Effect of water management practice on pesticide behavior in paddy water

Abstract

The fate and transport of three herbicides commonly used in rice production in Japan were compared using two water management practices. The herbicides were simetryn, thiobencarb and mefenacet. The first management practice was an intermittent irrigation scheme using an automatic irrigation system (AI) with a high drainage gate and the second one was a continuous irrigation and overflow drainage scheme (CI) in experimental paddy fields. Dissipation of the herbicides appeared to follow first order kinetics with the half-lives (DT50) of 1.6–3.4 days and the DT90 (90% dissipation) of 7.4–9.8 days. The AI scheme had little drainage even during large rainfall events thus resulting in losses of less than 4% of each applied herbicide through runoff. Meanwhile the CI scheme resulted in losses of about 37%, 12% and 35% of the applied masses of simetryn, thiobencarb and mefenacet, respectively.The intermittent irrigation scheme using an automatic irrigation system with a high drainage gate saved irrigation water and prevented herbicide runoff whereas the continuous irrigation and overflow scheme resulted in significant losses of water as well as the herbicides. Maintaining the excess water storage is important for preventing paddy water runoff during significant rainfall events. The organic carbon partition coefficient Koc seems to be a strong indicator of the aquatic fate of the herbicide as compared to the water solubility (SW). However, further investigations are required to understand the relation between Koc and the agricultural practices upon the pesticide fate and transport. An extension of the water holding period up to 10 days after herbicide application based on the DT90 from the currently specified period of 3–4 days in Japan is recommended to be a good agricultural practice for controlling the herbicide runoff from paddy fields. Also, the best water management practice, which can be recommended for use during the water holding period, is the intermittent irrigation scheme using an automatic irrigation system with a high drainage gate.

Full text

Page 1

This article was originally published in a journal published by
Elsevier, and the attached copy is provided by Elsevier for the
author’s benefit and for the benefit of the author’s institution, for
non-commercial research and educational use including without
limitation use in instruction at your institution, sending it to specific
colleagues that you know, and providing a copy to your institution’s
administrator.
All other uses, reproduction and distribution, including without
limitation commercial reprints, selling or licensing copies or access,
or posting on open internet sites, your personal or institution’s
website or repository, are prohibited. For exceptions, permission
may be sought for such use through Elsevier’s permissions site at:
http://www.elsevier.com/locate/permissionusematerial

Page 2

Author's personal copy
Effect of water management practice on pesticide behavior
in paddy water
Hirozumi Watanabea,*, My Hoang Tra Nguyenb, Komany Souphasayc, Son Hong Vua,
Thai Khanh Phonga, Julien Tournebized, Satoru Ishiharae
aTokyo University of Agriculture and Technology, 3-5-8 Saiwaicho, Fuchu, Tokyo 183-8509, Japan
bJICA Vietnam, 2 Thi Sach Street, District 1, Ho Chi Minh City, Vietnam
cWater Resources Coordination Committee, Prime Minister Office, Nahaidieo, Chanthabouly District, Vientiane, Laos
dCEMAGREF, Parc de Tourvoie, BP44, 92163 Antony, France
eNational Institue of Agro-Environmental Sciences, 3-1-1 Kannondai, Tsukuba, Ibaraki 305-8604, Japan
agricultural water management 88 (2007) 132–140
a r t i c l e i n f o
Article history:
Accepted 10 October 2006
Published on line 21 November 2006
Keywords:
Pesticide
Paddy field
Water management
a b s t r a c t
The fate and transport of three herbicides commonly used in rice production in Japan were
compared using two water management practices. The herbicides were simetryn, thioben-
carb and mefenacet. The first management practice was an intermittent irrigation scheme
usinganautomaticirrigationsystem(AI)withahighdrainagegateandthesecondonewasa
continuous irrigation and overflow drainage scheme (CI) in experimental paddy fields.
Dissipation of the herbicides appeared to follow first order kinetics with the half-lives
(DT50) of 1.6–3.4 days and the DT90(90% dissipation) of 7.4–9.8 days. The AI scheme had little
drainage even during large rainfall events thus resulting in losses of less than 4% of each
applied herbicide through runoff. Meanwhile the CI scheme resulted in losses of about 37%,
12% and 35% of the applied masses of simetryn, thiobencarb and mefenacet, respectively.
The intermittent irrigation scheme using an automatic irrigation system with a high
drainage gate saved irrigation water and prevented herbicide runoff whereas the contin-
uous irrigation and overflow scheme resulted in significant losses of water as well as the
herbicides. Maintaining the excess water storage is important for preventing paddy water
runoff during significant rainfall events. The organic carbon partition coefficient Kocseems
to be a strong indicator of the aquatic fate of the herbicide as compared to the water
solubility (SW). However, further investigations are required to understand the relation
between Koc and the agricultural practices upon the pesticide fate and transport. An
extension of the water holding period up to 10 days after herbicide application based on
the DT90from the currently specified period of 3–4 days in Japan is recommended to be a
good agricultural practice for controlling the herbicide runoff from paddy fields. Also, the
best water management practice, which can be recommended for use during the water
holding period, is the intermittent irrigation scheme using an automatic irrigation system
with a high drainage gate.
# 2006 Elsevier B.V. All rights reserved.
* Corresponding author. Tel.: +81 42 367 5889; fax: + 81 42 367 5889.
E-mail address: pochi@cc.tuat.ac.jp (H. Watanabe).
0378-3774/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.agwat.2006.10.009
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/agwat

