Impact-based Flood early warning for rural livelihoods in Uganda

Anticipatory actions are increasingly being taken before an extreme flood event to reduce the impacts on lives and livelihoods. Local contextualised information is required to support real-time local decisions on where and when to act and what anticipatory actions to take. This study defines an impact-based early warning trigger system that integrates flood forecasts with livelihood information, such as crop calendars, to target anticipatory actions better. We demonstrate the application of this trigger system using a flood case study from the Katakwi District in Uganda. First, we integrate information on the local crop cycles with the flood forecasts to define the impact-based trigger system. Second, we verify the impact-based system using historical flood impact information and then compare it with the existing hazard-based system in the context of humanitarian decisions. Study findings show that the impact-based trigger system has an improved probability of flood detection compared to the hazard-based system. The number of missed events are fewer in the impact-based system while the trigger dates are similar in both systems. In a humanitarian context, the two systems trigger anticipatory actions at the same time. However, the impact-based trigger system can be further investigated in a different context (e.g., for livelihood protection) to assess the value of the local information. The impact-based system could also provide a valuable tool to validate the existing hazard-based system, which builds more confidence in its use in informing anticipatory actions. The study findings should therefore open avenues for further dialogue on what the impact-based trigger system could mean within the broader Forecast-based Action landscape towards building the resilience of at-risk communities.