FN Archimer Export Format
PT J
TI Bounded nonlinear forecasts of partially observed geophysical systems with physics-constrained deep learning
BT 
AF Ouala, Said
   Brunton, Steven L.
   Chapron, Bertrand
   Pascual, Ananda
   Collard, Fabrice
   Gaultier, Lucile
   Fablet, Ronan
AS 1:1;2:2;3:3;4:4;5:5;6:5;7:1;
FF 1:;2:;3:PDG-ODE-LOPS-SIAM;4:;5:;6:;7:;
C1 IMT Atlantique; Lab-STICC, 29200 Brest, France
   University of Washington, USA
   Ifremer, Lops, 29200 Brest, France
   IMEDEA, UIB-CSIC, 07190 Esporles, Spain
   ODL, 29200 Brest, France
C2 IMT ATLANTIQUE, FRANCE
   UNIV WASHINGTON, USA
   IFREMER, FRANCE
   IMEDEA, SPAIN
   OCEANDATALAB, FRANCE
SI BREST
SE PDG-ODE-LOPS-SIAM
UM LOPS
IN WOS Ifremer UMR
   copubli-france
   copubli-europe
   copubli-int-hors-europe
IF 4
TC 5
UR https://archimer.ifremer.fr/doc/00815/92707/99013.pdf
LA English
DT Article
DE ;Partially-observed systems;Embedding;Boundedness;Deep learning;Neural ODE;Forecasting
AB The complexity of real-world geophysical systems is often compounded by the fact that the observed measurements depend on hidden variables. These latent variables include unresolved small scales and/or rapidly evolving processes, partially observed couplings, or forcings in coupled systems. This is the case in ocean-atmosphere dynamics, for which unknown interior dynamics can affect surface observations. The identification of computationally-relevant representations of such partially-observed and highly nonlinear systems is thus challenging and often limited to short-term forecast applications. Here, we investigate the physics-constrained learning of implicit dynamical embeddings, leveraging neural ordinary differential equation (NODE) representations. In particular, we restrict the NODE representation to linear-quadratic dynamics and enforce a global boundedness constraint, which promotes the generalization of the learned dynamics to arbitrary initial conditions. The proposed architecture is implemented within a deep learning framework, and its relevance is demonstrated with respect to state-of-the-art schemes for different case studies representative of geophysical dynamics. 
PY 2023
PD APR
SO Physica D-nonlinear Phenomena
SN 0167-2789
PU Elsevier BV
VL 446
UT 000995188800001
DI 10.1016/j.physd.2022.133630
ID 92707
ER

EF