Seminar - Modelling and estimation of extreme events observed in space and time

08/26/2013 - 2:00pm
Mesa Laboratory - Main Seminar Room

Prof. Dr. Claudia Klüppelberg, Center for Mathematical Sciences, Munich University of Technology

Link to Recording

Often, in modelling meteorological data like precipitation and wind fields, statistical methodology can be applied to reconcile the physical models with the data. For an adequate risk analysis, the statistical modelling of extreme events, such as heavy rain and floods, and severe wind gusts or storms is essential. A natural extension from uni- and multivariate extreme value theory is formed by so-called max-stable random fields. We suggest new statistical models for extreme data measured in space and time. We present the basic aspects and challenges of simulation and parametric and non-parametric estimation of max-stable spatio-temporal random fields. Finally we apply our model to heavy rainfall data in Florida.

This is joint work with Richard Davis and Christina Steinkohl.


[1] S.A. Padoan and M. Ribatet and S.A. Sisson (2009). Likelihood-based inference for max-stable processes. JASA 105, 263-277.

[2] M. Schlather (2002). Models for stationary max-stable random fields. Extremes 5(1), 33-44.

[3] R. L. Smith (1990). Max-stable processes and spatial extremes. Unpublished manuscript.

[4] R.A. Davis, C. Klüppelberg and C. Steinkohl (2011). Max-stable processes for modelling extremes observed in space and time. J. Korean Statistical Society. Online at

[5] R.A. Davis, C. Klüppelberg and C. Steinkohl (2013). Statistical inference for max-stable processes in space and time. J. Royal Stat. Soc., Series B. First published online: 20 March 2013, DOI: 10.1111/rssb.12012.

[6] R.A. Davis, C. Klüppelberg and C. Steinkohl (2013). Semiparametric stimation for max-stable space-time processes. In preparation.