Deep learning is transforming many scientific and engineering disciplines due to a convergence of large volumes of data, scalable algorithms, and new computer architectures. In the earth and space sciences, deep learning has been used to accelerate the identification of exoplanets, track cyclones and atmospheric rivers, and predict volcanic eruptions. With hundreds of petabytes of climate data on the horizon, climate science and weather prediction stand to benefit from deep learning algorithms that can detect complex, non-linear patterns in the climate system that are difficult, even impossible, for humans to identify and interpret. This session encourages contributions that apply data-driven and deep learning approaches to climate science and weather prediction. Submissions are welcome that use deep learning to improve our understanding of models and observations, and enhance the predictability of extreme events, including life-threatening hazards such as floods and wildfires.