AI
Theme
Introduction to Artificial Intelligence for Weather and Climate Modelling
📅
📍
Venue
AIMS, Kigali, Rwanda
🏛
Organiser
African Centre of Meteorological Applications for Development
About the Summer School
The African Summer School on Artificial Intelligence for Weather and Climate Modelling is an annual ACMAD capacity-development programme for African weather, climate, research and early-warning institutions.
Artificial intelligence and machine learning are rapidly emerging as complementary and, in some applications, surrogate approaches for weather forecasting, climate modelling, downscaling, bias correction, extreme-event detection and impact-based early warning.
These approaches offer important opportunities for Africa. They can complement numerical models, support forecast post-processing, improve local-scale information and help institutions develop practical applications where computing resources, specialised expertise and dense observational networks remain limited.
The 2026 Summer School will focus on building foundational knowledge and practical skills in artificial intelligence for weather and climate modelling. Participants will work with African datasets and explore how AI-enabled methods can be evaluated, adapted and responsibly integrated into institutional and operational workflows.
The programme contributes to the development of an African-led AI and climate-modelling community working closely with NMHSs, RCCs, universities, research institutions, relevant sectors and the global AI and climate community.
Overall purpose
To strengthen Africa’s capacity to apply artificial intelligence in weather and climate modelling in order to improve climate information services, early warning and disaster risk reduction.
Summer School objectives
The Summer School will help participants understand, evaluate and apply AI and machine-learning approaches in African weather and climate-service contexts.
1. Understand
Introduce the concepts and principles of artificial intelligence and machine learning for weather and climate modelling.
2. Connect
Explain how AI can complement numerical prediction systems, climate models, observations and expert forecasting.
3. Explore
Examine practical applications in forecasting, bias correction, downscaling, nowcasting and extreme-event monitoring.
4. Apply
Provide hands-on experience with AI and machine-learning workflows using African weather and climate datasets.
5. Evaluate
Build participants’ ability to assess model accuracy, bias, uncertainty, limitations and operational suitability.
6. Collaborate
Promote sustained collaboration among NMHSs, RCCs, universities, researchers and relevant climate-service sectors.
What participants will learn
By the end of the Summer School, participants should be able to:
- Explain the role, opportunities and limitations of AI in weather and climate services.
- Prepare weather and climate data for AI and machine-learning workflows.
- Work with CSV, NetCDF and GRIB data using basic Python tools.
- Describe applications in forecasting, bias correction, downscaling and early warning.
- Develop and evaluate a simple AI-based example relevant to African services.
- Prepare a practical institutional plan for applying the acquired knowledge.
How participants will learn
The Summer School combines expert-led lectures, keynote sessions, hands-on coding workshops, peer learning and group project work.
Expert lectures
Keynote sessions
Hands-on coding
African datasets
Group projects
Peer learning
Participants will work in teams to develop prototype solutions addressing real forecasting, climate-modelling, extreme-event or impact-based early-warning challenges.
Five-day programme
Select a day below to view its planned learning activities.
Day 1 — Foundations of AI for Weather and Climate
Morning session
- Inauguration and opening remarks
- Keynote on AI in weather and climate modelling
- Introduction to artificial intelligence and machine learning
- AI, machine learning and deep-learning fundamentals
Afternoon session
- Opportunities for African weather and climate services
- Limitations, uncertainty and responsible use
- How AI complements numerical models and expert forecasting
- Examples of operational and research applications
Day 2 — Weather and Climate Data for AI
Morning session
- Station observations
- Satellite products
- Reanalysis datasets
- Forecast datasets
- Climate-model outputs
- Data quality, gaps and representativeness
Afternoon session
- Introduction to CSV, NetCDF and GRIB
- Basic data exploration using Python
- Spatial and temporal subsetting
- Data cleaning and preprocessing
- Preparing predictors and target variables
Day 3 — Building a Machine-Learning Model
Morning session
- Defining a weather or climate prediction problem
- Selecting predictors and target variables
- Training, validation and testing datasets
- Model selection and baseline methods
- Building a simple machine-learning model
Afternoon group work
- Forecasting
- Bias correction
- Statistical downscaling
- Nowcasting
- Extreme-event monitoring
- Impact-based forecasting
Day 4 — Applied AI Workflows and Model Evaluation
Morning session
- AI workflow for temperature or rainfall prediction
- Feature preparation
- Model training and prediction
- Review and improvement of group prototypes
Afternoon session
- Accuracy and forecast skill
- Mean Absolute Error and Root Mean Square Error
- Forecast bias
- Precision, recall and F1 score
- Interpretation of results
- Uncertainty and model limitations
Day 5 — Projects, Institutional Plans and Certification
- Finalisation of group prototypes
- Group project presentations
- Peer and facilitator feedback
- Institutional application plans
- Participant reflections
- Post-training assessment
- Community-of-practice launch
- Course evaluation
- Certification and closing ceremony
Who should participate?
The Summer School will target approximately 30 trainees drawn from African institutions and programmes.
- National Meteorological and Hydrological Services
- Regional Climate Centres
- Universities and research institutions
- Climate and early-warning organisations
- Meteorology and climatology programmes
- Hydrology and environmental science programmes
- Computer science and data-science programmes
- Relevant climate-sensitive sectors
Selection will prioritise gender balance, regional representation, early-career professionals and applicants with a clear institutional role in weather, climate or early-warning services.
Admission criteria
- Affiliation with an NMHS, RCC, university, school, research institution or climate-related technical agency.
- Background in meteorology, climatology, hydrology, computer science, data science, geography or a related field.
- Basic knowledge of Python, or willingness to complete a pre-course Python module.
- Demonstrated interest in AI applications for weather, climate or early-warning services.
- Institutional support or a clear plan for applying the training after the Summer School.
- Commitment to participate in all sessions, practical exercises and group project activities.
Women and early-career professionals are strongly encouraged to apply.
Expected outputs and impact
30 participants trained
Improved foundational and practical capacity in AI-enabled weather and climate modelling.
Training resources archived
Curriculum, presentations, notebooks and practical materials made available through ACMAD eLearning.
AI prototypes developed
Participant teams develop practical examples relevant to African forecasting, climate modelling and early-warning services.
Community of practice launched
Post-training collaboration, mentorship and sharing of resources, methods and experiences.
Beyond the Summer School
The Summer School will not be treated as a one-off training event. ACMAD will establish a follow-up mechanism through an online community of practice, shared training resources, mentorship sessions and continued development of participant projects.
Lessons from the 2026 edition will inform future themes and curricula, contributing to a sustained continental learning platform on artificial intelligence for weather, climate and early-warning services.
