The complex and multi-scale coupling of aerosols and clouds combined with the large volume of data generated with each climate model simulations poses a significant interpretation challenge, even for the most experienced of climate scientists. This ESR will make use of an innovative new network analysis technique, δ-MAPS, to determine the regions that behave similarly in climate simulations, and the intensity and time lag with which they are linked. This approach reduces the highly complex and multidimensional climate simulation fields into vastly simplified dynamic net- work representation. The networks of each simulation and parameter (aerosol, cloud) capture the essential dynamics of these parameters in the evolv- ing climate, and can be used to understand how aerosol and cloud fields interact and their local impacts propagate through the climate system. Shifts in each network structure with model, aerosol-cloud process parameterization, simulation characteristics, emissions scenarios can be robustly detected and evaluated against networks derived from other models and observations. A variety of network metrics can be used to quantify similarity between networks, and understand differences the result from internal variability versus structural differences in climate. The largely reduced representation of the climate dynamics allows the networks themselves to be used as reduced-form models to represent the patterns of climate change and help determine causal links.
Athanasios Nenes (supervisor), Jakob Runge (co-supervisor), Alberto Arribas (MetOffice, non-academic advisor)
3 months at DLR & 3 months at Met Office Informatics Lab
Enrolment in Doctoral degree
PhD in Environmental Engineering at the Ecole Polytechnique Federale de Lausanne