Flood Early Warning Systems for Rural Populations in South Asia

Between 2011-20, floods caused over 45,000 deaths with most occurring in lower-income countries. Early warning systems (EWS) for floods can lower human and economic losses and improve post-flood recovery. But underdeveloped dissemination infrastructure in lower-income countries typically limits their reach and adoption by the most vulnerable: poor, less educated citizens in remote areas. Cost-effective infrastructure investments are also hindered by a lack of rigorous evidence on how best to relay flood alerts at scale. In collaboration with Google, researchers have been conducting a randomized evaluation of a flood EWS in 319 Bihari communities with more than 3.6 million people. It pairs cutting-edge forecasting and android-based alerting systems with incentivized grassroots volunteers trained in community outreach activities. Initial results show that treatment communities were 28 percent more likely to receive any flood alerts compared to control communities during the 2022 flood season. Furthermore, inland treatment households were 42.8 percent more likely to take protective actions before floods than inland control households. Building off of an earlier K-CAI-funded evaluation, researchers will extend their process monitoring and impact evaluation by three years and generate insights on the longer-term effectiveness of disaster warning systems in low-literacy rural contexts.