Report on fully coupled data assimilation in simplified systems with implications for Earth system reanalysis - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Rapport (Rapport De Recherche) Année : 2017

Report on fully coupled data assimilation in simplified systems with implications for Earth system reanalysis

Résumé

This report describes the Inria contribution to WP2 regarding a study of coupled data assimilation algorithms applied to academic models. This task was twofold: 1. Propose and study new coupled variational data assimilation algorithms. 2. Create a stand-alone coupled single column model (SCM) that mimics the ocean atmosphere behaviour and that can be used to validate the algorithms proposed in 1. Regarding subtask 1, a focus was made on ways to explicitly account for model coupling in the varia-tional optimisation problem, either as a strong contraint or as a weak constraint, or as a combination of both. It is then applied to both linear and non linear coupled problems, and leads to the conclusion that it does bring some noticeable benefit, but at a cost, both in time of development and computing time. The cost-benefit ratio has therefore to be studied for each given application. The coupled SCM has been developed in Fortran and interfaced with the OOPS data assimilation framework. It is documented in this report and a reference test case is given, so that it can be reused by partners of the project and beyond.
Fichier principal
Vignette du fichier
delivrable_D2.11.pdf (2.44 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01667509 , version 1 (19-12-2017)

Identifiants

  • HAL Id : hal-01667509 , version 1

Citer

Arthur Vidard, Rémi Pellerej, Florian Lemarié. Report on fully coupled data assimilation in simplified systems with implications for Earth system reanalysis. [Research Report] Inria Grenoble Rhône-Alpes. 2017. ⟨hal-01667509⟩
288 Consultations
119 Téléchargements

Partager

Gmail Facebook X LinkedIn More