BAYESIAN OPTIMIZATION OF THE METHOD OF IDENTIFICATION OF MESOSCALE OCEAN EDDIES IN THE LABRADOR SEA IN EDDY-RESOLVING MODELLING DATA

Authors

  • M. A. Kalinin Shirshov Institute of Oceanology, Russian Academy of Sciences
  • P. S. Verezemskaya Shirshov Institute of Oceanology, Russian Academy of Sciences
  • M. A. Krinitskiy Shirshov Institute of Oceanology, Russian Academy of Sciences; Moscow Institute of Physics and Technology National Research University)
  • M. A. Borisov Shirshov Institute of Oceanology, Russian Academy of Sciences; Moscow Institute of Physics and Technology National Research University)
  • N. D. Tilinina Shirshov Institute of Oceanology, Russian Academy of Sciences

DOI:

https://doi.org/10.29006/1564-2291.JOR-2024.52(4).4

Keywords:

Labrador Sea, subpolar North Atlantic, NEMO4, mesoscale eddies, Bayesian optimization, Tree-structured Parzen Estimator

Abstract

Deep convection in the Labrador Sea has a significant role in the formation of the Northern Hemisphere climate. The eddy activity in the Labrador Sea, represented by Irminger rings, influences the spatial and temporal heterogeneity of the mixed layer depth. Automated eddy identification methods are widely used as a tool to study eddy activity in statistically significant samples. However, the most commonly used method for finding local extrema is highly dependent on a variety of parameters chosen by the author or the user of the method. In this paper, a new algorithm for the identification of mesoscale anticyclonic eddies in numerical simulation data is developed. Using Bayesian optimisation method, the optimal values of hyperparameters of the developed eddy identification algorithm were selected. The identification quality in the F1-score measure is improved to 0.232 compared to 0.352 in the baseline configuration.

References


  1. Akiba, T., S. Sano, T. Yanase, T. Ohta, and M. Koyama, 2019: Optuna: A next-generation hyperparameter optimization framework. Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, 2623–2631.

  2. Bergstra, J., R. Bardenet, Y. Bengio, and B. Kégl, 2011: Algorithms for hyper-parameter optimization. Advances in neural information processing systems, 24.

  3. Chanut, J., B. Barnier, A. Lecointre, J. Le Sommer, J. M. Molines, T. Penduff, R. Bourdalle-Badie, A. M. Treguier, and G. Madec, 2008: Mesoscale eddies in the Labrador Sea and their contribution to convection and restratification. Journal of Physical Oceanography, 38 (8), 1617–1643.

  4. Holdsworth, A. M. and P. G. Myers, 2015: The influence of high-frequency atmospheric forcing on the circulation and deep convection of the Labrador Sea. Journal of Climate, 28 (12), 4980–4996.

  5. Katsman, C. A., M. A. Spall, and R. S. Pickart, 2004: Boundary current eddies and their role in the restratification of the Labrador Sea. Journal of Physical Oceanography, 34 (9), 1967–1983.

  6. Klein, A., C. Lemke, T. Blumensath, and C. Friedrich, 2017: Fast bayesian optimization of machine learning hyperparameters on large datasets. Artificial intelligence and statistics, PMLR, 528–536.

  7. Lilly, J. M., P. B. Rhines, M. Visbeck, R. E. Davis, F. Schott, and J. Fischer, 2003: Observations of the Labrador Sea eddy field. Progress in Oceanography, 59 (1), 75–176.

  8. Marsh, R., B. A. de Cuevas, A. C. Coward, Z. L. Jacobs, S. A. Josey, and A. J. G. Nurser, 2009: Recent changes in the North Atlantic circulation simulated with eddy-permitting and eddy-resolving ocean models. Ocean modelling, 28 (4), 226–239.

  9. Nencioli, F., C. Dong, T. Dickey, L. Washburn, and J. C. McWilliams, 2010: A vector geometry-based eddy detection algorithm and its application to a high-resolution numerical model product and high-frequency radar surface velocities in the Southern California Bight. Journal of atmospheric and oceanic technology, 27 (3), 564–579.

  10. Pennelly, C. and P. G. Myers, 2022: Tracking Irminger Rings’ properties using a sub-mesoscale ocean model. Progress in Oceanography, 201, 102735.

  11. Rossum, Guido van, and Fred L. Drake, 1995: Python reference manual. Vol. 111. Amsterdam, Centrum voor Wiskunde en Informatica.

  12. Snoek, J., H. Larochelle, and R. P. Adams, 2012: Practical bayesian optimization of machine learning algorithms. Advances in neural information processing systems, 25.

  13. Stouffer, R. J., D. Seidov, and B. J. Haupt, 2007: Climate response to external sources of freshwater: North Atlantic versus the Southern Ocean. Journal of Climate, 20 (3), 436–448.

  14. Verezemskaya, P., N. Koldunov, S. Danilov, C. Wekerle, M. Losch, C. Schrum, and J. Schroeter, 2023: A new regional model of the Subpolar Gyre based on NEMO4. Copernicus Meetings, EGU23-7251.

  15. Wu, J., X. Yue, Y. Wu, L. Wang, and Y. Tan, 2019: Hyperparameter optimization for machine learning models based on Bayesian optimization. Journal of Electronic Science and Technology, 17 (1), 26–40.

  16. Wolpert, D. H. and W. G. Macready, 1997: No free lunch theorems for optimization. IEEE transactions on evolutionary computation, 1 (1), 67–82.

Downloads

Published

2024-12-29

Issue

Vol. 52 No. 4 (2024): Journal of Oceanological Research

Section

Ocean physics and climate

License

Transfer of copyrights occurs on the basis of a license agreement between the Author and Shirshov Institute of Oceanology, RAS

Most read articles by the same author(s)