Chemosphere 74 (2009) 265–273

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

Aldicarb and carbofuran transport in a Hapludalf influenced by differentialantecedent soil water content and irrigation delay

H.V. Kazemi b, S.H. Anderson a,*, K.W. Goyne a, C.J. Gantzer a

a Department of Soil, Environmental, and Atmospheric Sciences, University of Missouri, 302 ABNR Building, Columbia, MO 65211-7250, United Statesb Karadj Agricultural College, University of Tehran, Karadj, Iran

a r t i c l e i n f o

Article history:Received 29 May 2008Received in revised form 8 September 2008Accepted 9 September 2008Available online 16 October 2008

Keywords:AldicarbCarbofuranPore water velocity (v)Dispersion coefficient (D)Retardation coefficient (R)Degradation rate (l)

0045-6535/$ - see front matter � 2008 Elsevier Ltd. Adoi:10.1016/j.chemosphere.2008.09.015

* Corresponding author. Tel.: +1 573 882 6303; faxE-mail address: andersons@missouri.edu (S.H. And

a b s t r a c t

Pesticide use in agroecosystems can adversely impact groundwater quality via chemical leaching throughsoils. Few studies have investigated the effects of antecedent soil water content (SWC) and timing of ini-tial irrigation (TII) after chemical application on pesticide transport and degradation. The objectives ofthis study were to investigate the effects of antecedent soil water content (wet vs dry) and timing of ini-tial irrigation (0 h Delay vs 24 h Delay) on aldicarb [(EZ)-2-methyl-2-(methylthio)propionaldehydeO-methylcarbamoyloxime] and carbofuran [2,3-dihydro-2,2-dimethylbenzofuran-7-yl methylcarbamate]transport and degradation parameters at a field site with Menfro silt loam (fine-silty, mixed, superactive,mesic Typic Hapludalf) soils. Aldicarb and carbofuran were applied to plots near field capacity (wet) ornear permanent wilting point (dry). Half of the dry and wet plots received irrigation water immediatelyafter chemical application and the remaining plots were irrigated after a 24 h Delay. The transport anddegradation parameters were estimated using the method of moments. Statistical significance deter-mined for SWC included averages across TII levels, and significance determined for TII included averagesacross SWC levels. For the dry treatment, aldicarb was detected 0.10 m deeper (P < 0.01) on two of thefour sampling dates and carbofuran was detected at least 0.10 m deeper (P < 0.05) on all of the samplingdates compared to the wet treatment. Pore water velocity was found to be higher (P < 0.10) in the dry vswet treatments on three of four dates for aldicarb and two of four dates for carbofuran. Retardation coef-ficients for both pesticides showed similar evidence of reduced values for the dry vs wet treatments.These results indicate deeper pesticide movement in the initially dry treatment. For aldicarb and carbo-furan, estimated values of the degradation rate were approximately 40–49% lower in the initially dryplots compared to the initially wet plots, respectively. When the initial irrigation was delayed for 24 h,irrespective of antecedent moisture conditions, a 30% reduction in aldicarb degradation occurred. Thisstudy illustrates the deeper transport of pesticides and their increased persistence when applied to ini-tially dry soils.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Pesticide use for controlling insects, weeds, and pathogens isvital for efficient production of food, fiber and fuel crops. Assessingpesticide transport and degradation mechanisms as influenced byland management practices will assist in minimizing environmen-tal impacts of these agrichemicals.

Soils have natural features, such as planar voids (inter-aggre-gate pores), weather-induced shrink-swell cracks, and biologi-cally-induced, highly conductive macropores which act aschannels for water and chemical transport. Thomas and Phillips(1979) found that a 25.4 mm application of labeled tritiated waterwas detected at the 0.8 m soil depth in Maury silt loam (loess-de-

ll rights reserved.

: +1 573 884 5070.erson).

rived soil) with little or no increase of the water content of the soil.This indicates that a given rainfall or irrigation may not always wetthe root zone (the first 0.5 m of soil) before moving deeper in thesoil profile due to preferential flow paths.

The effect of macropores on chemical transport is not only man-ifested in the convective transport but also in chemical residencetime within the biologically active zone of the soil. Gish and Coff-man (1987), Kazemi et al. (2008b) presented field data on atrazineleaching which suggest that retention of atrazine in the soil may beless than values predicted from measured soil properties and theatrazine carbon partition coefficient (Rao and Jessup, 1982). Mal-one et al. (2004) found that because of preferential flow, the break-through time of herbicides was independent of their sorptiveproperties. This effect was also observed by Kung et al. (2000). Thisindicates that field experiments are necessary to verify transportand adsorption parameters for specific chemicals.

Table 1Selected soil physical and chemical properties for the Menfro silt loam at the studysite

Soil horizon Soil depth(m)

Silt (g kg�1) Clay (g kg�1) Organic C (g kg�1) pH

A 0.00–0.10 721 225 17.6 6.7AB 0.10–0.20 681 272 8.6 6.6Bt1 0.20–0.45 641 319 6.1 6.5Bt2 0.45–0.75 615 330 4.2 6.3

266 H.V. Kazemi et al. / Chemosphere 74 (2009) 265–273

Carbofuran is one insecticide that has been a groundwater chal-lenge due to leaching (Cogger et al., 1998). A survey of rural wellwater quality in Missouri by Sievers and Fulhage (1989) indicatedthat 17% of insecticide detections were carbofuran. Due to pollu-tion of water resources by carbofuran, some researchers have eval-uated encapsulated formulations. Marei et al. (2000) found that theleaching potential for three different encapsulations of carbofuranreduced transport more than nine times compared to the granularform.

