Unsupervised Learning Reveals Geography of Global Ocean Dynamical Regions. Issue 5 (21st May 2019)

• Record Type: Journal Article
• Authors: Sonnewald, Maike; Wunsch, Carl; Heimbach, Patrick
• Is Part Of: Earth and space science. Volume 6:Issue 5(2019)
• Page Start: 784
• Page End: 794
• Publication Date: 2019-05-21
• Subjects: machine learning -- global patterns -- ocean dynamics -- big data -- ocean modeling -- physical oceanography; Space sciences -- Periodicals; Geophysics -- Periodicals; 500.5
• DOI: 10.1029/2018EA000519
• Languages: English
• ISSNs: 2333-5084

Abstract

Abstract : Dynamically similar regions of the global ocean are identified using a barotropic vorticity (BV) framework from a 20‐year mean of the Estimating the Circulation and Climate of the Ocean state estimate at 1° resolution. An unsupervised machine learning algorithm, K ‐means, objectively clusters the standardized BV equation, identifying five unambiguous regimes. Cluster 1 covers 43 ± 3.3% of the ocean area. Surface and bottom stress torque are balanced by the bottom pressure torque and the nonlinear torque. Cluster 2 covers 24.8 ± 1.2%, where the beta effect balances the bottom pressure torque. Cluster 3 covers 14.6 ± 1.0%, characterized by a "Quasi‐Sverdrupian" regime where the beta effect is balanced by the wind and bottom stress term. The small region of Cluster 4 has baroclinic dynamics covering 6.9 ± 2.9% of the ocean. Cluster 5 occurs primarily in the Southern Ocean. Residual "dominantly nonlinear" regions highlight where the BV approach is inadequate, found in areas of rough topography in the Southern Ocean and along western boundaries. Plain Language Summary: A geography of the global ocean is presented of dynamical regions found using unsupervised machine learning techniques. A bottom‐up approach is used to identify emergent patterns in the modern global ocean. Their existence demonstrates commonalities that lead to methods for understanding the global circulation. Five areas vary from a depth coherent flow, quasi‐Sverdrupian, interior, interior with a strong vertical component, and an interior flow specific to the Southern Ocean. In some regions nonlinear terms are important, which will be the subject of future work at higher resolution. Key Points: Machine learning is applied to create a geography of global dynamical regions, emergent dynamics are discussed A barotropic vorticity framework is sufficient for 93% of the ocean, with nonlinear terms have a small extent but could impact circulation Robust and novel application of machine learning techniques are demonstrated to offer insight into large-scale ocean physical regimes