Article pubs.acs.org/est
The Role of Irrigation Techniques in Arsenic Bioaccumulation in Rice (Oryza sativa L.) Antonino Spanu,† Leonardo Daga,‡ Anna Maria Orlandoni,§ and Gavino Sanna‡,* †
Dipartimento Agraria, Università di Sassari, Via De Nicola 1, I-07100 Sassari, Italy Dipartimento di Chimica e Farmacia, Università di Sassari, Via Vienna 2, I-07100 Sassari, Italy § Stabilimento Polimeri Europa, ZI La Marinella, I-07041 Porto Torres (SS), Italy ‡
S Supporting Information *
ABSTRACT: The bioaccumulation of arsenic compounds in rice is of great concern worldwide because rice is the staple food for billions of people and arsenic is one of the most toxic and carcinogenic elements at even trace amounts. The uptake of arsenic compounds in rice comes mainly from its interaction with system soil/water in the reducing conditions typical of paddy fields and is influenced by the irrigation used. We demonstrate that the use of sprinkler irrigation produces rice kernels with a concentration of total arsenic about fifty times lower when compared to rice grown under continuous flooding irrigation. The average total amount of arsenic, measured by a fully validated ICP-MS method, in 37 rice grain genotypes grown with sprinkler irrigation was 2.8 ± 2.5 μg kg−1, whereas the average amount measured in the same genotypes grown under identical conditions, but using continuous flooding irrigation was 163 ± 23 μg kg−1. In addition, we find that the average concentration of total arsenic in rice grains cultivated under sprinkler irrigation is close to the total arsenic concentration found in irrigation waters. Our results suggest that, in our experimental conditions, the natural bioaccumulation of this element in rice grains may be completely circumvented by adopting an appropriate irrigation technique.
■
INTRODUCTION The tendency of rice to accumulate arsenic1−4 (As) represents a serious threat to public health, primarily in countries in which rice is the staple food.1,5 Total As concentrations higher than 2000 μg kg−1 have been found in rice grains produced in Bangladesh,3,6 and usually at least one-half of the total amount is found in inorganic forms (i.e., the most toxic forms).6 Chronic intake of a few μg/kg/day of inorganic As can result in a variety of serious health problems in humans,7−9 including cancer,7,9−11 skin diseases,9,12 vascular diseases,9,13 and diabetes.9,13 The uptake of As compounds in rice comes mainly from soil absorption in anaerobic conditions,1,2,14−16 like those typically found in a paddy field irrigated with continuous flooding. The As uptake in rice appears to increase when its concentration in agricultural soils or in irrigation waters increases,17−21 but it reaches levels of the highest concern for public health when both these conditions occur on the same site (e.g., the Bengal Delta Countries).2,16−19,21,22 The most common oxidation states of As in the terrestrial environment are +3 and +5. As(III) compounds are much more toxic, more soluble and more mobile than those of As(V).9 As a matter of fact, in a continuously flooded paddy field the bioavailability of As is more enhanced than in aerobic soils due to an increased presence of the As(III) forms. The factors involved in this phenomena are the reductive dissolution © 2012 American Chemical Society
of iron oxides and/or hydroxides and the consequent release of As soluble species, and the reduction of strongly adsorbed arsenate to more weakly adsorbed arsenite.23,24 Consequently, soluble forms of As are taken up by plants through the pathways for nutrients.14 In particular, arsenite is readily taken up through the Si pathway via aquaglyceroporin channels25 due to its physicochemical similarity with silicic acid, whereas arsenate is taken up via phosphate transporters. From an agronomic perspective, different approaches14,26−31 have been used to mitigate As bioaccumulation in rice. For example, a number of recent studies have ascertained that the tendency of rice to accumulate As varies between different genotypes.32−37 In addition, insights into the role that irrigation techniques may play in the bioaccumulation of As in rice grains have recently been published.5,29−31 Comparisons between different water management techniques, that is, traditional continuous flooding irrigation and “raised beds”,5 intermittent flooding,29 or deficit irrigation30 show that the rice grains harvested from fields irrigated with the latter techniques Received: Revised: Accepted: Published: 8333
