Estimating global ocean heat content from tidal magnetic signals with machine learning

Sketch for setup and training of the neural network. Global maps of the M2 tidal magnetic field are used as inputs for the neural network to derive gobal ocean heat content values.

Oceanic processes related to climate change can be tracked by magnetic signals that are generated by ocean tides. As such, the periodic magnetic signals of the M2 tide contain trends that are related to the continuously warming ocean. In turn, long-term observations of the tidally induced magnetic field can be used as an additional measure of the heat budget and the overall warming of the ocean. We utilize artificial neural networks as non-linear inversion schemes to investigate this topic. With the help of a neural network, which was trained with ocean temperature data and corresponding simulated tidal M2 magnetic fields over the time period 1990-2015, we could, for the first time, derive global ocean heat content values from recent Swarm satellite observations.

Reference

Irrgang, C., Saynisch, J., & Thomas, M. (2019). Estimating global ocean heat content from tidal magnetic satellite observations. Sci. Rep., 9, 7893. https://doi.org/10.1038/s41598-019-44397-8

Contact

Christopher Irrgang
Scientist
Dr. Christopher Irrgang
Earth System Modelling
Telegrafenberg
Building A 20, Room 313
14473 Potsdam
+49 331 288-2847
Profile