethylurea applied at the rate of 8.8 kg/ha of active ingredient. *LF horizon corrected for a 41% dilution factor. ~n= 1, where mean is defined by normal random number generator N(2,a2), using I M S L (International Maths and Stats Library, programme G G N M L ) ; means and std. deviations are as follows: LF, 12.2, 2.1, n = 8 3 9 ; Ae, 6.9, 2.6, n = 8 3 9 ; Bf, 25.3, 4.9, n = 8 3 9 ; BC, 14.7, 1.0, n--734; C, 9.8+, 1.9+, n--69. ~ Mean residue value of duplicate samples. ^ Horizon not sampled. n.d., None detected, < 0.05 ppm, limit of detection (0.05 ppm). n.a., not applicable.

Organic and mineral-soil horizons First, for all of the hexazinone (except day 382), tebuthiuron and picloram sample points, the high values of the LF distinctly contrast with the low values for the four mineral layers, Ae-C. For dicamba, this separation is distinct, though not as great as with the other compounds, and after day 14 the mineralsoil residues drop to below detectable levels making the separation extreme for the remainder of the sampling period. These high values in the LF matrix (see Tables 4-7) probably occur because of the adsorptive characteristic of humic substances and other soil organo-metallic complexes (Grover, 1971, pp. 417418; Khan, 1972, pp. 1-12; 1980, pp. 32-36; Grover and Smith, 1974, pp. 179186; Nearpass, 1976, pp. 272-277; Morrill et al., 1982, pp. 170-172; Weed and Weber, 1974 pp. 39-65). The organic content of soils is most often represented by their organic carbon content, which for this soil ranges as follows: LF, 81.2-

142

s. HERBERT

TABLE7 Picloram residues: m e a n horizon thickness, ppm, mg, percent of total mg per sample date, a n d percent of total herbicide applied

Day 5 LF:* AE: BF: BC: C: ^ Totals Day 11 LF:* AE: BF: BC: C: Totals Day 32 LF:* AE: BF: BC: C: Totals Day 64 LF:* AE: BF: BC: C: Totals Day 310 LF:* AE: BF: BC: C: Totals

(1) (2) M e a n horizon

(3)

(4) Residues

(5)

Depth (cm)

Range (cm)

Unadj. (ppm) ~

Adjusted (mg) ~

Count (n)

(6) Horiz. To~

(7) Herbicide % +

8 10 23 13 0 54

! ! v ! 0 4

n.a. n.a. n.a. n.a. 0 n.a.

5.500 0.080 0.020 n.d. 0.0 5.600

97.0 10.7 5.7 0.0 0.0 113.4

85.5 9.4 5.0 0.0 0.0 100.0

42.32 4.68 2.49 0.0 0.0 49.49

5 8 27 16 9 65

! ! ! ! ~ 5

n.a. n.a. n.a. n.a. n.a. n.a.

7.720 0.078 0.030 n.d.

83.8 8.4 10.0 0.0 0.0 102.2

82.0 8.2 9.8 0.0 0.0 100.0

36.56 3.65 4.38 0.0 0.0 44.59

9 6 21 14 9 59

37 37 37 37 11 159

9.9-28.0 0.0-14.2 12.7-34.9 10.3-19.7 3.0-18.0 n.a.

0.60 n.d. n.d. n.d. n.d. 0.60

12.2 0.0 0.0 0.0 0.0 12.2

100.0 0.0 0.0 0.0 0.0 100.0

5.33 0.0 0.0 0.0 0.0 5.33

7 7 15 13 8 50

57 57 57 57 11 239

7.7-15.5 2.0-20.7 5.7-25.7 10.0-19.9 5.0-14.0 n.a.

1.00 n.d. n.d. n.d. n.d. 1.00

16.3 0.0 0.0 0.0 0.0 16.3

100.0 0.0 0.0 0.0 0.0 100.0

7.10 0.0 0.0 0.0 0.0 7.10

8 10 28 14 12 72

44 44 44 44 12 188

8.0-20.1 1.2-25.9 12.5-47.6 5.5-19.6 3.0-20.0 n.a.

