REVERSE OSMOSIS AS A HAZARD: EVIDENCE FOR RO WASTE CONTAMINATION OF THE SOLE-SOURCE AQUIFER OF THE DRY ZONE, SRI LANKA

This study addresses unwitting potential increase in kidney disease risk from unregulated environmental mitigation efforts in the Dry Zone of Sri Lanka. For over 30 years, residents of this groundwater-dependent area have experienced a persistent chronic kidney disease of unknown origin (CKDu, a leading cause of death there). The strong spatial correlation of CKDu with the Dry Zone indicates an important environmental component to the disease. An ad hoc response has been installation of many small-capacity private reverse osmosis (RO) water treatment plants selling purified water to local residents. The plants discharge their waste directly to the surface, thereby potentially concentrating environmental factors contributing to CKDu.

Monitoring of temporal changes in water chemistry at 12 wells over 1.5 years and multiple monsoon-dry cycless reveals 2 hydrofacies and several wells with long term increasing trends in total dissolved solids (TDS) and fluoride (F, a suspected nephrotoxin in this area). One well within 900 m of an RO plant exhibited steady increase in both, as well as other indicators. Novel serendipitous tracers of RO waste such as TDS vs. Pb, Sr/K, Mg,K and Ba/B ratios, are consistent with a strong RO waste component at this well. At the other extreme are wells with high seasonal variability of TDS, and a distinct CaMg-HCO⁻³ chemistry that show little evidence of RO waste contamination. Most other wells exhibiting the more typical Na-HCO^-3 chemistry show weak trends in these potential RO waste tracers.

Groundwater in this area is extracted from a saprolite/metamorphic regolith aquifer developed primarily in poorly-outcropping charnockite terranes (likely origin of Na−HCO⁻³ water) with isolated carbonate (potential source of CaMg-HCO^-3 water) and impermeable quartzite. Groundwater basins constrained by quartzite appear to exhibit the strongest RO waste impact, with a minimal seasonal impact. Most other wells near RO plants in larger charnockite zones show milder RO waste impacts (somewhat elevated TDS and RO waste tracers) and mild seasonal changes. A small but significant group of wells show strong seasonal changes, but no indication of direct communication to the surface (e.g. bacterial contamination), and no discernible RO waste impact. This category of well might be the safest in terms of unknown but likely CKDu-related dissolved components. Wells near quartzite lenses may be much more strongly impacted, and appear to be becoming more hazardous as a result of uncontrolled RO waste disposal.