The «Extrêmes-Statistiques-Impacts-Régionalisation» LSCE team (Philippe
Naveau, Davide Faranda), near Paris, and the Climate and statistical
mechanics” group (ENS, Lyon, Freddy Bouchet) will be strongly involved in
the PhD supervision.
Collaborations with P. Ailliot (Math dept, UBO, Brest) and P. Tandeo (IMT
Atlantique, Brest) will be also established.
Global climate models are generally assessed on average behaviours and typical
fluctuations, and biases are thus determined. The aim of this PhD will be to
set up a new framework for the evaluation of model biases for extreme events.
The goal is to develop a theoretical framework to combine the relative merits
of different models according to their ability to reproduce specific extremes:
heatwaves and extreme precipitation. Extreme value theory and statistical
physics will be at the heart of the analysis and will be coupled with machine
learning techniques dedicated to the aggregation of biases and uncertainties.
Once the conceptual framework, the data from the CMIP experiments will be
analysed in detail. The algorithms developed may be useful for other types of
The PhD is funded by a CNRS 80 PRIME grant for a three year period.