0.160 0.007 n.d. n.d. n.d. 0.167

3.0 0.9 0.0 0.0 0.0 3.9

76.4 23.6 0.0 0.0 0.0 100.0

1.33 0.41 0.0 0.0 0.0 1.74

7.828

143

BEHAVIOUR OF FOUR SOIL-ACTIVE HERBICIDES IN A BOREAL PODZOL

TABLE 7 ( continued ) (1) (2) Mean horizon

(3)

(4) Residues

(5)

(6) Horiz. % ~

Depth (cm)

Count (n)

Range (cm)

Unadj. (ppm) ~

Adjusted (rag)

9 7 29 15 12 72

49 49 49 49 11 207

6.8-29.0 0.0-18.0 16.1-48.2 10.1-19.8 3.0-23.0 n.a.

0.164~ n.d.0.006~ n.d. n.d. 0.170

(7) Herbicide % +

Day 483

LF:* AE: BF: BC: C: Totals

3.3 0.0 2.2 0.0 0.0 5.5

60.7 0.0 39.3 0.0 0.0 100.0

1.46 0.0 0.94 0.0 0.0 2.40

*Expressed on a dry-weight basis and corrected for recoveries. ~'Estimate of total mg of residues expressed on a dry-weight basis by volume for the specified soil layer of 1 X 1-m area and 0.x-m thick (column 1 ). Figures in column (6) are expressed as a percent of column (6)'s total showing the herbicide's relative distribution within the profile. +Percentage of total applied: 4-amino-3,5,6-trichloropicolinicacid applied at the rate of 2.1 kg/ ha for active ingredient. *LF horizon corrected for a 41% dilution factor. !n= 1, where mean is defined by normal random number generator N(2,a2), using IMSL (International Maths and Stats Library, programme GGNML), means and std. deviations are as follows: LF, 12.2, 2.1, n=839; AE, 6.9, 2.6, n=839; Bf, 25.3, 4.9, n--839, BC, 14.7, 1.0, n=839; C, 9.8+, 1.9+, n=69. Mean residue value of duplicate samples ^ Horizon not sampled. n.d., None detected, < 0.005 ppm, limit of quantification (0.5 ppm ). n.a., not applicable.

95.8%; Ae, 0.8-2.2%; Bf, 4.1-9.1%; BC, 1.3-2.8%; and C, 0.8-7.7% (results of carbon loss-on-ignition). The contrast between the carbon content of the LF horizon, with a mean of 90.4%, and that of the Ae-C horizons, with a mean of 2.72%, is consistent with this pattern which continues throughout the entire experimental period of 486 days. In this case it is unlikely that clay is more important than organic matter as a substrate for herbicide adsorption because of the small amounts found in this soil: 8.22%, 8.81%, 12.7% and 14/41% clay in the Ae, Bf, BC and C horizons, respectively (results of particle-size analysis, hydrometer method), However, the role of clay may be significant because of the large surface-area: mass ratio. In addition, as the organic content decreases with depth the clay content increases, which may indicate that where organic matter contributes fewer adsorption sites with depth, clay is increasingly able to supply an alternative pool of adsorption sites.

144

S. HERBERT

Mass [low Second, the pattern where residue values for the lower mineral layers (Bf, BC and C) increase over time can best be described as a displacement process. Initially, as the herbicides begin to move through the soil, the greatest quantities are detected in the LF layer (see Tables 4-7). As time passes, values for the LF layer begin to decline (post day 5 for picloram, day 6 for dicamba and hexazinone, and day 64 for tebuthiuron) and for relatively short periods, values for the Bf, BC and C horizons generally tend to increase (6-42-day range). This pattern is upheld for dicamba (days 14-20 for the Bf layer, 10-48 for the BC layer, and 48 for the C layer), hexazinone (days 10-36 for the Ae, Bf, BC and C layers), tebuthiuron (days 5-32 for the Bf layer, 11-32 for the BC layer, and 32 for the C layer), and picloram {days 5-11 for the Bf layer). These short periods are characteristic of the movement of herbicide as a slug influenced by mass flow. When these periods end there is a change away from this dominant slug pattern to patterns representing other dissipation mechanisms. The actual shift away from the dominating process of mass flow is represented by the end-points which terminate these intervals (days 48, 36, 32 and 11 for dicamba, hexazinone, tebuthiuron and picloram, respectively).

