Irrigation management for arsenic mitigation in rice grain: Timing and
severity of a single soil drying
Daniela R. Carrijo
⁎, Chongyang Li
, Sanjai J. Parikh
, Bruce A. Linquist
Department of Plant Sciences, University of California-Davis, 387 North Quad, Davis, CA 95616, USA
Department of Land, Air and Water Resources, University of California-Davis, 387 North Quad, Davis, CA 95616, USA
•Continuously ﬂooded irrigation favors
the accumulation of arsenicin rice grain.
•Treatments with one soil drying period
differing in timing and severity were
•Across all timings, severe soil drying
(≤−71 kPa) decreased total As
•However, inorganic As (the most toxic
to humans) not always decreased.
•Irrigation management affects both
total As and As speciation within rice
Received 21 June 2018
Received in revised form 14 August 2018
Accepted 17 August 2018
Available online 18 August 2018
Editor: Jay Gan
The accumulation of arsenic (As) in rice grain is a public health concern since As is toxic tohumans; in particular,
inorganic Ascan cause many chronic diseases including cancer. Rice crops are prone to accumulating As, in part,
due to the anaerobic soil conditions triggered by the traditional continuously ﬂooded irrigation practice. The ob-
jective of this study was to determine how the severity and the timing (i.e. crop stage) of a single soil drying pe-
riod impacttotal As concentration and As speciation within the rice (bothwhite and brown) grain,compared to a
continuously ﬂooded (CF)control. Drying thesoil until the perched watertable reached 15 cm below thesoil sur-
face (same severity as in the “Safe Alternate Wetting and Drying”), which in this study corresponded to a soil
(0–15 cm) water potential of ~0, did not decrease grain As concentrations, regardless of timing. Drying the soil
to Medium Severity [MS: soil (0–15 cm) water potential of −71 kPa] or High Severity [HS: soil (0–15 cm)
water potential of −154 kPa] decreased total As by 41–61%. However, inorganic As did not always decrease be-
cause the severity and the timing of soil drying affected As speciation within the grain. Overall, the soil had to be
dried to HS and/or late in the growing season (i.e., at booting or heading instead of at panicle initiation) to de-
crease inorganic As concentration inthe rice grain. This studyindicatesthat the imposition of a single soil drying
period within the growing season can mitigate As accumulation in rice grain, but it depends on the severity and
timing of the drying period. Further, irrigation management affects As speciation within the rice grain and this
must be considered if regulations on inorganic As are based on a percentage of total As measured.
Alternate wetting and drying
Rice is the primary staple food for more people on Earth than any
other crop and provides one quarter of the global calorie intake
Science of the Total Environment 649 (2019) 300–307
E-mail addresses: firstname.lastname@example.org (D.R. Carrijo), cyﬂi@ucdavis.edu (C. Li),
email@example.com (S.J. Parikh), firstname.lastname@example.org (B.A. Linquist).
0048-9697/© 2018 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
(GRiSP, 2013). However, rice can be a signiﬁcant route of human expo-
sure to arsenic (As), especially in populations with high rice consump-
tion (Meharg, 2004;Bhowmick et al., 2018). Different forms of As are
found in rice grains, the most common being the inorganic species
) and arsenate (As
) and the organic species
monomethylarsonic acid (MMA) and dimethylarsinic acid (DMA).
While the toxicities of the organic As species are considered to be low
(Hirano et al., 2004), exposure to inorganic As is associated with many
types of cancer, in addition to other non-carcinogenic diseases such as
diabetes and hypertension (Bjørklund et al., 2017). Inorganic As accu-
mulates in the bran, causing paddy and brown rice to have higher inor-
ganic As concentration than white rice (Sun et al., 2008). Maximum
levels for inorganic As of 0.35 and 0.2 mg kg
in paddy and white
rice, respectively, have been adopted by CODEX, a joint commission of
the World Health Organization and the Food and Agriculture Organiza-
tion (CODEX, 2018).
Arsenic in rice grain originates from the soil, where it can be natu-
rally present (e.g., parent rock) or carried over to via irrigation
(i.e., water with high As levels) or other sources (e.g., arsenical pesti-
cides, urban residue) (Kumarathilaka et al., 2018). Arsenic tends to ac-
cumulate in rice for two reasons. First, the uptake of As
and, in part,
MMA are mediated by the silicon uptake pathway; rice being a silicon
accumulator is therefore naturally effective in taking up As from the
soil (Suriyagoda et al., 2018). Second, rice is commonly grown under
ﬂooded conditions for most of the growing season. Anaerobic soil condi-
tions increase As bioavailability in the soil because it triggers the reduc-
tion of As
, which is more mobile in the soil, and of Fe
dissolving iron hydro(oxides) that bindto As and, subsequently, releas-
ing As to the soil solution (Meharg and Zhao, 2012). In addition, the high
irrigation input to rice contributes to As building up in the soil if the ir-
rigation water has high As levels (Kumarathilaka et al., 2018).
Since soil ﬂooding is a major contributor to As accumulation in rice
grains, irrigation practices that include one or more periods of soil dry-
ing within the growing season [e.g., alternate wetting and drying
(AWD), intermittent ﬂooding, mid-season drain] have been proposed
as mitigation strategies (Bakhat et al., 2017). However, the impact of
these irrigation practices on grain As concentration is highly variable,
with decreases of 0 to 90% in total grain As being reported (Arao et al.,
2009;Linquist et al., 2015;Honma et al., 2016;Yang et al., 2016;
Norton et al., 2017a;Carrijo et al., 2018). This variability may be attrib-
uted, at least in part, to differences in the severity and timing (i.e. crop
stage) of soil drying.
The severity of soil drying affects soil redox potential (Eh), which is
intrinsically related to soil As bioavailability, and subsequently As up-
take. Grain As concentration decreases sharply with an increase in soil
Eh from −200 to −100 mV and tends to plateau at very small concen-
trations (b0.02 mg kg
) as soil Eh increases above 0 (Honma et al.,
2016). Carrijo et al. (2018) found that total grain As concentration de-
creased with increasing soil drying severity, although drying the top
soil (0–15cm) to a water potentiallower than −33 kPa did nottranslate
into a further decrease in grain As concentration. However, their study
was limited to total As concentrations and, given that As speciation
may be affected by soil drying (Yamaguchi et al., 2014), their conclu-
sions may not apply to all As species present in the grain.
The effect of timing of soil drying on grain As concentration has been
scarcely investigated, with the exception of a few pot studies (e.g., Arao
et al., 2009;Li et al., 2009). Soil drying would be most effective in min-
imizing grain As concentration if imposed when As uptake is highest
under ﬂooded conditions. However, predicting temporal As uptake is a
difﬁcult task considering the many variables regulating As availability
and uptake. Reports that the expression of Lsi1, a root transporter in-
volved in the uptake of As
, is enhanced around the heading stage
(Yamaji and Ma, 2007;Ma et al., 2008), suggest that this could be a
key stage for As uptake. In contrast, Li et al. (2015) reported that of
the total As present in aboveground tissues at harvest, 64% had been
taken up at the jointing (~panicle initiation) and booting stages, and
they attributed that to enhanced nutrient uptake and root size during
this period. In addition, As uptake is strongly inﬂuenced by the forma-
tion of iron plaques on the surface of rice roots, and their formation
and capacity of sequestering As is dependent on crop stage (Garnier
et al., 2010;Awasthi et al., 2017;Yu et al., 2017). For example, Mei
et al. (2012) found that plants at the bolting (~booting) stage showed
higher root radial oxygen loss and higher root porosity than plants at
the tillering stage, and this translated into higher As sequestration in
the root plaque and lower As uptake.
In this study, we sought to quantify how the severity and the timing
of soil drying impact total As and As speciation within the rice grain. We
hypothesized that grain arsenic concentration (total and individual spe-
cies) would decrease with increasing soil drying severity independent
of crop stage.
2. Material and methods
2.1. Study site characteristics
Aﬁeld experiment was conducted at the Rice Experiment Station
(39°27′47″N, 121°43′35″W) in Biggs, California, USA, during the sum-
mer of 2016. The soil at the site is a Vertisol, comprised of ﬁne, smectitic,
thermic, Xeric Epiaquerts and Duraquerts, with a soil texture of 29%
sand, 26% silt and 45% clay, a pH of 5.3, 1.06% organic C and 0.08%
total N (Pittelkow et al., 2012). Total As concentration in the soil was
3.85 mg kg
(Carrijoet al., 2018) and in the irrigation water it averaged
in the growing season. Total As in irrigation water was mea-
sured by collecting water samples from the main irrigation canal on
four sampling dates (July 8th and 25th, August 9th and September
16th). Samples were immediately ﬁltered [through glass microﬁber ﬁl-
ter paper (Whatman GF/F)] and acidiﬁed with nitric acid (67–70%, trace
metal grade) to a pH of 2 prior to storage at 4°C. Total As in water was
quantiﬁed by inductively coupled plasma (ICP) mass spectrometry
(MS), following the same methodology described for rice grains in
Section 2.6.2. The climate is Mediterranean, and the total precipitation
and average daily temperature over the growing season (May through
October) was 10.1 mm and 21.7° C, respectively (CIMIS, 2018).
There were ten irrigation treatments: a continuously ﬂooded (CF)
control, which was maintained ﬂooded (i.e., standing water maintained
at ~12 cm above the soil surface) from sowing to three weeks before
harvest (pre-harvest drain), and nine treatments in which a single soil
drying period was imposed, and that represented a combination of
three timings and three severities of soil drying. The three timings,
which determined the onset of the soil drying period, were: at panicle
initiation, during booting and at 50% heading (this extended into the
early grain ﬁlling period for most treatments). We did not include soil
drying timings that were earlier than panicle initiation due to the risk
of potential for high fertilizer nitrogen losses (LaHue et al., 2016). The
severities were: Low Severity (LS - reﬂooded when the perched water
table reached 15 cm below the soil surface), Medium Severity (MS -
reﬂooded when the soil volumetric water content at the 0–15 cm soil
depth reached 35%), and High Severity (HS - reﬂooded four days after
the MS treatments). The choice on severity treatments aimed at
representing a wide range of severities, from the LS, whichis the sever-
ity used in a common form of AWD known as “Safe-AWD”(widely
adopted in some Asian countries and considered to not limit rice yields)
(Bouman et al., 2007;Lampayan et al., 2015),to the HS, which is consid-
ered to limit rice yields (Carrijo et al., 2018). All soil drying treatments
underwent a single drying period according to their respective timing/
severity, and except for this period, followed the same water manage-
ment as in the CF. Within each timing, all plots started drying together
and plots of the same severity were reﬂooded at the same time when,
on average, the targeted severity was reached across plots. No
301D.R. Carrijo et al. / Science of the Total Environment 649 (2019) 300–307
precipitation occurred during any of the drying events. Detailed infor-
mation about the irrigation treatments and general management prac-
tices are reported in Table 1.
2.3. Experimental design and general management practices
The experiment was laid out in a completely randomized design
with four replications. Plots were comprised of polyvinyl chloride cylin-
ders (30 cm in height and 76cm in diameter –total area of 0.46 m
were buried 20 cm deep in the soil inside a rice ﬁeld (0.3 ha basin) and
2 m apart. A fertilizer blend of mono-ammonium phosphate, urea, am-
monium sulfate and muriate of potash was banded at 5–7 cm below the
soil surface, which provided a total of 171, 45 and 25 kg ha
of N, P
O, respectively. Following fertilization, seeds of the medium grain
variety M-206 were broadcasted at the rate of 168 kg ha
, and ﬂood-
water was applied. Pesticides were applied to all cylinders as necessary.
Irrigation was managed in individual cylinders by having two holes
drilled in each cylinder just above the soil surface (which could be
plugged or unplugged for holding or draining water, respectively) and
a drip irrigation line mounted on top of the cylinders to supply water
(same irrigation source as for the rest of ﬁeld) when needed. At the
start of each soil drying treatment, the main ﬁeld was drained along
with the desired treatment cylinders. The cylinders destined to be
kept ﬂooded were pluggedand ﬂoodwater was maintained usingirriga-
tion from the drip line. Except for when there was a soil drying treat-
ment, the main ﬁeld was maintained ﬂooded as in the CF to maintain
plant growth outside the cylinders, thus preventing border effects. In
addition, whenever the ﬁeld was ﬂooded (i.e. all treatments ﬂooded),
cylinders were kept unplugged and, except for during the ﬁrst three
weeks after sowing, ﬂoodwater height was maintained above the top
of the cylinders (i.e., ∼12 cm above the soil surface) to allow ﬂoodwater
exchange between cylinders and ﬁeld.
2.4. Soil moisture measurements
Soil volumetric water content (VWC) at 0–15 cm depth was moni-
tored using capacitance sensors (10HS, Decagon Devices Inc., Pullman,
WA) connected to data loggers (Em50, Meter Group Inc., Pullman,
USA). The sensors were installed vertically in the soil with their centers
at 7.5 cm soil depth, and had a volume of inﬂuence of 1 L, which
spanned from 0.5 to 14.5 cm soil depth. Soil water potential (WP) at
0–15 cm depth was monitored using electrical resistance sensors (Wa-
termark 200SS, Irrometer Co Inc., Riverside, CA) connected to data log-
gers (900 M Monitor, Irrometer Co Inc., Riverside, CA). The sensors
were installed vertically in the soil with their centers at 7.5 cm soil
depth. One VWC and one WP sensor wasinstalled in all soil drying treat-
ment plots and, for comparison, one VWC and one WP sensor was
installed in one of the CF plots. The perched water table (PWT) was
measured at the end of the drying period in all the LS treatment plots
using perforated tubes. In each LS treatment plot, a 30 cm long, 5 cm di-
ameter polyvinyl chloride tube perforated with 1 cm diameter holes
spaced approximately 2 cm apart was inserted 25 cm deep into the
soil after drilling a hole of the exact same diameter.
2.5. Yield, yield components and grain milling for As quantiﬁcation
When plants were mature, all plants were manually harvested from
within each plot. The total number of tillers was counted and the num-
ber of panicles was counted from 50 representative tillers to estimate
the percentage of unproductive tillers. Grain and straw were dried at
65 °C until constant weight for the determination of yield and harvest
index, and 1000 grains were counted using a seed counter (Model
750-2C, International Marketing and Design Corp., San Antonio, TX)
and weighed. Yield and 1000-grain weight were adjusted for 14%mois-
ture. Brown rice was obtained by removing the husk from paddy grain
using a laboratory dehusker (Model FC2K-Y, Yamamoto Co. Ltd., Yama-
gata, Japan). White rice (i.e. polished rice) was obtained by removing
the husk and the bran from paddy grain using a laboratory mill (Paz-
1/DTA, Zaccaria USA, Anna, TX). Grains were ball milled to pass a 250
μm sieve and stored in the dark at 4 °C prior to As analysis.
2.6. As quantiﬁcation
All water used for analysis was 18.2 MΩ∙cm (Barnstead Nanopure).
Trace metal grade nitric acid (67–70%), ammonium phosphate dibasic
(≥99%) and ammonium hydroxide (28–30%) were from Fisher Chemical
(USA). Stock standards of arsenite (1001 mg L
) and arsenate
(998 mg L
) were from Spex Certiprep (USA) and of DMA (≥98%)
and MMA (≥98.5%) were from ChemService (USA). Rice ﬂour certiﬁed
reference material (CRM 1568b) was from National Institute of Stan-
dards and Technology (NIST).
2.6.2. Total As
Total As in grains was determined followingthe method of Sun et al.
(2008) with some modiﬁcations described as follows. Samples of 0.5 g
(two analytical replicates per plot) were digested in glass digestion
tubes by adding 5 mL of nitric acid and allowing it to dissolve overnight
at room temperature. Samples were further digested in a heating block
at 105 °C until the cessation of a brown fog, and then at 120 °C until
complete dryness. The ash was re-dissolved with 10 mL of
0.28 mol L
nitric acid and ﬁltered using a syringe ﬁlter (0.45 μm), tak-
ing care to discard the ﬁrst 1 mL of the ﬁltrate. The extract was then di-
luted 5-fold with water.
Total As in grain samples (followingthe digestion method described
above) and irrigation water samples (as collected in the ﬁeld, without
further preparation) was quantiﬁed by inductively coupled plasma
mass spectrometry (ICP-MS 7900, Agilent Technologies, Santa Clara,
CA, USA) with a detection limit of 0.01 μgL
. As was monitored at m/
zof 75 and selenium was also monitored (m/z77, 78 and 82) to check
Cl interferences on m/z75.
Summary of management practices.
Crop development and general management practices Date DAS
Fertilization May 26 −1
Sowing and initial soil ﬂooding May 27 0
Jul 11 45
Aug 15 80
Pre-harvest drain Sep 21 117
Harvest Oct 12 138
Water management in soil drying treatments Date DAS
Start of panicle initiation drying period
Jul 12 46
LS reﬂooded Jul 15 49
MS reﬂooded Jul 22 56
HS reﬂooded Jul 26 60
Start of booting drying period Jul 28 62
LS reﬂooded Jul 29 63
MS reﬂooded Aug 5 70
HS reﬂooded Aug 9 74
Start of heading drying period Aug 15 80
LS reﬂooded Aug 17 82
MS reﬂooded Aug 26 91
HS reﬂooded Aug 30 95
Abbreviations: DAS = daysafter sowing; LS = lowseverity; MS= medium severity; HS =
Determined according to theUniversityof California Agriculture and Natural Resources
degree day model for California rice varieties available at http://rice.ucanr.edu/
When 50% of the panicles in the ﬁeld had at least partially exserted from the boot.
The start day of a soil drying period was considered the day when the perched water
table was at the soil surface.
302 D.R. Carrijo et al. / Science of the Total Environment 649 (2019) 300–307
2.6.3. As speciation
As speciation in grains was determined as in FDA (2012). In brief,
samples of 1 g (two and three analytical replicates per plot for white
and brown rice, respectively) were digested in plastic centrifuge tubes
with 4 mL of 0.28 mol L
nitric acid at 95 °C for 90 min. The digested
sample was centrifuged at 5858gfor 15 min and the supernatant was
neutralized with the mobile phase (10 mmol L
phate dibasic, pH of 8.25 adjusted with ammonium hydroxide) and am-
monium hydroxide to a target pH of 6.0 to 8.5. The resulting solution
was ﬁltered (0.45 μm nylon) prior to analysis.
Four As species (As
, DMA and MMA) were quantiﬁed by high
performance liquid chromatography (HPLC 1200 series, Agilent Tech-
nologies) coupled with ICP-MS 7900. Arsenic species (elution order:
, DMA, MMA and As
) were separated using an anion-exchange col-
umn (Hamilton PRP-×100, 250 mm ×4.1 mm ×10 μm) with isocratic
mobile phase at 1.0 mL min
and then quantiﬁed by ICP-MS following
the same procedure as for total As, but with a detection limit of 0.1
. The percentage of inorganic As in grains (grain inorganic As%)
was calculated as in Eq. (1):
Grain inorganic As%¼100 "AsIII þAsV
AsIII þAsVþDMA þMMA ð1Þ
, DMA and MMA are the grain concentrations of these
species as quantiﬁed by the As speciation analysis.
2.6.4. Quality control
In both totaland speciation analyses, at least one blank, one fortiﬁed
sample (for the total As analysis, As
was spiked at 0.1 mg As kg
grain; for the speciation analysis, all four species were spiked at
0.1 mg As kg
grain), and one reference material (CRM 1568b) were
included with every 10 samples analyzed. Analytical replicates were ac-
cepted when their coefﬁcient of variation waswithin 15%. For each plot,
mass balance was performed between the sum of the four species deter-
mined by HPLC-ICP-MS and the total As determined by ICP-MS. Sample
grain moisture was measured and As concentrations are presented on a
dry mass basis. Recoveries of fortiﬁed samples and reference material
and mass balances were calculated as described in FDA (2012) and are
presented in Table S1, Appendix A.
2.7. Statistical analyses
All statistical analyses were performed in R software (R Core Team,
2016) and all variables were ﬁt to a linear model before being subjected
to analysis of variance (ANOVA). A one-way ANOVA was performed on
grain As concentration (separately for brown and white rice and total As
and individual As species), yield and yield components, with treatment
as a ﬁxed effect, followed by Tukey means separation. A two-way
ANOVA was performed on soil VWC and WP with timing and severity
(as well as their interaction) as ﬁxed effects, followed by Tukey means
separation; the CF control was excluded from this analysis since these
measurements were taken in only one CF plot. A one-way ANOVA was
performed on PWT with timing as a ﬁxed effect, followed by Tukey
means separation. Grain As concentration data from brown and white
rice were combined and a two-way ANOVA was performed on grain in-
organic As% with treatment and grain milling (including their interac-
tion) as ﬁxed effects. Grain milling means were separated by Tukey
means separation. The effect of treatment was analyzed using a set of
single-degree-of-freedom orthogonal contrasts to provide information
on the factorial part of the experiment (i.e. maineffects of timing and se-
verity and their interaction), and the Sidak correction was used to con-
trol the familywise error rate. The set of contrasts excluded all LS
treatments for reasons discussed in the results Section 3.4.
3.1. Soil moisture in the soil drying treatments immediately before
To ensure that there were no confounding effects between timing
and severity of soil drying, we tested the hypothesis that soil moisture
measured at the end of the drying period was the same across
treatments of the same severity (Table 2). Although there were small
differences in the number of drying days required to achieve a desired
soil moisture, soil WP was the same within each severity
(i.e., independent of timing) and increased in the order: HS bMS bLS,
as expected. Similarly, there was no difference in soil VWC within
each severity except for the MS (p = 0.015), where soil VWC was
lower in the treatments dried at panicle initiation than at booting. The
PWT, measured only in the LS treatments, was the same across all tim-
ings and averaged 17 cm below the soil surface.
Given the inaccuracy of the soil WP sensor in the 0 to −10 kPa range
(Irmak and Haman, 2001;Shock and Wang, 2011), we assume that any
value in this range is ~0 kPa (Table 2). Averaged across timings, soil
VWC was 48%, 36% and 33% and soil WP was −1 (~0), −71 and
−154 kPa at the end of the drying period in the LS, MS and HS treat-
ments, respectively (Table 2). For comparison, soil VWC and WP in
the CF control averaged 48% and −9 kPa (~0) throughout the season,
excluding the pre-harvest draining period. The severities achieved in
the LS and MS treatments were close to what was targeted (i.e., PWT
of −17 cm vs. −15 cm targeted for LS and VWC of 36% vs. 35% targeted
for MS). Averaged across timings, the drying period lasted 2, 10 and
14 days in the LS, MS and HS treatments, respectively.
3.2. As in white rice
In all the LS treatments, the concentration of total As (Fig. 1A) and
various As species (Fig. 1C) was similar to the CF control. For both MS
and HS treatments, independent of timing, total As concentration de-
creased by 42–61% compared to the CF control. Of the four As species
(i.e., organic MMA and DMA, and inorganic As
) quantiﬁed in
this study, only As
and DMA were detected in white rice. Drying the
soil to HS (independent of timing) or to MS at booting (but not at
other timings) decreased As
concentration by 31–46% compared to
the CF control. DMA concentration decreased by 61–71% when the soil
was dried down to MS or HS at panicle initiation or booting, but not at
heading, compared to the CF control.
3.3. As in brown rice
Similar to what was observed for white rice, total As (Fig. 1B) and As
speciation (Fig. 1D) in brown rice were similar in the LS and CF treat-
ments. Total As concentration in brown rice decreased in all the MS and
HS treatments, compared to the CF control, similar to what was observed
for white rice. However, the decrease was dependent on the timing of the
drying period. Drying the soil to HS at booting decreased total As by 59%,
whereas drying the soil to MS at panicle initiation or heading decreased
total As by 41–42%. There was no MMA or As
detected in brown rice of
any treatment. In the MS and HS treatments, As
41–56% lower than in the CF control at booting or heading, but not at pan-
icle initiation. DMA concentration decreased by 71–81% when the soil
was dried down to MS or HS at panicle initiation or booting, but not at
heading, compared to the CF control. Averaging across all treatments,
total As, As
and DMA concentrations were 24 and 50% higher and 12%
lower in brown rice than in white rice, respectively.
3.4. Grain inorganic As%
Both grain milling and treatment affected grain inorganic As% (per-
centage of inorganic As relative to the sum of all As species present in
303D.R. Carrijo et al. / Science of the Total Environment 649 (2019) 300–307
grain), although there was not an interaction between them (Table 3).
White rice (63%) had lower grain inorganic As% than brown rice
(75%). The effect of treatment is represented by a set of orthogonal con-
trasts. The purpose of this analysis was to test the hypothesis that de-
creases in grain total As concentration caused by soil drying are
accompanied by changes in grain As speciation. Since drying the soil
to LS had no effect on grain As concentration (Fig. 1), all LS treatments
were excluded from this analysis. The estimates obtained from the set
of contrasts indicate that, in brown and white rice, grain inorganic As
%:1) is higher in the MS and HS treatments compared to CF (on average,
Soil moisture(soil volumetricwater content,VWC, and water potential,WP, at 0-15 cm soil depth, andperched water table, PWT)immediately beforereﬂooding, and soildrying duration.
Soil VWC and WP were monitored in one CFplot for comparison.
Treatment VWC WP WP mean PWT
Severity Timing % kPa kPa cm days
CF 48 −9- - -
LS Panicle Initiation 50 (1)
a 0 (0) −1C −917.4 (3) 3
Booting 48 (0.4) a −2 (2) −913.0 (3) 1
Heading 47 (0.5) a 0 (0) −919.8 (2) 2
MS Panicle Initiation 33 (1) b −57 (5) −71 B - 10
Booting 38 (1) a −69 (13) - 8
Heading 37 (2) ab −87 (3) - 11
HS Panicle Initiation 33 (1) a −173 (26) −154 A - 14
Booting 35 (1) a −163 (28) - 12
Heading 31 (2) a −127 (13) - 15
ANOVA Timing ns ns - ns -
Severity *** *** - - -
Timing:Severity * ns - - -
Abbreviations: CF=continuously ﬂooded, LS = low severity; MS = medium severity; HS = high severity; *** = p b0.001; * = p b0.05; ns=not signiﬁcant.
Numbersin parenthesis are standard error of means. Differentuppercase lettersin a column indicate signiﬁcantdifferences (p b0.05). Differentlowercase letters within each severity
indicate signiﬁcant differences (p b0.05).
Negative PWT indicates that it is below the soil surface.
The duration of a drying period was the number of days from when the PWT was atthe soil surface to when the soil was reﬂooded.
Medium Severity High SeverityLow Severity Medium Severity High SeverityLow Severity
Medium Severity High SeverityLow Severity Medium Severity High SeverityLow Severity
Fig. 1. Total As and As species concentration in brown and white rice. As
and MMA were not detected in any of the samples. Error bars represent standard error of means. Different
lowercase or uppercase letters indicate signiﬁcant (p b0.05) differences between treatments. Abbreviations: CF = continuously ﬂooded, PI = panicle initiation, Boot = booting, Head
304 D.R. Carrijo et al. / Science of the Total Environment 649 (2019) 300–307
68 vs. 57% for white and 80 vs. 68% for brown rice), 2) is higher if the soil
is dried during panicle initiation or booting compared to heading (on
average, 72 vs. 60% for white and 85 vs. 72% for brown rice), and 3) is
not affected by an interaction between timing and severity of soil
3.5. Yield and yield components
There were no differences between treatments with respect to yield,
number of tillers per m
, percentage of unproductive tillers, and 1000-
grain weight. However, the harvest index was higher in CF (53%) than
when the soil was dried to HS at booting (47%) (Table S2, Appendix A).
Our results indicate that the severity of soil drying plays an impor-
tant role in mitigating the accumulation of As in rice grain. In contrast
to the more severe MS andHS treatments, drying the soil to LS(severity
used in the “Safe-AWD”) had noeffect on grain total As concentration or
As speciation, independent of timing (Fig. 1). This was likely because in
the LS treatments the soil moisture never dropped below saturation
(i.e., soil WP was ~0 kPa at the end of the drying period) (Table 2),
thus, the soil was not aerated sufﬁciently to alter As soil bioavailability.
In contrast, in soils of the same textural class as the one in this study
(i.e., clay), others have reported a decrease in grain As concentration
when the soil was dried to LS multiple times (Islam et al., 2017;
Norton et al., 2017b). This would suggest that the different results ob-
served by Islam et al. (2017) and Norton et al. (2017b) were caused
by the higher number of soil drying periods imposed in their studies.
However, another study conducted in an adjacent ﬁeld reported that
when two severe (i.e., comparable to MS and HS here) drying periods
were imposed (ﬁrst at panicle initiation, second at booting), total As
concentration in white rice decreased by 56–68%, compared to a CF con-
trol (Carrijo et al., 2017); similar reductions were obtained here with a
single drying period imposed at panicle initiation (46–55%) or booting
(57–61%). This suggests that the number of drying periods imposed in
the growing season may not be an important factor inﬂuencing grain
As concentrations. Differences in soil properties (other than textural
class) may explain the discrepancy between the results of our study
and the studies of Islam et al. (2017) and Norton et al. (2017b), given
that the relation between PWT and soil WP is soil-speciﬁc(Lampayan
et al., 2015). That said, decreases in grain As concentration reported in
the studies of Norton et al. (2017b) and Islam et al. (2017) with LS are
generally low (16–35%) compared to what was observed here with
the higher severities MS and HS.
While all MS (i.e., soil WP of −71 kPa after 10 days drying) and HS
(i.e., soil WP of −154 kPa after 14 days drying) treatments decreased
total grain As concentration compared to the CF control, this did not al-
ways translate into a decrease in As
and DMA concentrations (Fig. 1C
and D), because grain As speciation varied with the timing and severity
of soil drying (Table 3). Higher grain inorganic As% was observed when
severe drying periods (MS and HS) were imposed compared to CF, and
when the drying period was imposed earlier (i.e., panicle initiation or
booting) rather thanlater (i.e., heading) in the cropping season.As a re-
sult, there were soil drying treatments (e.g., MS-panicle initiation) that
were effective in decreasing total As, but not As
concentration. This is
important given that inorganic As is the most toxic form of As and the
primary focus of public health regulations (Signes-Pastor et al., 2016).
Based on these ﬁndings, reports on the effect of soil drying on grain
total As (e.g., Carrijo et al., 2018, and LaHue et al., 2016) likely overesti-
mate the health beneﬁts of soil drying in reducing grain As. Further,
these results indicate that there is not a ﬁxed conversion factor from
total As to inorganic As concentration.For example, for white ricethe in-
organic As% ranged from 52 to 75% depending on the irrigation treat-
ment (Table 3). If, due to the high cost and complexity of As
speciation analysis, regulations targeting inorganic As are made based
on an estimated percentage from total As, irrigation practices must be
taken into account.
Differences in grain inorganic As% observed between the MS and HS
treatments and the CF control (Table 3) may be a result of As transfor-
mations inthe soil triggered by soil drying.Soil microbes can convert in-
organic As into organic As and vice-versa (i.e., methylation and
demethylation reactions, respectively) and aerobic conditions favor As
demethylation over methylation reactions (Frohne et al., 2011;
Meharg and Zhao, 2012;Reid et al., 2017). In addition, there is strong
evidence that rice plants cannot methylate As and that methylated As
species found in grain originate from the soil (Lomax et al., 2012;Jia
et al., 2013;Mishra et al., 2017). These ﬁndings suggest that soil drying
increases the proportion of inorganic As in the soil (relative to the total
As), which is then taken up by the plant and transported to the grain.
This explains the higher grain inorganic As% in the MS and HS treat-
ments compared to CF (Table 3), as others have observed (Xu et al.,
2008;Somenahally et al., 2011;Das et al., 2016).
The lower grain inorganic As% observed when the soil was dried at
heading, compared to panicle initiation and booting (Table 3), may be,
in part, explained by differences in As species translocation within the
Percentage of inorganic arsenicrelative to all arsenic speciespresent in grain (inorganic As
%) in white and brown rice.The concentrationof inorganic arsenic(inorganic As, mg kg
is represented for comparison (replicated data from Figure 1). The treatment effect ex-
cluded all low severity treatments, since none of them affected grain As concentration.
Grain inorganic As% (relative to the sum of all As species in grain)
Treatment White rice Brown rice
CF 57 (5)
2 LS Panicle Initiation 52 (3) 65 (5)
3 Booting 54 (5) 61 (7)
4 Heading 56 (4) 69 (6)
5 MS Panicle Initiation 75 (4) 83 (3)
6 Booting 69 (2) 85 (3)
7 Heading 58 (3) 70 (6)
8 HS Panicle Initiation 73 (3) 88 (2)
9 Booting 70 (1) 82 (3)
10 Heading 62 (6) 74 (7)
Grain milling ⁎⁎⁎
Grain milling Means
White rice 63 b
Brown rice 75 a
Treatment Orthogonal contrast estimate
CF vs. (MS+HS) (-) **
(1) vs. (3+4+6+7+9+10)
MS vs. HS ns
(3+6+9) vs. (4+7+10)
Panicle Initiation vs. Booting ns
(3+4) vs. (6+7)
(Panicle Initiation+Booting) vs. Heading (+) ***
(3+4+6+7) vs. (9+10)
Interaction (Panicle Initiation vs. Booting) ns
(3+7) vs. (4+6)
Interaction (Panicle Initiation +Booting vs.
(3+6+10) vs. (4+7+9)
Abbreviations: CF = continuously ﬂooded,LS = low severity, MS = medium severity, HS
= high severity, * pb0.05, ** pb0.01, *** pb0.001, ns = not signiﬁcant.
Numbers are for treatment identiﬁcation referred to in the set of contrasts.
Numbers in parenthesis are standard error of means.
Means were averaged across treatments since there was not a signiﬁcant (p b0.05)
interaction between grain milling and treatment. Different lettersindicate signiﬁcant(p b
A positiveor a negative sign indicates that the ﬁrstterm of the contrastis, respectively,
higher or lower than the second term.
305D.R. Carrijo et al. / Science of the Total Environment 649 (2019) 300–307
plant. Zheng et al. (2011) found that almost all DMA present in the ma-
ture grain was transported to the ovary before ﬂowering while most of
the inorganic As was unloaded during grain ﬁlling. Their ﬁndings sug-
gest that any effort made to decrease DMA uptake after ﬂowering
would not translate into a decrease in DMA concentration in grains, al-
though that is not true for As
. In agreement, all the MS and HS treat-
ments imposed at heading did not decrease DMA but decreased (with
the exception of the MS-heading treatment for white rice) As
concentration, compared to the CF control. In a pot study, Arao et al.
(2009) also found that aerobic treatment after heading was more effec-
tive in decreasing As
than DMA grain concentration, compared to aer-
obic treatment before heading. It is worth noting that although soil
drying at heading may be effective in decreasing grain As
, it is likely
a risky strategy as rice is very sensitive to water stress during ﬂowering
and yields and grain quality may be reduced (Sarvestani et al., 2008;
Gunaratne et al., 2011).
Grain As speciation also varied with grain milling (Table 3), with
brown rice (75%) having higher grain inorganic As% than white rice
(63%). This has been previously observed and is attributed to As
being less mobile in the plant than DMA, which results in As
less efﬁciently translocated from the bran to the endosperm of the
grain (Sun et al., 2008;Carey et al., 2010;Carey et al., 2011).
While grain As concentration was impacted by soil drying, yield
(and yield components, except for harvest index) remained similar to
the CF control, independent of the timing or severity of soil drying
(Table S2, Appendix A). Based on a meta-analysis, yields are generally
reduced when the soil WP at root depth is allowed to drop below
−20 kPa (Carrijo et al., 2017). However, in this study in both the MS
and HS treatments the soil WP dropped below −20 kPa (on average,
−71 kPa and −154 kPa for the MS and HS treatments, respectively)
and yet yields were not reduced. In a study we conducted in an adjacent
ﬁeld, similar yields were observed between severe soil drying treat-
ments and a CF control, and this was attributed to the availability of
water and the presence of roots at deeper soil layers (Carrijo et al.,
2018). This suggests that for at least some conditions such as those in
this study, yields can be maintained while at the same time As grain
concentrations can be reduced.
This ﬁeld study indicates that a single soil drying event during the
season has the potential to mitigate As accumulation in rice grains.
However, the severity and the timing of soil drying will determine if,
and to what extent, this potential can be realized. Grain As concentra-
tion was not affected unless the soil was dried to achieve negative soil
water potentials; in this study, this meant that the soil had to be dried
further than it is for “Safe-AWD (alternate wetting and drying)”
(i.e., perched water table at 15 cm below the soil surface). In white
and brown rice, the lowest grain As concentrations for all species de-
tected were observed when the soil was severely dried [i.e., soil
(0–15 cm) water potential of −154 kPa] at the booting stage. Impor-
tantly, the timing and the severity of soil drying affected grain As speci-
ation, which resulted in some soil drying treatments being effective at
decreasing total As but not inorganic As concentration. Overall, in
white and brown rice, the soil had to be dried severely and/or later in
the season (at booting or heading instead of at panicle initiation) to de-
crease inorganic As concentration in grain, compared to a continuously
ﬂooded control. This effect of irrigation management on grain As speci-
ation needs to be considered if regulations targeting grain inorganic As
are based on an estimated inorganic As percentage from measured total
This work was supported by: CAPES –Coordenação de
Aperfeiçoamento de Pessoal de Nível Superior –Brazil; the Department
of Plant Sciences, University of California –Davis; and the California Rice
Research Board. We would liketo thank Nadeem Akbar, Cesar Abrenilla,
Ray Stogsdill and Peter Green for their fundamental help with data col-
lection and analysis and ﬁeld work.
Appendix A. Supplementary data
Supplementary data to this article can be found onlineat https://doi.
Arao, T.,Kawasaki, A., Baba,K., Mori, S., Matsumoto, S., 2009.Effects of water management
on cadmium and arsenic accumulation and dimethylarsinic acid concentrations in
Japanese rice. Environ. Sci. Technol. 43, 9361–9367. https://doi.org/10.1021/
Awasthi, S., Chauhan, R., Srivastava, S., Tripathi, R.D., 2017. The journey of arsenic from
soil to grain in rice. Front. Plant Sci. 8, 1007. https://doi.org/10.3389/fpls.2017.01007.
Bakhat, H.F., Zia, Z., Fahad, S., Abbas, S., Hammad, H.M., Shahzad, A.N., Abbas, F., Alharby,
H., Shahid,M., 2017. Arsenic uptake, accumulation andtoxicity in rice plants: possible
remedies for its detoxiﬁcation: a review. Environ. Sci. Pollut. Res., 1–17 https://doi.
Bhowmick, S., Pramanik, S., Singh, P., Mondal, P., Chatterjee, D., Nriagu, J., 2018. Arsenic in
groundwater of West Bengal, India: a review of human health risks and assessment
of possible intervention options. Sci. Total Environ. 612, 148–169. https://doi.org/
Bjørklund, G., Aaseth, J., Chirumbolo, S., Urbina, M.A., Uddin, R., 2017. Effects of arsenic
toxicity beyond epigenetic modiﬁcations. Environ. Geochem. Health https://doi.org/
Bouman, B.A.M., Lampayan, R.M., Tuong, T.P., 2007. Water Management in Irrigated Rice:
Coping with Water Scarcity. International RiceResearch Institute, Manila, Philippines,
Carey, A.-M., Scheckel, K.G., Lombi, E., Newville, M., Choi, Y., Norton, G.J., Charnock, J.M.,
Feldmann, J., Price, A.H., Meharg, A.A., 2010. Grain unloading of arsenic species in
rice. Plant Physiol. 152, 309–319. https://doi.org/10.1104/pp.109.146126.
Carey, A.M., Norton Gareth, J., Deacon, C., Scheckel Kirk, G., Lombi, E., Punshon, T.,
Guerinot Mary, L., Lanzirotti, A., Newville, M., Choi, Y., Price Adam, H., Meharg
Andrew, A., 2011. Phloem transport of arsenic species from ﬂag leaf to grain during
grain ﬁlling. New Phytol. 192, 87–98. https://doi.org/10.1111/j.1469-
Carrijo, D.R., Lundy, M.E., Linquist, B.A., 2017. Rice yields and water use under alternate
wetting and drying irrigation: a meta-analysis. Field Crop Res. 203, 173–180.
Carrijo, D.R., Akbar, N., Reis, A.F.B., Li, C., Gaudin, A.C.M., Parikh, S.J., Green, P.G., Linquist,
B.A., 2018.Impacts of variable soil dryingin alternate wettingand drying rice systems
on yields, grain arsenic concentration and soil moisture dynamics. Field Crop Res.
222, 101–110. https://doi.org/10.1016/j.fcr.2018.02.026.
CIMIS (California Irrigation Management Information System), 2018. Accessed on Febru-
ary 10, 2018 at https://cimis.water.ca.gov/WSNReportCriteria.aspx
CODEX, 2018. Committee on Contaminants in Foods, Working Document for Information
and Use in Discussions Related to Contaminants and Toxins in the GSCTFF. Utrecht,
Netherlands, March 2018. Accessible at:. https://www.fsis.usda.gov/wps/portal/fsis/
Das, S., Chou,M.-L., Jean, J.-S., Liu, C.-C., Yang, H.-J., 2016. Water management impacts on
arsenic behavior and rhizosphere bacterial communities and activities in a rice agro-
ecosystem. Sci. Total Environ. 5 42, 642–652. https://doi.org/10.1016/j.
scitotenv.2015.10.122 Part A.
FDA (US Food and DrugAdministration), 2012. ElementalAnalysis Manual, 4.11 - Arsenic
Speciation in Rice and Rice ProductsUsing High Performance Liquid Chromatography
–Inductively Cou pled Plasma-Mas s Spectrometric Determination. Accessed at:.
Frohne, T., Rinklebe, J., Diaz-Bone, R.A., Du Laing, G., 2011. Controlled variation of redox
conditions in a ﬂoodplain soil: impact on metal mobilization and biomethylation of
arsenic and antimony. Geoderma 160, 414–424. https: //doi.org/10. 1016/j.
Garnier, J.M., Travassac, F., Lenoble, V., Rose, J., Zheng, Y., Hossain, M.S., Chowdhury, S.H.,
Biswas, A.K., Ahmed, K.M., Cheng, Z., van Geen, A., 2010. Temporal variations in arse-
nic uptake by rice plants in Bangladesh: the role of iron plaque in paddy ﬁelds irri-
gated with groundwater. Sci. Tota l Environ. 408, 4185–4193. https://doi.org/
GRiSP (Global Rice Science Partnership), 2013. Rice Almanac. 4th edition. International
Rice Research Institute, Los Baños, Philippines (283 pp).
Gunaratne,A., Ratnayaka, U.K., Sirisena, N., Ratnayaka, J., Kong, X., Arachchi, L.V.,Corke, H.,
2011. Effect of soil moisture stress from ﬂowering to grain maturity on functional
properties of Sri Lankan rice ﬂour. Starch 63, 283–290. https://doi.o rg/10.1002/
Hirano, S., Kobayashi, Y., Cui, X., Kanno, S., Hayakawa, T., Shraim, A., 2004. The accumula-
tion and toxicity of methylated arsenicals in endothelialcells: important roles of thiol
compounds. Toxicol. Appl. Pharma col. 198, 458–467. https://doi.org/10.1016/j.
306 D.R. Carrijo et al. / Science of the Total Environment 649 (2019) 300–307
Honma, T., Ohba, H., Kaneko-Kadokura, A., Makino, T., Nakamura,K., Katou, H., 2016. Op-
timal soil eh, ph, and water managementfor simultaneously minimizing arsenic and
cadmium concentrations in rice grains. Environ. Sci. Technol. 50, 4178–4185. https://
Irmak, S., Haman, D.Z., 2001. Performance of the watermark® granular matrix sensor in
sandy soils. Appl. Eng. Agric. 17, 787–795.
Islam, S., Rahman, M.M.,Islam, M.R., Naidu, R., 2017. Effect of irrigation and genotypes to-
wards reduction in arsenic load in rice. Sci. Total Environ. 609, 311–318. https://doi.
Jia, Y., Huang, H., Zhong, M., Wang, F.-H., Zhang, L.-M., Zhu, Y.-G., 2013. Microbial arsenic
methylation in soil and rice rhizosphere. Environ. Sci. Technol. 47, 3141–3148.
Kumarathilaka, P., Seneweera, S., Meharg, A., Bundschuh, J., 2018. Arsenic speciation dy-
namics in paddy rice soil-waterenvironment: sources, physico-chemical, and biolog-
ical factors - a review. Water Res. 140, 403–414. https:// doi.org/10.10 16/j.
LaHue, G.T.,Chaney, R.L., Adviento-Borbe, M.A.,Linquist, B.A., 2016. Alternate wetting and
drying in high yielding direct-seeded rice systems accomplishes multiple environ-
mental and agronomic objectives. Agric. Ecosyst. Environ. 229, 30–39. https://doi.
Lampayan, R.M., Rejesus, R.M., Singleton, G.R., Bouman, B.A.M., 2015. Adoption and eco-
nomics of alternate wetting and drying water management for irrigated lowland
rice. Field Crop Res. 170, 95–108. https://doi.org/10.1016/j.fcr.2014.10.013.
Li, R.Y., Stroud, J.L., Ma, J.F., McGrath, S.P., Zhao, F.J., 2009. Mitigation of arsenic accumula-
tion in ricewith water management and siliconfertilization. Environ. Sci.Technol. 43,
Li, R., Zhou,Z., Zhang, Y., Xie, X., Li,Y., Shen, X., 2015. Uptake and accumulation character-
istics of arsenic and iron plaque in rice at different growth stages. Commun. Soil Sci.
Plant Anal. 46, 2509–2522. https://doi.org/10.1080/00103624.2015.1089259.
Linquist, B.A., Anders, M.M., Adviento-Borbe, M.A., Chaney, R.L., Nalley, L.L., da Rosa, E.F.,
van Kessel,C., 2015. Reducing greenhouse gasemissions, water use,and grain arsenic
levels in rice systems. Glob. Chang. Biol. 21, 407–417. https://doi.org/10.1111/
Lomax, C., Liu, W.J., Wu, L., Xue, K., Xiong, J., Zhou, J., McGrath, S.P., Meharg, A.A., Miller,
A.J., Zhao, F.J., 2012. Methylated arsenic species in plants originate from soil microor-
ganisms. New Phyt ol. 193, 665–672. https://doi.org/10.1111/j.1469-
Ma, J.F., Yamaji, N., Mitani, N., Xu, X.-Y., Su, Y.-H., McGrath, S.P., Zhao, F.-J., 2008. Trans-
porters of arsenite in rice and their role in arsenic accumulation in rice grain. Proc.
Natl. Acad. Sci. 105, 9931. https://doi.org/10.1073/pnas.0802361105.
Meharg, A.A., 2004. Arsenic in rice –understanding a new disaster for South-East Asia.
Trends Plant Sci. 9, 415–417. https://doi.org/10.1016/j.tplants.2004.07.002.
Meharg, A.A., Zhao, Fang-Jie, 2012. Arsenic and Rice. Springer https://doi.org/10.1007/
Mei, X.Q., Wong, M.H., Yang, Y., Dong, H.Y., Qiu, R.L., Ye, Z.H., 2012. The effects of radial
oxygen loss on arsenic tolerance and uptake in rice and on its rhizosphere. Environ.
Pollut. 165, 109–117. https://doi.org/10.1016/j.envpol.2012.02.018.
Mishra, S.,Mattusch, J., Wennrich, R., 2017.Accumulation and transformation of inorganic
and organic arsenic in rice and role of thiol-complexation to restrict their transloca-
tion to shoot. Sci. Rep. 7, 40522. https://doi.org/10.1038/srep40522.
Norton, G.J., Shafaei, M., Travis, A.J., Deacon, C.M., Danku, J., Pond, D., Cochrane, N.,
Lockhart, K., Salt, D., Zhang, H., Dodd, I.C., Hossain, M., Islam, M.R., Pr ice, A.H.,
2017a. Impact of alternate wetting and drying on rice physiology, grain production,
and grainquality. Field Crop Res. 205, 1–13. https://doi.org/10.1016/j.fcr.2017.01.016.
Norton, G.J., Travis,A.J., Danku, J.M.C.,Salt, D.E., Hossain,M., Islam, M.R., Price, A.H., 2017b.
Biomass and elemental concentrations of 22 rice cultiva rs grown under alternate
wetting and drying conditions at three ﬁeld sites in Bangladesh. Food Energy Secur.
6, 98–112. https://doi.org/10.1002/fes3.110.
Pittelkow, C.M., Fischer, A.J.,Moechnig, M.J., Hill, J.E., Kofﬂer, K.B., Mutters, R.G.,Greer, C.A.,
Cho, Y.S., van Kessel, C., Linquist, B.A., 2012. Agronomic productivity and nitrogen re-
quirements of alternative tillage and crop establishment systems for improved weed
control in direct-seeded rice. Field Crop Res. 130, 128–137. https://doi.org/10.1016/j.
R Core Team, 2016. A Language and Environment for Statistical Computing. R Foundation
for Statistical Computing, Vien, Austria.
Reid, M.C., Maillard,J., Bagnoud, A., Falquet, L., Le Vo, P., Bernier-Latmani, R.,2017. Arsenic
methylation dynamics in a ricepaddy soil anaerobic enrichment culture. Environ. Sci.
Technol. 51, 10546–10554. https://doi.org/10.1021/acs.est.7b02970.
Sarvestani, Z.T., Pirdashti, H., Sanavy, S.A., Balouchi, H., 2008. Study of water stress effects
in different growth stages on yield and yield components of different rice (Oryza
sativa L.) cultivars. Pak. J. Biol. Sci. 11, 1303–1309. https://doi.org/10.3923/
Shock, C.C., Wang, F.-X., 2011. Soil water tension,a powerful measurement for productiv-
ity and stewardship. Hortscience 46, 178–185.
Signes-Pastor, A.J., Carey, M., Meharg, A.A., 2016. Inorganic arsenic inrice-based products
for infants and young children. Food Chem. 191, 128–134. https://doi.org/10.1016/j.
Somenahally, A.C., Hollister, E.B., Yan, W., Gentry, T.J., Loeppert, R.H., 2011. Water man-
agement impacts on arsenic speciation and iron-reducing bacteria in contrasting
rice-rhizosphere compartments. Environ. Sci. Technol. 45, 8328–8335. https://doi.
Sun, G.-X.,Williams,P.N., Carey, A.-M., Zhu, Y.-G., Deacon, C., Raab, A., Feldmann, J., Islam,
R.M., Meharg, A.A., 2008. Inorganic arsenic in rice bran and its products are an order
of magnitudehigher than in bulk grain. Environ. Sci. Technol.42, 7542–7546. https://
Suriyagoda, L.D.B., Dittert, K., Lambers, H., 2018. Mechanism of arsenic uptake, transloca-
tion and plant resistance to accumulate arsenic in rice grains. Agric. Ecosyst. Environ.
253, 23–37. https://doi.org/10.1016/j.agee.2017.10.017.
Xu, X.Y., McGrath, S.P., Meharg, A.A., Zhao, F.J., 2008. Growing rice aerobically markedly
decreases arsenic accumulation. Environ. Sci. Technol. 42, 5574–5579. https://doi.
Yamaguchi, N., Ohkura, T., Takahashi, Y., Maejima, Y., Arao, T., 2014. Arsenic distribution
and speciation near rice roots inﬂuenced by iron plaques and redox conditions of
the soil matrix. Environ. Sci. Technol. 48, 1549–1556. https:// doi.org/10.10 21/
Yamaji, N., Ma, J.F., 2007. Spatial distribution and temporal variation of the rice silicon
transporter Lsi1. Plant Physiol. 143, 1306–1313. https://doi.org/10.1104/
Yang, J., Zhou, Q., Zhang,J., 2016. Moderatewetting and drying increases riceyield and re-
duces water use, grain arsenic level, and methane emission. Crop J. 5, 151–158.
Yu, H.Y., Wang, X., Li, F.,Li, B., Liu, C., Wang, Q., Lei, J., 2017.Arsenic mobility and bioavail-
ability in paddy soil under iron compound amendments at different growth stages of
rice. Environ. Pollut. 224, 136–147. https://doi.org/10.1016/j.envpol.2017.01.072.
Zheng, M.-Z., Cai, C., Hu, Y., Sun, G.-X.,Williams, P.N., Cui, H.-J., Li, G., Zhao, F.-J., Zhu, Y.-G.,
2011. Spatial distribution of arsenic and temporal variation of its concentration in
rice. New Phytol. 189, 200–209. https://doi.org/10.1111/j.1469-8137.2010.03456.x.
307D.R. Carrijo et al. / Science of the Total Environment 649 (2019) 300–307