Co-Producing New Sub-Seasonal Weather Forecasts in Africa

By: Linda Hirons

Weather-related extremes affect the lives and livelihoods of millions of people across tropical Africa. Access to reliable, actionable weather information is key to improving the resilience of African populations and economies. Specifically, at the extended sub-seasonal timescale (forecasts of 1-4 weeks ahead), improved weather information could be transformational in building better early warning systems for the extreme events which cause infrastructural and societal damage. However, the uptake and availability of accurate weather information and services on these extended timescales remain very low across the continent.

Recent scientific advances have improved our understanding of what drives changes in weather on these timescales (e.g., the Madden Julian Oscillation (MJO); Zaitchik 2017) and subsequent modelling advances have enabled us to better represent these drivers (e.g., Vitart et al 2017) and their local impacts across Africa (e.g., de Andrade et al 2021). While these scientific and modelling advances are necessary to improve forecasts it is becoming increasingly clear that they are not sufficient to translate advances in knowledge into real tangible societal benefits. This requires a more collaborative and iterative approach where knowledge from scientists is combined with knowledge from local forecasters and knowledge from the specific decision-making context of forecast users to jointly co-produce (e.g., Vincent et al 2018) bespoke weather and climate services which can be truly effective.

Figure 1: The building blocks (a) and principles of good co-production (b) introduced in Carter et al. (2019) 

Through a Real-Time Pilot Initiative of the WMO Sub-seasonal to Seasonal Prediction Project, the GCRF African-SWIFT and ForPAc projects ran a two-year, sub-seasonal forecasting testbed (Hirons et al 2021) – a forum where prototype forecast products were co-produced and operationally trialled in real-time. Launched in November 2019 in Kenya, the testbed brought together national meteorological services, universities and forecast users from across tropical Africa, to use a co-production approach (Figure 1; Carter et al 2019) to improve the appropriate use of sub-seasonal forecasts. This testbed made real-time, sub-seasonal forecast data from the European Centre for Medium-Range Weather Forecasts (ECMWF) available to users in a range of sectors, including energy, health, agriculture, disaster risk reduction and food security across tropical Africa.

The sub-seasonal testbed has been providing co-produced, tailored forecast products and advisories to weather-sensitive sectors across Africa (Hirons et al 2021). Examples here from users in the energy sector in Kenya and the health sector across the Sahel exemplify the local application and benefits of new testbed forecast products.

In Kenya, sub-seasonal forecasts co-produced by the Kenya Meteorological Department and the Kenya Electricity Generating Company (KenGEN), which is responsible for supplying more than 70% of Kenya’s electricity, supported improved hydropower planning. Hydropower accounts for approximately 45% of KenGEN’s total supply and fills the gaps when other sources like solar or wind are unreliable. It uses fast-moving water to produce electricity so Kenya relies on key dams for sufficient water storage. Previously dam levels would have been systematically lowered before the start of the rainy season in anticipation of significant rainfall. However, if rains failed, drought could cause considerable interruptions to the power supply and increase reliance on diesel generators. Through the Testbed KenGEN has been incorporating the sub-seasonal rainfall information into their dam management decisions enabling them to maximise dam levels without overflowing and causing downstream flooding. During the Testbed Kenya has experienced uninterrupted power, even through periods of drought, and has eliminated emergency diesel generators from the national electricity grid entirely.

Figure 2: Example of the vigilance map for the emergence of meningitis outbreaks in Africa co-produced with GCRF African SWIFT project and WHO.

Across the Sahel GCRF African SWIFT researchers and forecast producers have been working closely with the World Health Organisation (WHO) to supply bespoke, multi-variable sub-seasonal forecast information for meningitis vigilance across 26 countries in the meningitis belt. It is well known that meningitis outbreaks are more likely in warm, dry conditions, particularly after dust events. Previously the observed environmental conditions were used to determine the likelihood of outbreaks. However, by combining forecasts of temperature, relative humidity and wind speed and direction with dust forecasts, the sub-seasonal testbed has extended the lead time of the existing vigilance maps by up to 2 weeks (Figure 2). Working closely with the WHO has shown that this information has huge implications for improving preparedness action and making timely life-saving interventions to prevent outbreaks.

The GCRF African SWIFT sub-seasonal testbed is coming to an end this year and the focus will be on ensuring that the knowledge co-produced through these collaborative partnerships can be institutionalised and become part of in-country standard operational procedure to ensure project-initiated services are sustained. However, continuing to provide these new services requires national meteorological agencies in Africa to continue to have access to sub-seasonal data in real-time. Surely these direct and tangible societal benefits are enough to convince data providers?


Carter, S., Steynor, A., Waagsaether, K., Vincent, K., Visman, E., 2019. Co-production of African weather and climate services. Manual, Cape Town: SouthSouthNorth.

de Andrade, F. M., Young, M. P., MacLeod, D., Hirons, L. C.Woolnough, S. J. and Black, E. (2021) Subseasonal precipitation prediction for Africa: forecast evaluation and sources of predictability. Weather and Forecasting, 36 (1). pp. 265-284. ISSN 0882-8156 doi:

Hirons L., Thompson, E., Dione, C., Indasi, V.S., et al. Using co-production to improve the appropriate use of sub-seasonal forecasts in Africa. Climate Services, 23. 100246. ISSN 2405-8807 (2021)

Vincent, K., Daly, M., Scannell, C., Leathes, B., 2018. What can climate services learn from theory and practice of co-production? Climate Services. 12, 48-58.

Vitart, F., Ardilouze, C., Bonet, A., Brookshaw, A., Chen, M., Codorean, C.,  2017. The sub‐seasonal to seasonal prediction (S2S) project database. Bull. Am. Meteorol. Soc. 98, 163–173

Zaitchik, B.F., 2017. Madden-Jullian Oscillation impacts on tropical African precipitation. Atmospheric Research.184, 88-102.

This entry was posted in Climate, Co-production, Energy meteorology, Forecasting Testbed, Madden-Julian Oscillation (MJO), Predictability, Renewable energy, Seasonal forecasting, subseasonal forecasting, Tropical convection, Weather forecasting and tagged , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *