Preprints
https://doi.org/10.5194/esd-2021-12
https://doi.org/10.5194/esd-2021-12

  08 Apr 2021

08 Apr 2021

Review status: a revised version of this preprint was accepted for the journal ESD and is expected to appear here in due course.

Is time a variable like the others in multivariate statistical downscaling and bias correction?

Yoann Robin1, and Mathieu Vrac2, Yoann Robin and Mathieu Vrac
  • 1Météo-France, 42 avenue Gaspard-Coriolis, 31057, Toulouse, France
  • 2Laboratoire des Sciences du Climat et de l’Environnement (LSCE-IPSL), CEA/CNRS/UVSQ, Université Paris-Saclay, Centre d’Etudes de Saclay, Orme des Merisiers, 91191 Gif-sur-Yvette, France
  • These authors contributed equally to this work.

Abstract. Bias correction and statistical downscaling are now regularly applied to climate simulations to make then more usable for impact models and studies. Over the last few years, various methods were developed to account for multivariate – inter-site or inter-variable – properties in addition to more usual univariate ones. Among such methods, temporal properties are either neglected or specifically accounted for, i.e., differently from the other properties. In this study, we propose a new multivariate approach called “Time Shifted Multivariate Bias Correction” (TSMBC), which targets to correct the temporal dependency in addition to the other marginal and multivariate aspects. TSMBC relies on considering the initial variables at various times (i.e., lags) as additional variables to correct. Hence, temporal dependencies (e.g., auto-correlations) to correct are viewed as inter-variable dependencies to be adjusted and an existing multivariate bias correction (MBC) method can then be used to answer this need. This approach is first applied and evaluated on synthetic data from a Vector Auto Regressive (VAR) process. In a second evaluation, we work in a “perfect model” context where a Regional Climate Model (RCM) plays the role of the (pseudo-) observations, and where its forcing Global Climate Model (GCM) is the model to be downscaled/bias corrected. For both evaluations, the results show a large reduction of the biases in the temporal properties, while inter-variable and spatial dependence structures are still correctly adjusted. However, increasing too much the number of lags to consider does not necessarily improve the temporal properties and a too strong increase in the number of dimensions of the dataset to correct can even imply some potential instability in the adjusted/downscaled results, calling for a reasoned use of this approach for large datasets.

Yoann Robin and Mathieu Vrac

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on esd-2021-12', Anonymous Referee #1, 18 May 2021
  • RC2: 'Comment on esd-2021-12', Anonymous Referee #2, 19 May 2021

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on esd-2021-12', Anonymous Referee #1, 18 May 2021
  • RC2: 'Comment on esd-2021-12', Anonymous Referee #2, 19 May 2021

Yoann Robin and Mathieu Vrac

Yoann Robin and Mathieu Vrac

Viewed

Total article views: 679 (including HTML, PDF, and XML)
HTML PDF XML Total Supplement BibTeX EndNote
528 134 17 679 40 12 13
  • HTML: 528
  • PDF: 134
  • XML: 17
  • Total: 679
  • Supplement: 40
  • BibTeX: 12
  • EndNote: 13
Views and downloads (calculated since 08 Apr 2021)
Cumulative views and downloads (calculated since 08 Apr 2021)

Viewed (geographical distribution)

Total article views: 601 (including HTML, PDF, and XML) Thereof 601 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 23 Oct 2021
Download
Short summary
We propose a new multivariate downscaling / bias correction approach called “Time Shifted Multivariate Bias Correction”, which aims to correct temporal dependencies in addition to inter-variable and spatial ones. Our method is evaluated in a “perfect model experiment” context where simulations are used as pseudo-observations. The results show a large reduction of the biases in the temporal properties, while inter-variable and spatial dependence structures are still correctly adjusted.
Altmetrics