Background

As a regional partner of Accelerating Impacts of CGIAR Climate Research for Africa (AICCRA) in West Africa, AGRHYMET Regional Centre plays a leading role in climate services and agro-hydro-meteorological monitoring in the region. In recent years, AGRHYMET has committed to strengthening its forecasting capacities in line with World Meteorological Organization (WMO) Decision 9 (EC-72). This decision emphasizes the need for objective, operational, transparent, and scientifically rigorous seasonal forecasting methods. The guidelines provided by WMO in 2020 on operational practices for objective seasonal forecasting offer the necessary framework to guide regional centers and National Meteorological and Hydrological Services (NMHSs) in transitioning towards more standardized and automated forecasting systems.

To achieve these objectives, AGRHYMET has initiated efforts to improve the reproducibility of seasonal forecasts by automating key steps in the forecasting process. This transition is crucial to ensuring that the forecasting process becomes not only more consistent but also more accessible for future evaluations and improvements. 

Phase I of AICCRA has played a central role in this transition, and AGRHYMET has leveraged the project to develop and implement innovative forecasting tools. One of the key outcomes of these efforts is the development and deployment of PyCPT (Python Climate Predictability Tool) which automates the statistical methods used for seasonal forecasts. PyCPT is a powerful tool that enhances the traceability and reproducibility of forecasts, facilitating the evaluation and refinement of forecasting methodologies.

About the training

This training workshop is designed to address the technological capacity-building need by training participants from 17 West African countries on the full range of seasonal and sub-seasonal forecasting methodologies, and to explore the role of artificial intelligence as a way of enhancing accuracy and speed.

Thanks to PyCPT, seasonal forecasts no longer rely on manual adjustments and subjective consensus processes, leading to more consistent and scientifically defensible results. Beyond automation, AGRHYMET is also exploring new technological opportunities to further improve forecasting capabilities. The use of artificial intelligence (AI) and machine learning (ML) is currently being explored as a way to enhance the accuracy and speed of seasonal and sub-seasonal forecasts. AI-based approaches offer the ability to analyze large datasets, recognize complex patterns, and make forecasts that are more adaptive to changing climate conditions. These technologies promise to complement existing methods and provide more robust and reliable forecasts, especially in a region as complex and variable as West Africa and the Sahel.

AGRHYMET’s ongoing work includes a comprehensive assessment of these new methods in comparison to existing ones. This evaluation aims to measure the effectiveness of AI and machine learning tools, as well as traditional statistical methods, in generating accurate and actionable seasonal forecasts. Additionally, the possibility of combining these new methods with traditional ones is being explored to create consolidated forecasts that leverage the strengths of each approach. The goal is to produce forecasts that are not only more reliable but also more useful for end-users, such as farmers, water resource managers, and policymakers across the region.

In response to these advancements, AGRHYMET recognizes the need to strengthen the capacities of its partners across West Africa and the Sahel. Many countries in the region depend on seasonal forecasts to make critical decisions that affect their economies and societies, particularly in the areas of agriculture, disaster risk management, and water resource planning. It is essential that NMHSs and other relevant institutions are equipped with the knowledge and skills needed to use both traditional and new forecasting methods to achieve broader goals of climate resilience and sustainable development in the region. 

The workshop will provide participants with practical experience in using WAS_S2S tools along with tools like PyCPT and Xcast, while also introducing them to AI-based approaches.

Additionally, the training will focus on evaluating the different methodologies, helping participants understand the strengths and limitations of each approach and how they can be applied in their respective national contexts.

Workshop objectives

• Evaluate seasonal forecasts from global climate centers;
• Introduce and train participants on new methods and tools currently being explored for seasonal and sub-seasonal forecasting;
• Provide hands-on experience in the comprehensive evaluation of new methods to facilitate an understanding of their strengths and weaknesses in the context of forecasting for West Africa and the Sahel;
• Master the combination of the most effective methods to generate more consolidated and accurate seasonal forecasts for the region; and
• Equip participants with the skills necessary to implement these forecasting methodologies operationally within their institutions, thereby improving forecasting services and climate services across the region.