Leaching Third, the Ae layer generally exhibits very low values which indicates that, under environmental conditions favourable to the leaching of soluble organic compounds, these herbicides are removed from the A horizon and transported to the B horizon. Specifically, the leached A layer does not retain the herbicides, much as it loses ions to the Bf horizon. This pattern of eluviation/illuviation commences some time between day 3 and 6 and continues until day 48 for dicamba, after day 1 until day 382 for hexazinone, from between day 5 and 11 until the last sample day (483) for tebuthiuron, and from between day 5 and 11 until a point before day 32 - with some evidence that the process may continue up to day 483 - for picloram.

Anomalies The adjusted residue data revealed some anomalously high values. These extremes, the result of a number of conditions, were manifested as values exceeding 100% of the application rate (column (7), Tables 4-7). For some of these cases it was possible to determine a measure for correcting results which were then incorporated into the computations. However, for the remaining cases this was either not possible or practicable. The presence of some values greater than 100% indicates that some of the indeterminably variant conditions are still in effect. Detection levels in excess of the application rates occur

BEHAVIOUROF FOUR SOIL-ACTIVEHERBICIDES IN A BOREAL PODZOL

145

on days 6 and 10 for dicamba and on days 5, 11, 32, 64 and 310 for tebuthiuron. All but one of these seven high values occur in the LF horizon, the one exception being in the Bf horizon on day 6 for dicamba. Column (7), Table 4, shows that 118% of the total applied herbicide was detected on day 6 in the Bf layer, which also has the second-high mean organic-carbon content of the five horizons sampled (6.0%, n = 12, range = 4.1-9.1% ). The conditions which are identified as factors influencing some of these extreme values and whether or not they were determinable, and therefore incorporated into the corrections, are as follows: 1. The use of any type of sprayer will result in at least some small a m o u n t of herbicide overlap along parallel spray swaths because of the uncontrollable drift t h a t occurs (see Fig. 5); Indeterminable. 2. A mist-blower device is an inappropriate herbicide applicator as it is difficult to control droplet size and the delivery of a precisely uniform application (see Fig. 5 ); Indeterminable. 3. Dilution/concentration problems due to variant nature of sample unit

Fig. 5. Solo m i s t s p r a y e r a n d 1-m s p r a y s w a t h s .

146

s. HERBERT

Fig. 6. Non-uniform surface geometry and non-soil surface material.

distributions relative to the nonuniformity of horizon thickness and form, especially with respect to the LF horizon; Indeterminable. 4. Reduction in plot areas due to configuration of plots and non-soil surfaces within experimental areas (see Fig. 6). Rock outcrops, tree stumps and skid tracks do not constitute a relatively undisturbed soil surface and they may contribute to chemical hot and cold spots. Hence, these objects effectively reduce the sample area of the plots because they cannot be sampled; Determinable. 5. The assumption of a uniform application is misleading in all pesticide/ soil studies where the surface geometry is nonuniform (see Fig. 6). Hummocks and depressions change the total area of a plot and the distribution of the spray solution because of catchment and shadow effects, and may also contribute to chemical hot and cold spots; Indeterminable. 6. The inability of analytical laboratories to establish consistent recovery rates, as clearly expressed in Table 8, which shows how the recovery levels contrast between the laboratory spikes and the experimentor spikes; Limited determinability based on laboratories' recovery results. 7. The problem of extrapolating results from a 10-40-g analytical subsample representative of a larger composite sample {600-1800 g fresh weight), to an even greater estimated volume of soil (10-359 kg; 1 m2× 0.x-m deep), when sample variability is unknown, increases the risk of a Type-II fl error; Indeterminable.

Herbicide persistence In general, there are three trends to note on the second graph of each of Figs. 1-4 (1) a lag period of 6-64 days from the time of application to the time of

BEHAVIOUROF FOUR SOIL-ACTIVEHERBICIDES IN A BOREAL PODZOL

147

TABLE 8 Recovery results of laboratory spikes vs. those of experimenter: (a double-check) Fortified recovery levels Laboratory spiked Herbicide Dicamba

Soil Type p p m

Organic Mineral Hexazinone $ Mineral Mineral T e b u t h i u r o n Organic Mineral Picloram Organic Mineral

