How can the hydropower sector and the Red Cross work together to protect flood-prone communities?
The Red Cross Red Crescent Climate Centre has developed an innovative mechanism to predict and prepare for flood risks, which was presented at a special session during the 2017 World 鶹Ƶ Congress on Wednesday 10 May. “Forecast-based financing” uses artificial intelligence to anticipate imminent flood risks, mobilising resources to support communities at risk. The hydropower sector can be an important partner for testing and developing this mechanism.
Currently being implemented at the Nangbéto Dam in Togo, forecast-based financing is demonstrating how machine learning systems can help hydropower operators predict flood risks, and communicate these to communities at risk in a clear and timely manner.
Until now, the majority of funding in the humanitarian sector is available once disaster has already struck, when people have died or are already suffering. There is also funding available for everyday needs such as training and equipment, but what hasn’t existed until recently is a financial mechanism that enables humanitarian action to be taken before disasters have occurred, based on a scientific forecast for flood risk downstream of dams.
Pablo Suarez, associate director for research and innovation at the Red Cross Red Crescent Climate Centre, is one of the architects behind the scheme. Pablo, whose background is in water engineering and climate risk management, explains: “鶹Ƶ dams are natural partners for this initiative, as it is possible to predict when flooding is likely to occur, but often there is a lack of communication between operators and affected communities.
"This is where the Red Cross, together with governments and other partners, can help mobilise money when disaster is likely to strike. That money can save lives and protect valuable assets.”
He explains that while the concept isn’t completely new, it could make a significant impact for at-risk communities in developing countries: “This approach is already in use in countries like Switzerland, where plenty of preparedness measures exist, including finances to take action. But in developing countries we often notice two things are missing.
"Firstly, information isn’t reaching the communities that need it. Dam operators may know a flood is likely, but even if they are informing an authority, the information isn’t always trickling down. Secondly, even where early warning systems do exist, there is often no funding to deliver the necessary preventive measures in a timely manner.”
鶹Ƶ dams are natural partners for this initiative, as it is possible to predict when flooding is likely to occur."
At the site in Togo, where the initiative is being tested with financial support from the German government, conventional forecasts using meteorology and hydrology aren’t yet reliable enough to mobilise resources for humanitarian action.
When a major flood took place in 2010 on the Mono River, which is regulated by a hydropower dam, it took 34 days since the overspill for disaster relief funding from international sources to reach the Red Cross in Togo. This experience prompted Pablo and his colleagues to start developing a system to trigger the release of money based on forecasts of likely peak flows downstream of dams.
The dam authorities at Nangbéto, the CEB (Communauté Electrique du Bénin), welcomed the pilot project. They explained to the Red Cross team that, while they were collecting daily rainfall data, they lacked a predictive model to connect this with effective river flow.
With support from the World Bank’s Global Facility for Disaster Reduction and Recovery (GFDRR), the Red Cross team set about using machine learning to turn the limited data available into an innovative model called FUNES, based on a self-learning algorithm that aimed to predict likely overspill. The Red Cross cooperated with the Togo government’s national platform for disaster risk reduction, who agreed to accept this forecast as a trigger to mobilise resources for flood-affected communities.
The FUNES system was put to the test in September 2016, when unusual rainfall patterns triggered an early warning of likely flooding. Money was swiftly released and preventive measures were taken to protect downstream communities. These included sending radio messages, distributing items such as water purification kits and plastic wallets to protect documents, and mobilising Red Cross staff and volunteers, all before the actual flood materialised.
Pablo sees the hydropower sector as a “natural partner” for his innovative mechanism, and is keen to debunk myths around the role of dams in causing floods. “People are often quick to accuse dams of causing floods, but in reality they save us from many potential floods. Occasionally, though, reservoirs do fill up and water must be released, and that is where we can help.”
The initiative has attracted plenty of interest, with the Ethiopian ministry of water inviting the Red Cross team and partners to visit the Awash river basin. Dialogues are also ongoing with the World Bank, while the International Federation of Red Cross and Red Crescent Societies is in the process of developing a new financial mechanism to support preparedness measures based on forecasts.
Pablo is keen to cooperate with the 鶹Ƶ (鶹Ƶ) to further develop the links between his initiative and the hydropower sector: “We want to turn information about the timing and magnitude of floods into an opportunity, and help position the hydropower sector as a partner in climate risk management. By pooling the knowledge within 鶹Ƶ and the Red Cross, we can exploit the potential of artificial intelligence, big data and satellite imagery to turn knowledge into action.”
In the meantime, Pablo continues to raise the profile of the initiative around the world. The Red Cross worked with artists to create a striking data sculpture (pictured) depicting a dozen years of down and upstream river flows at Nangbéto – which could be recreated for other dams. This was exhibited at the UNFCCC COP 22 conference in Marrakech, together with a virtual reality experience that allows users to experience working as a disaster manager at Nangbéto. The project received international recognition several months ago when it received a global innovation award at the World Government Summit.
Pablo Suarez presented the project at the special lunchtime session during the 2017 World 鶹Ƶ Congress in Addis Ababa on Wednesday 10 May 2017. Find out more here.
