SWIFT has pioneered the co-production of bespoke forecast services through real-time testbeds

Co-production is the process of bringing together different knowledge sources, experiences and working practices to jointly develop new and combined knowledge for addressing societal problems of shared concern. GCRF African SWIFT has pioneered the use of co-production to develop weather forecasting tools by bringing together researchers, forecast producers, and forecast users to co-create real-time forecasts together.

SWIFT has enabled African weather services to implement co-production by organising collaborative projects and events bringing together researchers, forecast producers and forecast users. Throughout SWIFT’s lifetime, three forecasting “testbeds” (forums where prototype forecast products are co-produced and operationally trialled in real-time) were organised. Each testbed co-developed forecast products for a different timescale: two focused on nowcasting (very short term prediction of the weather for the next few hours) and one on sub-seasonal to seasonal (S2S) forecasting (forecast predictions beyond two weeks but less than a season).

Benefits of Co-Production

By facilitating co-production, SWIFT has advanced the use of new meteorological research in operational forecasting products. The products developed through SWIFT co-production efforts (building on previous work by other projects) have already had positive applications in sectors such as agriculture, energy and disaster management. The project has strengthened links between operational forecasting institutions and research organisations, making future collaborations easier and more sustainable. SWIFT’s application of a co-production approach has set an example for future projects.

We summarise here some of the notable outcomes of the co-production approach pioneered through SWIFT:

Supporting the operationalisation of new and improved services

Figure 1: Co-exploring forecast users’ needs; S2S testbed kickoff workshop

Products co-produced through SWIFT are already in operational use across Africa. For example, S2S forecasting products developed through the SWIFT S2S testbed and building on previous work by the Strengthening Climate Information Partnerships project (SCIPEA), are used to predict meningitis outbreaks and to support hydropower generation in Kenya. The S2S Testbed also produced services tailored for specific energy and agriculture companies, which highlighted the opportunity for national operational centres to make new services sustainable by generating income through co-developing bespoke products for the private sector. Through collaborations between several projects, nowcasting products developed through the nowcasting testbeds are used in disaster management in Kenya, and have been developed into a public app for delivering predictions of the short term development of storms, which is already available in Kenya. The Nigerian Meteorological Agency, NiMet, has used collaborations developed through SWIFT to deliver customised forecasts to the agricultural sector in Nigeria.

Investing in building common ground

SWIFT co-production has enabled knowledge-sharing between research and operational communities and has developed a shared awareness of possibilities around co-production. Initial testbed workshops strengthened decisionmakers understanding of key climate concepts, and confidence in appropriate use of forecasts, while also enabling forecasters and climate researchers to better appreciate the specific decision-making contexts they are seeking to support. In the S2S testbed, users appreciated having probabilities provided with forecasts, which forecast producers are now more aware of and more likely to respond to. Co-production was noted to have enhanced users’ and researchers’ appreciation of the routine challenges facing operational centres.

The knowledge sharing carried out through co-production also illustrated the need for further training. In the nowcasting testbed it was requested that training for users on climate concepts and forecast application should be deepened to encourage engagement with forecasts. This is particularly important for impact-based forecasting, where user inputs are vital to assessing potential impacts.

Sustainable climate services

Developments made through co-production are often long-lasting and sustainable.

  • By strengthening capacity in partner institutions through training in testbed events, SWIFT has enabled partners to implement further developments independently in future;
  • Long lasting partnerships between institutions have been formed throughout the co-production process;
  • Much of the user feedback on forecasts will continue beyond testbeds;
  • Due to Covid restrictions, the nowcasting testbed was carried out in four separate centres, which showed stakeholders that co-production could also take place in a similar, less resource intensive way in future.

Enabling user feedback and real-time verification of new and improved forecast products

Figure 2: Co-developing new products; S2S testbed hackathon

User feedback from co-production activities has been used to verify the effectiveness of forecasts and even in some cases has encouraged further research. In the nowcasting testbed, social media and user engagement was used to assess the quality of forecasts. The development of meningitis warnings within the S2S Testbed encouraged research into drivers of dry conditions across sub-Saharan Africa.

Strengthening in-country links between national operational centres and academic institutions to support sustainable improvement in forecasting capacities

Forecasters have been taught new forecasting techniques and learnt to access new datasets while participating in testbeds. Staff and students from national research centres travelled to operational centres to take part in the nowcasting testbed, in some instances resulting in a Memorandum of Understanding between the national operational centre and partnering national universities.

Challenges and the future of co-production

SWIFT has set an example for future co-production projects by documenting its work in a policy brief on co-production (DOI: https://doi.org/10.48785/100/99) . The brief can be found online, and outlines the lessons learned from SWIFT’s co-production work, including many of the challenges faced by the project.

The project would like to emphasise the need for further skills and infrastructure for African specialists to effectively carry out co-production and exploit the many benefits. In particular, consistent access to international data is required.

SWIFT has demonstrated that co-production is resource intensive. As forecast users have widely varying needs, it can be challenging to co-produce bespoke forecast services at scale. The project has shown that few staff in weather services are employed to focus on stakeholder engagement, which can cause challenges for co-production, for example when trying to source continuous user feedback in the S2S testbed.

Further developments required for appropriate uptake of weather forecasting tools have been illustrated by SWIFT. At KMD, products developed through SWIFT are used to provide early warnings of floods and droughts. SWIFT has noted that there remains work to be done in training decision makers to act on these warnings. By recognising this shortage, SWIFT has provided direction for future co-production efforts.

SWIFT has made significant progress in the advancement of co-production in African forecasting. Co-production through SWIFT has had many tangible successes, and has facilitated the production of customised forecasts for energy, agriculture, health and disaster management. The project has clearly documented the potential and challenges of co-production, and established pathways and protocols to support the normalisation of co-production in the development of user-relevant climate services.


Martin Parker, Linda Hirons, Emma Visman