Preprints
https://doi.org/10.5194/egusphere-2022-105
https://doi.org/10.5194/egusphere-2022-105
17 May 2022
 | 17 May 2022

Causal deep learning models for studying the Earth system: soil moisture-precipitation coupling in ERA5 data across Europe

Tobias Tesch, Stefan Kollet, and Jochen Garcke

Abstract. The Earth system is a complex non-linear dynamical system. Despite decades of research, many processes and relations between Earth system variables are still poorly understood. Current approaches for studying relations in the Earth system may be broadly divided into approaches based on numerical simulations and statistical approaches. However, there are several inherent limitations to current approaches that are, for example, high computational costs, reliance on the correct representation of relations in numerical models, strong assumptions related to linearity or locality, and the fallacy of correlation and causality.

Here, we propose a novel methodology combining deep learning (DL) and principles of causality research in an attempt to overcome these limitations. The methodology combines the recent idea of training and analyzing DL models to gain new scientific insights in the relations between input and target variables with a theorem from causality research. This theorem states that a statistical model may learn the causal impact of an input variable on a target variable if suitable additional input variables are included. As an illustrative example, we apply the methodology to study soil moisture-precipitation coupling in ERA5 climate reanalysis data across Europe. We demonstrate that, harnessing the great power and flexibility of DL models, the proposed methodology may yield new scientific insights into complex, nonlinear and non-local coupling mechanisms in the Earth system.

Journal article(s) based on this preprint

20 Apr 2023
| Highlight paper
Causal deep learning models for studying the Earth system
Tobias Tesch, Stefan Kollet, and Jochen Garcke
Geosci. Model Dev., 16, 2149–2166, https://doi.org/10.5194/gmd-16-2149-2023,https://doi.org/10.5194/gmd-16-2149-2023, 2023
Short summary Executive editor

Tobias Tesch et al.

Interactive discussion

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on egusphere-2022-105', Matthew Knepley, 12 Sep 2022
  • RC2: 'Comment on egusphere-2022-105', Anonymous Referee #2, 22 Sep 2022

Interactive discussion

Status: closed

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on egusphere-2022-105', Matthew Knepley, 12 Sep 2022
  • RC2: 'Comment on egusphere-2022-105', Anonymous Referee #2, 22 Sep 2022

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload
AR by Tobias Tesch on behalf of the Authors (30 Nov 2022)  Author's response   Author's tracked changes   Manuscript 
ED: Referee Nomination & Report Request started (28 Dec 2022) by Richard Mills
RR by Chaopeng Shen (22 Jan 2023)
ED: Publish as is (08 Feb 2023) by Richard Mills
AR by Tobias Tesch on behalf of the Authors (25 Mar 2023)  Manuscript 

Journal article(s) based on this preprint

20 Apr 2023
| Highlight paper
Causal deep learning models for studying the Earth system
Tobias Tesch, Stefan Kollet, and Jochen Garcke
Geosci. Model Dev., 16, 2149–2166, https://doi.org/10.5194/gmd-16-2149-2023,https://doi.org/10.5194/gmd-16-2149-2023, 2023
Short summary Executive editor

Tobias Tesch et al.

Model code and software

Causal deep learning models for studying the Earth system: soil moisture-precipitation coupling in ERA5 data across Europe - Software Code Tobias Tesch, Stefan Kollet, Jochen Garcke https://doi.org/10.5281/zenodo.6385040

Tobias Tesch et al.

Viewed

Total article views: 917 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
715 183 19 917 17 8
  • HTML: 715
  • PDF: 183
  • XML: 19
  • Total: 917
  • BibTeX: 17
  • EndNote: 8
Views and downloads (calculated since 17 May 2022)
Cumulative views and downloads (calculated since 17 May 2022)

Viewed (geographical distribution)

Total article views: 844 (including HTML, PDF, and XML) Thereof 844 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 20 Apr 2023
Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Many papers are currently being published applying deep learning to geoscientific applications. However, most of them only offer proof of concept results on highly idealised scenarios. This paper combines deep learning approaches with the structural causal models popularized by the work of Judea Pearl, and it applies this methodology to a real problem, analyzing soil moisture-precipitation coupling in climate reanalysis data. In contrast to many papers in this field, this promises actual insight in the scientific application of the work.
Short summary
A recent statistical approach for studying relations in the Earth system is to train deep learning (DL) models to predict Earth system variables given one or several others, and use interpretable DL to analyze the relations learned by the models. Here, we propose to combine the approach with a theorem from causality research to ensure that the deep learning model learns causal rather than spurious relations. As an example, we apply the method to study soil moisture-precipitation coupling.