Page 3

Author's personal copy
holding period in Japan needs a detailed investigation on
water management practices specific to the Asian monsoon
climate.
The physico-chemical properties of pesticides such as the
solubility (SW) and the soil water partitioning coefficient
normalized for the organic carbon content (Koc) are important
parameters for the preliminary assessment of the pesticide
fate and exposure risks. Several studies have suggested that
1.Introduction
Herbicide has brought a great benefit to rice cultivation
(Matsunaka, 2001). Therefore, a large number of herbicides of
variousformulationsareusedinJapan,placingthecountryin
second place in terms of the intensive use of pesticides
worldwide (Sabik et al., 2000). However, monitoring studies
for pesticide concentrations in river systems in Japan
detected several herbicides commonly used in paddy fields
(Nakamura, 1993; Ebise et al., 1993; Nagafuchi et al., 1994).
Toxicology studies also showed the negative impact of
herbicides on aquatic biota (Ueji and Inao, 2001; Hatakeyama
et al., 1999).
Previous monitoring studies have demonstrated that the
discharge of pesticide reaches its peak during the period
shortly after the pesticide application time (Ebise and Inoue,
2002; Sudo et al., 2002). The surface drainage/runoff of paddy
water containing appreciably high concentrations of pesti-
cides is obviously responsible for this pollution.
Moreover, significant losses of pesticides used in the paddy
fields seem to occur after significant rainfall events. While the
pesticide loss without an intense rainfall after application was
less than 5%, it reached 20–30% if a significant rainfall event
follows pesticide application (Nagafuchi et al., 1994). A single
rainfall event can cause substantial pesticide loss to the
surface water (Flury, 1996). In Japan, Ebise and Inoue (2002)
also indicated that the surface runoff from paddy fields
increased during heavy storm events. Meanwhile, Vu et al.
(2004) reported increased discharge from the paddies after
rainfall events exceeding 1.5 cm day?1. The excess water
storage in a paddy to accommodate the excess precipitation
can be an applicable solution for this matter. Mishra et al.
(1998) reported that almost 100% of the intense rainfall can be
storedin arain-fedpaddyplothavingtheweirheightof30 cm.
Watanabe et al. (2006a) reported that excess water storage
created by the high drainage gate prevented herbicide runoff
during significant rain events.
The water holding period seems to be a key practice for
controlling the pesticide discharge from paddy fields. In
California, the water holding requirement after pesticide
application has successfully reduced the concentrations of
rice pesticides in streams (Newhart, 2002). However in Japan,
the water holding period, which is recommended and written
on pesticide labels, is only 3–4 days after pesticide application
without a proper extension or specific water quality program.
TheimportanceofextendingthewaterholdingperiodinJapan
has been previously discussed (Ishii et al., 2004; Watanabe
et al., 2006a). It should be noted that the popular practice in
monsoonpaddiesisshallowwaterponding,henceirrigationis
still performed during the water holding period to keep the
appropriate water level while maintaining an excess water
storage capacity. Therefore, imposing the proper water
the magnitude of the pesticide loss depends on the solubility
of the pesticides in the paddy water (Ebise et al., 1993; Ueji and
Inao,2001;Sudoetal.,2002).Nevertheless,anumberofstudies
have also pointed out a correlation between the sorption
behaviors of pesticides (Kocor log Pow) with their loss (Fajardo
et al., 2000; Nakano et al., 2004). More studies should thus be
carried out to determine the correlation of the pesticide loss
with their intrinsic properties.
In this study, we monitored the fate and transport of three
commonly applied rice herbicides having distinctive physico-
chemical properties. Two water management practices, one
intermittent irrigation scheme using an automatic irrigation
system with a high drainage gate and one continuous
irrigation overflow drainage scheme with a low drainage gate
were used. The objective of this study was to investigate the
effect of the water management practice and pesticide
chemical properties on their behaviors in paddy fields so that
a good water management practice for reducing herbicide loss
from paddy fields can be proposed.
2.Materials and methods
2.1.Field experiment
The pesticide fate and transport monitoring was conducted in
two paddy plots (27.9 m ? 49.0 m) on the experimental farm of
the Tokyo University of Agriculture and Technology (TUAT) in
Fuchu, Tokyo, from 27 May to 30 June 2003 (Fig. 1). The
physico-chemical properties of the paddy soil in these plots
are listed in Table 1.
One plot was assigned to the intermittent irrigation
scheme with a high drainage gate (denoted as AI plot) using
an automatic irrigation system (Rakutaro1, Nihon System
Kaihatsu Co. Ltd., Saitama). Irrigation was set to start when
the paddy water level is below about 2 then continue up to
5 cm. The height from the paddy soil to the bottom of the
notch of the drainage gate for the AI plot was set at 7.5 cm in
order to minimize the paddy water drainage. The other was
assigned to the continuous irrigation and overflow drainage
schemewithalowerdrainagegate(denotedasCIplot).TheCI
plotwaskeptirrigatedatalltimes,buttheflowratedepended
on the pressure head in the pipeline. The corresponding
heightofthedrainagegatefromthesoilsurfacefortheCIplot
was 2.5 cm.
Fig. 1 – Layout of the experimental plots.
agricultural water management 88 (2007) 132–140
133

