Response Framework: Tidal analysis and prediction through physics-informed ML

Thomas Monahan, Tianning Tang, Stephen Roberts, Thomas A. A. Adcock

Abstract

<title>Abstract</title> Worldwide, tidal flows pose significant operational and engineering challenges to numerous industries and are critical drivers of many biological, chemical, and geophysical processes. Accurate tidal predictions are important for modeling these phenomena at both global and regional scales. Conventionally, tidal prediction is carried out using harmonic analysis (HA) which places severe restrictions on the minimum length of tidal records and cannot separate oceanography from astronomy. While a physical framework was devised by Munk and Cartwright, the method failed due to the significant burden of realistic input function selection. Here, we develop a new framework for tidal analysis and prediction based on embedding universal function approximators within Munk and Cartwright’s response method. The new Response Framework overcomes the challenges imposed by the original method and demonstrates superior predictive accuracy over HA using 90\% less data. We also devise a method for obtaining physical insights from the learned model and apply this to explain oceanic resonance, something that is not possible in traditional tidal analysis. By explicitly separating oceanography from astronomy, the Response Framework provides a basis for the study of non-gravitational forcing, including anthropogenic climate change. Our approach is readily adapted to modern oceanographers’ needs and shows promise in terms of application to upcoming satellite missions and storm-surge modeling. An open-source Python package is provided.
BibTeX
@inproceedings{monahan2024response,
  author = {Thomas Monahan and Tianning Tang and Stephen Roberts and Thomas A. A. Adcock},
  title = {{Response Framework: Tidal analysis and prediction through physics-informed ML}},
  booktitle = {Research Square},
  year = {2024},
  doi = {10.21203/rs.3.rs-3289185/v2},
}

Loading related content...

Figure