Persistence of pesticides is also an important factor for estimat-ing transport processes. Persistence is influenced by factors such astemperature, moisture content, soil physical and chemical proper-ties, and chemical properties, such as solubility. Zindahl and Clark(1982) reported a series of half-lives for selected organic chemicalsas influenced by temperature and moisture. When soil moisturewas kept constant, a change of temperature from 10 to 30 �C in-creased the degradation rate. When temperature was kept con-stant, a change of moisture content from 20% to 80% alsoincreased the degradation rate. Rao et al. (1985) presented a sum-mary of biochemical half-lives for a number of pesticides andfound large differences between laboratory and field-measuredvalues. Other researchers (Beulke et al.,2000, 2005) have indicatedthat laboratory-estimated pesticide degradation in soil often over-estimates field persistence of pesticides suggesting the need formore field investigations.

To investigate the effect of soil macropores on pesticide trans-port, an experiment was proposed in which pesticides were ap-plied to soil profiles near field capacity (wet) and the permanentwilting point (dry) prior to pesticide application. It is postulatedthat pesticides can move to deeper depths in the soil profile whensoils are dry (voids open to transport) compared to wet profiles(voids full of fluid). However, this preferential movement may alsodepend upon the timing of water application (e.g., irrigation or pre-cipitation) subsequent to chemical application. We hypothesizedthat irrigation immediately following pesticide application wouldresult in greater chemical leaching and possibly longer persistencerelative to a one day Delay in the first irrigation application. Lessleaching may occur due to greater time for sorption processes.

To test these hypotheses, two systemic carbamate insecticideswere chosen for investigation: aldicarb and carbofuran The objec-tives of the experiment were: (i) to measure aldicarb and carbofu-ran concentrations as a function of time and soil depth at a fieldsite with antecedent soil water content at initially wet and dryconditions combined with an initial irrigation immediately afterpesticide application and with a one day Delay, and (ii) to estimateand compare transport and degradation properties of theseinsecticides.

2. Materials and methods

2.1. Treatments and field experimental details

Experiments were conducted at a field site, 0.02 ha in size, lo-cated at the Horticulture and Agroforestry Research Center nearNew Franklin, Missouri. The soil was Menfro silt loam (fine-silty,mixed, superactive, mesic Typic Hapludalf). Soil properties at thesite are shown in Table 1. The Menfro series consists of deep, welldrained, moderately permeable soils formed in thick loess depositson ridgetops, hills, and bluffs in the uplands adjacent to the Mis-souri and Mississippi Rivers. The slopes of these soils range from2% to 20%. The site had been undisturbed for several years, and astand of perennial grasses was growing on-site.

The experimental design was a completely randomized factorialexperiment with four replicates and four treatments (two factorsat two levels): initial soil water content and irrigation delay. Thelevels of initial soil water content were wet (near field capacity)

and dry (near wilting point). The irrigation delay levels were initialirrigation immediately after application (0-h Delay) and initial irri-gation one day after application (24-h Delay).

To achieve the design, the site was divided into 16 plots having1.5 m by 1.5 m dimensions. Plots had 1.0 m borders on all sides.Rainout shelters were constructed (0.75 m high with extensions0.5 m beyond plot borders) over eight plots (dry) 75 d before appli-cation of chemicals which allowed grasses to deplete the water fromthe soil profile. The other eight plots (wet) received about 200 mm ofwater in addition to natural precipitation (328 mm over the 75 dperiod) before pesticide application. This water was applied to as-sure the complete wetness of the soil profile compared to the soilin the dry plots. To avoid lateral flow of water and chemicals, plasticbarriers were installed to a 0.5-m depth which was well into the arg-illic subsoil horizon. Plastic barriers were installed on all fours sidesof each plot and soil horizon materials replaced in the order of exca-vation. The barriers, which provided restrictions to lateral flow, ex-tended above the soil surface 0.15 m to prevent plot runoff.

The site had received no chemical applications for several yearsprior to initiation of the experiments. Glyphosate [N-(phospho-nomethyl)glycine] was sprayed to kill the grasses, and after 3 dthe grass residues were removed from the soil surface prior tochemical application. Aldicarb and carbofuran (Table 2) were bothapplied at a rate of 2.98 kg ha�1 to all 16 plots (8 dry and 8 wet)using a 1.5 m long hand sprayer boom. Water (92.4 mm) was ap-plied immediately to eight plots (4 dry and 4 wet) and soil sampleswere subsequently removed 1 d after chemical application. Waterwas applied 24 h after chemical application to the other eight plots(4 dry and 4 wet) and soil samples were subsequently removed afew hours after irrigation for the first sampling (1 d after chemicalapplication). After the initial irrigation water was applied, plotswere covered with plastic to prevent any additional entry or re-moval of water. A depth of 46.2 mm of water was applied to theplots before the second sampling date (14 d after application).After the second sampling date, the plots were left uncovered to re-ceive natural rainfall.

2.2. Soil sample collection and analysis

Soil samples were removed on four selected dates: 1, 14, 42, and120 d after chemical application. A truck-mounted hydraulic soilprobe was used for sampling to a depth of 1.0 m. After removal,each core was segmented into 0.10 m increments. Holes createdby soil core removal were re-filled with bentonite and subsequentcores were separated by a 1.0-m sampling distance. To avoid sam-ple contamination, only the inner portion of each depth incrementwas retained for analysis. The samples were placed in glass con-tainers for storage.

