Evaluation 1950 spin-up simulation

[tsemmler05]

In this issue we describe the characteristics of the results of the 1950 spin-up simulation that according to the HighResMIP protocol lasts for 50 years to make sure that at least the upper ocean is in quasi equilibrium.

[tsemmler05]

The 1950 spin-up simulation runs from nominal years 1900 to 1949 with constant 1950 greenhouse gas concentrations and constant aerosol forcing.

Performance index 1900-1909 (reference: CMIP6 models):

Performance index 1910-1919:

Favorable performance compared to CMIP6 models.

[tsemmler05]

We have a strong AMOC:

[tsemmler05]

This is the sea ice extent in the Northern and Southern Hemisphere in winter and summer:

Looks reasonable with slight overestimation of Southern Hemisphere winter extent (should be 18e6 km2 but is at the end of the period 21e6 km2). Northern Hemisphere summer extent is underestimated (should be 6e6 km2 but is at the end of the period 4e6 km2). Southern Hemisphere summer extent and Northern Hemisphere winter extent are well represented (3e6 km2 and 15e6 km2).

[tsemmler05]

However, the Northern Hemisphere sea ice volume starts to decline a lot after 20 years of the spin-up simulation:

Maybe this is due to the strong AMOC?

[tsemmler05]

The SST bias looks generally good - especially noteworthy is that the Southern Ocean SST bias is lower than 1K in large areas (except south and southwest of Australia):

[tsemmler05]

This is the performance index for the years 1920 - 1929 - generally good.

[tsemmler05]

Also the performance index for 1925-1949 looks good:

[tsemmler05]

The AMOC seems to weaken and the Antarctic sea ice at the same time to decline in the last 1-2 decades, at least in winter:

[tsemmler05]

This is the surface temperature bias in the years 26-50 (nominal years 1925-1949) in the spin-up simulation:

Compared to earlier years the Southern Ocean hardly changes while the Northern North Atlantic is warmer.

[tsemmler05]

This is the cross section of the AMOC averaged over years 26-50:

[tsemmler05]

While the energy imbalance was very close to 0 in the first few years of the simulation, the ocean is taking up energy when looking at the later years of the 50 years. Here as an example the energy imbalance calculated over years 26-50 (nominal years 1925-1949) of the 1950-spinup simulation:

TOA: 0.69 W/m2 SEB: 0.59 W/m2 atm: 0.10 W/m2

[tsemmler05]

Seems like that at least part of the problem with the strong AMOC and the missing deep cell was just a problem of the analysis with only every 2nd level output. But still the North Atlantic Deep Water does not seem to spread far enough southward. This is year 1 of a test simulation that is a continuation of the 50 year spin-up run - the only difference being that every level is output:

[tsemmler05]

This is the AMOC using Patrick Scholz's program tripyview. He also uses vertical velocity for the computation of the AMOC.

It is not exactly the same compared to the plot above (made with fdiag also based on vertical velocity) but generally consistent.

[tsemmler05]

Hovmöller diagram of globally averaged temperature by depth: