Surface Energy Balance

[danabasoglu]

I have been looking at the surface energy balance in 349 using a script created by Gustavo. The plot here shows the annual-mean time series of surface fluxes that each component sees. The SUM is the sum of all these fluxes and its magnitude is as expected, that is, small. Isla is checking what the TOA is during this period. If Peter's bug fix is working then we should match the average atm fluxes over this period.

What is peculiar is rather constant heat fluxes for land ice and land components. While the former is gaining heat, the latter is losing heat. I would not expect such steady behavior over a 60 year period. Any thoughts? Are we supposed to consider land-ice and land surface budgets in a combined way?

Also there is a steady heat loss from the sea-ice model, particularly evident in the Southern Hemisphere. Meaning that we should see sea ice growth. Is this the case?

[islasimpson]

@cecilehannay Gokhan was intending to post this to discussions instead of issues but didn't have permissions for some reason. do you have privileges to move this over to Discussions? I don't think I have the permission to do so.

[gustavo-marques]

@islasimpson; I was able to covert it to a discussion.

[islasimpson]

Here is the TOA

Note: -1 times our usual convention to match what's in Gokhan's plots.

[danabasoglu]

@gunterl, @wwieder, @dlawrenncar, @dabail10 ...... any thoughts on the land, land-ice, and sea-ice surface time series?

[duvivier]

@danabasoglu, @dabail10 is out on PTO this week and Marika and I are both at a meeting for the rest of the week. We haven't looked at 349, but my understanding is that this run is similar to others. If so, there has not been a trend in NH or SH sea ice volume that would indicate consistent growth in either hemisphere. We can get diagnostics run on 349 early next week.

[danabasoglu]

Gustavo and I just looked at the surface heat flux time series from 346. They are very similar to those of 349. If we do not see trends in the sea-ice volumes, then there is an issue. Similarly, I expect to see energy loss from the land model equivalent to what is shown in these plots. ... and glacier melt at a constant pace. Or, possibly, I am missing something?

[wwieder]

It's been a while since I looked at long diagnostics from land, but in 316 we're globally losing snow (below)

We're also increasing global terrestrial water storage over time (which I'm struggling to align with global decreases in snow depth)

I suspect that the changes in water storage on land are actually more important for the energy balance?
Moreover, these global perspective masks regional differences. For example accumulating snow in the Arctic, and especially in the Canada and Russian Arctic where CESM likes to build up deep snowpacks, but losing it in the Antarctic

I'll let others speculate on how long-term trends in terrestrial water storage may influence the land energy budget, and will run diagnostics on newer runs to share.

,

[gustavo-marques]

Yes, I think it would be very helpful.

[wwieder]

FWIW, I think the terrestrial water storage (TWS) plot that I shared above should be doing this already, as it's the change in land water stocks over time. This should reflect the residual of the fluxes that you requested? @swensosc, can you weigh in here?

[wwieder]

here's a selection of a bunch of variables from 347 (blue) and 349 (orange). I think 349 branched from year 113 of of run 347, so the years show are roughly overlapping. Trends here seem less clear to me that what I'd posted above for TWS, but to my eye there still seems to be changes in snow depth?

FWIW I've uploaded diagnostics to /project/diagnostics/external/BMT1850/b.e30_alpha09b.B1850C_MTso.ne30_t233_wgx3.349/lnd/, but they aren't showing up online. Are the CGD servers still down?

[danabasoglu]

Please note that the energy loss from the land model is rather steady at roughly 0.03 W / m^2. So, I would expect an energy loss of the same amount showing up in something which may be as a linear trend in accumulation in a field.

[wwieder]

I can squint and draw a linear trend through the snowdp plot (lower left above). Does a 1 mm / y trend in snow provide a -0.03 W/m2 energy flux?

Is the question year about energy exchange between land (-0.03 w/m2) and land ice (+0.01 w/m2) models? We're using a data glacier model here right? How does energy exchange between land and data glacier model work? How does this influence energy balance calcultions?

[danabasoglu]

Will be plotting the net surface heat flux times from the land side to make sure that it matches the one from the coupler diagnostics. Also please note that these flux values are globally scaled. This means that a flux of -0.03 W / m^2 is actually -0.03 / 0.3 = -0.1 W / m^2. Similarly, a flux of 0.01 would be 0.01 / 0.02 (?) = 0.5 W / m^2 over glacier covered regions. So, these are relatively large.

[wwieder]

To do this I think I need the following:

(FSA+FIRE) - (EFLX_LH_TOT + FSH_TO_COUPLER + EFLX_GNET)

This is the net short and long wave radiation minus the sum of latent, sensible and ground heat fluxes. This net ground heat flux (EFLX_GNET) is the sum of soil, snow, surface water and lake heat fluxes. EFLX_GNET is not on by default in the CLM history files.

Looking at our energy balance checks which are passing, it seems like we could also use EFLX_SOIL_GRND. I'll note that this variable has a misleading name since it's calculated over lake patches. EFLX_SOIL_GRND is also not on my default in the history files.

Stepping back at what I think the coupler is seeing in nuopc/lnd_import_export.F90 the only energy terms are net shortwave, net longwave, sensible, and latent heat flux. I'll try summing those sums for now.

[wwieder]

One more observation looking at the diag.log files. The stub glacier model is "gaining" energy by passing runoff (presumably of ice) to the ocean, which perfectly loses the same amount of energy. I don't know what's going into the calculations in these log files, but it seems that (at least for the stub glacier model) this calculation is somewhat an artifact of not tracking the temperatures (or energy) of water fluxes being sent to the ocean. Could something similar be happening on land?

[wwieder]

It seems like in this example land is also gaining energy through runoff, which is also being balanced by the ocean ? Can you point @gustavo-marques' script at these hiroff fluxes to see if they are responsible for the gaps in component energy balance terms?

[olyson]

This is what I've used in the past to replicate the coupler energy balance diagnostics (now found in the diag log file):

global_avg = (-A(0)+A(1)+A(2)-A(3)-A(4)-A(5)+A(6))*cpl_landfrac

where

A(0) = SNOW_FROM_ATM * latice
A(1) = FSA
A(2) = FLDS
A(3) = FIRE
A(4) = FSH_TO_COUPLER
A(5) = EFLX_LH_TOT
A(6) = QRUNOFF_ICE_TO_COUPLER * latice

latice = 3.337d5 ; latent heat of fusion (J/kg)
cpl_landfrac can be found in the diags log, I believe in recent simulations it is 0.29251054.

[dabail10]

Sorry, I was out last week. We are conserving heat instantaneously in the sea ice at least to around 10e-10 or 10e-11 relative error. This is what we typically get. Here are diagnostics from 349/347 for the SH. Maybe a slight trend in SH volume? Kind of hard to tell.