The samples were transported to the laboratory and stored at�5 �C. Before chemical extraction, samples were allowed to equil-ibrate to room temperature. Aldicarb and carbofuran were ex-tracted using dichloromethane. The extract was concentratedusing an evaporator and dissolved in methanol. The dissolved sol-utes were analyzed by HPLC with an ultraviolet detector at 210/

Table 2Selected properties of the pesticides investigateda

Chemical Molecular weight (g mol�1) SWb (mg L�1) logKOW

c KOCd Field half-life (days)

Aldicarb 190.25 5900 1.41 26 50

NH

CH3

O

O

N

S CH3

CH3

CH3

(5900–6000) (0–80) (7–84)

Carbofuran 221.25 350 1.13 46 41CH3

CH3OO

O

NHCH3

(350–700) (9–158) (17–90)

United States Department of Agriculture, Agricultural Research Service, 2008. The ARS pesticide properties database [online]. Available at http://www.ars.usda.gov/Services/docs.htm?docid=14199 (verified 20 May 2008).

a Data were obtained from the USDA, ARS Pesticide Properties Database (2008). Individual values represent a recommended value for a select property and values inparentheses correspond to the range of values reported.

b SW, water solubility at 25 �C.c KOW, octanol–water partition coefficient.d KOC, organic carbon partition coefficient.

H.V. Kazemi et al. / Chemosphere 74 (2009) 265–273 267

280 nm on a C-18 column with acetonitrile/water gradient mobilephase (Buyanovsky et al., 1993).

2.3. Batch adsorption and degradation experiments

Batch aldicarb and carbofuran adsorption coefficients (Kd) weredetermined by reacting 10 g soil samples with 10 mL of pesticidesolution in 40-mL Teflon centrifuge tubes over a range of pesticideconcentrations. Ring-labeled 14C aldicarb and carbofuran solutionswere used in these batch experiments. After preparation of the sus-pensions, tubes were capped with rubber screw-in stoppers linedwith 0.025-mm teflon sheets and mounted on a shaker in a con-stant temperature for 24 h. Samples centrifuged at 15000 rpm for30 min after reaction. Supernatant solution (1-mL) was mixed withscintillation cocktail solution (4-mL) in a 10-mL glass vial prior tosample analysis with a Packard (TRI-CARB 1900 TR) liquid scintil-lation analyzer. Each sample was analyzed for 14C activity for10 min. The batch adsorption coefficient was determined for bothpesticides in soil samples from the field site by calculating theslope of the linear regression line (with zero intercept) betweenadsorbed and solution concentrations at equilibrium.

To estimate aldicarb and carbofuran degradation, the half-livesof these chemicals were assessed using incubation experimentswith 14C-labeled and non-labeled pesticides. In experiments withlabeled pesticides, 14CO2 evolved in the process of pesticide degra-dation and the remaining amount was measured periodically. Deg-radation rates were approximated from these data and residualradioactivity of incubated material was measured at the end ofincubation. In experiments with non-labeled pesticides, some ofthe replicates were analyzed after a period of incubation (usuallyafter 10 and 30 d).

2.4. Water balance

The daily amount of rainfall was monitored at the site. Dailyevaporation was determined using the Thornthwaite method(Ward and Trimble, 2003) and compared with pan evaporationdata. The amount of storage within the soil profile was estimatedusing water content measurements obtained with a neutron probe

along with gravimetric water content measured at the time of sam-ple collection. A neutron probe access tube was installed in eachplot with measurements being taken at 0.15 m increments to adepth of 1.20 m. Cumulative drainage was calculated using therainfall, evaporation, and storage data. The daily amount of drain-age was computed using the following relationship.

Drainage ¼ irrigationþ precipitation� runoff

� evaporation� storage ð1Þ

Runoff water was negligible due to installation of plastic berms(0.15 m high above soil surface) around each plot.

2.5. Transport parameter estimation

Transport parameters were estimated for aldicarb and carbofu-ran with the method of moments using chemical concentrations.The parameters of the convection–dispersion equation (CDE; Porroand Wierenga, 1993) were determined. The CDE equation withterms for solute adsorption and degradation is as follows:

RoCot¼ D

o2Coz2 � v

oCoz� lC ð2Þ

where R is the solute retardation factor, C is the volume averagedsolute concentration in the liquid phase, t is time after solute appli-cation, z is the vertical distance from the soil surface, D is the dis-persion coefficient, v is the average pore water velocity and l isthe solute degradation coefficient.

Since rainfall is a stochastic meteorological process, infiltrationand field-averaged soil water flux were not steady state (Gish andCoffman, 1987). Transient field problems can be approximatelyaccommodated by transforming the temporal variable to a newvariable defined as the ratio of the cumulative drainage to the aver-age net drainage. The transformation converts the real time to aflux time as follows:

t� ¼tmax

R t0hJwitdt

R tmax

0 hJwitdtð3Þ

268 H.V. Kazemi et al. / Chemosphere 74 (2009) 265–273

where t* is the flux time, t is the real time, hJwi is the spatiallyaveraged daily net water drainage, and tmax is the duration ofthe experiment (Gish and Coffman, 1987). Flux times were

120

105

90

75

60

45

30

15

00.0 0 .1 0.2 0 .3 0.4 0 .5

Soil

Dep

th (

cm)

Day 0Dry Wet

120

105

90

75

60

45

30

15

0

Soil

Dep

th (

cm)

Day 14 Dry Wet

120

105

90

75

60

45

30

15

0

Soil

Dep

th (

cm)

Day 120 Dry Wet

Volumetric Wa

Fig. 1. Volumetric water content of the initially dry and wet treatments at 0, 1, 14, 42 aapproximately at the wilting point and wet plots approximately at field capacity.

estimated for the aldicarb and carbofuran sampling datesand transport parameters were reported as a function of fluxtime.

