Detecting climatic change in AMOC observations

The detection of trends and variations in the Atlantic Meridional Overturning Circulation (AMOC) is an important and at times controversial topic. On average, CMIP6 models project a 1 Sv/decade decrease in the strength of the AMOC in response to anthropogenic climate change. Atlantic subpolar decadal sea surface temperature variations of 0.5 indicate an associated change in AMOC strength of 2 Sv. These are challenging thresholds of signal detection for AMOC observing. Estimates of the AMOC streamfunction, such as those from the RAPID array, have a number of sources of variability ranging from short term Ekman transport to variations in the strength of North Atlantic Deep Water associated with deep water formation that have a slower timescale. Climate model studies have shown that Ekman transport contributes little to the signal of future AMOC decline. We look at the nearly 20 years of data from the RAPID array from a signal to noise perspective. Fluctuations associated with Ekman transport are the largest contribution to noise in the AMOC estimates and hold no signal of low frequency change. Deeper layers show more of the low frequency signal. We amplify this low frequency signal by removing the impact of noise derived from the Ekman transport on the deep temperature and salinity. Finally, we show that the best place for detection of low frequency, climatic changes in AMOC is in the deepest North Atlantic Deep Water, with the noise of the wind removed.