Preprints
https://doi.org/10.5194/gmd-2021-221
https://doi.org/10.5194/gmd-2021-221

Submitted as: development and technical paper 17 Aug 2021

Submitted as: development and technical paper | 17 Aug 2021

Review status: this preprint is currently under review for the journal GMD.

Training a supermodel with noisy and sparse observations: a case study with CPT and the synch rule on SPEEDO - v.1

Francine Janneke Schevenhoven1,2 and Alberto Carrassi3,4 Francine Janneke Schevenhoven and Alberto Carrassi
  • 1Geophysical Institute, University of Bergen, Bergen, Norway
  • 2Bjerknes Centre for Climate Research, Bergen, Norway
  • 3Dept. of Meteorology and NCEO, University of Reading, United Kingdom
  • 4Mathematical Institute, Utrecht University, Utrecht, the Netherlands

Abstract. In alternative to using the standard multi-model ensemble (MME) approach to combine the output of different models to improve prediction skill, models can also be combined dynamically to form a so-called supermodel. The supermodel approach allows for a quicker correction of the model errors. In this study we focus on weighted supermodels, in which the supermodel state is a weighted superposition of different imperfect model states. The estimation, “the training”, of the optimal weights of this combination is a critical aspect in the construction of a supermodel. In our previous works two algorithms were developed: (i) cross pollination in time (CPT-based technique), and, (ii) a synchronization based learning rule (synch rule). Those algorithms have been so far applied under the assumption of complete and noise-free observations. Here we go beyond and consider the more realistic case of noisy data that do not cover the full system's state and are not taken at each model's computational time step. We revise the training methods to cope with this observational scenario, while still being able to estimate accurate weights. In the synch rule an additional term is introduced to maintain physical balances, while in CPT nudging terms are added to let the models stay closer to the observations during training. Furthermore, we propose a novel formulation of the CPT method allowing for the weights to be negative. This makes it possible for CPT to deal with cases in which the individual model biases have the same sign, a situation that hampers constructing a skilful weighted supermodel based on positive weights. With these developments, both CPT and the synch rule have been made suitable to train a supermodel consisting of state-of-the-art weather or climate models.

Francine Janneke Schevenhoven and Alberto Carrassi

Status: open (until 02 Nov 2021)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on gmd-2021-221', Anonymous Referee #1, 17 Sep 2021 reply

Francine Janneke Schevenhoven and Alberto Carrassi

Model code and software

Supermodel training: CPT and the synch rule on SPEEDO Francine Schevenhoven https://doi.org/10.5281/zenodo.5034370

Francine Janneke Schevenhoven and Alberto Carrassi

Viewed

Total article views: 218 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
182 32 4 218 1 0
  • HTML: 182
  • PDF: 32
  • XML: 4
  • Total: 218
  • BibTeX: 1
  • EndNote: 0
Views and downloads (calculated since 17 Aug 2021)
Cumulative views and downloads (calculated since 17 Aug 2021)

Viewed (geographical distribution)

Total article views: 185 (including HTML, PDF, and XML) Thereof 185 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 28 Sep 2021
Download
Short summary
In this study, we present a novel formulation to build a dynamical combination of models, the so-called supermodel. Supermodels need to be trained based on data. Previously, we assumed complete and noise-free observations. Here, we move towards a realistic scenario and develop adaptations to the training methods in order to cope with sparse and noisy observations. Results are very promising and shed light on how to apply the method with state-of-the art GCMs.