0.0 0 .1 0.2 0 .3 0.4 0 .5

Day 1Dry Wet

Day 42 Dry Wet

ter Content (v/v)

nd 120 d after experiment initiation. At initiation of the experiment, dry plots were

100

200

300

400

500

a

D

E

SDep

th o

f Wat

er (

mm

)

P + I

H.V. Kazemi et al. / Chemosphere 74 (2009) 265–273 269

The method of moments was used to estimate solute transportparameters of the CDE model utilizing the first and second mo-ments of the resident solute concentrations (Jury and Sposito,1985; Jury and Roth, 1990)

zn ¼R1

0 znCdzR1

0 Cdzð4Þ

where zn is the nth depth moment. The D and v parameters wereestimated by the solution of the following equations

z�1 ¼Dvþ vt� ð5Þ

and

z�2 ¼ 4Dt� þ v2t�2 ð6Þ

where z�1 and z�2 are the first and second moments, respectively (Juryand Roth, 1990). The D and v parameters were estimated using themethod of moments with concentrations of aldicarb and carbofuransamples. The R parameter was estimated at 42 days (75 flux days)after pesticide application using the ratio of the v parameter esti-mated from the water balance to the v parameters estimated fromaldicarb and carbofuran as follows:

R ¼ vwater

valdicarbor R ¼ vwater

vcarbofuranð7Þ

The R parameter was also estimated from soil properties with thefollowing relationship

R ¼ 1þ KdDb

hð8Þ

where Db is soil bulk density, and h is the soil water content. The Rparameter was estimated for the site using parameters in the upper0.20 m profile.

The degradation rate constant was estimated using the summedaldicarb and carbofuran masses throughout the profile for eachmeasurement date [M(t)]. These data were compared to the aldi-carb and carbofuran concentrations measured at application (M0)to estimate the degradation rate. First-order degradation wasassumed:

MðtÞ ¼ M0e�lt ð9Þ

0

200

300

400

500

b

D

epth

of W

ater

(m

m)

P + I

2.6. Statistical analysis

The experiment was set up as a completely randomized designwith four treatment levels and four replicates. Treatments con-sisted of two factors: antecedent soil water content and initial irri-gation delay. Antecedent soil water content levels were dry andwet and the initial irrigation delay levels were 0 and 24 h. One coresample was removed per plot on each sampling date. Comparisonswere made between the four treatments using analysis of variancewith the General Linear Models procedure from SAS (SAS Institute,2005).

0 20 40 60 80 100 120

0

100E

S

D

Time After Initiation (d)

Fig. 2. Depth of cumulative water input (P + I, P = precipitation, I = irrigation),cumulative evaporation (E), storage relative to initial storage (S), and cumulativedrainage (D) for (a) the initially dry and (b) initially wet treatments. At initiation ofthe experiment, dry plots were approximately at the wilting point and wet plotsapproximately at field capacity.

3. Results and discussion

3.1. Water balance

Data in Fig. 1 illustrate the initial soil water content of dry andwet plots prior to chemical application and on each sampling date.As expected, much of the water (precipitation + irrigation) reach-ing the soil in dry plots is used in recharging the soil profile. Inthe wet plots, soil water content does not change significantlythroughout the experiment which allows most of this water todrain.

Distributions of cumulative water input (P + I), evaporation (E),storage relative to initial storage (S), and drainage (D) as a functionof days after chemical application for both the dry and wet plotsare shown in Fig. 2. The initial steep rise in the cumulative waterinput curve is due to the addition of 92.4 mm of water after chem-ical application. The steep rise during the next two wk is a result ofthe addition of 46.2 mm of irrigation water. Subsequent water in-put after the first two wk was from rainfall. Cumulative E was lowduring the first two wk since plots were covered. The amount of Ewas low over the duration of the experiment since the chemicalswere applied during autumn.

The net drainage over 120 d was 312 mm of water for the wetplots (233 mm in dry plots) which results in an average daily fluxof 2.60 mm d�1 (1.94 mm d�1 for dry plots). To obtain the averagepore water velocity, the flux is divided by the average volumetricsoil water content for the initially dry and wet plots (0.31 and0.35 m3 m�3, respectively). Therefore, the average daily pore watervelocity for this site was 6.3 and 7.4 mm d�1 for the initially dryand wet plots, respectively.

3.2. Aldicarb and carborfuran transport

Aldicarb concentration as a function of soil depth for thefour treatments on selected dates is shown in Fig. 3. Aldicarb

270 H.V. Kazemi et al. / Chemosphere 74 (2009) 265–273

concentration in all treatments is highest at the soil surfacethroughout the experiment. At 1 d after application, aldicarb con-centration in the initially dry plots was not significantly differentcompared to values in the initially wet plots. However, a slightlyhigher (not significant) aldicarb concentration was detected inthe initially wet plots at the 250 mm depth within 24 h after appli-cation. This suggests a possibility of more downward movement ofaldicarb in the initially wet plots compared to the initially dry plotsduring this time period. At 14, 42, and 120 d after application, thealdicarb concentration was approximately two times higher (sta-tistically significant) at the 5, 15 and 25 cm depths in the initiallydry plots compared with the initially wet plots.

