Environ. Sci. Technol. 2011, 45, 971–976
Arsenic Dynamics in Porewater of an Intermittently Irrigated Paddy Field in Bangladesh L I N D A C . R O B E R T S , * ,†,‡ S T E P H A N J . H U G , * ,† ANDREAS VOEGELIN,† J E S S I C A D I T T M A R , ⊥,‡ RUBEN KRETZSCHMAR,‡ BERNHARD WEHRLI,‡ GANESH C. SAHA,§ A. BORHAN M. BADRUZZAMAN,| AND M. ASHRAF ALI| Eawag, Swiss Federal Institute of Aquatic Science and Technology, Ueberlandstrasse 133, CH-8600 Duebendorf, Switzerland, Institute of Biogeochemistry and Pollutant Dynamics, Department of Environmental Sciences, ETH Zurich, CH-8092 Zurich, Switzerland, Department of Civil Engineering, Dhaka University of Engineering and Technology, Gazipur-1700, Bangladesh, and Department of Civil Engineering, Bangladesh University of Engineering and Technology, Dhaka-1000, Bangladesh
Received August 21, 2010. Revised manuscript received November 19, 2010. Accepted November 25, 2010.
In Bangladesh, irrigation of dry season rice (boro) with arseniccontaminated groundwater is leading to increased As levels in soils and rice, and to concerns about As-induced yield reduction. Arsenic concentrations and speciation in soil porewater are strongly influenced by redox conditions, and thus by water management during rice growth. We studied the dynamics of As, Fe, P, Si, and other elements in porewater of a paddy field near Sreenagar (Munshiganj), irrigated according to local practice, in which flooding was intermittent. During early rice growth, As porewater concentrations reached up to 500 µg L-1 and were dominated by AsIII, but As release was constrained to the lower portion of the soil above the plow pan. In the later part of the season, soil conditions were oxic throughout the depth range relevant to rice roots and porewater concentrations only intermittently increased to ∼150 µg L-1 AsV following irrigation events. Our findings suggest that intermittentirrigation,currentlyadvocatedinBangladeshforwatersaving purposes, may be a promising means of reducing As input to paddy soils and rice plant exposure to As.
Introduction Dry season rice (boro) production has greatly increased over the last decades in Bangladesh and now amounts to more than 50% of national rice production (1). Since the irrigation * Address correspondence to either author. Phone: +41-448235418 (L. C. R.); 41-44-8235454 (S. J. H.). Fax: +41-44-8235210 (L. C. R.); +41-44-8235210 (S. J. H.). E-mail:
[email protected] (L. C. R.);
[email protected] (S. J. H.). † Eawag. ‡ ETH Zurich. § Dhaka University of Engineering and Technology. | Bangladesh University of Engineering and Technology. ⊥ Current address: Department of Environmental Earth System Science, Stanford University, Stanford, California 94305. 10.1021/es102882q
2011 American Chemical Society
Published on Web 12/17/2010
demand of boro rice is mainly met by groundwater from shallow tube wells, often containing high As concentrations (1), on average ∼0.4 kg ha-1 of As are added to arable soils each year (2, 3). This is leading to increased As contents in paddy soils (4-6), rice straw and grain (5, 7, 8), and could cause significant yield reductions associated with As toxicity (9-11). Strategies to mitigate As input to soils and to reduce rice plant exposure to As are therefore urgently needed and require a detailed understanding of As behavior under the range of conditions suitable for rice growth, which in principle span continuously submerged to drained soils (12, 13). In flooded soils, As is mobilized into porewater due to reductive dissolution of FeIII(hydr)oxides and to arsenate (AsV) reduction to the less competitively sorbing arsenite (AsIII) (14, 15). By contrast, As concentrations in porewater are markedly lower under oxic conditions and generally dominated by AsV (16, 17). Rice plants take up AsV through phosphate transporters and AsIII via the silicate uptake system (18). Concentrations of phosphate and silicate in soil porewater therefore play an important role in regulating As uptake by rice plants under oxic and reducing conditions. Aerobic rice cultivation is receiving considerable attention as a means of reducing plant exposure to As. Compared to continuous flooding, aerobic conditions maintained throughout (17) or during part of the growing season (19) significantly reduced As uptake in greenhouse studies. Similarly, raised bed cultivation reduced plant As uptake and alleviated yield reductions across a gradient of increasing soil As contents in the field in Bangladesh (20). Continuous cropping of aerobic rice, however, leads to severe yield declines in the tropics and is therefore not a feasible option at present (21). By contrast, alternate wetting and drying (AWD), an intermittent irrigation practice in which water level is allowed to drop to ∼15 cm below the soil surface between irrigation events, is considered yield-robust and ready for practice (12). In Bangladesh, intermittent irrigation practices are already common in some areas and AWD is currently disseminated as a water-saving technique (22, 23). This makes it crucially relevant to understand the behavior of As in intermittently irrigated rice paddies subject to varying soil redox conditions. It was the aim of this study to asses porewater dynamics of As and other elements (Fe, Mn, P, S, Si) influencing As mobility and/or As uptake by rice, at a field site in Munshiganj (Bangladesh) where farmers allow the soil surface to become exposed to air between irrigation events (24, 25). The study field is characterized by a spatial As gradient in irrigation water and soil (25, 26), which is also reflected in higher grain and straw As contents in plants sampled near the irrigation water inlet (27). We assessed soil porewater dynamics with high vertical resolution during two different stages of boro rice growth: (i) vegetative plant growth and (ii) grain-filling, and investigated the effect of lateral concentration gradients in irrigation water (25) and paddy soil (26).
Materials and Methods Field Site. The field site (Figure 1) is located 30 km south of Dhaka and 5 km north of the river Ganges and is subject to intense monsoon flooding between mid-June and late October (14, 28). Rice production is therefore limited to groundwater-irrigated boro rice, grown between late December and May. The local irrigation well delivers anoxic groundwater containing 400 µg L-1 As (84% AsIII), 11 mg L-1 Fe, 2 mg L-1 P and 20 mg L-1 Si (25). The well was constructed in the early 1990s, and the studied field (2778 m2) has since been exclusively used for boro production. The field is VOL. 45, NO. 3, 2011 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
971
FIGURE 1. (a) Map of the study field showing locations where porewater samplers (PS) were deployed during the irrigation season 2005, 2006, and 2007. (b) Enlargement of the field corner near the irrigation inlet with the exact positions of the porewater samplers deployed in this location. (c) Periods of sampler deployment, indicated with gray bars for each sampling year. The vertical blue bars mark irrigation events. puddled annually before rice transplantation and has a carbonate-free silty clay loam topsoil confined by a plow pan at 20-25 cm depth (24, 26). The irrigation water is delivered by a shallow channel and enters the field through a breach (inlet) in the field bund. The inlet remained in the same location until 2006, but was shifted 15 m to the left in January 2007 following a change in field tenancy (Figure 1). The field is irrigated with ∼1.0 m of water per season (24, 25). Between irrigation events, which last 2-3 h and add 3-10 cm of water to the field, water levels decrease gradually, mainly due to infiltration through the unplowed bunds, preferential flow through cracks in the plow pan and evapotranspiration (24, 29). In general, the next irrigation is not initiated before the moist soil surface is exposed, with scattered water-filled depressions in wetter and soil cracks in dryer parts (local farming practice). Sampling and Analyses of Field Water and Soil Porewater. In the following, irrigation water sampled above the field soil is referred to as field water, whereas water sampled from the soil matrix is termed porewater. Field water was sampled as detailed in (25). Porewater was sampled during vegetative plant growth in February 2006 and during the grain-filling stage toward the end of the boro season in April 2005 and 2007, using samplers which yielded profiles of the upper 40-60 cm of soil with a 1.5 cm depth resolution. We focus our discussion primarily on the depth range above the plow pan (0-20 cm), coinciding with the main depth of rice root growth and therefore of plant As uptake (26, 30). The porewater samplers were deployed over periods of ∼2 weeks in February 2006 and April 2007, and over 5 days in April 2005 (Figure 1c). Rice growth stages during sampler deployment are documented in Supporting Information (SI) Figure S1. In February 2006, one sampler was 972
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 45, NO. 3, 2011
placed 3 m away from the irrigation inlet and one at a distance of 3 m from the opposite field corner (Figure 1a and b), that is, at the extremities of the As gradient in field water (25) and topsoil (26). All other porewater samplers were deployed in the vicinity of the irrigation inlet 2005/2006, where topsoil As contents are highest (26). The porewater samplers, described in detail in (14) and in the SI, were deoxygenated and transported to the field in oxygen-impermeable bags. Upon retrieval from the soil, samples were collected within 2-3 h under ambient (2005-2006) or N2-atmosphere (2007). Three subsamples (i) untreated, (ii) after 0.2 µm filtration, and (iii) after filtration through a 0.2-µm filter combined with a modified As speciation cartridge (25) were added to preacidified polyethylene vials, allowing for differentiation between dissolved and total element contents and between dissolved AsIII and AsV. Discrepancies between unfiltered and filtered samples indicated the formation of colloids within the sampler (see SI). Arsenic was determined by hydride generation atomic fluorescence spectrometry (HG-AFS; PS Analytical Ltd., U.K.); Fe, Mn, P, Si, Ca, Mg, Na, K, and S by inductively coupled plasma optical emission spectrometry (ICP-OES; Spectro Ciros CCD) as described in (14, 25).
Results and Discussion Porewater Dynamics during Early Rice Growth (Vegetative Stage). All porewater samplers deployed during the irrigation season were characterized by the presence of brown precipitates on the membranes and inside the sampler chambers within certain depth ranges (SI Figure S2). The precipitates were visible upon retrieval of the samplers from the soil, that is, did not result from Fe oxidation during the sampling process but identified depth ranges in which FeIII(hydr)oxide particles formed while the samplers were in situ. In these
FIGURE 2. Porewater profiles near the irrigation inlet of Field I during vegetative boro growth (11 February 2006, sampler PS-3 (a-d)) and during the grain-filling stage (26 April 2007, sampler PS-6 (e-h)). The depth ranges in which brown precipitates were visible upon sampler retrieval are shaded in brown; the plow pan is indicated by a hatched bar in all panels. (a, e) As(tot), As(diss) and AsIII(diss); (b, f) Fe(tot) and S(tot); (c, g) Mn(tot) and P(tot); (d, h) Ca(tot) and Si(tot). depth ranges, conditions therefore changed from Fe-reducing to Fe-oxidizing at least once during sampler deployment. During early rice growth in February 2006 brown precipitates formed in the topmost portion of the soil (SI Figure S2), most likely due to cyclic redox changes related to decreasing water levels between irrigation events (24). The field was irrigated shortly before and twice while the samplers were in situ, with 5-day periods in between (Figure 1c). When the samplers were retrieved, 3-4 days had passed since the last irrigation, indicating that the soil was retransitioning toward more oxidizing conditions. In the sampler deployed near the irrigation inlet (PS-3, Figure 1b) the zone of brown coloration extended over the top ∼9 cm (SI Figure S2). Porewater As concentrations were
