The Statistics of Climate Extreme

Date:  Tuesday 23rd November 2021. 

Time:

6:30pm (AEDT) Philippe Naveau - 30 minutes + short Q&A

7:05pm (AEDT) Jenny Wadsworth

- 30 minutes + short Q&A

7:35pm (AEDT) Questions/Discussions

Zoom link: https://qut.zoom.us/j/87051262671?pwd=ZndlT3hFbnJSTkNRT0l0enZEUW81QT09

Password: 494108

Philippe Naveau:

Title: Non-parametric multimodel Regional Frequency Analysis applied to climate change detection and attribution

joint work with Philomène Le Gall, Anne-Catherine Favre (Univ. Grenoble Alpes,  France) and Alexandre Tuel (University of Bern, Switzerland) 

Abstract: A recurrent question in climate risk analysis is determining how climate change will affect heavy precipitation patterns. Dividing the globe into homogeneous sub-regions should improve the modelling of heavy precipitation by inferring common regional distributional parameters. In addition, in the detection and attribution (D&A) field, biases due to model errors in global climate models (GCMs) should be considered to attribute the anthropogenic forcing effect.

Within this D&A context, we propose an efficient clustering algorithm that, compared to classical regional frequency analysis (RFA) techniques,  is covariate-free and accounts for dependence. It is based on a new non-parametric dissimilarity  that combines both the RFA constraint and the pairwise dependence.  We derive asymptotic properties of our dissimilarity estimator, and we interpret it for generalised extreme value distributed pairs.

As a D&A application, we cluster annual daily precipitation maxima of 16 GCMs from the coupled model intercomparison project.  We combine the climatologically consistent subregions identified for all GCMs.  This improves the spatial clusters coherence  and outperforms methods either based on margins or on dependence.  Finally, by comparing the natural forcings partition  with the  one with all forcings, we assess the impact of anthropogenic forcing on precipitation extreme patterns. 

Bio: Since 2004, Philippe Naveau has worked as a CNRS research scientist in the ESTIMR (“Extremes-Statistics-Impacts-Regionalization”) team of the Laboratoire des Sciences du Climat et de l’Environnement, near Paris. He has extensive experience in the statistical modelling of extremes in environmental and climate sciences.  Before 2004, he was assistant professor in the  Applied Math Department of Colorado University, USA. In 1998, he received a PhD in statistics from the department of statistics of Colorado State University, USA.

Jenny Wadsworth:

Title: Towards higher-dimensional spatial extremes

Abstract: The past decade has seen a huge effort in modelling the extremes of spatial processes. Significant challenges include the development of models with an appropriate asymptotic justification for the tail; ensuring model assumptions are compatible with the data; and the fitting of these models to (at least reasonably) high-dimensional datasets. I will review basic ideas of modelling spatial extremes, and introduce a more scalable approach via conditioning on a single site being extreme, which can also be applied in the space-time context. As we move towards being able to model extremes at more locations, we must also learn to deal with the complex structures of larger datasets, such as long-range independence and spatial non-stationarity in the extremal dependence. Time permitting, these issues will also be discussed.

Bio: Jenny is currently a Senior Lecturer in the Department of Mathematics and Statistics at Lancaster University. Her main research interest is extreme value theory, with a particular focus on multivariate and spatial problems, and applications in the environmental sciences.

Contact: Boris Beranger (b.beranger@unsw.edu.au), Kate Saunders (kate.r.saunders@qut.edu.au)