Origin and modelling of water salinization in a coastal aquifer of the Bay of Bengal: The Kaluvelly watershed, Tamil Nadu, India

Over-pumping of the Vanur sandstone aquifer has led to a lowering of the piezometric surface and degradation of water quality through increased salinization. This aquifer salinization is a good example of many similar problems in India and other parts of the world. The Vanur Formation is the main aquifer of a multilayered system bordered by the sea on the eastern side and partly overlaid by the brackish waters of the Kaluvelly swamp in the north. The origin of the salinity, which apparently is not simply seawater intrusion, and its dynamics are examined, using chemical and isotopic tools together with ID hydrological modelling of the movement of the seawater/fresh water interface. The content of major elements and some trace elements as well as isotopic ratios (O-18/O-16, D/H and Sr-87/Sr-86) were measured in groundwater, surface and rainwater during five sampling surveys (January 1999 to October 2001). Available data on rainfall, piezometric and hydrogeologic records were used. We identified human contamination by F, Li in parts of the aquifer, which invalidated their use as tracers. The chemical composition of water from the Vanur aquifer shows a classical chemical evolution from the recharge area to the deeper confined area, consisting in increased water-rock interaction and a subsequent increase of solute species. However, the range of major compound concentration ratios for some wells does not follow this general trend. The non-consistent points are located in the most depressed area of the aquifer (-20 m amsl in June 2000), except in the north where the brackish water of the Kaluvelly swamps seems to enter the aquifer. In the depressed area, the sulphate signature corresponds to a mixing with a mineralized and sulphate-rich water body, likely to be Ramanathapuram sandstone water. It is due to the upward leakage from this underlying formation. A seasonal evaporation signature recorded by stable isotopes (delta O-18, delta D) suggests the addition of return irrigation flow. The 1D hydrodynamical model of the seawater/fresh water movement was built with the available geological and hydraulic data. Hydrodynamic calculations show that seawater intrusion can be expected to occur within 3 to 20 years after the year 2000, depending on the value of unknown parameters (porosity) or boundary conditions (recharge, pumped volume). But we cannot rule out that a lithologic or tectonic barrier prevent any seawater intrusion inland; future geological investigation has to be done to confirm or infirm this hypothesis.