Page 4

Author's personal copy
methylthio-1,3,5-triazine-2,4-diamine), mefenacet (2-(1,3-
benzothiazol-2-yloxy)-N-methylacetanilide) and thiobencarb
(S-4-chlorobenzyl diethyl (thiocarbamate)) were applied as a
granular formulation, KumishotSM1(4.5% simetryn, 4.5%
mefenacet, 15.0% thiobencarb, 2.4% MCPB), at a rate of
10 kg ha?1on 27 May 2003, 21 days after transplanting. The
paddy water samples were taken at 1, 3, 7, 14, 21, and 35 days
after herbicide application (DAHA). At each sampling, five
Following the general paddy field preparation and water
ponding, the paddy soil was puddled and leveled by several
passes of a rotary tiller under a few centimeters ponding-
water condition. After the soil preparation, about 20-day-old
rice seedlings (Oryza sativa L. cv. var. kogane-mochi) were
transplanted with a spacing of 16 cm ? 30 cm on 12 May 2003.
The water balance variables included irrigation, surface
drainage, evapotranspiration and percolation. Precipitation
data were collected from a meteorological station in TUAT.
The volume of the irrigation water in each treatment was
monitored with a flow meter connected to a data logger. The
depth of the paddy water was monitored with a water level
sensor (LSP-100, UIJIN Co. Ltd., Tokyo) and the volume of the
surface runoff through the V-notch weir was calculated using
the paddy water level data from the equation described by Rao
and Muralidhar (1963). Evapotranspiration (ET) was observed
using a water level sensor in a lysimeter box (35 cm ?
50 cm ? 30 cm) containing 15 cm of puddled soil in a flood
condition with four growing rice plants. The total percolation
including lateral seepage was calculated from the remaining
monitored hydrological data in the following water balance
equation:
dhPW
dt
¼ RAIN þ IRR ? DRAIN ? PERC ? ET(1)
where hPWis the depth of water in a paddy field (cm), t is the
time (day), RAIN is the rainfall (cm day?1), IRR is the irrigation
(cm day?1), DRAIN is the paddy water discharge from a paddy
plot including the overflow from the drainage gate (cm day?1),
PERC is the percolation (cm day?1) including the vertical per-
colation through the paddy soil and the seepage through the
levees and plot borders, and ET is the evapotranspiration
(cm day?1). In addition, the distribution of the vertical perco-
lation through paddy soil was measured using PVC rings with
a diameter of 16 cm for 53 spots in each plot.
Threeactiveingredients,simetryn(N2,N4-diethyl-6-
100-ml samples of the paddy water taken from five spots
(Fig. 1) weremixedtogetherto makeone composite sample.At
the same time, 500 ml water samples were taken at the
drainage gate. The samples were kept frozen until the
chemical analysis.
2.2.Gas chromatography analysis
Mefenacet, thiobencarb and simetryn in the water samples
were extracted by liquid–liquid extraction with dichloro-
methane. The thawed samples were filtered through 1.2 mm
glass micro-fiber filters (GF/C, Whatman, Maidstone, UK) then
the filtered water samples were mixed with 30 g of NaCl and
extracted twice with 400 ml of dichloromethane. The samples
with a high concentration (1, 3, 7 DAHA) were diluted with
deionized water before extraction. The dichloromethane
solution was dehydrated by sodium sulfate and filtered with
silicontreatedfilterpaper(1PS,Whatman,Maidstone,UK).The
dichloromethane in the filtrate was evaporated by a rotary
evaporatoruptoavolumeof1 mlandthentocompletedryness
underagentlenitrogenstream.Theresiduewasresuspendedin
5 ml of acetone in an ultrasonic device. All of the solution was
transferredtoatesttubeandkeptat4 8CuntiltheGCanalysis.A
gas chromatographic system (GC-17A, Shimazu, Kyoto, Japan)
was used for the analysis. The column was a DB-5 column
(30 m ? 0.25 mm ? 0.32 mm) (J&W Scientific, Rancho Cordova,
USA). The temperature was programmed as follows: 60 8C
(2 min) ramped up to 140 8C at 10 8C min?1then to 270 8C at
5 8C min?1.Thetemperaturewasthenheldat270 8Cfor4 min.A
splitless injection mode was used with an injection volume of
4 ml. The carrier gas pressure was set at 40 kPa for 2 min then
increased to 64 kPa at 3 kPa min?1and continued to ramp at
1.5 kPa min?1to 103 kPa which was maintained for 4 min. The
herbicides were detected by the flame thermoionic detector
(FTD).Thedeterminationlimit
0.016 mg l?1for simetryn, mefenacet and thiobencarb, respec-
tively. The recovery values were 118%, 136% and 119% for
simetryn, mefenacet and thiobencarb, respectively.
were0.024,0.039and
3. Results and discussion
3.1.Water balance
The monitored water balance in the two plots (AI and CI) are
presentedinTable2.Theprecipitationforthestudyperiodof2
Table 1 – Physico-chemical properties of paddy soil (0–
15 cm) in experimental plot
Physico-chemical propertiesValue
pH (H2O)
Organic carbon content (%)
Total carbon content (%)
Total nitrogen content (%)
Cation exchange capacity
(cmolckg?1)
Particle density (g cm?3)
Sand (%)
Silt (%)
Clay (%)
Soil texture (ISSS)
6.5
3.96
4.77
0.44
22.5
2.50
37.6
31.8
30.6
Light clay (LiC)
ISSS: International Society of Soil Science.
Table 2 – Water balance in paddy plots
AI plotCI plot
Input (cm)
Irrigation
Precipitation
Total
34.8
17.6
52.4
56.5
17.6
74.1
Output (cm)
Drainage
Percolation
ET
Total
3.5
34.1
15.5
53.1
24.5
34.1
15.5
74.2
agricultural water management 88 (2007) 132–140
134

