The frequency of extreme weather events has increased significantly due to the intensifying climate change globally, impacting regions on a large scale and costing states billions of dollars annually. According to German reinsurer Munich Re, natural disasters caused $280 billion in damage in 2021, compared to $210 billion in 2020 and $166 billion in 2019.

In this context, many companies, States, and institutions are turning to technologies, particularly artificial intelligence, to model and predict climate change and its consequences on human activities. According to a BCG study published in September 2022, 87% of decision-makers believe that AI is essential in the fight against climate change. 

Here’s an explanation of how AI can support climate change!

A Digital Twin of the Earth at the European Level

This is particularly the case of the European Commission, which launched the “Destination Terre” initiative in March 2022. This digital twin of Earth is on a mission to help fight climate change and protect nature. The initiative receives initial support of €150 million under the Digital Europe program until mid-2024. It will help understand climate change and find solutions at all levels-local, regional and global.

As part of this initiative, two digital twins will be designed by the ECMWF (European Center for Medium-Range Weather Forecasts). The first will be devoted to risks induced by meteorological conditions and geophysical risks. It will focus on extreme weather events, such as heat waves, droughts, and floods, and phenomena like tsunamis, earthquakes, volcanic eruptions, etc. In the event of floods, for example, it will help local and regional authorities to take the right actions, saving lives and reducing material damage.

The second digital twin will be dedicated to adaptation to climate change. It will support mitigation plans related to climate change by providing stimulation and observation capabilities in support of climate change mitigation activities and scenarios. To help achieve carbon neutrality, information from different areas, such as sustainable agriculture, energy security, and biodiversity protection, will be made available.

United Nations Initiatives to Better Manage Floods, Forest Fires and Other Climate-Related Disasters

Another global initiative is that of the United Nations and its Satellite Center (UNOSAT), which initially aimed to manage climatic disasters using AI in partnership with Nvidia to gather and assess geospatial info, providing real-time information on wildfire, floods, and other climatic disasters.

In addition, an educational module launched by UNOSAT aims to generate accurate flood detection models through deep learning methods. This model is built on DLI (Nvidia Deep Learning Institute) course.

Can AI and Machine Learning Really Help Humanity to Face Climate Change?

A study entitled Tackling Climate Change with Machine Learning published in 2019, recalls that since the first prediction linked to global warming, made in 1896 by the Swedish chemist Svante August Arrhenius (who later estimated that fossil fuels could release enough CO2 to warming of the Earth by 5◦C), has become much more accurate.

The opportunities for machine learning to advance the state of the art in climate prediction are plentiful. The creation of petabytes of climate observation data through satellites, the generation of petabytes of simulated climate data through massive climate modeling projects, and ML methods becoming faster have turned the focus of climatologists towards ML techniques and creating exciting new applications, teaming up with computer scientists.

While climate models cannot predict specific dates of future climate events, they can predict long-term changes. For example, they can help assess the intensity and frequency of climatic events and enable municipalities, businesses, and individuals to make informed infrastructure decisions and disaster response plans. More and more businesses have broken into this field, including risQ, foldAI, Dynamhex, Tomorrow, Blue Sky Analytic, to name a few, to make AI-based climate forecasts more usable by translating them into risk scores for different human activities.

Final Thoughts!

More specific and local forecasts mean more actionable plans. Thus, AI and machine learning can be widely used against climate change. Several research groups are working on translating high-resolution climate forecasts into risk scenarios. There’s a need for combined efforts to remove barriers to scaling up established and emerging technologies and consider how they can help in the fight against climate change.