Relating excess and redistributed temperature to the Turner Angle in the subtropical North Atlantic using GO-SHIP observations and Machine Learning

The ocean takes up 93% of the warming in the climate system. Here, we develop methods to isolate this warming signature using multidecadal observations in the North Atlantic. As part of GO-SHIP, repeat ship-based CTD hydrographic observations have been made across the A05 section in the North Atlantic at 24.5˚N. These are climate quality observations of relatively high spatial resolution along the section, providing a unique opportunity to monitor the state of Atlantic physical properties and biogeochemistry. The A05 section has been occupied approximately every 5 years since 1992. Temperature and salinity variability across A05 is influenced by several factors, including air-sea interaction and the effects of anthropogenically driven climate change. Excess temperature is a measure of the amount of extra temperature in the ocean due to post-industrial atmospheric CO2, whereas redistributed temperature quantifies the reorganisation of ocean temperature structure by ocean circulation and mixing. Existing methods to decompose the excess and redistributed temperature changes rely on estimates of the anthropogenic carbon change. The Turner angle, which represents the angle between the theta-s curve and an isopycnal in theta-s space, provides an index of the relative contributions of temperature and salinity on stratification, and thus, on water column stability. Using data from A05, we explore how temporal shifts in temperature and salinity affect the Turner angle, with the aim of using this relationship to separate the excess and redistributed components of change without relying on estimates of anthropogenic carbon. We will establish the relationship between excess and redistributed temperature and Turner angle using Machine Learning tools and the known link between anthropogenic carbon and excess temperature. This approach will enable the use of the Turner angle-based method in areas without any carbon data.