Page 5

Author's personal copy
months (May and June) was 17.6 cm. This amount was similar
to the 22-year average precipitation (1979–2000) in the region.
The average rates of percolation during the monitoring period
were 0.97 cm day?1for both the AI and CI plots. Although the
percolation rate was obtained from the rest of the measured
variables, the obtained values were still in the range of the
typical Japanese paddy field reported by Nakagawa (1967). The
similar percolation rates between the two plots may be due to
the consistent soil preparation and the equal ponding water
depth (about 4 cm).
Fig. 2 shows the spatial distribution of the measured
percolation rates. The measured vertical percolation followed
a log normal distribution with the mean and standard
deviation of ?0.57 (0.57 for normal value) and 0.73, respec-
tively. Note that there were two spots with extremely high
percolation (24.2 cm day?1at the Northwest corner of the AI
plot and 25.7 cm day?1at the Southwest corner of CI plot)
therefore they were excluded from the statistical analysis.
However, the average values of the measured percolation
rates including these hot spots coincided with the monitored
data. Some points with a high rate were also observed along
the concrete banks of the plots that means the seepage and/or
under-bund percolation may contribute a significant portion
to the water flow at these points. According to Sharma and De
Datta (1985), puddling of rice fields before transplanting or
direct seeding reduces the percolation losses. However, for
machinery puddling, the edges and corners along the banks
are usually left undisturbed because the plow cannot reach
these areas. Tuong et al. (1994) reported that unpuddled soil
(about 1% of field area) increased the field percolation by a
factor of five and under-bund percolation may cause a further
two- to five-fold increase.
Fig. 3 shows the daily changes in precipitation, irrigation,
runoff/drainage,andpaddywaterdepthfortheAIandCIplots. Due to the continuous irrigation and overflow drainage
scheme, the CI plot discharged a significant amount of water
through the surface runoff. Also, the CI plot maintained a
relatively deep paddy water (about 4–6 cm) although it had a
lower drainage gate due to the discharge characteristics of the
Vnotchgate.ThedischargefromtheVnotchgateintheCIplot
was significant and that raised the paddy water level to drain
excess water during the monitoring period. Runoff from the CI
plot especially increased after significant rainfall events. The
total depth of drainage in the CI plot contributed about 33% of
the total output. Meanwhile, the AI plot with a high drainage
gate had almost no surface runoff. Since the water depth set
for the irrigation system in the AI plot was 4 cm while it had a
7.5 cm drainage gate, the AI plot basically had a 3.5-cm excess
water storage depth (EWSD). This EWSD in the AI plot
prevented the surface runoff during the rainfall events except
for two cases of extremely large rainfall.
As a consequence, during the monitoring period the CI plot
required 60% more irrigation water as compared to the AI plot.
The excessive amount of irrigation water was wasted as
runoff.ThisresultconfirmedtheconclusionofWatanabeetal.
(2006a) that important factors in water management for
saving water used in paddy rice production are
(1) reduction of overflow drainage from excess irrigation and
(2) efficientutilization ofprecipitationby storingas muchrain
water as possible in the fields.
Fig. 2 – Spatial distribution of the percolation rate within
the studied plots.
Fig. 3 – Observed water balance including precipitation
(Pre), irrigation (Irr), drainage (Drain) and paddy water
depth (Hpw) in AI plot (above) and CI plot.
agricultural water management 88 (2007) 132–140
135

