**Chaire Geolearning**I am the Principal Investigator of the Geolearning Chair, with a funding from 2023 to 2027 This Chair brings together two research groups: the Geostatistics team from the__Centre de Géosciences (CG) at Mines Paris__, and the__Biostatistics and Spatial Processes Unit (BioSP__), belonging to the MathNum division at INRAE. The general objective of the Geolearning Chair is to develop and apply methods originating from the recent revolution in data science to challenges induced by the climate and ecological transitions our world is facing. More specifically, we will design models and methods in geostatistics, extreme events theory and Machine Learning with applications to environmental, climate and risk sciences.Sébastien Coube-Sisqueille, currently post-doc at the Basque Center for Applied Mathematics visited me to work on multivariate spatio-temporal models in Bayesian context, with an application to air pollution.

Thomas Opitz is now the coordinator of the research network on risks, extreme events and spatio-temporal statitstics

**RESSTE (Risques, Extrêmes et****StatistiquesSpatio-Temporelles)**. RESSTE is funded by the Mathematics and digital technologies division (MathNum) of INRAE. It gathers more than 60 researches from about 20 research teams in France and abroad. We organize seminars and workshops and we support all sorts of actions actions in view of developping models and methods for analyzing space-time data. Feel free to contact me or Thomas if you wish to be on the RESSTE mailing list.- A
**RESSTE**team, with Denis Allard (BioSP, INRAE), Lucia Clarotto (Geostatistics, Mines ParisTech), Thomas Opitz (BioSP, INRAE) and Thomas Romary (Geostatistics, Mines ParisTech) won the**"****2021 KAUST Competition on Spatial Statistics for Large Datasets****"**organized by the Spatio-Temporel Statistics & Data Science group of KAUST. 22 teams from all over the world took part in this competition.

The competition was organized in 4 challenges consisting in estimating the parameters of several Gaussian fields and in making spatial predictions (kriging) under different conditions. The size of the datasets ranged from 90,000 data to 900,000 data. The RESSTE team came out on top of 3 exercises and placed themselves on the podium in the fourth exercise.

All details are available on the competition website at https://cemse.kaust.edu.sa/stsds/2021-kaust-competition-spatial-statistics-large-dataset

The

*Journal de la Société Française de Statistique*recently published a Special Issue Statistics for spatial and spatio-temporal data and RESSTE Network, containing very interesting contributions by several RESSTE network members.- The paper
, Math. Geosc. Volume 48(3): 305–328, doi: 10.1007/s11004-015-9594-x., co-authered with Rachid Senoussi from BioSP and Emilio Porcu from University of Newcastle was**Anisotropy Models for Spatial Data****awarded best 2016 paper by the Journal****Mathematical Geoscience**. The paper offers a full characterization of anisotropic variograms, in terms of both regularity and range. It is first shown that, if the regularity parameter is a continuous function of direction, it must necessarily be constant, whereas the scale parameter can vary in a continuous or discontinuous fashion with direction. As a second result, it is then established that all valid anisotropies for the range parameter can be represented as a directional mixture of zonal anisotropies. This representation makes it possible to build a very large class of anisotropic variograms, far more flexible than the classical anisotropies. A turning band algorithm for the simulation of Gaussian anisotropic random fields, derived from the mixture representation, is then presented and illustrated.

Recent conferences I (co-) organized and/or I have been involved with:

- METMA 2018 is the ninth of a series of workshops on the topic of Spatio-Temporal Modelling, which have been held every two years since 17 years. We are proud to organize the first "French" edition of this event in Montpellier. This workshop aims to promote the development and application of spatial, temporal, and mainly
**spatio-temporal statistical methods**in different fields related to the**environment**. It seeks to bring together practitioners and researchers of different areas and countries all over the world.The scientific program features sessions covering topics on the latest advancements in theory, methods and applications. - L’Université d’Avignon (UAVP) accueillera les Journées de Statistique 2017 du
**29 mai au 2 juin**sur le campus Hannah Arendt (anciennement Sainte-Marthe) au centre-ville d’Avignon. L'événement est co-organisé par le laboratoire de mathématique d'Avignon, BioSP, le laboratoire d'Informatique d'Avignon et par l'UMR ESPACE. - The 2015 edition of the Spatial Statistics Conference took place in Avignon, 9 - 12 June, 2015. It was co-chaired by Denis Allard (BioSP, INRA) and Alfred Stein (ITC). It was sponsored by the Applied Mathematics and Computer Science division of INRA.
- BioSP hosted the Workshop on Stochastic Weather Generators from 17 - 19 september, 2014. This workshop brought together a wide range of researchers, practitioners, and graduate students whose work is related to the stochastic modelling of meteorological variables and stochastic weather generators. Presentations can be found here.
- The 9th edition of the French-Danish Workshop took place in May 2012 in Avignon, France. It was jointly organized by the Biostatistics and Spatial Processes research unit (INRA) and the Dpt. of Mathematics, LANLG (University of Avignon). It was devoted to spatial statistics and image analysis and their applications in biology (agriculture, aquaculture, ecology, economy, environment, health, medicine, ...). Presentations can be found here.