This study reexamines the notion that extensive As mobilization in anoxic groundwater of Bangladesh is intimately linked to the dissolution of Fe oxyhydroxides on the basis of analyses performed on a suite of freshly collected samples of aquifer material. Detailed sediment profiles extending to 40 to 70 m depth below the surface were obtained at six sites where local groundwater As concentrations were known to span a wide range. The sediment properties that were measured include (1) the proportion of Fe(II) in the Fe fraction leached in hot 1.2 N HCl, (2) diffuse spectral reflectance, and (3) magnetic susceptibility.In parallel with local concentrations of dissolved As ranging from Ͻ5 to 600 g/L, Fe(II)/Fe ratios in shallow (gray) Holocene sands tended to gradually increase with depth from values of 0.3 to 0.5 to up to 0.9. In deeper (orange) aquifers of presumed Pleistocene age that were separated from shallow sands by a clay layer and contained Ͻ5 g/L dissolved As, leachable Fe(II)/Fe ratios averaged ϳ0.2. There was no consistent relation between sediment Fe(II)/Fe and dissolved Fe concentrations in groundwater in nearby wells. The reflectance measurements indicate a systematic linear relation (R 2 of 0.66; n ϭ 151) between the first derivative transform of the reflectance at 520 nm and Fe(II)/Fe. The magnetic susceptibility of the shallow aquifer sands ranged from 200 to 3600 (x 10 -9 m 3 /kg SI) and was linearly related (R 2 of 0.75; n ϭ 29) to the concentrations of minerals that could be magnetically separated (0.03 to 0.79% dry weight). No systematic depth trends in magnetic susceptibility were observed within the shallow sands, although the susceptibility of deeper low-As aquifers was low (up to ϳ200 ϫ 10 -9 m 3 /kg SI).This set of observations, complemented by incubation results described in a companion paper by van Geen et al. (this volume), suggests that the release of As is linked to the transformation of predominantly Fe (III) oxyhydroxide coatings on sand particles to Fe(II) or mixed Fe(II/III) solid phases with a flatter reflectance spectrum such as siderite, vivianite, or magnetite, without necessarily resulting in the release of Fe to groundwater. The very low As/Fe ratio of magnetically separated minerals compared to the As/Fe of bulk acid leachate (2 vs. 40 10 -6 , respectively) suggests that such a transformation could be accompanied by a significant redistribution of As to a mobilizable phase on the surface of aquifer particles.

Groundwater drawn daily from shallow alluvial sands by millions of wells over large areas of South and Southeast Asia exposes an estimated population of over 100 million to toxic levels of arsenic (1). Holocene aquifers are the source of widespread arsenic poisoning across the region (2, 3). In contrast, Pleistocene sands deposited in this region more than ~12,000 years ago mostly do not host groundwater with high levels of arsenic. Pleistocene aquifers are increasingly used as a safe source of drinking water (4) and it is therefore important to understand under what conditions low levels of arsenic can be maintained. Here we reconstruct the initial phase of contamination of a Pleistocene aquifer near Hanoi, Vietnam. We demonstrate that changes in groundwater flow conditions and the redox state of the aquifer sands induced by groundwater pumping caused the lateral intrusion of arsenic contamination over 120 m from Holocene aquifer into a previously uncontaminated Pleistocene aquifer. We also find that arsenic adsorbs onto the aquifer sands and that there is a 16–20 fold retardation in the extent of the contamination relative to the reconstructed lateral movement of groundwater over the same period. Our findings suggest that arsenic contamination of Pleistocene aquifers in South and Southeast Asia as a consequence of increasing levels of groundwater pumping have been delayed by the retardation of arsenic transport.

Whereas serious health consequences of widespread consumption of groundwater elevated in As have been documented in several South Asian countries, the mechanisms responsible for As mobilization in reducing aquifers remain poorly understood. We document here a previously unrecognized and consistent relationship between dissolved As concentrations in reducing groundwater and the phosphate-mobilizable As content of aquifer sediment for a set of precisely depth-matched samples from across Bangladesh. The relationship holds across nearly 3 orders of magnitude in As concentrations and suggests that regional as well as local patterns of dissolved As in shallow groundwater are set by the solid phase according to a remarkably constant ratio of ∼250 μg/L dissolved As per 1 mg/kg P-mobilizable As. We use this relationship in a simple model of groundwater recharge to propose that the distribution of groundwater As in shallow aquifers of the Bengal Basin could primarily reflect the different flushing histories of sand formations deposited in the region over the past several thousand years.