The effect of the initial timing of irrigation did not appear tocause differences in aldicarb concentrations for the first three sam-pling dates. There is a slight trend for the first two sampling datesof the 0-h Delay to have lower concentrations at the surface com-

Aldicarb Conce

60

50

40

30

20

10

00 2 4 6 8

Soil

Dep

th (

cm)

Day 1Dry, imm.Dry, delayWet, imm.Wet, delay

60

50

40

30

20

10

00.0 0.5 1.0 1.5 2.0

Soil

Dep

th (

cm)

Day 42Dry, imm.Dry, delayWet, imm.Wet, delay

Fig. 3. Average aldicarb and carbofuran concentration as a function of soil depth at 1, 14,initial irrigation Delay of 0 h (imm.) and 24 h (delay) treatments. Concentrations are pewilting point and wet plots approximately at field capacity.

pared to the 24-h Delay. Interestingly, there appears to be a moresignificant effect of this treatment on the last day of sampling forthe first two sampling depths with higher concentrations appear-ing in the 24-h Delay treatment level. It is noted that concentrationlevels overall are much lower by this time.

Average carbofuran concentration as a function of soil depth forall treatments at selected sampling dates is shown in Fig. 3. Carbo-furan concentrations up to 42 d after application are highest at thesoil surface. At 1 d after application, carbofuran concentrations atthe soil surface in the dry plots are similar to those in the initiallywet plots. However, below the 150 mm depth carbofuran concen-trations are higher in the initially wet plots compared to the ini-tially dry plots. In the initially wet plots for the for 0-h Delay,carbofuran moved to a depth of 450 mm. At 14, 42, and 120 d afterapplication, the initially dry plots had higher carbofuran concen-trations than the initially wet plots to the 250 or 350 mm depth.

ntration (µg cm-3)

0 1 2 3 4 5

Day 14Dry, imm.Dry, delayWet, imm.Wet, delay

0.0 0.1 0.2 0.3 0.4 0.5

Day 120Dry, imm.Dry, delayWet, imm.Wet, delay

42, and 120 d after experiment initiation for the initially dry and wet treatments andr soil volume. At initiation of the experiment, dry plots were approximately at the

60

50

40

30

20

10

00 2 4 6 8

Soil

Dep

th (

cm)

Day 1Dry, imm.Dry, delayWet, imm.Wet, delay

0 1 2 3 4

Day 14Dry, imm.Dry, delayWet, imm.Wet, delay

60

50

40

30

20

10

00.0 0.5 1.0 1.5 2.0

Soil

Dep

th (

cm)

Day 42Dry, imm.Dry, delayWet, imm.Wet, delay

0.0 0.1 0.2 0.3 0.4 0.5

Day 120Dry, imm.Dry, delayWet, imm.Wet, delay

Carbofuran Concentration (µg cm-3)

Fig. 3 (continued)

H.V. Kazemi et al. / Chemosphere 74 (2009) 265–273 271

In the wet plots with an initial 24-h irrigation Delay, higher carbo-furan concentrations occurred below 350 mm at 120 d after appli-cation. Timing of initial irrigation did not show significant effectson carbofuran concentrations.

For the dry treatment, aldicarb was detected on the average0.10 m deeper (P < 0.01) on 14 and 42 d after application and0.05 m deeper (P < 0.10) on 120 d after application compared tothe wet treatment. This comparison used the deepest depth ofdetection for the pesticide for each plot. This suggests evidenceof more slightly deeper movement in the initially dry treatmentcompared to the wet treatment. At 120 d after pesticide applica-tion, aldicarb was detected 0.08 m deeper (P < 0.05) in the 24-hirrigation Delay treatment compared to the 0-h Delay treatment(averaged across both dry and wet treatments). A possible reasonfor this latter result is the potential slower degradation of aldicarbwhen it diffuses into soil pores when irrigation water is delayed for24 h at initiation of the experiment; however, the chemical maydiffuse out later and be leached.

Carbofuran was also observed deeper in the profile for the dryrelative to the wet plots. It was detected on the average 0.10 mdeeper (P < 0.05) 1 d after application and 0.15 m deeper(P < 0.01) 14 and 42 d after application in dry compared to wet

plots. At 120 d after application, this insecticide was detected at0.13 m deeper (P < 0.10) in the dry vs wet treatments. As with aldi-carb, evidence of deeper movement for carbofuran was observed inthis experiment.

Dousset et al. (2004) found evidence of higher herbicide trans-port in both silt loam (Aquic Hapludalf) and silty clay (Vertic Eut-rochrept) soils during the drier year of a two year study. Theyattributed the smaller loss in the wetter year to a longer time inter-val between the herbicide treatment and the first drainage event.Research by Brown et al. (2001) did not find any effects of differen-tial timing of water added to a heavy clay soil prior to herbicideapplication on subsequent herbicide losses. Transport studies con-ducted using tile-drained soil by Kung et al. (2000) showed that thewetter the soil profile, the longer the water dynamics of preferen-tial flow paths dominate contaminant transport which was differ-ent compared to the results of this study.

