Machine Learning‐Derived Inference of the Meridional Overturning Circulation From Satellite‐Observable Variables in an Ocean State Estimate

Abstract The oceanic Meridional Overturning Circulation (MOC) plays a key role in the climate system, and monitoring its evolution is a scientific priority. Monitoring arrays have been established at several latitudes in the Atlantic Ocean, but other latitudes and oceans remain unmonitored for logis...

Full description

Bibliographic Details
Main Authors: Aviv Solodoch, Andrew L. Stewart, Andrew McC. Hogg, Georgy E. Manucharyan
Format: Article
Language:English
Published: American Geophysical Union (AGU) 2023-04-01
Series:Journal of Advances in Modeling Earth Systems
Subjects:
Online Access:https://doi.org/10.1029/2022MS003370
_version_ 1797663732647591936
author Aviv Solodoch
Andrew L. Stewart
Andrew McC. Hogg
Georgy E. Manucharyan
author_facet Aviv Solodoch
Andrew L. Stewart
Andrew McC. Hogg
Georgy E. Manucharyan
author_sort Aviv Solodoch
collection DOAJ
description Abstract The oceanic Meridional Overturning Circulation (MOC) plays a key role in the climate system, and monitoring its evolution is a scientific priority. Monitoring arrays have been established at several latitudes in the Atlantic Ocean, but other latitudes and oceans remain unmonitored for logistical reasons. This study explores the possibility of inferring the MOC from globally‐available satellite measurements via machine learning (ML) techniques, using the ECCOV4 state estimate as a test bed. The methodological advantages of the present approach include the use purely of available satellite data, its applicability to multiple basins within a single ML framework, and the ML model simplicity (a feed‐forward fully connected neural network (NN) with small number of neurons). The ML model exhibits high skill in MOC reconstruction in the Atlantic, Indo‐Pacific, and Southern Oceans. The approach achieves a higher skill in predicting the model Southern Ocean abyssal MOC than has previously been achieved via a dynamically‐based approach. The skill of the model is quantified as a function of latitude in each ocean basin, and of the time scale of MOC variability. We find that ocean bottom pressure generally has the highest reconstruction skill potential, followed by zonal wind stress. We additionally test which combinations of variables are optimal. Furthermore, ML interpretability techniques are used to show that high reconstruction skill in the Southern Ocean is mainly due to (NN processing of) bottom pressure variability at a few prominent bathymetric ridges. Finally, the potential for reconstructing MOC strength estimates from real satellite measurements is discussed.
first_indexed 2024-03-11T19:19:00Z
format Article
id doaj.art-4ca5d254295a44efac7877a426a6e706
institution Directory Open Access Journal
issn 1942-2466
language English
last_indexed 2024-03-11T19:19:00Z
publishDate 2023-04-01
publisher American Geophysical Union (AGU)
record_format Article
series Journal of Advances in Modeling Earth Systems
spelling doaj.art-4ca5d254295a44efac7877a426a6e7062023-10-07T19:57:25ZengAmerican Geophysical Union (AGU)Journal of Advances in Modeling Earth Systems1942-24662023-04-01154n/an/a10.1029/2022MS003370Machine Learning‐Derived Inference of the Meridional Overturning Circulation From Satellite‐Observable Variables in an Ocean State EstimateAviv Solodoch0Andrew L. Stewart1Andrew McC. Hogg2Georgy E. Manucharyan3Department of Atmospheric and Oceanic Sciences University of California in Los Angeles Los Angeles CA USADepartment of Atmospheric and Oceanic Sciences University of California in Los Angeles Los Angeles CA USAResearch School of Earth Sciences Australian National University Canberra ACT AustraliaSchool of Oceanography University of Washington Seattle WA USAAbstract The oceanic Meridional Overturning Circulation (MOC) plays a key role in the climate system, and monitoring its evolution is a scientific priority. Monitoring arrays have been established at several latitudes in the Atlantic Ocean, but other latitudes and oceans remain unmonitored for logistical reasons. This study explores the possibility of inferring the MOC from globally‐available satellite measurements via machine learning (ML) techniques, using the ECCOV4 state estimate as a test bed. The methodological advantages of the present approach include the use purely of available satellite data, its applicability to multiple basins within a single ML framework, and the ML model simplicity (a feed‐forward fully connected neural network (NN) with small number of neurons). The ML model exhibits high skill in MOC reconstruction in the Atlantic, Indo‐Pacific, and Southern Oceans. The approach achieves a higher skill in predicting the model Southern Ocean abyssal MOC than has previously been achieved via a dynamically‐based approach. The skill of the model is quantified as a function of latitude in each ocean basin, and of the time scale of MOC variability. We find that ocean bottom pressure generally has the highest reconstruction skill potential, followed by zonal wind stress. We additionally test which combinations of variables are optimal. Furthermore, ML interpretability techniques are used to show that high reconstruction skill in the Southern Ocean is mainly due to (NN processing of) bottom pressure variability at a few prominent bathymetric ridges. Finally, the potential for reconstructing MOC strength estimates from real satellite measurements is discussed.https://doi.org/10.1029/2022MS003370Meridional Overturning Circulationocean circulationsatellite sensingclimate variabilitymachine learningobserving systems
spellingShingle Aviv Solodoch
Andrew L. Stewart
Andrew McC. Hogg
Georgy E. Manucharyan
Machine Learning‐Derived Inference of the Meridional Overturning Circulation From Satellite‐Observable Variables in an Ocean State Estimate
Journal of Advances in Modeling Earth Systems
Meridional Overturning Circulation
ocean circulation
satellite sensing
climate variability
machine learning
observing systems
title Machine Learning‐Derived Inference of the Meridional Overturning Circulation From Satellite‐Observable Variables in an Ocean State Estimate
title_full Machine Learning‐Derived Inference of the Meridional Overturning Circulation From Satellite‐Observable Variables in an Ocean State Estimate
title_fullStr Machine Learning‐Derived Inference of the Meridional Overturning Circulation From Satellite‐Observable Variables in an Ocean State Estimate
title_full_unstemmed Machine Learning‐Derived Inference of the Meridional Overturning Circulation From Satellite‐Observable Variables in an Ocean State Estimate
title_short Machine Learning‐Derived Inference of the Meridional Overturning Circulation From Satellite‐Observable Variables in an Ocean State Estimate
title_sort machine learning derived inference of the meridional overturning circulation from satellite observable variables in an ocean state estimate
topic Meridional Overturning Circulation
ocean circulation
satellite sensing
climate variability
machine learning
observing systems
url https://doi.org/10.1029/2022MS003370
work_keys_str_mv AT avivsolodoch machinelearningderivedinferenceofthemeridionaloverturningcirculationfromsatelliteobservablevariablesinanoceanstateestimate
AT andrewlstewart machinelearningderivedinferenceofthemeridionaloverturningcirculationfromsatelliteobservablevariablesinanoceanstateestimate
AT andrewmcchogg machinelearningderivedinferenceofthemeridionaloverturningcirculationfromsatelliteobservablevariablesinanoceanstateestimate
AT georgyemanucharyan machinelearningderivedinferenceofthemeridionaloverturningcirculationfromsatelliteobservablevariablesinanoceanstateestimate