Preprints
https://doi.org/10.5194/gmd-2021-205
https://doi.org/10.5194/gmd-2021-205

Submitted as: methods for assessment of models 12 Aug 2021

Submitted as: methods for assessment of models | 12 Aug 2021

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

Emulation of high-resolution land surface models using sparse Gaussian processes with application to JULES

Evan Baker1, Anna Harper2, Daniel Williamson2, and Peter Challenor2 Evan Baker et al.
  • 1Living Systems Institute, University of Exeter, Exeter EX4 4QD, UK
  • 2Department of Mathematical Sciences, University of Exeter, Exeter EX4 4QF, UK

Abstract. Land surface models are typically integrated into global climate projections, but as their spatial resolution increases the prospect of using them to aid in local policy decisions becomes more appealing. If these complex models are to be used to make local decisions, then a full quantification of uncertainty is necessary, but the computational cost of running just one simulation at high resolution can hinder proper analysis.

Statistical emulation is an increasingly common technique for developing fast approximate models in a way that maintains accuracy but also provides comprehensive uncertainty bounds for the approximation. In this work, we develop a statistical emulation framework for land surface models which acknowledges the forcing data fed into the model, providing predictions at a high resolution. We use The Joint UK Land Environment Simulator (JULES) as a case study for this strategy, and perform initial sensitivity analysis and parameter tuning to showcase its capabilities. JULES is perhaps one of the most complex land surface models, and so our success here suggests incredible gains can be made for all types of land surface model.

Evan Baker et al.

Status: open (until 03 Nov 2021)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CEC1: 'Comment on gmd-2021-205', Astrid Kerkweg, 13 Aug 2021 reply
  • RC1: 'Comment on gmd-2021-205', Anonymous Referee #1, 15 Sep 2021 reply

Evan Baker et al.

Viewed

Total article views: 308 (including HTML, PDF, and XML)
HTML PDF XML Total Supplement BibTeX EndNote
241 61 6 308 14 1 1
  • HTML: 241
  • PDF: 61
  • XML: 6
  • Total: 308
  • Supplement: 14
  • BibTeX: 1
  • EndNote: 1
Views and downloads (calculated since 12 Aug 2021)
Cumulative views and downloads (calculated since 12 Aug 2021)

Viewed (geographical distribution)

Total article views: 261 (including HTML, PDF, and XML) Thereof 261 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 28 Sep 2021
Download
Short summary
We have adapted cutting-edge machine learning techniques to build a model for the land surface (vegetation) in Great Britain. This model was trained using a very complex land-surface model called JULES, but our model is much faster at producing simulations and predictions. With this speed, our model can be used to investigate many different scenarios, which can be used to improve our understanding of the climate and could also be used to help make local decisions.