UKCEH scientists are contributing to a project to enhance early weather warning systems in Southern Africa. Dr Jawairia Ahmad tells us about a recent testbed in Zambia that is helping to refine forecasting techniques for extreme weather events.
The Early Warnings for Southern Africa (EWSA) project, part of the UK Met Office’s Weather and Climate Information Services (WISER) programme, aims to improve access to early weather warning systems for communities in South Africa, Zambia and Mozambique. As part of this, a testbed was recently completed in Lusaka, Zambia organized in collaboration with the Zambian Meteorological Department (ZMD).
Building on lessons learned from last year’s testbed, my colleague Chris Taylor and I took part in this year’s event to learn more about Southern Africa’s weather systems and evaluate and refine UKCEH’s nowcasting products. Nowcasting—a short-term forecasting technique that delivers real-time weather predictions—was a key focus of the testbed.
The event included participants from UK universities as well as different regional meteorological organizations. The goal was to share information regarding best practices in nowcasting and how to improve current forecasting capabilities. Impact-based forecasts were generated and shared with the community.
The participants were rotated between three teams: Nowcasting; Synoptic forecasting; and Evaluation. The nowcasting team predicted the short-term weather using satellite observations (eg, wind direction, precipitation, cloud top temperature etc) and issued nowcasts for the next two and six hours. The synoptic team developed a forecast for the next one to three days, while the evaluation team tested the accuracy of the nowcasts issues previously. This workflow created an effective feedback loop such that the quality of short-term and synoptic forecasts was improved based on continuous evaluation.
The generated nowcasts were shared with three pre-selected communities in Zambia, Mozambique, and South Africa, selected based on their vulnerability to extreme weather, in particular flooding. Different methods were used to inform the community members about the daily weather forecasts including a document highlighting the expected weather conditions, as well as text and voice messages in multiple languages.
This testbed provided a unique opportunity to test the utility of models and datasets produced by UKCEH for nowcasting in Africa. Our Nowcasting portal showcases different observed and modelled datasets that are used for nowcasting across the African continent. These range from lightning observed by a satellite, to rainfall estimations in real time.
Among them is an AI-based model that predicts the probability of convective cores at different time periods in the future. Convective cores represent deep convective clouds usually associated with high precipitation. Our model uses thermal infrared imagery from satellite and a simple deep learning algorithm to estimate the probability of a convective core being present at a resolution of approximately 3km. The model evaluation yielded encouraging results, and we are now working on further development to improve longer lead time nowcasts.
Our Nowcasting portal is designed to help regional forecasters access relevant weather information at different scales on the same user interface. It reduces the effort and time required to obtain the same information on different platforms and includes tools to build nowcasting maps in seconds. The different information layers can be overlaid, and textual information can also be added according to local nowcasting requirements. The portal is still evolving based on UKCEH’s partnerships with different regional meteorological agencies in Africa and their nowcasting requirements.
Strengthening early warning for all
Early warning systems are necessary for southern Africa where storms can leave devastation and cause loss of live and livelihood. Every year the region suffers socioeconomic loss due to extreme weather. Improving the quality of weather information and access to reliable weather forecasts can help in containing these annual losses.
Public trust in early warning systems is crucial. To be effective, these systems must be designed with direct input and feedback from at-risk communities. Only then can product developers understand what is needed to ensure forecasts are accessible and actionable. As extreme weather events become more frequent, collaborative efforts like EWSA are critical for building climate resilience across Southern Africa.
The WISER-EWSA project is led by the University of Leeds (Prof. Doug Parker) with support from UKCEH, World Meteorological Organization, National Institute of Meteorology (INAM) Mozambique, South Africa Weather Services, Finnish Meteorological Institute and Kulima Integrated Development Solutions. The Met Office’s WISER programme is funded by the UK government’s Foreign, Commonwealth and Development Office (FCDO).
Jawairia Ahmad