With 35% of coastal areas at high to very high risk for landslides, the hilly regions of Bandarban and Chittagong put families at risk and cause immense loss of assets. Populations in remote settlements are particularly vulnerable to these sudden disasters because conventional early warning systems fail to reach communities. This failure occurs due to poor mobile coverage, power outages, heavy rains, and disrupted communication, leaving households unprepared.
To address this, we piloted a localized anticipatory action model with technical partner RIMES, local partners YPSA and ASHIKA, and support from ECHO. Coordinated through the Disaster Management Committee (DMC)-led Learning Lab, meteorological warnings were translated into Marma, Chak, and Tripura languages, simplified, and converted into clear, actionable steps/ action for communities in both Chattogram and Bandarban districts. These steps included reaching even the most remote households with warnings, coordinating volunteers and resource pools for safe evacuation, shifting people to safe locations, storing essential supplies, and protecting children. Activating communities was central to this approach. Interpreter pool members, youth volunteers, community leaders, women’s groups, and religious institutions helped deliver warnings effectively. A survey of 497 households showed that 97% received early warnings, and over 80% took anticipatory actions.
Our experience demonstrates that culturally and linguistically adapted warnings, youth and children as multipliers, community-led hubs for last-mile dissemination, and hybrid approaches combining technology with in-person outreach are essential for reaching remote communities. While challenges such as limited shelters and water access remain, this model proves that trusted local warnings can turn alerts into lifesaving action, protecting lives and assets even in the hardest-to-reach areas.
