Lab Sea Freezing

[gustavo-marques]

I've been trying to understand why 154 freezes while 150 does not. By increasing the area over which liquid and frozen runoff is halved, the total amount of salt in the Arctic is reduced:

I do not know what could explain this difference, perhaps sea ice formation -- @dabail10?

One hypothesis is that less salt in the Arctic will eventually affect the Labrador Sea. The following plot suggests it takes ~ 30 years for the "fresh" signal from the Arctic to reach the Lab Sea:

The freshening of the Lab Sea in 154 compared to 150, first happens at the sub-surface (around ~ 500 m, year 25ish):

The response in sea surface salinity occurs after:

I don't fully understand what's going on, but reducing runoff in the Arctic is having an effect opposite of what we want at the sub-surface salinity in the Labrador Sea.

[dabail10]

I will take a look at the sea ice plots in the next couple of days. There is the so-called great salinity anomaly that has big pulses of fresh water coming through the Fram Strait. If you have reduced fresh water into the central Arctic in 154, then perhaps? I guess I don't understand how this reduction influences the sea ice-ocean flux where there is ice. I might also need to talk more with Alice and Marika when they are back next week.

[gustavo-marques]

I am only reducing the liquid and frozen runoff -- these are seen only by the ocean.

The liquid runoff into the Arctic is slightly reduced in 154:

It will be interesting to see how the sea ice is responding to this difference.

[dabail10]

Here are two figures up to year 50 before the Lab Sea officially freezes over. The CGD web server is down. The ice is actually thinner in most of the central Arctic in 154 (red) However, it is thicker in Baffin Bay and the Lab Sea. Looks like there might be thicker ice coming around Greenland in the East Greenland Current. However, the second plot shows congelation growth which is local bottom growth. It seems like most of this is local to Davis Strait and Labrador Sea and is much higher in 154 (blue). Do you have a map of where the freshwater is being reduced in the two runs?

[dabail10]

Tagging @duvivier.

[duvivier]

Said in person, but maybe 154 that has less runoff than 150 has thinner ice actually leads to less ice formation (hence thinner ice) in the Central Arctic. So this could impact subsurface salinity that later makes its way to the Lab Sea? This is my only hypothesis at this point.

[dabail10]

I talked with Frank a bit at lunch. As Alice highlights, the ice is much thinner in the central Arctic in 154 compared to 150 (red). However, in the East Greenland Current there is thicker ice in 154 (blue). So, one hypothesis is that the ice is melting in the EGC and arrives as liquid around Greenland. This might be getting mixed into the central Labrador Sea causing excessive freshening. I would think this would be more at the surface. @gustavo-marques, do you have a surface salinity map?

[dabail10]

I found the salinity maps. This is at 20m depth, but you can definitely see the freshwater in the EGC and coming around the tip of Greenland. This is more obvious in 154 than in 150. So, this is a two-headed monster. The sea ice melt in the EGC is definitely contributing to the Lab Sea freshening, but then hitting it with a runoff flux is a big part of the problem. I will try a run where I reduce the sea ice albedo and see if I can reduce this freshwater flux from the EGC.

[dabail10]

For curiosity, here is a paper led by Marcus Jochum from CCSM3 where we also had Lab Sea freezing issues. This was solved when he turned off the Smagorinsky viscosity.

Journal of Geophysical Research Oceans - 2008 - Jochum - Ocean viscosity and climate.pdf

[swrneale]

We looked at all CCSM/CESM-type variants around the time of the CESM2 problems, and freezing is not that rare. CCSM3 varied across the 20th century. I wonder if we ran a 20th C whether the ice would improve ??

[dabail10]

Normally even when starting from a frozen Lab Sea, then a 20th century run does indeed have an ice free Lab Sea. I think E3SM at one point just lived with a a frozen Lab Sea for PI runs. Anyhow, I the key point here is that we need to keep the fresh water in the boundary currents somehow.

[duvivier]

I know I haven't been around for all the conversations, but is there consensus about if the fresh bias that leads to freezing is primarily originating at depth or the surface? Or maybe both? It looks like 154 has pretty clear fresh biases well below 200m. Doesn't 154 decrease only the surface aspect?