An Ekman Transport Mechanism for the Atlantic Multidecadal Oscillation

Abstract:

Multidecadal global climate since 1850 consists of the expected greenhouse warming and two cycles of a fluctuation commonly associated with the AMO that so far has not been satisfactorily explained.

In GC53C-06 at AGUFM13 we compared land and sea temperatures during the global warmings of 1860-1880 and 1910-1940 and inferred that heat flowed sea to land, ruling out aerosol-based external forcings and indicating an internal source such as an instability in the AMOC.

Length of day during the past century has varied by ~4 ms inversely with the AMO. Noting that the ocean floor is some five times thinner than the continental crust, we propose here that Earth’s rotation regulates heat flux through the ocean floor. One mechanism for this is centrifugal force pulling plates apart, particularly along the Mid-Atlantic Ridge and around the Ring of Fire, increasing flux by an amount that would easily pass unnoticed in the 1930s. Another mechanism, perhaps less strong, is stress from rotational acceleration increasing the thermal conductivity of the young rocks comprising the ocean floor.

A difficulty is that the ocean would absorb the fluctuations before reaching the surface. We overcome this difficulty via Ekman transport. This mechanism acts on a 50 m deep layer at the surface to drive it polewards from the ITCZ at 3 cm/sec or 1000 km/yr, orders of magnitude faster than the MOC which therefore cannot interfere. This creates a suction at the ITCZ and a downwards pumping action at 30°. In order to close this cycle there must be a flow equal in volume rate towards the ITCZ at depth. We propose that the heat entering the ocean bottom between 30° S and 30° N enters these two "Ekman cells", which carry it to the surface via the ITCZ.

To evaluate feasibility, take the area of the participating 50m surface layer to be 10¹⁴ m², making the volume of the top and bottom layers 10¹⁶ m³. Only 10²² J of heat is needed to warm or cool this by 1/3.85 = 0.26 °C. Over the 30 years 1910-1940 or 10⁹seconds, this represents 10 TW of warming, raising Earth's total geothermal release from a cold 40 TW to 50 TW (80 to 100 mW/m²), subsequently declining to today's 45-47 TW, an entirely feasible amount. An additional 10 TW to warm the whole OML is still plausible.

If this mechanism is correct, varying LOD would forecast varying AMO several years in advance.