From calibration to parameter learning: Harnessing the scaling effects of big data in geoscientific modeling

TL;DR

This paper introduces a differentiable parameter learning framework that leverages big data in geosciences, significantly improving calibration efficiency, physical coherence, and generalizability while reducing computational costs.

Contribution

The novel dPL framework enables scalable, data-efficient parameter learning in geoscientific models, outperforming traditional methods and integrating deep learning with process-based models.

Findings

dPL outperforms existing calibration methods in soil moisture and streamflow modeling

dPL requires only ~12.5% of training data to achieve similar performance

dPL exhibits better physical coherence and generalizability with larger datasets

Abstract

The behaviors and skills of models in many geosciences (e.g., hydrology and ecosystem sciences) strongly depend on spatially-varying parameters that need calibration. A well-calibrated model can reasonably propagate information from observations to unobserved variables via model physics, but traditional calibration is highly inefficient and results in non-unique solutions. Here we propose a novel differentiable parameter learning (dPL) framework that efficiently learns a global mapping between inputs (and optionally responses) and parameters. Crucially, dPL exhibits beneficial scaling curves not previously demonstrated to geoscientists: as training data increases, dPL achieves better performance, more physical coherence, and better generalizability (across space and uncalibrated variables), all with orders-of-magnitude lower computational cost. We demonstrate examples that learned from soil moisture and streamflow, where dPL drastically outperformed existing evolutionary and regionalization methods, or required only ~12.5% of the training data to achieve similar performance. The generic scheme promotes the integration of deep learning and process-based models, without mandating reimplementation.