Greenhouse gas concentrations are at record levels, fuelling temperature increase into the future. The emissions gap between aspiration and reality remains high. Under current policies, there is a two-thirds likelihood of global warming of 3 °C this century, according to the United in Science report. 

United in Science offers much-needed grounds for hope. It explores how advances in natural and social sciences, new technologies, and innovation enhance our understanding of the Earth’s system. It could be a game changer for climate change adaptation, disaster risk reduction, and sustainable development. 

“We need urgent and ambitious action to support sustainable development, climate action and disaster risk reduction. The decisions we make today could be the difference between a future breakdown or a breakthrough to a better world,” said WMO Secretary-General Celeste Saulo.  

“Artificial Intelligence and machine learning have emerged as potentially transformative technologies that are revolutionizing weather forecasting and can make it faster, cheaper and more accessible. Cutting-edge satellite technologies and virtual realities that bridge the physical and digital worlds are opening new frontiers in, for instance, land and water management,” said Celeste Saulo. 

“However, Science and technology alone are not enough to address global challenges such as climate change and sustainable development alone. In an increasingly complex world, we must embrace diverse knowledge, experiences and perspectives to co-create solutions together,” she said. 

The United Nations Summit of the Future provides a once-in-a-generation opportunity to revitalize and reboot our collective commitment to global goals, says the report, which was compiled by a consortium of United Nations agencies, meteorological organizations, and scientific and research bodies. It also embraces input from young people and early-career scientists who are agents of change for the future.  

Artificial intelligence and Machine Learning: revolutionizing weather forecasting 

Thanks to rapid progress, Artificial Intelligence (AI) and Machine Learning (ML) can make skilful weather modelling faster, cheaper and more accessible to lower-income countries with limited computational capacities. 

Traditionally, weather forecasting relies on physics-based models through a process known as numerical weather prediction. AI/ML models are trained on reanalysis and observational datasets, making weather forecasting faster and cheaper. Some evaluations have shown the potential of AI/ML for forecasting hazardous events such as tropical cyclones and longer-term predictions of El Niño and La Niña. 

There are tremendous opportunities but many challenges, as well as minimal data quality and availability. Current AI/ML models do not include harder-to-predict variables related to the ocean, land, cryosphere, and carbon cycle. 

Strong global governance is needed to ensure AI/ML serves the global good. Enhanced transparency will be necessary to build trust and develop standards for responsible use. 

For instance, engaging scientists, policymakers, practitioners and local and Indigenous communities from the outset enriches understanding of climate change impacts on the ground and offers a more complete perspective. 

It also strengthens trust in National Meteorological and Hydrological Services (NMHSs) institutions. 

World Meterological Organization