Abstract

Global climate models are complex tools that embody much of our understanding about the Earth system. Yet we know that they have their shortcomings and that different models are better for different purposes. Moreover, the complexity of these models often makes it hard to understand where efforts to improve the models should be targeted. In this talk, I will cover some of recent and ongoing projects where data science methods have been used, both to improve our understanding of complex model behaviour (with emulation) and to make more nuanced projections from multi-model ensembles (model weighting).

Coming from the atmospheric science domain, my talk will focus more on application rather than the detail of data science methods. However, I will also discuss my experience of interdisciplinary working, including insight into what works within the large Lancaster/UKCEH Data Science of the Natural Environment (DSNE) project.