February 15, 2012 July 5, 2012 July 5, 2012 July 5, 2012 dx.doi.org/10.1021/es300636d | Environ. Sci. Technol. 2012, 46, 8333−8340
Environmental Science & Technology
Article
irrigation techniques does not significantly affect productivity. The 10 most productive rice genotypes, at 13% humidity in the final product, yielded between 12.4 and 14.1 t ha−1an average of 13.0 t ha−1with continuous-flooding irrigation, and between 11.6 and 13.6 t ha−1an average of 12.2 t ha−1 with sprinkler irrigation. For each genotype and each plot, 100 g of paddy rice was dried at 32 °C, mechanically husked and bleached. Reagents. The 68% aqueous solution of HNO3 and 30% aqueous solution of H2O2 were both Suprapur reagents from Merck, Milan, Italy. A 2% HNO3 standard aqueous solution of As (1000 mg dm−3) was purchased from Fluka, Milan, Italy. High-purity water (type I, resistance >18 MΩ) was produced using a Milli-Qplus System (Millipore, Vimodrone, Italy) device. Unless noted, this water was used for all of the analytical phases of the study. The ICP-MS tuning solution (a 2% HNO3 solution in water containing 10 mg dm−3 of each of the following elements: Li, Y, Ce, Co, and Tl, code 8500−6943) and the Rh standard solution (a 2% HNO3 solution in water containing 10 mg dm−3 of Rh, code 8500−6945) were both purchased from Agilent, Cernusco sul Naviglio, Italy. The certified rice flours NIST SRM 1568a (As = 290 ± 30 μg kg−1), NCS ZC 73008 (As = 102 ± 8 μg kg−1), and IRMM 804 (As = 49 ± 4 μg kg−1) were produced by the National Institute of Standards and Technology, U.S., by the China National Analysis Centre, China, and by the Institute for Reference Materials and Measurements, Belgium, respectively. Instrumentation. The samples were digested by acid/ oxidant-assisted microwave irradiation using a microwave oven model Ethos 900 from Milestone, Sorisole, Italy. The redox potentials are measured according Pansu and Gautheyrou41 using an mV meter by Thermo Orion, model 210A, connected to a home-assembled Pt electrode and an Ag/AgCl reference electrode. The As determination was accomplished using an ICP-MS spectrometer model 7500a from Agilent, Cernusco sul Naviglio, Italy, running the Windows XP operating system. The optimized conditions used during the ICP-MS measurements are summarized in Table 1.
contain less As; this result is likely a consequence of higher aerobic conditions.31 The sprinkler irrigation technique was optimized for rice in Sardinia, Italy,38,39 where its adoption in place of flooding, applied to several rice genotypes over a number of crop years, has produced no significant differences in yield.40 In addition, sprinkler irrigation has demonstrated several agronomic advantages: the water requirements are halved, the number of herbicide treatments needed to control weeds is greatly reduced, no preliminary leveling is required in the soils irrigated by sprinkler, and only commonly used agricultural machinery (i.e., equipment typically used for autumn−winter crops) is required. Moreover, it appears reasonable that sprinkler irrigation may maximize the aerobic conditions needed to reduce the As concentration in rice. For this reason, the principal aim of this study is to verify whether the irrigation method used in rice cultivation (sprinkler irrigation vs flooding irrigation) affects the concentration of total As measured in rice kernels.
