Submitted as: model evaluation paper 02 Mar 2021

Submitted as: model evaluation paper | 02 Mar 2021

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

DecTree v1.0 – Chemistry speedup in reactive transport simulations: purely data-driven and physics-based surrogates

Marco De Lucia1 and Michael Kühn1,2 Marco De Lucia and Michael Kühn
  • 1GFZ German Research Centre for Geosciences, Telegrafenberg, 14473 Potsdam, Germany
  • 2University of Potsdam, Institute of Geosciences, Hydrogeology, Potsdam, Germany

Abstract. The computational costs associated with coupled reactive transport simulations are mostly due to the chemical subsystem: replacing it with a pre-trained statistical surrogate is a promising strategy to achieve decisive speedups at the price of small accuracy losses and thus to extend the scale of problems which can be handled. We introduce a hierarchical coupling scheme in which full physics, equation-based geochemical simulations are partially replaced by surrogates. Errors on mass balance resulting from multivariate surrogate predictions effectively assess the accuracy of multivariate regressions at runtime: inaccurate surrogate predictions are rejected and the more expensive equation-based simulations are run instead. Gradient boosting regressors such as xgboost, not requiring data standardization and being able to handle Tweedie distributions, proved to be a suitable emulator. Finally, we devise a surrogate approach based on geochemical knowledge, which overcomes the issue of robustness when encountering previously unseen data, and which can serve as basis for further development of hybrid physics-AI modelling.

Marco De Lucia and Michael Kühn

Status: open (until 27 Apr 2021)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on gmd-2020-445', Anonymous Referee #1, 08 Mar 2021 reply
  • RC2: 'Comment on gmd-2020-445', Glenn Hammond, 31 Mar 2021 reply
  • RC3: 'Comment on gmd-2020-445', P. Trinchero, 06 Apr 2021 reply

Marco De Lucia and Michael Kühn

Marco De Lucia and Michael Kühn


Total article views: 324 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
248 68 8 324 4 2
  • HTML: 248
  • PDF: 68
  • XML: 8
  • Total: 324
  • BibTeX: 4
  • EndNote: 2
Views and downloads (calculated since 02 Mar 2021)
Cumulative views and downloads (calculated since 02 Mar 2021)

Viewed (geographical distribution)

Total article views: 237 (including HTML, PDF, and XML) Thereof 237 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
Latest update: 19 Apr 2021
Short summary
DecTree evaluates a hierarchical coupling method for reactive transport simulations in which pre-trained surrogate models are used to speedup the geochemical sub-process, and equation-based full physics simulations are called only if the surrogate predictions are implausible. Furthermore, we devise and evaluate a decision-tree surrogate approach designed to inject domain knowledge the surrogate by defining engineered features based on law of mass action or stoichiometric reaction equations.