Affiliation:
1. Geophysical Institute University of Bergen and Bjerknes Centre for Climate Research Bergen Norway
2. NORCE Norwegian Research Centre Bjerknes Centre for Climate Research Bergen Norway
Abstract
AbstractWe evaluate changes in dissolved inorganic carbon (DIC) in the Greenland Sea between 2002 and 2016, a period characterized by increasing convection depths. We find a mid‐depth maximum in anthropogenic carbon (Cant) accumulation that occurred as waters at these depths were rejuvenated by deeper reaching convection; broadly, these waters have caught up with the atmospheric CO2 rise that had happened between the last time they were ventilated and 2002 and also tracked the atmospheric CO2 rise 2002–2016. The overlying waters only tracked the atmospheric CO2 rise 2002–2016. The mid‐depth maximum in Cant accumulation was not evident in estimates generated with commonly used multiple linear regression (MLR) methods. We analyze the reasons why and show that the eMLR(C*) method may not fully capture nonsteady state changes in Cant when applied along a single hydrographic section as done here. This nonsteady component equates to redistribution of C*, whose spatial gradients in the Greenland Sea are dominated by Cant. We also show that the regular extended multiple linear regression method is sensitive to loss of spatial DIC gradients, which now happens as more and more Cant enters the ocean. Our findings demonstrate that MLR‐based estimates of the Cant accumulation rate should not be taken at face value in highly dynamical ocean regions, such as the Greenland Sea, and the need for also considering the total change in DIC and how this is affected by natural processes. Further investigations into the ability of MLR methods to reproduce nonsteady state changes in Cant are encouraged.
Funder
Norges Forskningsråd
HORIZON EUROPE Framework Programme
Publisher
American Geophysical Union (AGU)