Preprints
https://doi.org/10.5194/hess-2024-297
https://doi.org/10.5194/hess-2024-297
02 Oct 2024
 | 02 Oct 2024
Status: this preprint is currently under review for the journal HESS.

Can causal discovery lead to a more robust prediction model for runoff signatures?

Hossein Abbasizadeh, Petr Maca, Martin Hanel, Mads Troldborg, and Amir AghaKouchak

Abstract. Runoff signatures characterize a catchment's response and provide insight into the hydrological processes. These signatures are governed by the co-evolution of catchment properties and climate processes, making them useful for understanding and explaining hydrological responses. However, catchment behaviours can vary significantly across different spatial scales, which complicates the identification of key drivers of hydrologic response. This study represents catchments as networks of variables linked by cause-and-effect relationships. We examine whether the direct causes of runoff signatures can explain these signatures across different environments, with the goal of developing more robust, parsimonious, and physically interpretable predictive models. We compare predictive models that incorporate causal information derived from the relationships between catchment, climate, and runoff characteristics. We use the Peter and Clarck (PC) causal discovery algorithm, along with three prediction models: Bayesian Network (BN), Generalized Additive Model (GAM), and Random Forest (RF). The results indicate that among models, BN exhibits the smallest decline in accuracy between training and test simulations compared to the other models. While RF achieves the highest overall performance, it also demonstrates the most significant drop in accuracy between the training and test phases. When the training sample is small, the accuracy of the causal RF model, which uses causal parents as predictors, is comparable to that of the non-causal RF model, which uses all selected variables as predictors. This study demonstrates the potential of causal inference techniques in representing the interconnected processes in hydrological systems in a more interpretable and effective manner.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.
Hossein Abbasizadeh, Petr Maca, Martin Hanel, Mads Troldborg, and Amir AghaKouchak

Status: open (until 27 Nov 2024)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
Hossein Abbasizadeh, Petr Maca, Martin Hanel, Mads Troldborg, and Amir AghaKouchak
Hossein Abbasizadeh, Petr Maca, Martin Hanel, Mads Troldborg, and Amir AghaKouchak

Viewed

Total article views: 173 (including HTML, PDF, and XML)
HTML PDF XML Total Supplement BibTeX EndNote
131 25 17 173 10 0 1
  • HTML: 131
  • PDF: 25
  • XML: 17
  • Total: 173
  • Supplement: 10
  • BibTeX: 0
  • EndNote: 1
Views and downloads (calculated since 02 Oct 2024)
Cumulative views and downloads (calculated since 02 Oct 2024)

Viewed (geographical distribution)

Total article views: 161 (including HTML, PDF, and XML) Thereof 161 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 13 Oct 2024
Download
Short summary
Here, we represented catchments as networks of variables connected by cause-and-effect relationships. By comparing the performance of statistical and machine learning methods with and without incorporating causal information to predict runoff properties, we showed that causal information can enhance models' robustness by reducing accuracy drop between training and testing phases, improving the model's interpretability, and mitigating overfitting issues, especially with small training samples.