By Ahmadul Hassan, PhD
Advances in predictive analytics, using data and scientific models, allow the humanitarian community to mount a response before a predictable shock manifests into humanitarian needs. Taking such an anticipatory approach where possible, will lead to a response that’s faster, more efficient, and more dignified.
In late June 2020, the Resident Coordinator in Bangladesh and the Emergency Relief Coordinator pre-approved and endorsed an anticipatory action framework and the corresponding Central Emergency Response Fund (CERF) projects. Building on existing structures and experiences by the IFRC, WFP, and the Government, the pilot was set up within two months.
The framework pre-established when and on what basis financing and action will be triggered ahead of a specific monsoon flooding peak; how much funding would go to which agency; and what activities the funding is being used for in what timeframe.
On 4 July 2020, severe floods were forecasted for 18 July onwards, the framework triggered, and CERF funding was released to agencies, enabling humanitarian assistance to reach people before peak flood. With everything pre-agreed, CERF launched the fund’s fastest-ever allocation within 4 hours.
By the time the water reached life-threatening levels, more than 220,000 people had already received assistance through WFP, FAO, and UNFPA, working together with the IFRC, 10 local NGOs, and the Government.
An integral part of the pilot was to document evidence and learning. This report is a contribution to that learning process on evaluating the forecast and triggers to contribute to learning.
Forecasts at different lead times tend to predict different variables (e.g., seasonal rainfall vs 3-day rainfall totals), and it is critical to understand the relationship between the forecast itself and the hazard of interest (e.g., floods).
A trigger should be based on the action the humanitarian community can take in advance of a disaster that mitigates the impact of the hazard. Based on this action, responders can determine both (1) which forecasted impact is the most relevant level to intervene, and (2) what probability of this impact should be used to trigger action.
To determine this impact-based forecast, five steps are taken, as indicated in Figure 1. The first two steps are to create a risk analysis and an inventory of forecasts. The third step defines hazard magnitudes. The fourth step creates an impact-based forecast by either combining the risk analysis with the forecasts using expert knowledge or using elementary or more advanced modeling and historical impact data; the fourth generates an impact-based forecasting intervention map.
- UN Office for the Coordination of Humanitarian Affairs
- To learn more about OCHA's activities, please visit https://www.unocha.org/.