salt or germs

Problem statement

In the south-west coastal districts of Khulna, Bagerhat and Satkhira the aquifers are frequently saline, so households harvest rainwater in the monsoon and, when storage runs out in the dry months, drink from ponds — which are faecally contaminated (98% of pond samples in the icddr,b study exceeded 10 MPN/100 mL faecal coliforms) and are periodically inundated by tidal surges and cyclones. Drinking brackish water is associated with hypertension and with pre-eclampsia in pregnancy. Managed aquifer recharge (MAR) — designed by the University of Dhaka's Geology Department and piloted with UNICEF (20 sites 2009–12, then 75 more in 2013–14) — pumps sand-filtered pond water into the aquifer to dilute salinity and store it safely underground; it succeeded at the microbiology (81% of MAR samples had no or low-risk faecal coliforms) but the recovered water had roughly 1.7 times the conductivity of pond water (1,624 vs 974 µS/cm), only 42% of households used it exclusively, and in a 16-community stepped-wedge trial access to MAR raised systolic blood pressure by about 2 mmHg and urinary sodium by about 10 mmol/day relative to the pond-water counterfactual. The authors concluded their findings "do not support the scale-up of MAR systems as a routine drinking-water source." No available option — rainwater, pond, pond-sand filter, MAR, or desalination — currently delivers low-sodium, pathogen-free, year-round water at village cost.

Why this matters

Millions of people live in Bangladesh's salinity-affected coastal belt and the choice they face is literally between salt and germs: the low-sodium source carries diarrhoeal disease risk and fails in storms; the microbiologically safer engineered source carries a cardiovascular dose. The blood-pressure effect is small per person but population-wide and chronic, and it falls on pregnant women in particular. Because MAR was the flagship "climate-resilient" adaptation for the region and had passed its salinity pilot (16 of 20 sites reached ≤2 mS/cm), the trial result removes the main scalable option from the table without replacing it — leaving agencies to keep installing rainwater tanks whose limited storage is exactly why people revert to ponds. The problem is worth attention because it is a design problem with a clear, measurable target rather than a knowledge gap: sodium, E. coli, dry-season continuity and cost per litre are all quantifiable.

What’s been tried and why it hasn’t worked

Rainwater harvesting is promoted and widely practised but "limited storage capacity requires them to revert to pond water during dry seasons." Pond-sand filters reduce turbidity and some microbes but do not remove sodium and are vulnerable to the same tidal inundation as the ponds. MAR was engineered against the wrong comparator: it was designed to bring water below the salinity of the brackish tubewells (and did, at 16 of 20 pilots), but the households it served were mostly drinking pond water, which is less saline than the recovered MAR water once infiltrated freshwater mixes with the native aquifer — so the intervention that improved microbiology worsened the sodium exposure of its adopters. Uptake was also partial (42% exclusive use; 60% still predominantly pond water), MAR systems need a dedicated caretaker, monthly sand-filter cleaning and pumping fuel or electricity, and the icddr,b team notes that the region's pond-sand filters "were successful for initial periods but broke down later and left behind by the communities," with MAR facing "similar management challenges." Reverse osmosis and solar desalination plants exist as more capital- and energy-intensive options that have not been shown to work at village scale on a sustained basis (a general characterisation, not from the two cited sources). Each option was optimised for one axis and evaluated on that axis alone.

What would unlock progress

Progress requires treating the household's dry-season drinking supply as a joint sodium–pathogen–continuity–cost optimisation rather than a single-hazard fix: for example, MAR operated as seasonal storage for excess rainwater (low sodium in, low sodium out) with recovery restricted to the freshwater lens, blended supply schemes that meter conductivity, or expanded community rainwater storage sized from actual dry-season demand. Two enabling tools are missing: a cheap conductivity/sodium indicator that households or caretakers can use to choose the lower-sodium source day to day, and a village-scale evaluation protocol that scores every candidate option on all four axes so that agencies stop scaling solutions that pass one test and fail another. Adjacent precedent: small-island states use similar rainwater–groundwater conjunctive schemes with lens-monitoring rules, and point-of-use water-quality indicators have precedent in arsenic field kits.

Entry points for student teams

A design team could model and prototype a rain-fed MAR variant (recharge only with harvested rainwater, recovery from a shallow freshwater lens with conductivity cut-off) and estimate the achievable dry-season sodium and volume for a typical village using published hydrogeology; a sensing team could build and field-test a low-cost conductivity indicator (colour-coded, sub-$10) for household source choice, validating it against a lab meter across pond, MAR and rainwater samples; an analytics team could construct the four-axis (sodium, E. coli, days-of-supply, cost per litre) scorecard for existing options using the published icddr,b data and identify the Pareto frontier and the biggest gap. Relevant skills: hydrogeology and water engineering, low-cost sensing, epidemiology and decision analysis, and participatory design with coastal communities.