William Ingram (Univ. of Reading/Met Office): "Sensitivity of climate feedbacks to vertical resolution in a GCM"
Hansi Singh (Univ. of Victoria): "Climate sensitivity is sensitive to changes in ocean heat transport"
Nadir Jeevanjee (GFDL): "Robustness of the longwave clear-sky feedback: From pencil-and-paper to coupled GCMs"
09:17:58 From Gavin Schmidt : Big correlation btw AMOC decline and ECS no?
09:19:38 From Mark England : Great talk, is the prescribed Q flux seasonally varying or annual mean?
09:19:55 From Marian Osei : what informed the thresholds of +30%, +15%, -30%, -15%? Were other values used? Did they have any significant impact?
09:19:56 From Saravanan, Ramalingam : How is the atmospheric compensation in midlatitudes partitioned between transient and stationary eddy heat transport?
09:20:06 From Bjorn Stevens (MPI-Hamburg) : does the extra tropical RH increase?
09:20:15 From Bjorn Stevens (MPI-Hamburg) : with more OET
09:20:22 From Thorsten Mauritsen : @Gavin, CanESM is high sensitivity, but at the low end…
09:20:29 From Zhihong Tan : If the OHT is mainly due to the wind-driven gyres, then the OHT change is coupled to the atmospheric energy transport because both are linked to the circulation strength and surface temperature gradient — I have some difficulty to think about a scenario where OHT can change independently.
09:21:47 From Chris Wells (he/him) : UKESM1 has very high ECS and in lower half of AMOC change too - perhaps not a very clear correlation
09:21:56 From Gavin Schmidt : Fair point.
09:23:10 From Gavin Schmidt : The OHT is also affected by the anomalous freshwater that is coming off Antarctica and Greenland- not included in any of these models….
09:23:13 From Hansi Singh (she/her) : @Mark — these are monthly-varying Q-fluxes
09:23:36 From Nicholas Lutsko (he/him) : @Hansi, I might have missed it, but are these aquaplanet simulations?
09:24:17 From Kyle Armour : @Gavin: I think the correlation between AMOC and ECS depends a lot on the timescale over which you look and what you mean by ECS. AMOC decline definitely acts to reduce warming (e.g., see work by Winton and others), so it’s not surprising that AMOC decline would be correlated with transient warming under 4xCO2, but unclear to me what happens on longer timescales.
09:26:04 From Hansi Singh (she/her) : @Marian — these are idealized values that we came up with. We did not use any others.
09:26:54 From Jonah Bloch-Johnson : @Hansi do you end up finding that the long wave clear-sky feedback differs depending on the ocean heat transport?
09:28:12 From Hansi Singh (she/her) : @Ramalingam — I haven’t looked carefully at the partitioning! But this would be interesting to check — it seems that the atmosphere has a hard time keeping up with the ocean in the NH, which may be because the stationary eddies are less effective at keeping up with (opposing) the OHT changes
09:28:19 From Schmittner, Andreas : It is well known from both model simulations and paleo data that AMOC collapse/reduction affects the ITCZ and thus precipitation.
09:29:02 From Hansi Singh (she/her) : @Andreas — yes! It does affect the ITCZ. But it also seems to impact the hemispheric precipitation sensitivity! Which is probably because of it’s impact on surface evaporation.
09:30:33 From Hansi Singh (she/her) : @Nicholas — no, these have continents.
09:33:52 From Hansi Singh (she/her) : @Bjorn — the RH increases with increasing OHT (and decreases with decreasing OHT). I also have calculations of the RH feedbacks and their temperature contributions. These don’t change the attribution, though does really show that when you move more ocean heat away from the deep tropics, there’s a large increase in RH that accompanies it.
09:34:28 From Bjorn Stevens (MPI-Hamburg) : @Hansi … thanks, that helps me understand it.
09:35:31 From Hansi Singh (she/her) : @Zhihong — yes, the ocean won’t change without the atmosphere! However, we are trying to figure out if the changes in the ocean (which happen as part of the fully-coupled system’s response) impact how much the climate warms. In order to do this, we’ve artificially prescribed the OHT in order to see what the impact of those OHT changes are.
09:36:12 From Chris Rentsch : Is the spectrally resolved feedback in the window also ignoring ozone?
09:36:43 From Hansi Singh (she/her) : @Jonah — yes, it does! Because of water vapour.
09:37:20 From Chris Rentsch : Thanks.
