Citation:
Solodoch A, Stewart AL, Hogg AMC, Manucharyan GE. Machine Learning-Derived Inference of the Meridional Overturning Circulation from Satellite-Observable Variables in an Ocean State Estimate. JAMES [Internet]. 2023.
Abstract:
| The ocean’s Meridional Overturning Circulation (MOC) plays a key role in the climate system, and thus 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’s simplicity (a feed-forward fully connected neural network with small number of neurons). The ML model exhibits high skill in reconstructing the abyssal overturning cells in the Indo-Pacific and Southern Oceans, as well as the Atlantic intermediate MOC cell (AMOC). In particular, the approach achieves a higher skill in predicting the model Southern Ocean deep 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, as well as a function of the time scale of MOC variability. Further tests are conducted to examine which combination of satellite-observable variables are optimal, and to explore how spatial coarsening of the input variables influences the model skill. ML interpretability methodologies are applied to evaluate the locations which contribute the most to reconstruction skill. Finally, the potential for reconstructing MOC strength estimates from real satellite measurements is discussed. |