0.04-40.0 0.04-40.0 0.42 0.04-1.00 ~ ~ * *

E x p e r i m e n t e r spiked

Average %

Range %

ppm

Average %

Range %

68.3 72.5 72.5 84.6 ~ ~ 90 90

63.5-80.9 59.6-80.0 66-79 70-100 ~ ~ * *

0.0 1.0 0.0 1.0 0.0 1000.0 0.0 1.0

1 35 n.d. 102 n.d. 175 n.d. 149

n.a 27-43 n.a. 93-120 n.a. 110-130 n.a. 107-179

(n = 3) (n=4) (n=2) (n=16)

(n = 1 ) (n=2) (n=l) (n=4) (n=l) (n=4) (n=2) (n=7)

n, Sample size. n.a., N o t applicable, because organic soil was n o t spiked. n.d., N o n e detected, which is correct because organic soil was n o t spiked. Fortified recoveries n o t reported b u t residues have been adjusted by Elanco Laboratories, levels unknown. *Residue report states a n average 90% recovery w i t h o u t stating which horizons spiked or at what levels the samples were fortified. SWhich mineral horizons were spiked is unknown.

peak (the time of highest residue values detected), with the exception of picloram which does not display this pattern; (2) a relatively quick decline of residue values in 4-64 days following the time to peak (except for tebuthiuron); and (3) a slower decline of residue values over the balance of the sample period (more than 400 days), as indicated by the minor decrease in slope. The first phase of the distribution shows the lag where, for dicamba and hexazinone, there is a 6-day period to peak. The period to peak takes longer for tebuthiuron (64 days), and is not apparent for picloram unless it occurs before day 5. The second phase of the distribution, which shows the steep slope following the peak, lasts 4 days for hexazinone, 8 days for dicamba, 53 days for picloram, and does not occur for tebuthiuron. The third phase of the distribution exhibits the trend of a slow decline of residue values over the balance of the sample periods, which take at least 469 days for dicamba, 476 days for hexazinone, 451 days for tebuthiuron and 419 days for picloram. Table 3 records the empirically derived index (in days) for a 50, 75, 90 and 95% reduction in soil residues for the soil profile sampled as a whole. The disappearance times (td's) given are based on two levels of concentration: (1) td is based on the one-time herbicide application at the onset of the experimental period; (2) td is based on the total amount of herbicide (sum of LF-C

148

S. HERBERT

horizons) for a particular day, taken as the maximum detected over all sample days. This allows for a comparison and contrast of values expected (initial rate of application ) and values observed (maximum rate detected). It is most likely that the true values, though an unknown identity, fall in the interval between the two calculated td's. It should be noted that the td's are only relevant for the concentration of herbicides given, the environmental conditions under which the experiment was conducted, and the time period for which the herbicides were monitored (Hamaker, 1972, pp. 276-279). Dicamba. The calculated tso'S of 9 and 25 days compare well with Corbin and Upchurch's (1967, p. 373), tso of approximately 21 days using prepared and highly organic soils from North Carolina (Corbin and Upchurch, 1967, p. 370 ). The calculated td's of 0.4 and 1.2 months are within the less than 2 months cited by Kearny et al. (1969, p. 139 ). Hexazinone. The calculated tso'S of 4 and 6 months compare well with the tso'S of Rhodes (1980, p. 313) for 2 of 3 sites but sharply contrast with the ts0's of Sung (1982, p. 21). Rhodes (1980, p. 313) calculated t~o'S of 3-4, 6-7 and 1012 months for three different silt loam soils in Illinois, Delaware and Mississippi, respectively. Sung (1982, p. 21 ) calculated tso'S of less than 4 weeks for two soils (loamy sand and a sandy clay loam) in Alabama each at a low and high rate of application, 2 and 4 kg a.i. ha-1. Considering the high rate of application (Table 1 ), the td's may be more comparable knowing that a downward though undetermined adjustment is necessary. Tebuthiuron. The only calculable disappearance time possible is based on the maximum rate detected at the 50% level (ts0= 10 months). At the one-time application rate all the calculated values fall outside the range of given data, and therefore indicate an indeterminate upper limit greater than 16 months. The calculated tso of approximately 10 months, possibly upwards to 16 months, includes the 12-15 months half-life in areas receiving 100-150 cm annual rainfall (Anonymous, n.d., p. 3). This report states that the half-life is considerably greater in low-rainfall areas and in highly organic soils regardless of rainfall. As the total precipitation regime is comparable to that of central Newfoundland, it would not be unrealistic to think that the majority of td's outside the range of values reflects the longer half-life of residues remaining in the highly organic LF layer. Of the residues detected in the five horizons samples, those found in the LF horizon account for approximately 92% of the total per sample period (Table 6, column (6), with a mean thickness of 7 cm), the 8% balance of residues detected found in the remaining four mineral horizons. Picloram. The calculated t~o'S of 7 and 16 days are not very different from Lutz et al. (1973, p. 487) who reported for three soils (Fannin clay loam, Chandler