■
MATERIALS AND METHODS Soil Cultivation. The study was conducted in Sardinia at the experimental field (39°59′ N and 8°40′ E) of the University of Sassari. The soil, classified as a Typic Eutric Haplic Fluvisol (World Reference Base for Soil Resources), has in its surface layers a sandy-clay texture, a subalkaline pH, a very low amount of carbonates, a low amount of total N and organic carbon and adequate amounts of assimilable phosphorus and exchangeable potassium. The soils undergoing analysis were sampled at depths of 0−20 cm and 20−40 cm. The sampling of the soil for analytical purposes was accomplished following a systematic, mixed and casual grid scheme. The final sample was obtained using established quartering techniques and was dried and sieved prior to analysis, which was performed on a granulometric fraction of less than 2 mm in diameter. The irrigation waters were only from Lake Omodeo, an artificial basin located in the center of Sardinia along the Tirso River. A total of 37 rice genotypes (29 from the japonica subspecies and 8 from the indica subspecies) were grown in two separate fields. The first of these fields was irrigated by continuous flooding, whereas the second underwent sprinkler irrigation. It is important to underline that the paddy field irrigated by continuous flooding had also been used in this manner for the previous thirty-two years. The chosen experimental design followed a randomized block design with four replications for each genotype. First, the fields to be sown were fertilized with the following elemental rates: 80 kg ha−1 of N, 39.3 kg ha−1 of P, and 41.5 kg ha−1 of K applied as urea, calcium dihydrogen phosphate and potassium sulfate, respectively. Both fields were sown on May 21 and harvested on October 13, 2010. During the period between tillering and the end of stem elongation, the sprinkler- and flooding-irrigated fields were respectively fertilized with 105 and 135 kg ha−1 of nitrogen. Pre-emergence weed control was performed using 1.5 kg ha−1 of Pendimethalin for the rice irrigated by sprinkler, whereas the rice irrigated with continuous flooding was treated with 1.0 kg ha−1 of Pendimethalin for pre-emergence weed control and a mixture of Penoxsulam (43.8 g ha−1), Triclopyr (133.2 g ha−1), and 2methyl-4-chlorophenoxyacetic acid (61.2 g ha−1) for postemergence weed control. Further agronomic details of sprinkler irrigation are provided in the Supporting Information. Harvesting and Sampling of Rice. The harvesting was performed with a plot combine. The adoption of different
Table 1. ICP-MS Optimized Operating Conditions in the As Determination in Rice, Soils and Water Samples instrumental characteristics spectrometer nebulizer interface RF generator power output argon flow, l min−1 optimization data acquisition analytical mass, amu
operating conditions Agilent 7500a (Agilent, Milano, Italy) Babington type sampler (1 mm id) and skimmer (0.4 mm id) cones of Ni alloy 1300 W 15.00, plasma; 1.20, nebulizer; 0.10, auxiliary. on masses of 7Li, 89Y, 205Tl dwell time of 100 ms, 3 points per peak, acquisition time of 10 s. 75 As
ICP-MS Measurements in Rice Grains, Soils and Irrigation Waters. Each sample was analyzed three times, and each analytical datum is the average of five replicated ICPMS measurements. Determination of Total As in Rice GrainsGeneral Features. Digestion was performed using a basket containing ten PTFE vessels. Each digestion cycle consisted of seven rice samples, two method blanks, and one certified reference 8334
dx.doi.org/10.1021/es300636d | Environ. Sci. Technol. 2012, 46, 8333−8340
Environmental Science & Technology
Article
Table 2. Validation Parameters for the ICP-MS Determination of Total As in Rice repeatabilityb
linearity a
LoD (μg kg−1)
0.32
LoQ (μg kg−1)
1.06
concentration range (μg kg−1): 1 − 1500 Y = (a ± sa)X + (b ± sb) a = 1.0002 sa = 0.0054 b = 0.0175 sb = 0.0083 R2 = 0.9996
reproducibilityc
accuracy
rice genotype
CV%exp,rd
HorRatre
sample
CV%exp,Rf
HorRat,Rg
CRM
recovery (% ± s)h
SAPISE 164 ARSENAL
1.28 2.63
0.04 0.07
NIST 1568a IRMM 804
3.16 5.68
48.9 64.4
NIST 1568a NCS ZC 73008
87 ± 10 84 ± 10
ARPA
2.89
0.08
IRMM 804
84 ± 9
a
Following Currie (1999).47 bEvaluated by analyzing 3 samples of rice 10 times within the same analytical session. cEvaluated by analyzing 3 CRM rice flours 17 times at different concentrations of total As. dCV%exp,r is the experimental coefficient of the variation of repeatability. eHorRatr is the ratio between CV%exp,r and the theoretical repeatability data (CV%H,r) according to Horwitz’s theory.48 fCV%exp,R is the experimental coefficient of the variation of reproducibility. gHorRatR is the ratio between CV%exp,R and the theoretical reproducibility data (CV%H,R) according to Horwitz’s theory.48 hStandard deviation.
material (CRM): NIST SRM 1568a rice flour for rice samples under flooding irrigation, IRMM 804 rice flour for rice samples under sprinkler irrigation. The absence of significant differences (criteria: two-tail t test, p = 0.95) between the slopes of the regression plots obtained using external calibration and multiple standard additions for the same concentration interval support the lack of any matrix effect. Based on this finding, quantification was accomplished using external calibration. The interference of the 40As35Cl species in the determination of the 75 As isotope is not observed in the method used, as evidenced by the lack of any significant overestimation bias in the results of a recovery test performed on three CRM rice flours containing different total amounts of As. Determination of Total As in Rice GrainsProcedures. The analytical method, based on those described by Caroli et al.,42−44 was optimized in terms of the quantity of the sample and the analytical procedure. Approximately 0.7 g of the sample was weighed in a PTFE vessel on an analytical balance and then treated with 2 cm3 of 68% HNO3 Suprapur, 1 cm3 of 30% H2O2 Suprapur and 2 cm3 of high purity H2O. After closing, the vessel underwent the relevant digestion cycle: 1 min 40 s at 250 W, 2 min at 0 W, 6 min at 250 W, and then 8 min at 400 W. At the end of the digestion cycle, the temperature inside the vessel was typically no higher than 120 °C. The closed vessel was then cooled under ice for at least 2 h and then opened. The content was diluted with up to 20 cm3 of water. Prior to the analysis, the solution was filtered using a 0.45 μm polypropylene filter. The quantification was accomplished by external calibration using a 1 mg dm−3 Rh solution as an internal standard. Detailed analytical data are reported in Table 1. Determination of Extractable As in Soil Samples. The soil digestion procedure used prior to the ICP-MS determination of extractable As (EAs) is an optimization of the EPA 3050b method.45 Approximately 0.3 g of soil (a sieved fraction with diameter under 2 mm) was weighed in a PTFE vessel on an analytical balance and then treated with 5 cm3 of 68% HNO3 Suprapur, 1 cm3 of 30% H2O2 Suprapur, and 2 cm3 of high purity H2O. After closing, the vessel underwent the relevant digestion cycle: 2 min at 250 W, 2 min at 0 W, 6 min at 250 W, 4 min at 400 W, and 5 min at 500 W. At the end of the digestion cycle, the temperature inside the vessel was typically not higher than 130 °C. The closed vessel was cooled under ice for at least 2 h and then opened. The solution was then separated using Whatman 41 filter paper, and the solid fraction
was washed three times with small volumes of water. The washing waters were combined with the filtrate, which was further diluted by 100 cm3 of water. Prior to the analysis, the solution was filtered using a 0.45 μm polypropylene filter. Quantification was accomplished by external calibration using a 1 mg dm−3 Rh solution as an internal standard. Total As in Irrigation Waters. The determination was accomplished following Birke et al.46 Validation of the Analytical Methods. The main validation parameters for the ICP-MS method used for the total As determination in the rice samples are reported in Table 2. The proposed method is characterized by a very low LoD (0.32 μg kg−1)47 and excellent levels of linearity, which spans more than 3 orders of magnitude, and precision (repeatability better than 3%, reproducibility better than 6%). The acceptability of the precision values has also been successfully verified according to Horwitz’s theory.48 Trueness was evaluated by repeated analyses on three CRM rice flours at different total As concentrations, ranging between 290 ± 30 μg kg−1 (NIST SRM 1568a) and 49 ± 4 μg kg−1 (IRMM 804). The recovery values obtained ranged between 84% and 87%. These data indicated a slight underestimation bias, which is acceptable for the concentration levels of the analyte, according to the AOAC Peer Verified Methods.49
■
RESULTS Figure 1 reports the results of the ICP-MS determination of total As in grains of 37 rice genotypes grown with continuous flooding irrigation (white points) or with sprinkler irrigation (blue points). The total As concentration measured in rice grains grown by continuous flooding irrigation ranged between 95 ± 2 μg kg−1 (Arsenal genotype) and 235 ± 3 μg kg−1 (Sapise 164), with an average value of 163 ± 23 μg kg−1. These data are within the range commonly reported in many studies4 and are well below the alarming values reported by Islam et al. in 2004.3 Also in agreement with findings in the literature,32−37 our data support a wide variability in the total As concentration as a function of the rice genotype; the ratio between the highest and the lowest concentrations measured is approximately 2.5. The total As concentration in rice genotypes under sprinkler irrigation is very low in comparison to the previously reported data, ranging between 1.3 ± 0.3 μg kg−1 (Vulcano) and 5.1 ± 0.3 μg kg−1 (Ulisse), with an average value of 2.8 ± 2.5 μg kg−1. The 8335
dx.doi.org/10.1021/es300636d | Environ. Sci. Technol. 2012, 46, 8333−8340
Environmental Science & Technology
Article
Figure 1. The total As concentration in 37 rice genotypes irrigated by either continuous-flooding irrigation or sprinkler irrigation. ○ Rice samples cultivated by continuous flooding irrigation (left Y axis); ■ Rice samples cultivated by sprinkler irrigation (right Y axis); indica genotypes along the X axis are highlighted in bold.
WHO upper limit for As concentration in drinking water (10 μg dm−3).7
average reduction in total As by switching irrigation technique from continuous flooding to sprinkler for the 37 genotypes is 98%, with every genotype experiencing at least a 95.5% reduction in As (Arsenal genotype). Even at these low levels of total As, the effect of the genotype is evident; in this case, the ratio between the highest and the lowest concentrations measured is approximately 3.9. To explain the levels of As observed in the rice, the amount of EAs (i.e., the amount of the element released after the acid/ oxidant digestion of the matrix without the use of hydrofluoric acid) in the soils and the total concentration of As in the irrigation waters was measured. Table 3 shows the typical EAs concentration in the soil samples, based on the soil depth and the irrigation method
■
DISCUSSION The rigorous planning for this study resulted in the production of very reliable and high quality data. Our conclusions are supported by four strength points. To begin, all of the samples were obtained under the same agricultural conditions (geographic location, pedological characteristics, chemical composition of the irrigation water), with the only variable being the irrigation method used: sprinkler irrigation or continuous flooding irrigation. In addition, the samples are genetically uniform, controlling the variability based on genetic factors, and each sample is representative of a specific genotype and not as has been the case in many previous studiesof an unknown mixture of different genotypes grown under unknown conditions and in unspecified countries. Conversely, the wide genetic background investigated suggests that the result of this study is of general interest. Moreover, this study is a comparative trial among a very high number (37) of different rice genotypes; the data thus reflect not causality but a general behavior. Finally, the analytical methods used to evaluate the As in the rice and soils were specifically optimized for this study and completely validated, thus excluding a significant bias error from the data. For the first time, to the best of our knowledge, a relatively simple change such as the adoption of a different irrigation technique has resulted in the reduction of the concentration of total As in rice grains by around 2 orders of magnitude. The effect of sprinkler irrigation on the bioaccumulation of As in rice grains is substantial and was observed for all of the 37 examined rice genotypes. A number of water management regimens, alternative to continuous flooding, have been evaluated by various authors for their effect on the bioaccumulation of As in rice grains. The methods tested are intermittent flooding,29,30 saturation,30
Table 3. Extractable As (EAs) Concentration and the Redox Potential in Soils as a Function of the Sampling Depth and Irrigation Technique soil depth (cm) EAs (μg kg−1 ± sa) in sprinkler soils EAs (μg kg−1 ± sa) in flooding soils redox potential (mV ± sa) in sprinkler soils redox potential (mV ± sa)in flooding soils a
0−20 3190 4680 −200 130
± ± ± ±
20−40 20 45 20 30
2700 ± 10 5470 ± 80
Standard deviation.
