The Earth and humanity are facing an existential crisis related to climate and the overexploitation of the Earth’s resources. It is critical to better adapt and mitigate climate change. Better and more tailored future adaptation requires detailed climate projection going beyond projecting changes in global air temperature but requiring detailed information (e.g. floods, droughts, hurricane intensity and frequency, impact on crops and agriculture). Climate change projection is informed by improved monitoring using satellite imagery and modeling, using climate models. As such this is a very data rich domain, requiring not only to use state-of-the-art machine learning techniques but also to develop novel machine learning techniques (for instance to deal with uncertainties in the observations, the high dimensionality of the problem, disentangling cause and effects, respecting physical laws).
Climate change mitigation i.e. overturning the impact of climate change, also requires innovations such as on carbon capture and sequestration to accelerate the kinetics of chemical reactions or on the development of reduction of emissions across various sectors (automobile or aviation industry) or by developing renewable energies and integrating them onto the electrical grid. Machine learning is helping in the process too to optimize the mitigation process (e.g. grid optimization, optimization of chemical reactions).
Finally, Earth resources need to be better mined and optimized and machine learning is being used to optimize the process (e.g. water use of crops, mining extraction).
