Statistical modeling of climate extremes

Teacher(s)

Kiriliouk Anna;

Language

English

Content

The topics covered in the course are:

• Modelling maxima (annual, monthly, etc.) using the Generalized Extreme Value (GEV) distribution
• Modelling threshold exceedances using the Generalized Pareto (GP) distribution
• Seasonality and long-term non-stationarity
• Dependence between extremes and compound (climate) events
• Extremes of spatial data
• One or more of the following topics: spatial prediction (interpolation), forecasting of extremes, Bayesian hierarchical models

The examples and illustrations presented throughout the course will focus mainly on the modelling of climate data (i.e., time series of environmental data). Links will also be made with actuarial practice, showing how these extreme-value models can be useful for insurers.

Teaching methods

Seven lectures, complemented by two R tutorials.

Evaluation methods

Students are assessed on the basis of a compulsory project carried out in R using real-world data. The dataset and the main research questions will be selected by the student. The oral examination held during the examination period will cover both the presentation and discussion of this project and other aspects of the course content.
In accordance with Article 72 of the General Regulations for Studies and Examinations, the course instructor may recommend to the examination board that a student who has not submitted the required work by the specified deadline be refused registration for the examination.

Online resources

Slides on Moodle

Bibliography

• Coles (2001). An introduction to Statistical Modeling of Extreme Values. Springer.
• De Carvalho, Huser, Naveau & Reich (2026). Handbook on Statistics of Extremes. Chapman & Hall / CRC.
• Beirlant, Goegebeur, Segers & Teugels (2004). Statistics of Extremes: theory and applications. Wiley.

Faculty or entity

> LSBA