Kevin Schwarzwald (IRI Columbia): On the propagation of climate data uncertainty in projections of the impacts of climate change on society
Li-Wei Chao (TAMU): Cloud feedback components in observations and their representation in climate models
Ian Eisenman (UCSD/SIO): The radiative feedback continuum from Snowball Earth to an ice-free hothouse
00:42:08 Cristian Proistosescu: Hi Kevin. The quote about the want for Monte Carlo simulations piqued my interest. Do you think there is a role for Statistical emulators that can generate *very* large quantities of possible future data? And if so, what do you think that role is?
00:50:52 Kevin Schwarzwald (he/him): Yeah, I think that’s a good question. I think it depends how it’s used. I think they could be quite useful for sensitivity analyses, where having access to a complete exploration of the parameter space would allow really figuring out what factors actually matter for a given question. I worry about using them as inputs / drivers of final projections of climate impacts though, at least without a better understanding of how representative they truly are of a realistic distribution of future climate metrics. Basically, would there be a danger of over-estimating our certainty of what we think future climate distributions are? (This is one of my issues with uses of NEX GDDP, for example - because it’s a relatively easy-to-use product, it gets used a lot, and since it appears to many users as a plug-and-play product, I think some users tend to overestimate how good NEX GDDP actually represents local climate futures - and skip the hard but necessary work of climate model / data evaluations)
00:52:19 Xiaoli Zhou: Great talk Li-Wei, how does the aerosol trend look like from 2002-2014. to what degree the perturbation in aerosols from observation affects the “cloud feedback” computed from observation?
00:53:25 dennishartmann: Are observed high clouds going up faster than models as the climate warms? How does this compared to fixed cloud temperature?
00:56:27 McKim, Brett: Hi Li-Wei, It’s interesting that you find both the observed and simulated anvil cloud amount feedback to be small, in apparent contradiction to Sherwood 2020 and AR6. Do you have any thoughts on this? Also, have you assessed the tropical high cloud optical depth feedback?
00:56:49 Mark Zelinka: @Xiaoli: The AMIP runs do include changing aerosol loading (nominally) consistent with obs
00:57:29 Xiaoli Zhou: I see. Thanks Mark!
01:01:58 Chao, Li-Wei: @dennishartmann, we found that altitude of high-cloud increase as temperature increase in the observations, which leads to a positive observed feedback. But for those models that simulate negative high-cloud altitude feedback, the altitude of high-cloud decrease in response to increasing temperature. But when further looking into that, we find those models simulate increase in high-cloud altitude using cloud fraction from CALIPSO simulator. So the question is why those models have inconsistency in different cloud fraction product, and that require future investigations.
01:09:43 Chao, Li-Wei: @Brett, Thank you for pointing out that findings, we also find that interesting. The tropical anvil cloud feedback includes the amount and optical depth components from tropical marine ascent regions. Note that we are looking at short-term cloud feedback here, and Sherwood et al 2020 have also points out there are disagreements among previous studies, some have negative feedback while other suggest neutral or even weak positive feedback.
01:09:48 Kevin Schwarzwald (he/him): Are there any concerns about CESM2 being able to well simulate climates with GMST this far outside of the PI - historical range?
01:10:36 Dessler, Andrew E: Similar question: do you think the radiative transfer algorithm can handle such high CO2?
01:12:24 McKim, Brett: @Li-Wei Thanks!
01:13:23 Kyle Armour: @Kevin: Great question. Some features are probably robust, while others (e.g., the exact width of the U shape in feedbacks) are surely model dependent. We’d love to see this experiment done in other models too.
01:13:59 Eric DeWeaver: What happens to the Sherwood et al. result? Do we have to go back to the earlier ECS range?
01:14:01 Tim Merlis: Andy is presumably talking about the GCM RT
01:16:12 Kevin Schwarzwald (he/him): Reacted to "@Kevin: Great questi..." with 👍
01:16:22 Andreas Schmittner: @Ian I’m curious about AMOC changes in your experiments for the snowball and hothouse cases
01:17:08 Isaac Held: related to previous question, can you use existing LGM simulations to test the relevance of a no-ice sheet response assumption?
01:18:29 Kyle Armour: @Eric: Good question. These results suggest that the formulation of feedbacks as changing linearly with global temperature isn’t quite right, and that because of the U shape comparing feedbacks between two states depends sensitively on what the temperatures in those two states are. In Sherwood et al, they used a somewhat large value of alpha (large change in feedback in energy budget equation). These results suggest that that value might be too large when applied to the LGM since the U is more flat there. This would imply the LGM constraint would get a bit tighter compared to Sherwood et al.