Machine learning seminar series - Spatiotemporal complexity and time-dependent networks in mid- to late Holocene simulations

Europe/London
11:30 BST

11:30 BST

Description

Host

Magdalena Alonso Balmaseda

Speaker

Ritabrata DuttaDr Annalisa Bracco received her bachelor in Physics at the University of Torino in 1996 and a PhD in geosciences at the University of Genova in 2000. 

She is currently a professor in climate dynamics and oceanography in the School of Earth and Atmospheric Sciences at the Georgia Institute of Technology.  Her research focuses on the transport of physical, chemical and biological material in the ocean, and on tropical climate teleconnections, often using tools from computer science and applied mathematics. 

In 2011 she received the American Meteorological Society’s Nicholas Fofonoff award for her work on ocean turbulence and mesoscale transport.

Abstract

In climate science regime transitions include abrupt changes in modes of climate variability and shifts in the connectivity of the whole system. While important, their identification remains challenging. In this talk we present a new framework to investigate regime transitions and connectivity patterns in spatiotemporal climate fields. This framework first quantifies local regime shifts by means of information entropy and then infers a weighted, direct and time-dependent network between entropy "domains", i.e. areas formed by grid points that are homogeneous in terms of their entropy. 

The spatiotemporal variability in sea surface temperature (SST) in two simulations of the last 6000 years is investigated with this new approach. The largest regional regime shifts emerge as abrupt transitions from low to high-frequency SST oscillations, or vice versa, in both simulations. Generally, rapid and sudden transitions in the degree of connectivity of the system are observed in both simulations but, in most cases, at different times, with few exceptions. We focus, finally, on the relation between ENSO and the Indian Ocean Dipole, looking in more detail at their evolution from the mid- to late Holocene.

Registration
Machine learning seminar series
The agenda of this meeting is empty