Moderator:
Francesco Granella, CMCC – SEME Division
Abstract
This webinar demonstrates approaches to applying machine learning (ML) techniques for modelling large-scale datasets and analyzing questions relative to climate change adaptation at different scales. The presentation is framed around two main case studies: the first application demonstrates the use of ML on household survey data for understanding drivers and patterns of air conditioning adoption and utilization in the residential sector at a global scale and deriving future projections at a sub-national spatial scale along different scenarios. The second application illustrates ongoing work applying ML on multi-spectral satellite imagery to model street-level urban vegetation density indicators to map and keep track of the spatio-temporal evolution of urban green space, a key indicator for sustainable cities and passive cooling potential. A general discussion on future avenues, opportunities and challenges for using ML methods and big data to assess and project different dimensions of impacts and adaptation at multiple scales concludes the presentation.