09:38:20 From Bjorn Stevens (MPI-Hamburg) : @Nadir… great and very enjoyable talk. It said it all.
09:39:07 From Bjorn Stevens (MPI-Hamburg) : @jonah… that is why I asked about the RH
09:40:21 From Mark Zelinka : @Nadir, So FAT + Simpson’s law means there should not be much wiggle room for total LW feedbacks, would you agree?
09:41:58 From Zhihong Tan : @Hansi Thanks! I think the results are interesting. Just from a dynamical perspective, perhaps the OHT can be decomposed into two components that are directly coupled to the circulation (e.g., the wind-driven Ekman transport component) and one that is not directly coupled to it (e.g., AMOC), and perturb only the latter, while keeping the first component in accordance with atmospheric changes (such as Kang et al. [2018] ‘The partitioning of poleward energy transport response between the atmosphere and Ekman flux to prescribed surface forcing in a simplified GCM’).
09:43:30 From Mark Zelinka : @Nadir, thanks. The work of Yoshimori et al (DOI: 10.1175/JCLI-D-19-0108.1) has gone in this direction, combining Held and Shell (constant RH feedbacks) with FAT for LW cloud fbk
09:43:34 From Hansi Singh (she/her) : @Zhihong — it would be fun to do this decomposition!
09:44:13 From Jonah Bloch-Johnson : @Hansi thanks, thanks makes sense
09:46:09 From Mark Zelinka : In case anyone wants it, the updated figure is here (minus Gavin’s labeling): https://github.com/mzelinka/cmip56_forcing_feedback_ecs/blob/master/ECS6_histogram_gavin1.pdf
09:46:38 From Gavin Schmidt : On the vertical resolution, GISS-E2.1 and GISS-E2.2 have 40 and 110 layers respectively. No big change in ECS
09:47:26 From Brett McKim : @Mark Zelinka, we build a simple model for estimating the all-sky feedback (incorporating the T-FRAT thinking of Yoshimori) in our paper: https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2021GL094074
09:47:32 From Bjorn Stevens (MPI-Hamburg) : http://doi.wiley.com/10.1002/qj.546 … is to thanks for teaching us all again about Simpson
09:48:01 From Mark Zelinka : cool, thanks Brett
09:50:51 From Gavin Schmidt : (Sorry GISS-E2.2 has 106 layers, same horizontal resolution, and higher top. Roughly double vert res in the common region).
09:54:14 From Thorsten Mauritsen : @William, did you try less vertical levels? I mean 85 or 240 are in both cases quite a bit, and it wasn’t too long ago since people would happily use ~30 levels
09:55:30 From stephan de roode : @William In some models the cloud cover for fields of broken clouds is not a conserved variable for changes in the number of vertical levels (due to the kind of overlap assumptions), could that play a role?
09:55:45 From Gavin Schmidt : (Sorry again! 102 layers in GISS-E2.2).
09:56:21 From Schmittner, Andreas : @William: can you provide a reference for sensitivity to changes in vertical resolution at low vertical resolution that you mentioned. Where it matters.
09:57:12 From Bjorn Stevens (MPI-Hamburg) : @William question about an earlier talk … if there is time.
09:57:42 From Clara Orbe : https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2019JD032204
09:58:56 From stephan de roode : @william: thanks!
09:59:48 From Zhihong Tan : @William I may have missed it, but does the increased vertical resolution go preferentially in the lower troposphere or is the increase uniform through the column?
09:59:56 From Gavin Schmidt : Worth making clear that increasing vertical resolution (in line with increasing horizontal resolution) has a lot of benefits (QBO, Strat polar variability, planetary wave spectrum, marine stratus etc.) even if ECS doesn’t budge much.
09:59:56 From Thorsten Mauritsen : For what it is worth, there was an MPI-ESM in CMIP5 with 47 and 95 levels, though most of the added levels were in the stratosphere. Didn’t make much of a difference.
10:05:20 From Eric DeWeaver : Does anything bad happen if you increase vertical resolution dramatically without also changing the horizontal resolution?
10:05:29 From Nadir Jeevanjee : I have to sign off. Thanks everyone for great presentations and discussion.
10:09:25 From Adam Bauer : @Bjorn, seems like your viewpoint is we should use GCMs to inform other experiments with observations/theory, rather than taking them as ground truth?
10:09:35 From Bjorn Stevens (MPI-Hamburg) : yes!