The SAMA theme (Statistics for Analysis, Modeling and Assimilation) at IPSL aims at improving the data analysis of observations, the numerical outputs of climate models and their coupling. The final goal is to better represent the climate, geophysical fluids, their constituents and improve their forecast. These objectives can be achieved by leveraging recent mathematical and methodological developments. SAMA is historically grouped around 3 sub-themes: neural network (which has evolved into “Artificial Intelligence”), data assimilation, and statistics for climate and environmental sciences.
The relevant SAMA themes include:
- Analysis of massive, heterogeneous and complex data
- Quantification of uncertainty
- Non-linear processes
- Detection and attribution of climate change
- Extreme events
- Inverse problems
Cross-disciplinary approaches in terms of data science with SAMA foster methodological developments, in particular:
- Artificial intelligence: machine learning
- Combination of ensemble and variational methods for data assimilation
- Stochastic methods: Bayesian approaches, downscaling, stochastic weather generators, causality.
- SAMA organizes seminars (notably the bi-annual SAMA IA and Climate seminar https://ai4climate.lip6.fr/) and helps to define research projects relating to this theme.
- Master internships and visiting scholars are also promoted within SAMA.
Moderators of the theme
Adriana Coman
LISA-IPSL
Soulivanh Thao
LSCE-IPSL
Cécile Mallet
LATMOS-IPSL