used. Also, the typical redox potentials of topsoil used for growing rice by sprinkler and continuous flooding irrigation are reported in this table. The observations are bias-free, supported by the results of four different recovery tests and were performed in triplicate for each analytical sample. The recovery data obtained ranged between 95 ± 4% and 99 ± 2%, and precision (repeatability) is typically better than 3%. The average total concentration of As in the irrigation waters, periodically measured (n = 6) during the entire vegetative cycle of rice, is 1.9 ± 0.6 μg dm−3. This low value is well below the 8336
dx.doi.org/10.1021/es300636d | Environ. Sci. Technol. 2012, 46, 8333−8340
Environmental Science & Technology
Article
raising beds,5 and “aerobic conditions”;30,31 all methods were tested on a selected, single rice genotype and conducted on field5,29,30 or pot scale.5,31 Even considering the differences due to specificity of each experiment, there is a general agreement on the fact that the adoption of less reducing conditions in the soil during the rice growth causes the reduction of the total As amount in rice grains. This effect appears to be small (i.e., less than 5% of reduction of total As amount) when changing from continuous flooding to saturation irrigation,30 but becomes more significant (between 25 and 50% of reduction of total As amount) when traditional irrigation was replaced with raising beds,5 intermittent flooding30 or aerobic methods.30 However, in a pot experiment Xu, McGrath, Meharg and Zhao31 observed a substantial reduction of As in rice grains of close to 90% by switching from continuous flooding irrigation to an “aerobic” irrigation method (i.e., adding a volume of deionized water equal to 70% of the soil’s water holding capacity, contained in a pot with holes at the bottom). According the authors of that study, the “...soil chemical transformations occurring under the flooded conditions are the main reason for the much enhanced As accumulation in paddy rice”. This result seems to roughly indicate an independence of the bioaccumulation of As in rice from the increase of the As concentration in soil. This conclusion is in contrast with the claim substantiated by Duxbury5 and Somenhally,29 both reporting that increasing the As concentration in soil when passing from “reduced” to “oxidated” water management procedures results in increase of the bioaccumulation of As. Hence, our data strengthen and improve experimental evidence previously reported by Xu et al.31 We can attribute the further reduction of the As concentration in rice to (1) a low EAs concentration in our soils (ca. 3.2 mg kg−1 vs 15 mg kg−1 for the As-unamended soil) in comparison to those described by Xu et al;31 (2) a better constancy of a number of key parameters (e.g., pH, redox potential, humidity) usually achievable by passing from pot to field experiments. In addition, the wide variability in the reduction of the bioaccumulation of As in the various rice genotypes clearly suggests a genetic variability effect, whose extent is within the range defined by previous studies.32−35,37 Although we are currently unable to deeply interpret this effect due to a general unavailability of information for each genotype (almost all of them are covered by patents), it is possible propose a comparison between data in literature.32−37 Although the effect of genetic variability in the As bioaccumulation in rice grain is evident and well documented in literature, it is important to consider that it accounts for only for 9−10% of the observed variability in the amount of As in rice grains,35 confirming that the environment plays a more important role than genetics32 (69−80% of the variability of total As concentration in rice grain). Since the effect of genetic variability is the only one operating in both subsets of our study (sprinkler irrigation or continuous flooding irrigation), a convenient way to evaluate it is provided by the measure of the ratio, R, between the highest and the lowest As concentrations in rice grains of different genotypes. R is ca. 2.5 for rice cultivated using continuous flooding irrigation, whereas R is rises to ca. 3.9 in our samples using sprinkler irrigation. Our data also highlights the lack of a meaningful difference in the As bioaccumulation capabilities of the japonica and indica subspecies. The R ratio for each group of subspecies cultivated using continuous flooding and sprinkler irrigation is, respectively, 2.1 and 2 for indica genotypes, and 2.0 and 3.9 for japonica genotypes. Furthermore, the R values