Extreme spatial heterogeneity has emerged as a salient characteristic of groundwater arsenic in many complex fl uviodeltaic environments. Here we examine patterns of arsenic heterogeneity in the shallow (<23 m) groundwaters of a well-studied fl oodplain setting in Araihazar, Bangladesh. Ninetyfi ve augers and 200 shallow wells sampled at spacings of 10 1 -10 3 m in the 25 km 2 area indicate that the concentration of arsenic in shallow groundwater largely varies with the grain size, thickness, and distribution of fi ne-grained (<63 µm) sediments that overlie buried aquifer sands. The overall pattern shows that lower arsenic concentrations are typically found where aquifer sands outcrop at or near the surface, whereas higher arsenic levels typically underlie, or are adjacent to thicker, fi ne-grained deposits. Furthermore, chronostratigraphic reconstructions of aquifer sediments indicate that sediment distribution, and consequently the patterning of dissolved arsenic, is readily explained in the context of local river history and fl oodplain development within the past 1000 yr. An important implication is that complex patterns of groundwater arsenic in affl icted fl uviodeltaic settings can be better understood through reconstructions of local aquifer history. This fi nding is especially relevant because the village and tube-well locations are closely linked with surface landforms such as former levees and bars. An additional and worrisome fi nding is that the artifi cial fi lling of villages to protect from fl ooding can mimic the natural fi ne-grained stratigraphy commonly associated with high concentrations of arsenic.

[1] The high-degree of spatial variability of dissolved As levels in shallow aquifers of the Bengal Basin has been well documented but the underlying mechanisms remain poorly understood. We compare here As concentrations measured in groundwater pumped from 4700 wells <22 m (75 ft) deep across a 25 km 2 area of Bangladesh with variations in the nature of surface soils inferred from 18,500 measurements of frequency domain electromagnetic induction. A set of 14 hand auger cores recovered from the same area indicate that a combination of grain size and the conductivity of soil water dominate the electromagnetic signal. The relationship between pairs of individual EM conductivity and dissolved As measurements within a distance of 50 m is significant but highly scattered (r 2 = 0.12; n = 614). Concentrations of As tend to be lower in shallow aquifers underlying sandy soils and higher below finer-grained and high conductivity soils. Variations in EM conductivity account for nearly half the variance of the rate of increase of As concentration with depth, however, when the data are averaged over a distance of 50 m (r 2 = 0.50; n = 145). The association is interpreted as an indication that groundwater recharge through permeable sandy soils prevents As concentrations from rising in shallow reducing groundwater.

No abstract

One of the reasons the processes resulting in As release to groundwater in southern Asia remain poorly understood is the high degree of spatial variability of physical and chemical properties in shallow aquifers. In an attempt to overcome this difficulty, a simple device that collects groundwater and sediment as a slurry from precisely the same interval was developed in Bangladesh. Recently published results from Bangladesh and India relying on the needle-sampler are augmented here with new data from 37 intervals of grey aquifer material of likely Holocene age in Vietnam and Nepal. A total of 145 samples of filtered groundwater ranging in depth from 3 to 36 m that were analyzed for As (1–1000 μg/L), Fe (0.01–40 mg/L), Mn (0.2–4 mg/L) and S (0.04–14 mg/L) are compared. The P-extractable (0.01–36 mg/kg) and HCl-extractable As (0.04–36 mg/kg) content of the particulate phase was determined in the same suite of samples, in addition to Fe(II)/Fe ratios (0.2–1.0) in the acid-leachable fraction of the particulate phase. Needle-sampler data from Bangladesh indicated a relationship between dissolved As in groundwater and P-extractable As in the particulate phase that was interpreted as an indication of adsorptive equilibrium, under sufficiently reducing conditions, across 3 orders of magnitude in concentrations according to a distribution coefficient of 4 mL/g. The more recent observations from India, Vietnam and Nepal show groundwater As concentrations that are often an order of magnitude lower at a given level of P-extractable As compared to Bangladesh, even if only the subset of particularly reducing intervals characterized by leachable Fe(II)/Fe >0.5 and dissolved Fe >0.2 mg/L are considered. Without attempting to explain why As appears to be particularly mobile in reducing aquifers of Bangladesh compared to the other regions, the consequences of increasing the distribution coefficient for As between the particulate and dissolved phase to 40 mL/g for the flushing of shallow aquifers of their initial As content are explored.

Reductive dissolution of iron oxyhydroxides and reduction of arsenic are often invoked as leading causes of high dissolved As levels in shallow groundwater of Bangladesh. The second of these assumptions is questioned here by comparing the behavior As and phosphate (P), a structural analogue for As (V) which also adsorbs strongly to Fe oxyhydroxides but is not subject to reduction. The first line of evidence is provided by a detailed groundwater time-series spanning two years for three wells in the 6–9 m depth range showing removal of As(III) from shallow groundwater during the monsoon without of loss of P. The data indicate a loss of ~90% of the dissolved As from groundwater in the intermediate well relative to a level of 3 μmol/L As predicted by conservative mixing between groundwater sampled from the shallower and the deeper well. In contrast, P concentrations of ~30 μmol/L in the intermediate well closely match the prediction from conservative mixing. Reduction therefore appears to inhibit the release of As to groundwater at this site relative to P instead of enhancing it. A re-analysis of existing groundwater As and P data from across the country provides a broader context for this finding and confirms that, without reduction, elevated concentrations of As would probably be even more widespread in shallow aquifers of Bangladesh. Without providing definite proof, X-ray absorption spectroscopy of sediment from the time-series site and elsewhere suggests that the loss of As from groundwater may be coupled to precipitation of As sulfide. Further study is needed to assess the implications of these observations for shallow aquifers that have been subjected to increased withdrawals for irrigation in recent decades.