Page 6

Author's personal copy
3.2.Pesticide fate and transport in paddy water
Fig. 4 shows the changes in the concentrations in the paddy
water of the three studied herbicides in the AI plot, the CI plot
aswellasthewatercollectedatthedrainagegateoftheCIplot.
In general, the dissipation trend of the three herbicides was
similar. The applied herbicides reached their peak concentra-
tions in the day following the application date. These
concentrations then rapidly declined during the 2 weeks after
herbicideapplication.Thepeakconcentrationofsimetrynwas
the highest followed by thiobencarb and mefenacet with the
values of 0.951, 0.595 and 0.498 mg l?1, respectively. At the end
of the monitoring period (35 DAHA), the measured concentra-
tions of the studied herbicides were lower than the advisory
levels (0.06 mg l?1for simetryn, 0.009 mg l?1for mefenacet,
the level for thiobencarb is not available) for public surface
water quality in Japan (MOE, 1994). The highest concentration
at 35 DAHA is that of simetryn at the drainage gate with the
value of 0.01 mg l?1. The 3-year pesticide monitoring study at
NIAES (Fajardo et al., 2000) also provided a comparable
dissipation pattern of mefenacet in paddy water considering
that thepresentexperiment onlyhadathird oftheapplication
rate of the NIAES study. A similar pattern for simetryn was
also reported by Inao et al. (2001). However, the maximum
concentration observed in that study was only 0.3 mg l?1for
simetryn, while in both studies, the same application rate was
used.
Concerning the effect of the different water management
practices on the behavior of the applied herbicides, the
concentrations of herbicides in the CI plot were consistently
lower than the concentrations in the AI plot. This was
probably due to dilution and loss of the dissolved herbicides,
which were greater in the continuous irrigation and overflow
drainage scheme. As shown by the water balance, both the
water input as well as the water drainage in the CI plot were
much higher than those in the AI plot.
The herbicide concentrations in the drainage water from
the CI plot were always higher than the concentrations
measured in the composite samples of the same sampling
date (representing the average concentration of the plot). The
average difference was about 70% throughout the monitoring
period. This phenomenon resulted from the structure of the
plot with irrigation water coming in one side and drainage
water goingout theother side. Newlyirrigatedwatertherefore
swept the standing water that contained a high concentration
of herbicides toward the drainage gate creating a spatial
variation in the herbicide concentrations within the plot. The
evidence for this phenomenon was reported by Watanabe
et al. (2006b),whodiscussed theapplicationoftheELISAkitfor
the risk assessment of rice herbicides used in the same paddy
field. The bensulfuron-methyl concentrations in this paddy
field at 43 DAHA ranged from below 0.03 mg l?1at the sampling
pointneartheirrigationinlettoabout1 mg l?1atpointscloseto
the drainage gate.
Considering the cumulative losses of the three studied
herbicides (Fig. 5), the losses in the CI plot were significantly
higher than those in the AI plot for all the herbicides. From the
CI plot, the total herbicidelosses were about 37%, 12% and 35%
of the applied mass for simetryn, thiobencarb and mefenacet,
respectively. Meanwhile, almost no herbicide lost from the AI
plot since the drainage was very small. The total losses in the
AI plotwere3.8%,1.2% and2.7%for simetryn, thiobencarband
mefenacet, respectively. The losses of thiobencarb were
significantly lower than the other two herbicides. This is
possibly due to its lower Kocvalue as compared to the others.
More discussion on the herbicide properties and their fate will
be presented in a later section.
Most of herbicide mass was lost during the first week after
application (Fig. 5). It should be noted that about 60% of the
total loss in the CI plot and about 90% of the total loss in the AI
plot occurred at 4 DAHA in a significant rainfall event
consisting of two consecutive rain days. Therefore, pesticide
runoff control during the earlier period is extremely impor-
tant. As reported from the simulation of the rice herbicide fate
and transport (Watanabe and Takagi, 2000), a paddy plot
managed by a continuous irrigation and overflow drainage
scheme had 86% of the total mefenacet loss during the first
week. Especially, the pesticide losses by the runoff at 2 and 3
Fig. 4 – Concentration of the studied herbicides in paddy
water and drainage water of the studied plots AI and CI.
agricultural water management 88 (2007) 132–140
136

Page 7

Author's personal copy
Table 3 – Results of linear regression of herbicide concentrations in paddy water during the monitoring period
DAHA were 12% and 11%, respectively, since these days had
significant precipitations of 1.7 and 3.3 cm (Watanabe and
Takagi, 2000). Precipitation during this event was almost 8 cm
(Fig. 3), which resulted in an appreciable runoff. In such an
event, the height of the drainage gate plays a key role in
preventing runoff. Since the AI plot had a considerable excess
water storage created by a high drainage gate, the volume of
the runoff from the AI plot in this large rainfall event was
much less than that from the CI plot. Watanabe et al. (2006a)
also reported that a paddy plot and a similar set up with a high
drainagegate of 7.5 cm in 2001 hadneither paddy waterrunoff
nor herbicide runoff due to sufficient excess water storage
during the monitoring period.
The result implies that water management is important in
order to control the herbicide runoff from a paddy field
especially during the earlier period with a higher herbicide
concentration in the paddy water. A continuous irrigation and
overflow drainage scheme is not recommended during this
vulnerable period. The disadvantages of the continuous
irrigation and overflow drainage scheme have been reported
by previous studies and mainly focused on the significant
herbicide losses, up to about 60% of the applied mass,
depending on the volume of irrigation and precipitation
(Watanabe and Takagi, 2000; Inao et al., 2001; Watanabe
etal.,2006a,b).Inaddition,theuseofahigherdrainagegatefor
excess water storage should be encouraged to prevent runoff
from such a large rainfall event. Watanabe et al. (2006a) also
concluded from their previous study that an intermittent
irrigation scheme using an automatic irrigation system with a
high drainage gate was recommended as the best manage-
ment practice for controlling the herbicide losses from paddy
fields.
A water holding practice should be recommended espe-
cially during the earlier period when the herbicide concentra-
tions are high. Ishii et al. (2004) suggested that increasing the
water holding period from 3 to 7 days would reduce the
herbicide concentrations in the paddy water from 50% to 10%
of the maximum concentrations. However, the appropriate
holding time may need to be investigated for each herbicideas
in California where different water holding periods were set
depending on the herbicides (Newhart, 2002). Watanabe et al.
(2006a) discussed the advantages of extending the current
water holding period specified in Japan and suggested that the
water holding period to be at least 10 days according to the
DT90index. Note that the DT90values of the studied herbicides
ranged up to 9.8 days as discussed in the next section.
Plotting the natural logarithm of the herbicide concentra-
tions in paddy water versus DAHA indicated a first order
kinetic dissipation for each herbicide. Note that the herbicide
concentrations at the drainage gate were excluded from the
analysis. The regression equations and their R values for the
three herbicides in the two plots are presented in Table 3. All
the R2values showed a good reliability (P > 0.05) yet the R2
valuesoftheCIplotwerealwayslowerthanthecorresponding
number of the AI plot. The difference in the factors in the
equations and the R2values between the two plots may be
explained by the assumption that the herbicides in the CI plot
were more affected by physical factors such as dilution and
chemical loss from the excessive irrigation, and drainage thus
Fig. 5 – Cumulative losses of the studied herbicides in the
two studied plots.
AI plotCI plot
Equation
R2
Equation
R2
Simetryn
Thiobencarb
Mefenacet
y = ?0.301x + 0.065
y = ?0.309x + 0.045
y = ?0.235x ? 0.565
0.991
0.974
0.995
y = ?0.291x ? 0.350
y = ?0.257x ? 0.455
y = ?0.245x ? 0.880
0.936
0.977
0.946
agricultural water management 88 (2007) 132–140
137

