Data Assimilation, Predictability, Error Identification and Uncertainty Quantification in Geosciences
Co-organized as AS5.18/HS3.6/OS4.21
Convener: Olivier Talagrand | Co-conveners: Javier Amezcua, Natale Alberto Carrassi, Amos Lawless, Mu Mu, Wansuo Duan, Stéphane Vannitsem
| Fri, 12 Apr, 08:30–12:30
Room L2
| Attendance Fri, 12 Apr, 14:00–15:45
Hall X4

Many situations occur in Geosciences where one wants to obtain an accurate description of the present, past or future state of a particular system. Examples are prediction of weather and climate, assimilation of observations, or inversion of seismic signals for probing the interior of the planet. One important aspect is the identification of the errors affecting the various sources of information used in the estimation process, and the quantification of the ensuing uncertainty on the final estimate.

The session is devoted to the theoretical and numerical aspects of that broad class of problems. A large number of topics are dealt with in the various papers to be presented: algorithms for assimilation of observations, and associated mathematical aspects (particularly, but not only, in the context of the atmosphere and the ocean), predictability of geophysical flows, with stress on the impact of initial and model errors, inverse problems of different kinds, and also new aspects at the crossing between data assimilation and data-driven methods. Applications to specific physical problems are presented.

Solicited Speakers
Olivier Pannekoucke (Météo-France, Toulouse)
Manuel Pulido (University of Reading)