3.3. Estimation of transport parameters

Table 3 presents the first depth moments (approximate centerof mass) for the four treatments for aldicarb and carbofuran,respectively. Antecedent soil water content had a significant effect

Table 3Average first depth moment, pore water velocity, and dispersion coefficient values foraldicarb (±SD) and carbofuran (±SD) for treatments of the study at selected days (fluxtime) after application (n = 4)

Timea Initially wet Initially dry

0-h Delay 24-h Delay 0-h Delay 24-h Delay

First depth moment (mm)Aldicarb28 85 ± 21 85 ± 12 103 ± 30 89 ± 1754 110 ± 68 144 ± 47 127 ± 20 154 ± 9475 151 ± 41 127 ± 50 179 ± 28 233 ± 129120 31 ± 55 53 ± 42 18 ± 9 92 ± 54Carbofuran28 161 ± 27 152 ± 11 173 ± 35 152 ± 2954 202 ± 27 191 ± 27 227 ± 62 209 ± 7975 260 ± 118 281 ± 49 270 ± 69 320 ± 84120 303 ± 170 378 ± 223 364 ± 152 363 ± 51

Pore water velocity (mm d�1)Aldicarb28 1.81 ± 0.15 1.90 ± 0.19 4.59 ± 0.53 4.39 ± 0.4454 0.84 ± 0.37 0.84 ± 0.12 1.08 ± 0.07 1.11 ± 0.3775 0.84 ± 0.24 0.81 ± 0.31 1.16 ± 0.12 1.27 ± 0.25120 0.13 ± 0.16 0.27 ± 0.14 0.16 ± 0.07 0.51 ± 0.25Carbofuran28 3.18 ± 0.62 3.07 ± 0.15 9.69 ± 2.64 7.94 ± 2.0254 1.54 ± 0.18 1.48 ± 0.26 2.07 ± 0.59 1.98 ± 0.8175 1.46 ± 0.72 1.62 ± 0.27 1.76 ± 0.42 2.02 ± 0.46120 1.13 ± 0.68 1.39 ± 0.80 1.38 ± 0.53 1.42 ± 0.15

Dispersion coefficient (mm2 d�1)Aldicarb28 38 ± 7 42 ± 8 101 ± 24 92 ± 1854 27 ± 21 24 ± 7 34 ± 5 38 ± 2775 32 ± 17 31 ± 18 53 ± 11 66 ± 26120 2 ± 4 6 ± 4 11 ± 19 18 ± 15Carbofuran28 120 ± 46 108 ± 11 470 ± 231 314 ± 16154 79 ± 18 75 ± 26 131 ± 72 127 ± 9975 107 ± 77 114 ± 37 128 ± 59 168 ± 75120 98 ± 92 145 ± 126 128 ± 103 121 ± 26

a 28, 54, 75, and 120 flux days are equivalent to 1, 14, 42, and 120 days.

Table 4Average values of retardation coefficient, R (±SD), degradation rate, l (±SD), andapparent half-life, t1/2, for aldicarb and carbofuran as a function of treatments for thestudy (n = 4)

Parameter Initially wet Initially dry

0-h Delay 24-h Delay 0-h Delay 24-h Delay

AldicarbRa 16.9 ± 4.9 19.8 ± 12.0 6.4 ± 0.7 6.0 ± 1.2la 48 ± 10 35 ± 14 29 ± 6 20 ± 6t1/2

a 14 20 24 35CarbofuranR 12.6 ± 9.9 8.4 ± 1.5 4.4 ± 1.1 3.8 ± 0.8l 33 ± 21 26 ± 10 15 ± 3 15 ± 5t1/2 21 27 46 47

a Units: R (unitless), l � 1000 (d�1), t1/2 (d).

272 H.V. Kazemi et al. / Chemosphere 74 (2009) 265–273

(P < 0.10) for the first depth moment of aldicarb 75 flux days (42 d)after application (dry > wet). The aldicarb first depth moment wassignificantly higher (P < 0.05) for the 24-h Delay treatment com-pared to the 0-h Delay treatment 120 flux days (120 d) after appli-cation. Antecedent soil water content, initial irrigation delay andthe interaction of antecedent soil water content with initial irriga-tion delay did not have any significant effects on the first depthmoment for carbofuran on any sampling date. These results wereprobably due to high variation among replicates.

Comparing the two pesticides, the depth of the first momentwas consistently deeper for carbofuran than aldicarb in all treat-ments. Even though there were not significant differences for mostdates, there was a trend of slightly deeper movement in dry plotscompared to wet plots. This is an indication of deeper transportof pesticides through air-filled macropores present in initially dryplots.

The first and second depth moments were used to estimate theaverage pore water velocity and dispersion coefficients for the twopesticides. Table 3 presents the average pore water velocity foraldicarb and carbofuran, respectively. Since both aldicarb andcarbofuran are organic compounds, they undergo chemical andbiological degradation and sorption processes while being trans-ported by both convective and dispersive flow. However, theirvelocity values can be compared relative to each other.

The average pore water velocity of carbofuran was consistentlyhigher than that of aldicarb (Table 3). In most cases, carbofuranmoved approximately twice as fast as aldicarb. Also, the pore watervelocity of both chemicals varied significantly as a function of sam-

pling date. Velocity values of both pesticides decreased as a func-tion of time for all treatments.

Antecedent water content significantly affected pore watervelocity on the first three sampling dates for aldicarb and the firsttwo sampling dates for carbofuran; values were always higher forthe dry treatment. Initial timing of irrigation only affected the porewater velocity for aldicarb on the last sampling date with the 24-hDelay having a higher value (P < 0.01) compared to the 0-h Delay.

Dispersion coefficient values for aldicarb and carbofuran as afunction of sampling date are presented in Table 3. The carbofurandispersion coefficient values were significantly higher than thosefor aldicarb throughout the experiment. The variability of valuesfor carbofuran was higher than those of aldicarb. The aldicarb dis-persion coefficient values for the dry treatment were significantlyhigher compared to the wet treatment on all dates except 54 fluxdays (14 d); carbofuran values were significant only on the firstsampling date.

