New insights into today’s AMOC strength through directly measured data: is the slowdown only a myth?

Up until 2004, most of our understanding of the AMOC and its variability were based on model analysis (see last post). Since 2004, the enormous RAPID project provides real-time measurements of the flow strength in the deep Atlantic at 26°N (http://www.rapid.ac.uk/rw/index.php). Since then, very interesting findings were published.  (For more information see SCIENCE FACT on RAPID project)

First results presented by Meinen et al. (2006) showed unexpectedly high variability in the AMOC transport within one year. You can see in the graph that the amount of deep water going south varied between 10 and 90 Sv (1 Sverdrup = 1 mio m3/s) within one year!

Figure from Meinen et al. (2006) showing southwards (blue) and net (red dashed) transport of deep water between the continental shelf and 72°W. LADCP points refer to cruise start and end.

These results strongly question whether the model outcome of a weakening in the AMOC is true or not. If the yearly variability is this high, how high is the interannual variability? Maybe the reduction of strength which appears in all models is only part of natural variability?

With more data, Atkinson et al. (2010) were able to define the yearly variability of the AMOC, with faster flow in spring and autumn, and slower flow in summer and winter. They also identified the upper layer Ekman transport (see post ? for more information) as a potential driver for some of the variability observed at 26.5°N. According to Atkinson et al. (2010), there is a positive linear relationship between the North Atlantic Oscillation (NAO) and the Ekman transport at 26.5°N. This implies that a weaker NAO will weaken the Ekman transport, which in turn will weaken the Ekman induced flow-part of the AMOC. A slower AMOC will then leads to slightly colder sea surface temperatures (SST) in the North Atlantic which makes temperature differences in the North Atlantic region more extreme, and in return strengthens the NAO. This suggests that there might be a negative feedback system which partially leads to the large variability. In 2011, Johns et al. find other influential drivers, such as the Mid-Ocean and the Western Boundary Abaco flow, adding to Atkinson et al.(2010). This means that the AMOC or THC cannot be seen as a pushed water mass that flows north, but as a resulting flow mass from many influencing factors at one point in the ocean.

Figure from Johns et al. (2011) with 5 heat transport curves relative to 0°C at 26.5°N. The black curve is the total of all.

Think back to the first model of ocean circulation: the conveyor belt. Deep water is produced in the North and “pushes” the stream through all oceans. It upwells and flows back to the production centre forming the warm Gulf Stream. But obviously, this view is far too simplified, as shown above. Instead, the flow at any point is controlled by certain factors, such as other currents, winds, and eddy mixing. Some researches go almost as far as to say: there is no current. Every flow of water is just an addition of the upper factors (Lozier,2010; Zhang & Qiu, 2014; Ide&Wiggins, 2015).

This view of ocean circulation might seem very complicated, but this view allows for much more factor inclusion in a model. For example: moisture export from the tropical Atlantic to the Pacific. Remember, the evaporation of lots of water from the tropical Atlantic is greater than all freshwater flowing and raining into the tropical Atlantic, because the easterly trade winds blow all the moisture across Middle America into the Pacific. This leads to saltier water and possible deep water formation in the Arctic. Everyone is concerned about the large freshwater input into the North Atlantic under warming conditions. However, warming conditions also means stronger evaporation in the tropical Atlantic PLUS stronger winds that export more moisture to the Pacific (Richter& Xie, 2010)! This suggests that a warming world might actually stabilize the THC (Richter & Xie, 2010).

With the new insights Matei et al. (2012) managed to remodel the AMOC and predict flow strengths for 4 years after 2012. Their new model underlines the theories listed above: no weakening was projected until 2014, which was the prediction limit at that time.

We have passed the year 2014… were their predictions right?