Machine learning strategies for Marine Cloud Brightening

Marine Cloud Brightening (MCB) could offer a feasible way to yield a cooling effect over the oceans.

This project aims at using existing data sets and machine learning techniques to develop techniques that can reliably distinguish the number of particles in clouds and their effects, and their susceptibility to MCB, which in due course could help inform optimal MCB strategies.

In particular, it will focus on tackling challenges uniquely related to aerosol-cloud interaction: (i) satellite image resolution; (ii) sparsity and noisy data; (iii) spatial and temporal correlation. It will also tackle challenges related to integrating knowledge of ship journeys.