The laboratory-estimated carbon partition coefficients were51.4 and 57.5 cm3 g�1 for aldicarb and carbofuran, respectively.Using these values and measured soil properties of organic carboncontent, bulk density, and water content, computed retardationvalues for aldicarb were 5.23 and 4.75, and for carbofuran were5.73 and 5.19 for the initially dry and wet plots, respectively. Aldi-carb and carbofuran retardation values estimated by comparingpore water velocity values are given in Table 4. These estimatesfor aldicarb are about 3.8 times higher for the wet plots and about18% higher for the dry plots than those estimated from soil prop-erty values. Initial soil water content significantly influenced thefield estimated values (dry < wet) probably due to more transportin the dry plots (compare pore water velocity values in Table 3).

Retardation coefficient values for carbofuran estimated frompore water velocity data (Table 4) are 29% lower in the dry plotscompared to the soil property-estimated value. However, retarda-tion coefficient values for the wet plots are about 2 times higherthan the soil property-estimated value. Values for dry plots areabout 250% lower (P < 0.05) compared to the wet plots suggestingdeeper movement with lower initial soil water content.

The laboratory-measured half-life of aldicarb was 25 d usingradioactive-labeled material using the soil from 0 to 10 cm depth.With non-labeled material, the aldicarb half-life was found to beabout 6 d. Field-estimated values of the aldicarb degradation coef-ficient were 68% higher in the initially wet plots compared to theinitially dry plots (Table 4). Values in the 0-h irrigation Delay were43% higher compared to the 24-h irrigation Delay. This indicatesthat aldicarb persistence is substantially higher when the initialwater content is lower and a 1 d Delay of initial irrigation occurs.The half-life of aldicarb was a bit lower than the average value of50 d (Table 2), but was similar to the value obtained using radio-la-beled material in the laboratory (25 d). The degradation rate values

H.V. Kazemi et al. / Chemosphere 74 (2009) 265–273 273

estimated from the field data seem to be higher than those valuesfrom the literature. An apparent half-life of 70 d was reported foraldicarb by Jury et al. (1984) which is much slower than that foundin this study. Hornsby et al. (1990) reported an apparent half-life ofabout 14 d when turned into its metabolites (sulfoxide and sul-fone) which have a longer half-life. The apparent half-life of totalaldicarb (including the metabolites) was estimated to be about69 d. Guo et al. (1999) also evaluated differences in degradationin solution and adsorbed phases with the adsorbed phase beingmore persistent. The differences among these reports can be dueto influence of soil and environmental factors at this site.

For carbofuran, the laboratory-measured apparent half-life inthe 0–10 cm horizon was found to be approximately 70 d for bothradioactive-labeled and non-labeled pesticides. The field-esti-mated carbofuran degradation rate values for the initially dry plotswere significantly lower (P < 0.05) than those for the wet plotsindicating slower degradation in the dry plots (Table 4). The appar-ent half-life of carbofuran estimated in the laboratory was 70 dwhich gave a degradation coefficient value of 0.0099 d�1. Theapparent half-life of carbofuran in the literature is 41 d (Table 2)and 37 d (Jury et al., 1984). These values are similar to those mea-sured in the field experiment and are within the range of reportedvalues. Timing of the initial irrigation did not have a significantinfluence on the degradation rate for carbofuran. In addition tothe effect of soil water content, microbial rhizosphere populationsare critical in carbofuran degradation (Buyanovsky et al., 1995)which is a major factor in decreased carbofuran degradation atdeeper soil depths (Buyanovsky et al., 1993).

A similar study was conducted on claypan soils (subsurfacehorizon at the 0.3 m depth with > 40% smectitic clays) which eval-uated the effects of different values of antecedent water content onatrazine and alachlor transport (Kazemi et al., 2008a). Results fromthat study also showed deeper transport of these herbicides in aninitially dry treatment (near permanent wilting coefficient). How-ever, results in the claypan study showed deeper movement of her-bicides probably due to more pronounced shrinkage cracks havingformed prior to chemical application compared to the currentstudy. Kazemi et al. (2008a) also found slower degradation ratesin the dry treatment for atrazine compared to plots near fieldcapacity at the initiation of the experiment.

4. Conclusions

A field study was conducted to evaluate the effects of highlycontrasting antecedent soil water content and the timing of initialirrigation on insecticide transport and degradation. The initiallydry treatment (very dry conditions) showed evidence of deepertransport of the aldicarb and carbofuran, attributed to greater lev-els of air-filled pores (compared to the wet Treatment)at the initi-ation of the experiment. Aldicarb was detected 0.10 m deeper(P < 0.01) in the dry vs wet treatment on two of the four samplingdates and carbofuran was detected at least 0.10 m deeper (P < 0.05)on all of the sampling dates. The pore water velocity was found tobe higher (P < 0.10) in the dry vs wet treatments on three of foursampling dates for aldicarb and two of four dates for carbofuran.Results suggest evidence of deeper movement in the initially drytreatment. Degradation rate coefficients were approximately 40–49% lower in the dry compared to the wet plots for aldicarb andcarbofuran, respectively. When the initial irrigation was delayedfor 24 h, a 30% reduction in aldicarb degradation occurred. Implica-

tions of this study suggest that application of aldicarb and carbofu-ran on initially very dry soils will enhance transport andpersistence of these chemicals.

References

Beulke, S., Dubus, I.D., Brown, C.D., Bottesburen, B., 2000. Simulation of pesticidepersistence in the field on the basis of laboratory data – a review. J. Environ.Qual. 29, 1371–1379.

