loading page

Integrating recurrent neural networks with data assimilation for scalable data-driven state estimation
  • +4
  • Stephen Gregory Penny,
  • Timothy A Smith,
  • Tse-Chun Chen,
  • Jason Alexander Platt,
  • Hsin-Yi Lin,
  • Michael Goodliff,
  • Henry Don Isaac Abarbanel
Stephen Gregory Penny
Cooperative Institute for Research in Environmental Sciences at the University of Colorado Boulder

Corresponding Author:steve.penny@noaa.gov

Author Profile
Timothy A Smith
Cooperative Institute for Research in Environmental Sciences at the University of Colorado Boulder
Author Profile
Tse-Chun Chen
University of Maryland, College Park
Author Profile
Jason Alexander Platt
University of California San Diego
Author Profile
Hsin-Yi Lin
The Cooperative Institute for Research in Environmental Sciences (CIRES)
Author Profile
Michael Goodliff
University of Colorado
Author Profile
Henry Don Isaac Abarbanel
UC San Diego
Author Profile

Abstract

Data assimilation (DA) is integrated with machine learning in order to perform entirely data-driven online state estimation. To achieve this, recurrent neural networks (RNNs) are implemented as surrogate models to replace key components of the DA cycle in numerical weather prediction (NWP), including the conventional numerical forecast model, the forecast error covariance matrix, and the tangent linear and adjoint models. It is shown how these RNNs can be initialized using DA methods to directly update the hidden/reservoir state with observations of the target system. The results indicate that these techniques can be applied to estimate the state of a system for the repeated initialization of short-term forecasts, even in the absence of a traditional numerical forecast model. Further, it is demonstrated how these integrated RNN-DA methods can scale to higher dimensions by applying domain localization and parallelization, providing a path for practical applications in NWP.