Forecasting, thresholds, and triggers: Towards developing a Forecast-based Financing system for droughts in Mozambique

Abstract

To support Mozambique in the mitigation and management of droughts, the World Food Programme (WFP) is seeking to implement innovative approaches to protect people’s livelihood who face drought risk. The approach that has potential of closing the humanitarian funding gap is Forecast-based Financing (FbF). FbF enables anticipatory actions against droughts using seasonal forecasts, which are implemented to reduce impacts in the critical window between a forecast and an event. An important step for leveraging seasonal forecasting information to implement FbF is the development of an operational trigger system for drought anticipatory action. Our study describes WFP’s FbF system for droughts being tested in Mozambique using the ECMWF 7-month rainfall ensemble forecast. We showed that the forecasting information has skill to detect most of the drought events, measured by the Standardized Precipitation Index (SPI) with lead time varying from 6 to 1 month. We observed that information of different lead times produces a marginal improvement on the average skill of the predictions. After deriving probability values that can trigger anticipatory action, we found that the overall system yields a mean Hit Rate and False Alarm Ratio of 74 % and 59 %, respectively. However, there are opportunities for reducing the false alarm ratio and increasing the hit rate when focusing on key specific lead times. We believe that implementing FbF against droughts based on forecast information can help humanitarian organizations to reduce losses to livelihoods and cost of humanitarian operations while supporting communities in a more dignified manner.