BEHAVIOUR OF FOUR SOIL-ACTIVE HERBICIDES IN A BOREAL PODZOL

149

fine sandy loam and Chester loam) in North Carolina, at two application rates (2.24 and 4.48 kg h a - l ) , a t6o of 15 days; furthermore, he reported atgo of 100 days which falls between the calculated tgo'S of 55 and 119 days but closer to the upper limit of 119 days. However, Altom and Stritzke (1973, p. 559) reported that, at the end of 100 days, from 63 to 77% of the original concentration remained in three soils tested from forest and grassland in Oklahoma. Therefore, if 100 days is required for t23 to t37, then it may be deduced that tso will take longer than 100 days under the conditions set forth in Oklahoma. Meikle et al. (1973, p. 552) presented decomposed picloram as a percent of the applied concentration for a variety of soils (clay, clay loam, loam and sandy loam) from Texas, California and Mexico which indicate a wide range of values for the 423-day period. For this period, percentage of applied concentration that disappeared ranged from 15 % (loam, California ), 66 % (sandy loam, Texas), 75 % ( clay, Texas ), 89 % (clay loam, Texas ), 90 % ( sandy loam, Mexico ) through to 94% (loam, California). CONCLUSION

Dicamba leached through the first four horizons (LF, Ae, Bf and BC) down to a mean depth of 60 cm (n = 4 valid arguments) during the interval after day 3 and until a time not greater than day 312. Hexazinone leached through the five horizons sampled (LF, Ae, Bf, BC and C) down to a mean depth of 65 cm (n = 4 valid arguments) during the period after day 10 and until a time not greater than day 486. As a cautionary note, the depth of leaching more likely reflects the high hexazinone application rate than the nature of the chemical, and should not be considered here for any other purpose than what might occur if a spill took place. Tebuthiuron leached through the top four horizons sampled (LF, Ae, Bf and BC ) down to a mean depth of 51 cm (n-- 4 valid arguments) during the interval after day 1 and before day 5 until a time not greater than day 310. Picloram leached through the first three horizons (LF, Ae and Bf) down to a mean depth of 42 cm (n-- 3 valid arguments ) during the period between days 1 and 5 until a time not greater than day 32. This pattern specifically reappeared on day 483. The organic LF horizon is a very important residue sink with a high buffering capacity, as the evidence suggests: for dicamba a mean of approximately 78% ( n = 9 ) of residues are detected in the LF layer; this mean value increases to 90% if two extremely low cases are deleted (n--7 ). For hexazinone, a mean close to 83% (n--8) of residues is detected in the LF layer; again, if one low outlier is excluded this figure increases to 87% ( n = 7), even at the high rate of application. For both tebuthiuron and picloram where no cases are excluded (n-- 6), the mean percentages of residues detected in each respective LF horizon are approximately 93 and 84%.

150

S. HERBERT

All four herbicides are eluviated from the Ae horizon and illuviated in the Bf horizon as determined by the leaching component of podzolization, a pedogenic process found in cool, h u m i d climates under coniferous or mixed vegetation common in the boreal ecosystems of circumpolar countries. The four selected herbicides have 50% disappearance times which range as follows: 716 days for picloram; 9-25 days for dicamba; 123-186 days for hexazinone; and 296 or more days for tebuthiuron. ACKNOWLEDGEMENTS This research was made possible through support granted by the School of Graduate Studies and the D e p a r t m e n t of Geography at Memorial University of Newfoundland, the Newfoundland D e p a r t m e n t of Forest Resources and Lands, the Forest Pest M a n a g e m e n t Institute of the Canadian Forestry Service, Newfoundland Forestry Research Centre, Dow Chemicals of Canada, D u P o n t Chemicals of Canada, Elanco Chemicals of Canada and Velsicol Chemicals of Canada. I t h a n k individually Dr. R.J. Rogerson of Memorial University of Newfoundland, Bruce Roberts of the Newfoundland Forestry Research Centre of the Canadian Forestry Service, and especially Joseph Feng, research scientist with the N o r t h e r n Forestry Research Center in Edmonton, for his patience through m a n y hours of problem solving.

