Modeling of the North Atlantic Gyre's Meridional Overturning Circulation with Neural Nets

The effects of changing climate patterns have been readily seen in weather observations. Their effects on land are much more understood than their effects on the ocean, largely due to disparities in the number of observations and studies. Relatively little research has been done to analyze the effects of climate change on oceanic cycles by directly utilizing artificial intelligence, particularly regarding meridional overturning circulation. Here we report the progress of an undergoing effort to collect observation data on sea surface features of the North Atlantic Gyre to investigate how well this new tool performs in predicting the future spatial and temporal surface conditions and how this could help us better understand the upcoming transitions in climate patterns. Our models follow current observations to a high fidelity with a mean squared error of the mean sea surface temperature of the North Atlantic Ocean of less than <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$4\times 10^{-5}\ ^{\circ}\mathrm{C}$</tex>. Our preliminary results suggest a decrease in the relative temperature gradient between the warm- and cold-water regions, with this boundary persistently shifting farther north every summer. These models may prove advantageous for future predictions and provide an alternative approach to existing geophysical models on longer timescales. Finally, we make additional recommendations on potential future areas where data acquisition would be advantageous.