reported are well inside the range of variability defined by a number of studies on the effects of genetic variability on the As bioaccumulation in rice cultivated in open fields. Table 4 summarizes selected literature and our data. In particular, the R values from this study are inside the R range (from 1.7 to 6.6) of the subset of genotype groups cultivated in a soil/water system with amounts of As (less than 6 mg kg−1 for soils and less than 10 μg dm−3 for irrigation waters) similar to the As concentration of our soil/water system.34,35,37 Regarding the EAs concentration in soil, all the analyzed samples reveal an EAs concentration within the range typically measured in uncontaminated soils (between 100 and 95 000 μg kg−1).9 However, the type of irrigation technique used seems to affect the concentration of As in soil. In particular, soils irrigated for many years with continuous flooding showed higher EAs concentrations (between 145% and 202%) than in similar soils irrigated with the sprinkler method. Given the common pedological origin of both soils, these data can be tentatively interpreted to indicate a “concentration” effect of the irrigation technique since paddy rice are usually not involved in any crop rotation practice (e.g., in our case, these soils were used to grow rice with continuous flooding irrigation for 32 consecutive years). Finally, we compared a “historical” topsoil sample, collected from a site before it was implanted with a traditional paddy field, and soil samples from a field where rice was cultivated using the sprinkler irrigation; we found no significant difference in the samples’ EAs concentrations. The relationship between the EAs concentration in the soils and their sampling depth has been examined. The analysis revealed an inverted concentration gradient of EAs in the soils irrigated with sprinklers with respect to those irrigated by continuous flooding. In the first case, higher concentrations of As were found on the surface layer, whereas the opposite is true in the flooded soils. This behavior can be explained by the overall As balance in the system comprised by the irrigation waters, soil and rice. The irrigation waters ensure the constant feed of As to the soil-rice system. In the case of sprinkler irrigation, the quantity fed is low and concentrated only on the soil surface. On the other hand, due to the higher water quantities used for continuous flooding irrigation, the quantity of As introduced into the system is much higher in this case and extends beneath the topsoil. Table 1 also shows the very different redox conditions of the soils irrigated by continuous flooding and by sprinkler techniques. While the redox potential of continuously flooded soil is not very different from those measured in other studies performed under similar conditions,1,5 the redox potential of sprinkled soils is significantly higher, even higher than those reported by Duxbury et al.5 for a “raised beds” rice cultivation. These data led us to confirm, in agreement with previous reports,5,22,23 that rice grown in a reducing environment can more efficiently bioaccumulate As compounds than rice grown in an oxidizing environment. As a matter of fact, the reducing conditions of flooded paddy soil can promote two different mechanisms that lead to As mobilization. One is the reduction of arsenate to arsenite, which then desorbs from the adsorption surfaces of iron oxyhydroxides into the solution phase. The other is the reductive dissolution of iron oxyhydroxides, which can release the adsorbed/coprecipitated As compounds to the solution.23,24,50,51 Because roots take up As mainly from the soil solution, the concentration and the chemical nature of As in this matrix should reflect its bioavailability to plant roots. It is well-known that arsenite 8337
dx.doi.org/10.1021/es300636d | Environ. Sci. Technol. 2012, 46, 8333−8340
Environmental Science & Technology
Article
Table 4. Genetic Variability Effect on the As Bioaccumulation Capability in Rice Grain Cultivated in Open Field total As in soils (mg kg−1)
total As in irrigation waters (mg kg−1)
total As range in rice grain (mg kg−1)
genotype variability ratio Ra
number of genotypes
name/number of sites of cultivation
references
8.1 5.7 4.9 12.8 5.7 4.4 6.8 15.1 8.1 19.6 5.7 4.9 8.0 12.8 5.7 4.4 15.1 29.6 10.3 6.3 17.9 65.6 64.6 4.29 9.04 5.9 5.9 5.9 4.68 3.19
