Ensembles of Neural Surrogates for Parametric Sensitivity in Ocean Modeling

TL;DR

This paper introduces ensemble neural surrogates for ocean models that improve sensitivity analysis and uncertainty quantification, aiding better parameter tuning and decision making.

Contribution

It develops an ensemble-based approach to enhance the reliability of neural surrogates for ocean modeling sensitivities and uncertainty estimation.

Findings

Ensemble methods improve the accuracy of sensitivity estimates.

Uncertainty quantification enhances decision-making reliability.

Large-scale hyperparameter search optimizes surrogate performance.

Abstract

Accurate simulations of the oceans are crucial in understanding the Earth system. Despite their efficiency, simulations at lower resolutions must rely on various uncertain parameterizations to account for unresolved processes. However, model sensitivity to parameterizations is difficult to quantify, making it challenging to tune these parameterizations to reproduce observations. Deep learning surrogates have shown promise for efficient computation of the parametric sensitivities in the form of partial derivatives, but their reliability is difficult to evaluate without ground truth derivatives. In this work, we leverage large-scale hyperparameter search and ensemble learning to improve both forward predictions, autoregressive rollout, and backward adjoint sensitivity estimation. Particularly, the ensemble method provides epistemic uncertainty of function value predictions and their derivatives, providing improved reliability of the neural surrogates in decision making.