Day 1 - September 22 | Foundation Setting & AI Overview | Location MBZUAI
9:00 - 9:30
Registration
9:30 - 9:45
Welcome to MBZUAI
Tim Baldwin (MBZUAI)
9:45 - 10:00
Welcoming Remarks
Khalfan Al Matrooshi (UAE Presidential Court)
10:00 - 10:15
Aim for Scale Initiative Overview
Imara Salas (AIM for Scale) | How training fits within larger agricultural innovation goals
10:15 - 10:30
Goals for this week
Souhaib Ben Taieb (MBZUAI) | Week's objectives, learning tracks and joint exercises.
10:30 - 10:45
☕ Coffee Break
10:45 - 11:10
Climate Services for Agricultural Sector
Tufa Dinku (IRI Columbia) | Crop decision-making examples & forecast utility
11:10 - 11:30
AI Advances in Medium-range to S2S Forecasts
Katie Kowal (UChicago) | Recent advances in AI weather and S2S forecasts, opportunities and needs for benchmarking
11:30 - 12:00
NCM’s Climate Services: Bridging Science and Society
Dr. Koteswararao Kundeti (NCM) | Discussing the NCM’s services and the importance of AI from a weather/climate perspective
12:00 - 13:00
🍴 Lunch Break
13:00 - 13:30
AI Overview
Souhaib Ben Taieb | Overview of what AI is and how AI weather forecasting fits in
Technical Track
13:30 - 14:00
AI vs. NWP and Available Models
Pedram Hassanzadeh (UChicago) | Comparison of AI vs. NWP and overview of AI model internals
14:00 - 14:45
Demo 1: Setting Up Your AI Weather Lab
Tech team (UChicago/MBZUAI) | Hand out computers, check required libraries installation
14:45 - 15:00
☕ Coffee Break
15:00 - 15:30
Continue Demo 1: Setting Up Your AI Weather Lab
Tech team (UChicago/MBZUAI) | Hand out computers, check required libraries installation
Applied Track
13:30 - 14:45
Climate 101
Tufa Dinku | Foundational knowledge on weather, climate, and climate change concepts
14:45 - 15:00
☕ Coffee Break
15:00 - 15:30
Impact-based Forecasting
Erkin Isaev (WMO) | Impact-based forecasting introduction
15:30 - 17:00
Operational Example: Blending AI Models for Indian Monsoon Onset
Pedram Hassanzadeh / Amir jina (UChicago) / Niriksha Shetty (PxD) | AIFS/NCGM selection, blending & farmer messaging
Day 2 - September 23 | Hands-on Implementation & Benchmarking | Location NCM
Technical Track
9:00 - 10:15
Current Weather Models Used at NMHS
Country representatives, (template provided)
10:15 - 10:30
☕ Coffee Break
Applied Track
9:00 - 10:15
Existing agroclimate services
Country representatives, (templates provided)
10:15 - 10:30
☕ Coffee Break
10:30 - 12:00
NCM Operational Tour
Tour of NCM operational facilities
12:00 - 13:00
🍴 Lunch Break
13:00 - 13:30
Benchmarking as an ecosystem - motivations and ingredients for success
Katie Kowal | Summary of needs for benchmarking, common practices today, and key ingredients for operational benchmarking frameworks
Technical Track
13:30 - 14:00
Considerations for running AI models yourself
Adam Marchakitus (UChicago) | Some guidance on ways to run AI models
14:00 - 15:00
Demo 2: Running AI Weather Forecasts
Tech team (UChicago/MBZUAI) | Hands-on experience with running AI weather model locally
Applied Track
13:30 - 14:00
Compare Climate Services Practices
Group Discussion
14:00 - 14:30
From Models to Messages: Making Forecasts Work for Farmers
Niriksha Shetty (PxD) | Methods for converting model information into useful decision-making support
14:30 - 15:00
Institutional Landscape Mapping Exercise
Hands-on Exercise | Each country rep lists all the key players in current agricultural context & info they provide
15:00 - 15:15
☕ Coffee Break
15:15 - 16:00
Observational Data Quality and Verification
George Sserwadda (TAHMO) | Consider how the use of local observations and observational density can improve model skill
16:00 - 17:00
Demo 3: Metrics Matter - Benchmarking Platform Overview
Genevieve Flaspohler (Rhiza) | Talk followed by hands-on exercise with the AI weather model scorecard, comparing metrics for evaluation in an online benchmarking platform
19:00 - 21:00
Networking Dinner
