| Summary: | Groundwater geochemistry and flow have been studied in Gotra, West Bengal, India, where geogenic arsenic contaminates groundwater at levels above World Health Organization limits. The village is situated upon the natural levee of an abandoned channel, which terminates a fluvio-deltaic depositional sequence. The formerly prograding meander bend deposited point-bar sands during the Holocene that now comprise the 30m-thick shallow aquifer, while incising deeper Pleistocene sands and a shallow floodplain sequence. Hand-pumped tubewells are completed in point-bar sands, whereas irrigation wells are generally screened within Pleistocene materials. Hydraulic conductivity of the point-bar aquifer is approximately 5x10⁻⁴ m/s. A leaky-confining layer of levee/crevasse splay deposits overlies this aquifer beneath crop fields. Near the village, the shallow aquifer is confined by channel-fill silts that have an estimated conductivity of 1x10⁻⁷ m/s. The paleohorizon separating Holocene and Pleistocene sediments is marked locally by fine-grained material comprised of organic material and a hard clay. Groundwater elevations vary 7m over the course of a given year due to monsoonal climate and groundwater extraction. Convergent flow towards shallow pumping wells is a salient feature during irrigation season, and persistent localized downward gradients suggest that meteoric recharge and pumping are the predominant groundwater sources and sinks. A northeast trending geochemical gradient in the point-bar aquifer suggests that local flows transporting arsenic are directed away from the channel-fill silt. Concentration gradients of terminal electron acceptors and redox reaction products in the shallow aqueous profile also coincide with a flowpath originating from this unit. A numerical model of groundwater flow was developed based on a conceptual model derived from field observations to investigate controls of arsenic transport and to provide a context in which to interpret geochemical data. Short- (7-day) and long- (3-year) term transient simulations were implemented to simulate groundwater flowpaths, water balance, and average residence times. Modeling results support field observations that contamination is related to the channel-fill unit deposit, and suggest that the water balance has been significantly altered compared to pre-irrigation conditions. Results also suggest contaminant flushing will not occur over a human timescale.
|
|---|