REFERENCES

Altom, J.D. and Stritzke, J.F., 1973. Degradations of dicamba, picloram, and four phenoxy herbicides in soils. WeedSci., 21: 556-560. Anonymous,n.d. Technical Report On Tebuthiuron (Tech. Rep.). ELANCOdivisionof Eli Lilly and Co., Canada, Ltd., London, Ont., 5 pp. Anonymous,1975. Environmental Chemistry. Environmental Protection Agency,Officeof Pesticide Programs, Part VI. Fed. Reg., June, 40 (123) 26878-26896. Anonymous,1978.( The Canadian System Of Soil Classification.Canadian Soil SurveyCommittee Subcommitteeon Soil Classification,Ottawa: Can. Dep. Agric., Publ. no. 1646. Armson, K.A., 1977.Forest Soils: Properties and Processes.Universityof Toronto Press, Toronto, 39Opp. Banfield,C.E., 1981. Climaticenvironmentof Newfoundland.In: A.G.Macphersonand J.B. Macpherson (Editors), The Natural Environment of Newfoundland,Past and Present. Memorial University of NewfoundlandPress, St. John's, Nfld., pp. 83-153. Bjerke, E.L., 1973. Determination Of Residues Of Picloram In Soil By Gas Chromatography. Residue Research Ag-OrganicsDept., Dow ChemicalU.S.A., Midland,Mich. Tech. Rep. ACR 73.3 Brown, A.W.A., 1978. EcologyOf Pesticides. Wiley,New York, 525 pp. Chesters, G., Pionke, H.B. and Daniel, T.C., 1974. Extraction and analytical techniques for pesticides in soil. In: W.D. Guenzi (Editor), Pesticides In Soil and Water. Soil Science Societyof America, Inc., Madison, Wisc.,pp. 451-550.

