Publications

Salinity intrusion has rendered groundwater undrinkable for over 35 million people in coastal Bangladesh, forcing women and girls to walk hours for fresh water during the dry season. We evaluate a BRAC-led programme that provided household rainwater harvesting systems — 2,000-litre tanks with ceramic filtration, installed on rooftops or with purpose-built catchment structures where rooftop installation was not feasible — to vulnerable households in Mongla Upazila, using a difference-in-differences design with household fixed effects. The intervention substantially improves water access, lowers collection burdens, and reduces water costs during the monsoon and post-monsoon period. However, fixed tank capacity is exhausted before the dry season ends, leaving a majority of treatment households still water-insecure outside the monsoon window. Rainwater harvesting is a viable but calendar-bounded adaptation strategy — the monsoon can be caught, but not stretched.

The Ganges-Brahmaputra-Meghna delta is a landscape in perpetual flux, and for the coastal populations of Bangladesh the unpredictable nature of land accretion (Char formation) is both an existential threat and an economic opportunity that traditional hydrodynamic forecasting has struggled to translate into land policy. We present a multi-decadal analysis of the Lower Meghna Estuary using 35 years of satellite-derived annual water-presence maps (1991-2025) to benchmark a Random Forest against a deep-learning model, a persistence forecast that assumes next year looks like this year, and a linear-trend forecast, tested on eight held-out years. Skill beyond the no-change baseline is real but small, significant by one statistical test but not by two others we also report. Applying the validated forecast at native 100-meter resolution, we project 52,712 hectares of gross new land by 2030, concentrated in the eastern estuary, where the predictive skill is also concentrated. Checked against what actually happened in 2020-2025, total predicted area comes within a tenth of the true total, but only one in three individually flagged plots actually became land: the totals are more dependable than the plot-level map, and we present the resulting zoning as a screening tool, not an allocation tool. We set out the resulting phased protocol as Stability-Led Infrastructure Development (SLID), with each tier reported alongside the rate at which it actually verified.