Page 8

Author's personal copy
Table 4 – DT50and DT90of studied herbicides in two plots
shifted the dissipation mechanism away from the first order
kinetics.
The calculated dissipation times for the herbicide concen-
trations in the paddy water to reach 50% and 10% of the initial
concentration of the three herbicides (DT50 and DT90) are
shown in Table 4. The DT50was calculated from the data of an
1–7 DAHA period and the DT90was calculated from the data of
an1–14DAHAperiod.TheDT50andDT90ofmefenacetofabout
3 and10 dayswerein accordancewithother studies.Ishii et al.
(2004) reported the DT50 of 1.8 days for mefenacet while
Fajardo et al. (2000) reported 4.1 days. Ishii et al. (2004) also
reported that the DT90in the paddy water was 9.7 days for
mefenacet. For simetryn, the estimation of the DT50and DT90
from the data of Inao et al. (2001) also gave similar values.
Meanwhile, the DT50values of thiobencarb found in this
experiment were lower than that available in the literature.
Yusa andIshikawa (1977)determined the half-life (DT50) in the
paddywaterof thiobencarb to be6–9 dayswhile Ross andSava
(1986) predicted a half-life longer than 6 days in field water.
The difference may be due to the product formulation because
the peak concentration in the study of Ross and Sava (1986)
was observed at 4 DAHA. The remarkably shorter DT50of all
the herbicides in the CI plot than in the AI plot were due to the
herbicide losses through the runoff resulted from the
continuous irrigation and overflow drainage scheme.
All the above-mentioned studies indicated that the
herbicide concentrations in the paddy water were not
significantly reduced during the first week after herbicide
application. Watanabe et al. (2006a) also found similar
phenomena in their experiment. This is in contrast to the
current recommendation of the Japanese pesticide manufac-
turers of only a 4-day water holding period. In their monitored
paddy water, Watanabe et al. (2006a) reported significant
herbicide concentrations and concurrent losses were still
observed after 4 DAHA. Furthermore, as they reviewed the
literature, most of the DT90s of the rice herbicides ranged from
about 8–10 days. Therefore, it is essential to extend the
Japanese water holding period to prevent discharge of the
paddy water/pesticide after pesticide application.
For the present study, the estimated herbicide concentra-
tions at 4 DAHA in the AI plot were 22.3%, 31.0%, and 39.2% of
the maximum concentrations of simetryn, thiobencarb and
mefenacet, respectively, and the corresponding herbicide
losses at 4 DAHA were 22.2%, 7.1%, and 24.0% of the applied
masses, respectively. High herbicide concentrations during
the earlier period (longer than 3–4 days) have also been
observed for pretilachlor (Fajardo et al., 2000), thiobencarb
(Ross and Sava, 1986), carbofuran, molinate, tryclopyr, 2,4-D
(Johnson and Lavy, 1995; Johnson et al., 1995), and azimsul-
furon (Armbrust et al., 1999).
StudybyIshiietal.(2004)suggestedthatincreasingthewater
holding period would significantly reduce the herbicide con-
centrations. A longer water holding period of at least 10 days
after herbicide application based on the DT90 for further
dissipation might thus be a good agricultural practice for
controlling the herbicide runoff from paddy fields. Therefore,
the currently recommended water holding of 3–4 days in Japan
needs to be re-evaluated. In addition, the best water manage-
ment practice, which can be recommended for use during the
water holding period, is the intermittent irrigation scheme
using anautomatic irrigation systemwitha highdrainagegate.
The proper extension of the water management practice for
obtaining the optimum excess water storage during the water
holding period is very important in order to have an effective
control of the herbicide runoff from paddy fields. However, the
optimum excess water storage and its control of the water
management are dependent on hydrological parameters such
as precipitation, percolation and evapotranspiration, and a
detailed analysis may be required for future studies.
3.3.Herbicide properties and their fate
The peak concentration value of each active ingredient
correlated with its physico-chemical properties. Although
thiobencarbhasan application rate3 timeshigherthan thatof
mefenacet and simetryn, the highest thiobencarb concentra-
tion in the paddy water was only about half that of simetryn
and slightly higher than that of mefenacet. A negative
correlation was found between the maximum concentration
in the paddy water divided by the applied mass or the relative
peak concentration and its Kocvalue as shown in Fig. 6. The
linear regression indicated a weaker positive correlation
(Fig. 6) between the relative peak value and its water solubility
value (SW). However, SWhas been used as a parameter for
evaluating the pesticide fate and transport including the
prediction of herbicide losses in the field as well as in the
watershed (Maru, 1990; Sudo et al., 2002).
In the next step, the Kocas well as the SWof the herbicides
were also correlated with the dissipation time and the
herbicide loss through runoff. A similar observation was
recorded such that the SWhas a weaker correlation with the
DT50of the compound and the herbicide loss than does the Koc
(Fig. 7). In this study, the greater the Koc, the longer the DT50
and the smaller the herbicide loss by runoff. Also, the greater
DT50(day)DT90(day)
AICIAICI
Simetryn
Thiobencarb
Mefenacet
2.0
3.4
2.6
1.6
2.2
2.0
7.9
9.0
9.4
7.7
3.4
9.8
Fig. 6 – Linear regression of the relative peak concentration
of the studied herbicides vs. Koc(y(Koc)) and vs. SW(y(SW)).
agricultural water management 88 (2007) 132–140
138