Beulke, S., van Beinum, W., Brown, C.D., Mitchell, M., Walker, A., 2005. Evaluation ofsimplifying assumptions of pesticide degradation in soil. J. Environ. Qual. 34,1933–1943.

Brown, C.D., Fryer, C.J., Walker, A., 2001. Influence of topsoil tilth and soil moisturestatus on losses of pesticide to drains from a heavy clay soil. Pest. Manag. Sci.57, 1127–1134.

Buyanovsky, G.A., Gajda, A.M., Kremer, R.J., Peiczonka, G.J., 1993. Effect of soil depthon carbofuran and aldicarb degradation. In: Linn, D.M. (Ed.) Sorption andDegradation of Pesticides and Organic Chemicals in Soil. Soil Sci. Soc. Am. Spec.Publ. No. 32, Madison, WI, pp. 65–71.

Buyanovsky, G.A., Kremer, R.J., Gajda, A.M., Kazemi, H.V., 1995. Effect of corn plantpopulations on pesticide degradation. Bull. Environ. Contam. Toxicol. 55, 689–696.

Cogger, C.G., Stark, J.D., Bristow, P.R., Getzin, L.W., 1998. Transport and persistenceof pesticides in alluvial soils: II. Carbofuran. J. Environ. Qual. 27, 551–556.

Dousset, S., Babut, M., Andreux, F., Schiavon, M., 2004. Alachlor and bentazonelosses from subsurface drainage of two soils. J. Environ. Qual. 33, 294–301.

Gish, T.J., Coffman, C.B., 1987. Solute transport under no-till field corn. T. ASAE. 30,1358–1363.

Guo, L., Wagenet, R.L., Jury, W.A., 1999. Adsorption effects on kinetics of aldicarbdegradation: equilibrium model and application to incubation and transportexperiments. Soil Sci. Soc. Am. J. 63, 1637–1644.

Hornsby, A.G., Rao, P.S.C., Jones, R.L., 1990. Fate of aldicarb in the unsaturated zonebeneath a citrus grove. Water Resour. Res. 26, 2287–2302.

Jury, W.A., Roth, K., 1990. Transfer Functions and Solute Movement Through Soil:Theory and Applications. Birkhauser, Boston, MA.

Jury, W.A., Sposito, G., 1985. Field calibration and validation of solute transportmodels for unsaturated zone. Soil Sci. Soc. Am. J. 49, 1331–1341.

Jury, W.A., Farmer, W.J., Spencer, W.F., 1984. Behavior assessment model for traceorganics in soil: IV. Review of experimental evidence. J. Environ. Qual. 13, 580–586.

Kazemi, H.V., Anderson, S.H., Goyne, K.W., Gantzer, C.J., 2008a. Atrazine and alachlortransport in claypan soils as influenced by differential antecedent soil watercontent. J. Environ. Qual. 37, 1599–1607.

Kazemi, H.V., Anderson, S.H., Goyne, K.W., Gantzer, C.J., 2008b. Spatial variability ofbromide and atrazine transport parameters for a Udipsamment. Geoderma 144,545–556.

Kung, K.-J.S., Steenhuis, T.S., Kladivko, E.J., Gish, T.J., Bubenzer, G., Helling, C.S., 2000.Impact of preferential flow on the transport of adsorbing and non-adsorbingtracers. Soil Sci. Soc. Am. J. 64, 1290–1296.

Malone, R.W., Shipitalo, M.J., Wauchope, R.D., Sumner, H., 2004. Residual andcontact herbicide transport through field lysimeters via preferential flow. J.Environ. Qual. 2004; 33, 2141–2148.

Marei, A.S., Soltan, H.R., Mousa, A., Khamis, A., 2000. Leaching and mobility ofcarbofuran from granular and alginate controlled release formulations. J. Agric.Sci. 134, 405–412.

Porro, I., Wierenga, P.J., 1993. Transient and steady-state solute transport through alarge unsaturated soil column. Ground Water 31, 193–200.

Rao, P.S.C., Jessup, R.E. 1982. Sorption and movement of pesticides and other toxicorganic substances in soils. In: Nelson, D.W., Elrick, D.E., Tanji, K.K. (Eds.)Chemical Mobility and Reactivity in Soil Systems. Soil Sci. Soc. Am. Spec. Publ.No. 11, Madison, WI, pp. 183–201.

Rao, P.S.C., Hornsby, A.G., Jessup, R.E., 1985. Indices for ranking the potential forpesticide contamination of groundwater. Soil Crop Sci. Soc. Fl. 44, 1–8.

SAS Institute, Inc., 2005. SAS Online Documentation. Version 9.1.3. SAS Inst., CaryNC.

Sievers, D.M., Fulhage, C.D., 1989. Quality of Rural Well Water-North Missouri.Missouri Agricultural Experiment Station Special Report No. 402, University ofMissouri-Columbia.

Thomas, G.W., Phillips, R.E., 1979. Consequences of water movement in macropores.J. Environ. Qual. 8, 149–152.

United States Department of Agriculture, Agricultural Research Service, 2008. TheARS Pesticide Properties Database [Online]. Available at <http://www.ars.usda.gov/Services/docs.htm?docid=14199> (verified 20 May 2008).

Ward, A.D., Trimble, S.W., 2003. Environmental Hydrology, second ed. LewisPublishers, Boca Raton, FL.

Zindahl, R.L., Clark, S.K., 1982. Degradation of three acetanilide herbicides in soil.Weed Sci. 30, 545–552.