Bushra by Buddha-Bar | Bus pickup time from hotel is 7PM
Day 3 - September 24 | Advanced Applications & Localization | Location MBZUAI
Technical Track
9:00 - 10:00
Demo 4: Putting Your Forecast to the Test
Tech team (UChicago/MBZUAI) | Perform a customized evaluation of the forecast generated in Demo 2, look at how local analysis can provide insights
10:00 - 10:30
Demo 5: Ground Truth Challenge - Validating Models with Your Data
Tech team (UChicago/MBZUAI) | If data provided beforehand in usable format, countries can benchmark some models for their regional use cases, otherwise can explore how models perform for trigger thresholds with preloaded data - high/low rainfall events
10:30 - 10:45
☕ Coffee Break
10:45 - 12:00
Demos 4/5 continued
Tech team (UChicago/MBZUAI)
Applied Track
9:00 - 9:15
Experience from Extension Systems
Tufa Dinku | Overview of extension systems in different countries
9:15 - 10:00
Discuss Extension Systems
Discussion on country extension systems
10:00 - 10:15
Communications Overview
Niriksha Shetty | Summarize communication practices
10:15 - 10:30
Brainstorm Communications Pathways
Draft recommendations for communication pathway support
10:30 - 10:45
☕ Coffee Break
10:45 - 11:15
Continue Brainstorm Communications Pathways
Draft recommendations for communication pathway support
11:15 - 12:00
Empirically Testing Communications
Amir Jina / Niriksha Shetty | A/B message testing in Indian monsoon onset
12:00 - 13:00
🍴 Lunch Break
13:00 - 14:00
Downscaling with Generative AI
Jonatan Mestres, Adah Jepkemoi Kibet (G42) | What is possible with NVIDIA technology for downscaling, consideration for downscaling model applications - discussion on considerations for high spatial resolution forecasts
14:00 - 14:45
Localization seminar
Rendani Mbuvha / Elinor KruseSeminar | on localization project AfriClimate AI collaboration with Google Deepmind and vision for future work
14:45 - 15:15
☕ Coffee Break and Group Photo
15:15 - 16:15
Bris Powered by Forecast-in-a-Box – a WMO Pilot Project For Malawi
Vegard Bønes (MET Norway) / Corentin Carton de Wiart (ECMWF) | ECMWF and MetNorway integrate the Bris model into Forecast-in-a-Box for the WMO Pilot project in Malawi, delivering a complete forecasting system that can generate high-resolution local forecasts.
Day 4 - September 25 | End-to-End System Design | Location NCM
9:00 - 10:30
Applied Track Shares Recommendations for Forecast info/comms Pathways
Applied track shares a list of recommendations on content needed for farmer decision-making support.
10:30 - 10:45
☕ Coffee Break
10:45 - 12:00
Technical Track Shares What is Feasible to Implement for Forecasts
Technical track shares highlights of what they trialled with AI models and list of feasible content to produce to support an early warning system
12:00 - 13:00
🍴 Lunch Break
13:00 - 14:00
CONOPS Development Workshop
Erkin Isaev (WMO) | What it means to establish a concept of operations followed by Q/A.
14:00 - 15:30
Joint Work on Mapping the CONOPS Landscape
Countries work as a team to draft a CONOPS based on template, prepare to present takeaways.
15:30 - 15:45
☕ Coffee Break
15:45 - 16:45
NCM Facilities Deeper Dive
NCM | Additional insights into NCM facilities
Day 5 - September 26 | Synthesis & Community Building | Location MBZUAI
9:00 - 10:00
Countries Refine E2E Presentations
Working within countries, teams refine their final presentations on opportunities where they could possibly integrate AI.
10:00 - 10:30
Final Presentations
15 minute presentations per country, 10 min. for questions
10:30 - 10:45
☕ Coffee Break
10:45 - 12:00
Final Presentations (Continued)
15 minute presentations per country, 10 min. for questions
12:00 - 13:00
🍴 Lunch Break
13:00 - 14:00
Building a Community of Practice and Forward Looking Activities
Last session focused on building a community of practice - discuss follow-on activities.
16:00 - 22:00
Cultural Event
Desert Safari | Bus pickup time from hotel is 4PM