Page 9

Author's personal copy
However, further investigations are required to understand
the relation between Kocand agricultural practices upon the
pesticide fate and transport.
An extension of the water holding period to 10 days after
herbicide application based on the DT90from the currently
recommended period of 3–4 days in Japan is recommended to
be a good agricultural practice for controlling herbicide runoff
from paddy fields. Also, the best water management practice,
the Koc, the lower the relative peak concentration of the
applied herbicide. Among the three herbicides, thiobencarb,
which had the highest Koc, had the lowest runoff loss.
Consequently, it is indicated that Kocis a good indicator of
the fate of herbicides in paddy fields and can be used together
with SWto make a good prediction of pesticide fate. However,
the importance of Koc (and SW) for the herbicide fate and
transport in paddy fields and its relation to agricultural
practices, such as water management and soil preparation,
should be further studied for the pesticide risk assessment.
4.Conclusions
From this study, controlling paddy water runoff by the
intermittent irrigation scheme using an automatic irrigation
system with a high drainage gate saved irrigation water and
prevented herbicide runoff whereas the continuous irrigation
and overflow scheme lost significant amount of water as well
as herbicides from the paddy fields. Maintaining the excess
water storage is important for reducing paddy water runoff
during significant rainfall events. Koc seems to be a good
indicator of the aquatic fate of herbicides as compared to SW.
which can be recommended for use during the water holding
period, is the intermittent irrigation scheme using an auto-
matic irrigation system with a high drainage gate. The proper
extension of the water management practice for obtaining the
optimum excess water storage during the water holding
period is very important in order to have effective control of
the herbicide runoff from paddy fields.
Acknowledgements
This work was supported by grants-in-aid for scientific
research #571 from the Ministry of Education, Culture, Sports,
Science and Technology of Japan and partially supported by a
Domestic Research Fellowship from the Japan Science
Technology CorporationandJAPAN-FRANCEintegratedaction
program SAKURA by the Japan Society for the Promotion of
Science (JSPS, no. 1522). Special thanks to Nihon System
Kaihatsu Co. Ltd, Saitama, for providing the automatic
irrigation system. We are also indebted to Mr. T. Motobayashi
at TUAT for the field operation and to Dr. M. Ishihara at the
National Institute of Agro-Environmental Sciences for the
technical advice on pesticide analysis.
r e f e r e n c e s
Armbrust, K.L., Okamoto, Y., Grochulska, J., Barefoot, A.C., 1999.
Predicting the dissipation of bensulfuron methyl and
azimsulfuron in rice paddies using the computer model
EXAMS2. J. Pest Sci. 24, 357–363.
Ebise, S., Inoue, T., 2002. Runoff characteristics of pesticides
from paddy fields and reduction of risk to the aquatic
environment. Water Sci. Technol. 45, 127–131.
Ebise, S., Inoue, T., Numabe, A., 1993. Runoff characteristics and
observation methods of pesticides and nutrients in rural
rivers. Water Sci. Technol. 28, 589–593.
Fajardo, F., Takagi, K., Ishizaka, M., Usui, K., 2000. Pattern and
rate of dissipation of pretilachlor and mefenaset in plow
layer and paddy water under lowland field conditions—a 3-
year study. J. Pest Sci. 25, 94–100.
Flury, M., 1996. Experimental evidence of transport of pesticides
through field soils—a review. J. Environ. Qual. 25, 25–45.
Hatakeyama, S., Inoue, T., Suzuki, K., Sugaya, Y., Kasuga, S.,
1999. Assessment of overall herbicide effects on growth of
duckweed in flowthrough aquarium carrying pesticide
polluted river water. J. Environ. Toxicol. 2, 65–75.
Inao, K., Ishii, Y., Kobara, Y., Kitamura, Y., 2001. Prediction of
pesticide behavior in paddy field by water balance on the
water management using pesticide paddy field model
PADDY. J. Pestic. Sci. 26, 229–235.
Ishii, Y., Inao, K., Kobara, Y., 2004. Dissipation of some herbicide
in a flooded rice field and increase of water holding time
after application of herbicide. Bull. Natl. Inst. Agro-Environ.
Sci., Jpn. (No. 23), 15–25 (in Japanese with English summary).
Johnson, W.G., Lavy, T.L., 1995. Persistence of carbofuran and
molinate in flooded rice culture. J. Environ. Qual. 24, 487–
493.
Johnson, W.G., Lavy, T.L., Gbur, E.E., 1995. Persistence of
triclopyr and 2,4-D in flooded and nonflooded soil. J.
Environ. Qual. 24, 493–497.
Maru, S., 1990. Outflow of pesticides from paddy lysimeters as
affected by the water solubility. J. Pestic. Sci. 15, 385–394 (in
Japanese with English Abstract).
Fig. 7 – Correlations of Kocvs. cumulative loss and Kocvs.
DT50.
agricultural water management 88 (2007) 132–140
139

Page 10

Author's personal copy
Matsunaka, S., 2001. Historical review of rice herbicides in
Japan. Weed Biol. Manage. 1, 10–14.
Mishra, A., Ghorai, A.K., Singh, S.R., 1998. Rainwater, soil and
nutrient conservation in rainfed rice lands in Eastern India.
Agric. Water Manage. 38 (1), 45–57.
Ministry of the Environment of Japan (MOE), 1994. Water quality
evaluation guidelines for pesticide in public surface waters
(in Japanese). http://www.env.go.jp/water/dojo/noyaku/law
data/f406kansuido 0086.htm.
Nagafuchi, O., Inoue, T., Ebise, S., 1994. Runoff pattern of
pesticide from paddy fields in the catchment area of
Rikimaru reservoir. Jpn. Water Sci. Technol. 30 (7), 137–144.
Nakagawa, S., 1967. Suiden-Yosuiro-Chosakeikakuhou (Survey
and planning method of water requirement in paddy fields).
Hatachinogyo-Shinkokai (Association for Promotion of
Upland Agriculture), Tokyo (in Japanese).
Nakamura, K., 1993. Studies on behavior and fate of pesticides
in soil and aquatic environment of agricultural lands. Bull.
Saitama Agric. Exp. Station 46 (in Japanese with English
summary).
Nakano, Y., Miyazaki, A., Yoshida, T., Ono, K., Inoue, T., 2004. A
study on pesticide runoff from paddy fields to a river in
rural region-1: field survey of pesticide runoff in the
Kozakura River, Japan. Water Res. 38 (13), 3017–3022.
Newhart, K., 2002. Rice pesticide use and surface water
monitoring 2002. Department of pesticide regulation,
California Environmental Protection Agency. http://
www.cdpr.ca.gov/docs/empm/pubs/ehapreps/eh0209/
eh0209.pdf.
Rao, N.S.G., Muralidhar, D., 1963. Discharge characteristics of
weir of finite-crest width. La Houille Blanche 5, 537–545.
Ross, L.J., Sava, R.J., 1986. Fate and thiobencarb and molinate in
rice fields. J. Environ. Qual. 15 (3), 220–225.
Sabik, H., Jeannot, R., Rondeau, B., 2000. Multiresidue methods
using solid-phase extraction techniques for monitoring
priority pesticides, including triazines and degradation
products, in ground and surface waters. J. Chromatogr. A
885, 217–236.
Sharma, P.K., De Datta, S.K., 1985. Effects of puddling on soil
physical properties and processes. In: Soil Physics and Rice,
IRRI, Los Ban ˜os, Philippines, pp. 217–234.
Sudo, M., Kunimatsu, T., Okubo, T., 2002. Concentration and
loading of pesticide residues in Lake Biwa basin (Japan).
Water Res. 36, 315–329.
Tuong, T.P., Wopereis, M.C.S., Marquez, J.A., Kropff, M.J.,
1994. Mechanisms and control of percolation losses in
irrigated puddled rice fields. Soil Sci. Soc. Am. J.
58, 1794–1803.
Ueji, M., Inao, K., 2001. Rice paddy field herbicide and their
effects on the environment and ecosystems. Weed Biol.
Manage. 1, 171–179.
Vu, H.S., Ishihara, S., Watanabe, H., Ueji, M., Tanaka, H., 2004.
Monitoring pesticide fate and transport in surface water in
Japanese paddy field watershed. In: Proceedings of The
Conference on Challenges and Opportunities for
Sustainable Rice-Bases Production System, Torino, Italy,
September 13–15, pp. 509–521.
Watanabe, H., Takagi, K., 2000. Simulation of pesticide
concentrations in paddy field by PCPF-1 model—in case of
mefenacet and its control of runoff and leaching. In:
Proceedings of the XIV Memorial CIGR World Congress,
Tsukuba, Japan, November 28–December 1.
Watanabe, H., Kakegawa, Y., Vu, S.H., 2006a. Evaluation of
the best management practice for controlling pesticide
runoff from paddy fields using an intermittent and a
continuous irrigation schemes. Paddy Water Environ. 4,
21–28.
Watanabe, H., My, H.T.N., Komany, S., Vu, S.H., Phong, T.K.,
Asami, Y., Tournebize, J., 2006b. Application of ELISA in
pesticide monitoring to control runoff of bensulfuron-
methyl and simetryn from paddy fields. Jpn. J. Pest Sci.
31, 123–129.
Yusa, Y., Ishikawa, K., 1977. Disappearance of benthiocarb
herbicide in irrigation water. In: Proceedings of the 6th
Conference of Asian Pacific Weed Science Society, vol. 2.
pp. 596–602.
agricultural water management 88 (2007) 132–140
140