Panel Discussion: "Can we rule out internal variability as the main driver of recent tropical SST trends?"
Peter Huybers (Harvard) & Rob Wills (ETH Zurich)
12:28:59 From Gavin Schmidt : If the mechanism proposed is correct (due to double ITCZ) one would predict that the models with better or worse ITCZ will have predictable rates of EP warming/cooling, no?
12:29:17 From iPhone : How well does the LFCA method compare to the two shown?
12:29:55 From Dillon Amaya : Thoughts on forced changes in SST variability (mainly thinking of ENSO variability) as opposed to mean state changes?
12:30:25 From Laurent TERRAY : What about the HiResMIP models ?
12:30:34 From Robert Jnglin Wills : Sorry if I went over time, I did not see the 2-minute warning
12:30:51 From Kyle Armour : Replying to "If the mechanism pro..."
You also need Southern Ocean cooling to get the teleconnection to the tropics, and most models miss that as well. So even models with better ITCZs could miss the tropical cooling.
12:31:03 From Aurélien Ribes : How is the CESM2-HR GSAT warming strengthening over time (how non-linear is it?)?
12:31:59 From Suzana Camargo : Replying to "What about the HiRes..."
They do have a slightly better gradient pattern as shown in Fig. 3 of Sobel et al. 2023.
12:32:40 From Pedro DiNezio : @Gavin Schmidt - 99% of the models have a double ITCZ bias, hasn’t improved much since cmip3:
https://webext.cgd.ucar.edu/Multi-Case/CVDP_repository/cmip3.20c3m/pr.mean.mam.png
https://webext.cgd.ucar.edu/Multi-Case/CVDP_repository/cmip5.historical/pr.mean.mam.png
https://webext.cgd.ucar.edu/Multi-Case/CVDP_repository/cmip6.historical-A/pr.mean.mam.png
12:33:50 From Suzana Camargo : Replying to "What about the HiRes..."
https://www.pnas.org/doi/10.1073/pnas.2209631120
12:38:26 From Gavin Schmidt : In the fyfe et al figure the width of the observed trend was only sampling the hadcrut ensemble but did not include the full structural uncertainty or even the uncertainty in the OLS fit. This was discussed at the time.
12:41:48 From Tim Osborn, University of East Anglia : Are those simulations control runs or forced with natural forcings?
12:49:36 From Deirdre Des Jardins : La-Nina like pattern associated with increased frequency of reduced precipitation in western North America. For risk analysis in climate adaptation planning, evaluation of the likelihood of SST pattern being forced vs internal is essential.
12:50:32 From Michelle L'Heureux : Not a question but a comment: I hope that more higher resolution runs (that seem to have a more representative gradient) can be made… it seems quite important to understand whether the trends will reverse at some point. Can it be shown whether higher resolution runs do better at capturing changes in the zonal gradient historically?
12:51:48 From Suzana Camargo : Replying to "Not a question but a..."
The HighResMIP models do behave slightly better. See my respond above and a link to our paper that has a figure on that.
12:51:59 From Michelle L'Heureux : Reacted to "The HighResMIP model..." with 👍
12:56:44 From Deirdre Des Jardins : Wondering about the discrepancies between obs and models and the warming of mode and intermediate waters found by Li, England, & Groeskamp
12:57:28 From phuybers : Replying to "In the fyfe et al fi..."
Yes, thanks for noting this. My point is not that the Fyfe figure is wrong but that we are in similar situation with the pattern effect of it being difficult to quantity the full uncertainty.
12:59:17 From Christoph Schaer : I have to run leave for dinner. Thanks for the nice presentations, very interesting!
12:59:26 From Andrew Williams (he/him) : Reacted to "I have to run leave ..." with ❤️
12:59:30 From Qikai Hu : Reacted to "I have to run leave ..." with ❤️
13:01:16 From Laurent TERRAY : Thanks to both speakers, great talks !
13:01:26 From Ric Williams : fantastic session, well done
13:01:26 From Amy Clement (she/her) : Thank you both so much for your thoughtful presentations!
13:02:21 From Michelle L'Heureux : I’d support with this excellent “debate” being continued/updated roughly every 6-12 months. ;-)
13:02:59 From Malte Stuecker : The CESM1 high resolution coupled experiment has a reasonable ENSO variance for a present day control but ENSO variance decreases a lot in response to strong greenhouse gas forcing: https://www.nature.com/articles/s41558-021-01132-4
13:04:31 From Andrew Wittenberg : Question for Rob, regarding the statistical methods: Have you evaluated how well they can discriminate the forced/internal components in a single realization from a CGCM, where we can determine the actual “truth” (for the model) using a forced historical large ensemble?
13:05:23 From Tim Merlis : For Peter: The Laepple & Huybers map looked like there was a model—obs discrepancy in N. Atlantic SST variability of the opposite sense as the tropics. Is there a concern that attention is focused where models might simulate too little variability vs. nature because that ends up being ‘hot spots’ of apparent discrepancies between trends in models and obs?
13:05:27 From Andrew Williams (he/him) : Replying to "I’d support with thi..."
noted! ;)
13:05:49 From Deirdre Des Jardins : Reacted to "I’d support with thi..." with 👍
13:08:29 From Kevin Schwarzwald : Reacted to "I’d support with thi..." with 👍
13:09:24 From Andrew Wittenberg : Reacted to "Question for Rob, ..." with 👍
13:10:55 From Steve Bloom : Thanks to all!
13:11:01 From Deirdre Des Jardins : Furiously taking notes. 😁