BEHAVIOUR OF FOUR SOIL-ACTIVE HERBICIDES IN A BOREAL PODZOL

151

Cline, M.G., 1944. Principles of soil sampling. Soil Sci., 58: 275-286. Corbin, F.T. and Upchurch, R.P., 1967. Influence of pH on detoxification of herbicides in soil. Weed Sci., 15: 370-377. Damman, A.W.H., 1983. An ecological subdivision of the island of Newfoundland. In: G.R. South (Editor), Biogeography and Ecology of the Island of Newfoundland. Dr. Junk, Boston, pp. 163-206. Eschenroeder, A., Bonazountas, M. and Thomas, R., 1983. Models for pesticide behaviour in terrestrial environments. Residue Rev., 85: 245-255. Fletcher, Wm.W., 1960. The effects of herbicides on soil micro organisms. In: E.K. Woodford and G.R. Sagar (Editors), Herbicides and The Soil. Blackwell, Oxford, pp. 20-62. Foth, H.D. and Schafer, J.W., 1980. Soil Geography And Land Use. Wiley, Toronto, 484 pp. Grover, R., 1971. Adsorption of picloram by soil colloids and various other adsorbents. Weed Sci., 19: 417-418. Grover, R. and Smith, A.E., 1974. Adsorption studies with the acid and dimethylamine forms of 2,4-D and dicamba. Can. J. Soil Sci., 54: 179-186. Hamaker, J.W., 1972. Decomposition: quantitative aspects. In: C.A.I. Goring and J.W. Hamaker (Editors), Organic Chemicals in the Soil Environment (Vol. 1 ). Marcel Dekker, New York, pp. 253-340. Hance, R.J. (Editor), 1980. Interactions between Herbicides and the Soil. Academic, London, 349 pp. Helbert, S., 1986. Herbicide Behaviour in a Boreal Forest Podzol. Master's thesis, Memorial University of Newfoundland, Dept. of Geography, 180 pp., unpublished. Holt, R.F., 1981. Determination of hexazinone and metabolite residues using nitrogen-selective gas chromatography. J. Agric. Food Chem., 29 (1), 165-172. House, W.B., Goodson, L.H., Gadberry, H.M. and Dockter, K.W., 1967. Assessment of Ecological Effects of Extensive or Repeated Use of Herbicides (Final report sponsored by Advanced Research Projects Agency of the Dept. of Defense). Midwest Research Institute, Kansas City, MO, A.R.P.A. order no. 1086, 317 pp. Johnson, B.L., 1977. Determination of residues of picloram in soil by gas chromatography supplement for green forage, grain and straw. Residue/Environmental/Metabolism Research Agricultural Products Dept., Dow Chemical, Midland, Mich., Tech. Rep. ACR 73.3 S.2. Kearny, P.C., Woolson, E.A., Plimmer, J.R. and Isensee, A.R., 1969. Decontamination of pesticides in soils. Residue Rev., 29: 137-149. Khan, S.U., 1972. Adsorption of pesticide by humic substances: a review. Environ. Lett., 3: 1-12. Khan, S.U., 1980. Pesticides In The Soil Environment. Elsevier, Amsterdam, 240 pp. Kimmins, J.P., 1975. Review of the ecological effects of herbicide usage in forestry. Can. For. Serv., Pac. For. Res. Cent., Inf. Rep. Np. BC-X-139, 44 pp. Lutz, J.F., Byers, G.E. and Sheets, T.J., 1973. The persistence and movement of picloram and 2,4,5-T in soils. J. Environ. Qual., 2: 485-488. Meikle, R.W., Youngson, C.R., Hedlund, R.T., Goring, C.A.I.,Hamaker, J.W. and Addington, W.W., 1973. Measurement and prediction of picloram disappearance rates from soil. Weed Sci., 21: 549-555. Morrill, L.G., Mahilum, B.C. and Mohiuddin, S.H., 1982. Organic Compounds In Soils: Sorption, Degradation and Persistence. Ann Arbor Science Publishers, Ann Arbor, Mich., 326 pp. Moyeux, H.S., Jr., Richardson, C.W., Bovey, R.W., Burnett, E., Merkle, M.G. and Meyer, R.E., 1984. Dissipation of picloram in storm runoff. J. Environ. Qual., 13: 44-49. Nearpass, D.C., 1976. Adsorption of picloram by humic acids and humin. Soil Sci., 121: 272-277. Neary, D.G., Bush, P.B. and Grant, M.A., 1986. Water quality of ephemeral forest streams after site preparation with the herbicide hexazinone. For. Ecol. Manage., 14: 23-40. Nigg, H.N., Henry, J.A. and Stamper, J.H., 1983. Regional behavior of pesticide residues in the United States. Residue Rev., 85: 257-275.

152

s. HERBERT

Petersen, R.G. and Calvin, L.D., 1965. Sampling. In: C.A. Black (Editor), Methods of Soil Analysis, Part 1. American Society of Agronony, Madison, Wisc., pp. 54-72. Rhodes, R.C.,1980. Soil studies with 14C-labeled hexazinone. J. Agric. Food Chem., 28: 311-315. Roberts, C.A., 1983. Soils. In: G.R. South (Editor), Biogeography and Ecology of the Island of Newfoundland. Dr. Junk, Boston, pp. 107-161. Smith, A.E., 1982. Herbicides and the soil environment in Canada. Can. J. Soil Sci., 62: 433-460. Sung, Shi-J.S., 1982. Hexazinone persistence in soil and its effect on photosynthesis in pine. Doctoral dissertation, Depts. of Forestry/Agronomy and Soils, Auburn University, Alabama, 111 pp. (unpublished). Thornton, W.R., 1986. Environmental Impact Statement, Herbicide Application Program 198589. Stage 1, Part A: Literature Review. Dept. of Forest Resources and Lands/Abitibi-Price, Corner Brook Pulp and Paper Ltd., Government of Newfoundland and Labrador, Corner Brook, Nfld., 234 pp. Weed, S.B. and Weber, J.B., 1974. Pesticide-organic matter interactions. In: W.D. Guenzi (Editor), Pesticides In Soil And Water. Soil Science Society of America, Inc., Madison, Wisc., pp. 39-66. Wells, R.E. and Heringa, P.K., 1972. Soil Survey Of The Gander-Gambo Area, Newfoundland (Newfoundland Soil Survey 1 ) ~Research Branch, Canada Dept. of Agriculture, Ottawa.