Submitted by Andrea Ficchì, University of Reading
17 May 2021

The coproduction of the Early Action Protocol for floods in Uganda: A model of strong collaboration for the effective scale-up of FbF

To trigger effective anticipatory actions before a natural hazard hits a vulnerable community, humanitarians need clear and robust protocols. Such protocols, developed by the Red Cross and its partners, are based on quantitative analysis of the risk profile and the forecast skill, alongside an assessment of physical thresholds and triggers to act. All of these pieces of information are contained in an Early Action Protocol (EAP), alongside the Forecast-based Financing (FbF) capacities, responsibilities, and procedures to be followed to initiate appropriate actions.

Uganda was one of the nations who first pioneered FbF. The benefits of FbF in Uganda were demonstrated in 2015 when the Uganda Red Cross Society (URCS) triggered FbF actions in pilot communities in Kapelebyong based on a flood forecast.

More recently the URCS have made great steps to scale up the FbF by taking on a national-scale approach. The new Early Action Protocol for floods was submitted to the IFRC’s validation committee in September 2020, after several months of collaborative work to develop effective and sustainable forecast triggers. The EAP is now under final review.

In February 2020, the URCS and the Red Cross Red Crescent Climate Centre (RCCC) started fortnightly calls with several partners to facilitate a dialogue around the EAP development and the integration of inputs from different actors. The network of partners included practitioners and scientists from these national and international organisations: the Uganda’s Water Authority (DWRM), the Uganda National Meteorological Authority (UNMA), the Netherlands Red Cross Data & Digital team 510, and the University of Reading, through the UK-supported project Forecasts for Anticipatory Humanitarian Action (FATHUM). Evidence of flood forecast skill from scientists at the University of Reading and 510 was combined with local context knowledge and ground assessments from practitioners at the national hydro-meteorological services, DWRM and UNMA, to provide actionable information to URCS.

A partnership with national hydrometeorological services

The Uganda’s Ministry of Water and Environment, through the Directorate of Water Resources Management (DWRM), is responsible for all water resources monitoring, assessment, and management over the Country. Thus, their engagement with the development of the FbF protocol for floods was essential.

One way that national agencies like DWRM can seek to improve their flood forecasting capabilities is by utilising openly available global flood forecasting information, such as forecasts from the European Commission’s Global Flood Awareness System (GloFAS, a Copernicus EMS service), alongside in country knowledge. Where this is done in a partnership approach, the use of global models can help improve faster and effectively in-country capacity and develop skills that go beyond the specific model application. For example Douglas Mulangwa (hydrologist at DWRM) attended the University of Reading’s FATHUM research placement in 2019 where he received training in flood forecasting and on the use of the GloFAS system. Since returning to Uganda, Douglas has been investigating the benefits of using GloFAS compared to local hydrological models and has been working with the University of Reading on the evaluation of flood forecasts for Uganda.

This partnership has helped link GloFAS expertise from the University of Reading with highly valuable local knowledge and access to data. During the EAP development, DWRM made a great contribution in terms of providing and analysing river flow observations, as well as technical inputs for the evaluation of flood forecasts. DWRM also provided information about flood history and which catchments or areas were more likely to be hit by floods. This enabled the forecast skill of GloFAS to be compared with observations at 12 key points where river gauges from the national agency are operating and enough river flow data were available.

Academia and humanitarians

The forecast skill analysis focused mainly on the estimation of the probability of acting in vain, by calculating standard metrics called False Alarm Ratios. We analysed the hindcasts (historical forecasts) from GloFAS over the most recent 21 years to see how these scores evolve with the forecast lead time (i.e. number of days ahead of a predicted event) and trigger probability. This helped URCS identify the appropriate lead time to trigger actions and the locations where FbF could be scaled up in a sustainable and effective way.

“Through the skill analysis done by the FATHUM team, URCS are able to identify at which locations and at what trigger probabilities and false alarm ratios GloFAS forecasts should be used to take anticipatory actions.”

Emmanuel Ntale, an Early Warning Early Action Officer from the Uganda Red Cross Society was among this group of practitioners to attend the regular meetings, discussing flood forecast skill, FbF triggers, flood impacts and vulnerabilities.

Emmanuel reported that “the University of Reading through FATHUM has been very instrumental in supporting URCS and partners to define the trigger thresholds.” He adds, “Through the skill analysis done by the FATHUM team, URCS are now able to determine the GloFAS stations in Uganda with a good skill. This has enabled URCS to identify which locations and at what (trigger) percentages and false alarm ratios forecasts should be used to take anticipatory actions”.

To support operational implementation of the EAP, 510 is developing a dashboard that combines streams of data from GloFAS with exposure and vulnerability indicators. This dashboard helps URCS to determine when the protocol should be activated and provides a wealth of information that supports the implementation of the early actions.

Coproduction benefits everyone


“During the evaluation of flood forecasts from GloFAS, I learnt that communication and response to hydro-meteorological emergencies is always going to be multisectoral and will need a concerted involvement of all stakeholders from all walks of life.”

Douglas Mulangwa Hydrologist at DWRM Uganda

“This type of partnership work with local hydro-met agencies can benefit academics and global forecasting researchers by providing local data and knowledge to conduct well informed, locally relevant skill analysis”.

Linda Speight Postdoctoral researcher, FATHUM team, University of Reading

As Douglas noted, a lesson learnt from this EAP work is that there always needs to be a starting point, however small the start is, as he experienced during his summer placement in Reading in 2019. This lesson could be transferred to other developing countries that lack established local flood forecasts: interim solutions can be found using openly available global hydro-meteorological forecasts if these can help complement local resources and forecasts by the local mandated agencies. Any analysis needs to be co-produced to be usable and useful at the local level, considering all constraints and the relevant context.

The co-production is key for all partners and is a way to support the local national hydrometeorological services to create their own alternative local forecasting tools beyond any specific global system. Local longer-term solutions will benefit from the learnings achieved with any global interim solution. So, there is an urgent need to build such strong partnerships between government agencies, international forecasting centres, academia, and private entities engaged with flood risk management.


This blog was written by: 

  • Andrea Ficchì, Linda Speight, Hannah Cloke and Liz Stephens (University of Reading, UoR)
  • Douglas Mulangwa (Uganda’s Directorate of Water Resources Management, DWRM)
  • Irene Amuron (RCCC)
  • Bouke Pieter Ottow (510)
  • Emmanuel Ntale (Uganda Red Cross Society, URCS)

For more information on this work please get in touch with the lead author Andrea Ficchì and Liz Stephens at the University of Reading.