High levels of arsenic in well water are causing widespread poisoning in Bangladesh. In a typical aquifer in southern Bangladesh, chemical data imply that arsenic mobilization is associated with recent inflow of carbon. High concentrations of radiocarbon-young methane indicate that young carbon has driven recent biogeochemical processes, and irrigation pumping is sufficient to have drawn water to the depth where dissolved arsenic is at a maximum. The results of field injection of molasses, nitrate, and low-arsenic water show that organic carbon or its degradation products may quickly mobilize arsenic, oxidants may lower arsenic concentrations, and sorption of arsenic is limited by saturation of aquifer materials.

The origin of dissolved arsenic in the Ganges Delta has puzzled researchers ever since the report of widespread arsenic poisoning two decades ago. Today, microbially mediated oxidation of organic carbon is thought to drive the geochemical transformations that release arsenic from sediments, but the source of the organic carbon that fuels these processes remains controversial. At a typical site in Bangladesh, where groundwater-irrigated rice fields and constructed ponds are the main sources of groundwater recharge, we combine hydrologic and biogeochemical analyses to trace the origin of contaminated groundwater. Incubation experiments indicate that recharge from ponds contains biologically degradable organic carbon, whereas recharge from rice fields contains mainly recalcitrant organic carbon. Chemical and isotopic indicators as well as groundwater simulations suggest that recharge from ponds carries this degradable organic carbon into the shallow aquifer, and that groundwater flow, drawn by irrigation pumping, transports pond water to the depth where dissolved arsenic concentrations are greatest. Results also indicate that arsenic concentrations are low in groundwater originating from rice fields. Furthermore, solute composition in arsenic-contaminated water is consistent with that predicted using geochemical models of pond-water-aquifer-sediment interactions. We therefore suggest that the construction of ponds has influenced aquifer biogeochemistry, and that patterns of arsenic contamination in the shallow aquifer result from variations in the source of water, and the complex three-dimensional patterns of groundwater flow.

No abstract

Arsenic-rich groundwater from shallow tube wells is widely used for the irrigation of boro rice in Bangladesh and West Bengal. In the long term this may lead to the accumulation of As in paddy soils and potentially have adverse effects on rice yield and quality. In the companion article in this issue, we have shown that As input into paddy fields with irrigation water is laterally heterogeneous. To assess the potential for As accumulation in soil, we investigated the lateral and vertical distribution of As in rice field soils near Sreenagar (Munshiganj, Bangladesh) and its changes over a 1 year cycle of irrigation and monsoon flooding. At the study site, 18 paddy fields are irrigated with water from a shallow tube well containing 397 +/- 7 microg L(-1) As. The analysis of soil samples collected before irrigation in December 2004 showed that soil As concentrations in paddy fields did not depend on the length of the irrigation channel between well and field inlet. Within individual fields, however, soil As contents decreased with increasing distance to the water inlet, leading to highly variable topsoil As contents (11-35 mg kg(-1), 0-10 cm). Soil As contents after irrigation (May 2005) showed that most As input occurred close to the water inlet and that most As was retained in the top few centimeters of soil. After monsoon flooding (December 2005), topsoil As contents were again close to levels measured before irrigation. Thus, As input during irrigation was at least partly counteracted by As mobilization during monsoon flooding. However, the persisting lateral As distribution suggests net arsenic accumulation over the past 15 years. More pronounced As accumulation may occur in regions with several rice crops per year, less intense monsoon flooding, or different irrigation schemes. The high lateral and vertical heterogeneity of soil As contents must be taken into account in future studies related to As accumulation in paddy soils and potential As transfer into rice.