the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Revisiting the wintertime emergent constraint of the Southern Hemispheric midlatitude jet response to global warming
- 1Department of Physics, Imperial College London, London, United Kingdom
- 2Grantham Institute, Imperial College London, London, United Kingdom
- 3Department of Meteorology, University of Reading, Reading, United Kingdom
- 1Department of Physics, Imperial College London, London, United Kingdom
- 2Grantham Institute, Imperial College London, London, United Kingdom
- 3Department of Meteorology, University of Reading, Reading, United Kingdom
Abstract. Most climate models show a poleward shift of the southern hemispheric jet in response to climate change, but the inter-model spread is large. In an attempt to constrain future jet responses, past studies have identified an emergent constraint between the climatological jet latitude and the future jet shift in austral winter. However, we show that the emergent constraint only arises in the zonal mean, and not in separate halves of the hemisphere. This can be explained by the presence of a double jet structure in the Pacific region, making the zonal mean jet latitude a poorly defined quantity that does not represent the latitude of a zonally coherent structure during this season. The usefulness of the emergent constraint is therefore questionable. This finding can further explain the prior finding among CMIP5 and CMIP6 ensembles that the meridional structure of the zonal-mean zonal wind response does not change with climatological jet latitude but stays fixed.
Philipp Breul et al.
Status: closed (peer review stopped)
-
RC1: 'Comment on wcd-2022-42', Anonymous Referee #1, 21 Aug 2022
Revisiting the wintertime emergent constraint of the Southern Hemispheric mid-latitude jet response to global warming, by Breul et al
This paper revisits the wintertime emergent constraint on the Southern Hemisphere (SH) jet latitude, which relates the climatological jet position to the future jet shift in CMIP models. Specifically, previous work has argued that this constraint arises primarily as a result of a geometrical effect related to the fact that the wind anomalies in the zonal mean are anchored at the same location regardless of the climatological jet position. It is argued here that the climatological jet latitude differences across models are related to inter-model differences in the relative strength of the single Atlantic jet and the strengths of the two jets in the double jet Pacific structure. The authors argue that the zonal mean jet location is not very physically meaningful given that it is an average over these two distinct jet structures. I think this study is tackling an issue that needed to be resolved and I find the arguments somewhat convincing. I do think that some improvement on the connections between their toy model and the actual CMIP model behavior could strengthen the conclusions considerably. At the moment, the toy model is presented as being able to reproduce the relationships that are found in CMIP, but it might be nice if there were a way to connect the toy model to the behavior of individual models a bit more. I've made some suggestions along these lines below, but overall I think this manuscript is acceptable for publication after minor revisions.
General suggestions:
(1) Improving the linkage between the toy model and the behavior of the individual CMIP models. At the moment, the pieces of evidence for the authors argument are (a)
there is no local connection between the jet position and jet latitude in the two longitudinal sectors separately and (b) the toy model can exhibit similar behavior to the CMIP models when random values are added to the amplitude of a1 and a2. It's not totally clear to me what a1 and a2 represent but I'm assuming it's either the amplitude of the two pacific jets or the amplitude of one of the Pacific jets and the Atlantic jet (see comment below). Anyway, the piece that seems a bit missing is then linking this back to the behavior of CMIP. It seems like it should be possible to then show that there is a relationship between the CMIP zonal mean jet latitude and the amplitude of the relevant jets in a manner that is similar to the toy model. If possible, I suggest the authors investigate whether these aspects can be tied together a bit better e.g., is the climatological latitude of the jet in each model highly correlated with the amplitude of the Pacific southern jet?(2) I think the prior work of Bracegirdle et al 2013 doi:10.1002/jgrd.50153 deserves some discussion. They showed that the emergent constraint holds in the Pacific sector and is kind of there, albeit weaker, in the Atlantic and the Indian ocean sector. The big difference here is probably that you are looking at the winter while Bracegirdle et al used the annual mean, but I think it could still be worth discussing this prior work and why your conclusions differ.
Minor comments by line number:
l20: It sounds a bit strange to first cite Simpson and Polvani 2016 and Breul et al 2022 in the context of studies that link the jet shift to annular mode timescale and then cite them in the next sentence to say that these studies couldn't find that constraint. Suggest removing Simpson and Polvani 2016 and Breul et al 2022 from the first lot of citations.
l40: What has motivated this choice of 22 models? Why not use all the models?
l46: "chose" --> "chosen"
l48: I think Kidston and Gerber (2010) used this quadratic method to define the jet latitude before Barnes and Polvani (2013) did. Suggest citing them instead.
l55: Simpson et al (2021) might be the more relevant one here since they used CMIP6, while Simpson and Polvani (2016) used CMIP5.
l56: This correlation is also quite a lot higher than was found for CMIP6 by Simpson et al 2021 (they found -0.57). This is probably due to the different models being used, but some motivation for not including some models should probably be given here.l90: I'm not sure that "observed response" is the best phrasing here. It could be mixed up with the climate change signal or the climatological jet latitude in the observations. I think really you're referring to the CMIP behavior here?
l97 and 98: I may have missed it but I don't think you've defined a1 and a2. I'm assuming you're defining the amplitude of the three jets with a1, a2 and a3, but it's not clear to me which jets a1 and a2 correspond to.
l99: Something's not right about this sentence "and the therefore also" is not making sense.
-
RC2: 'Comment on wcd-2022-42', Anonymous Referee #2, 18 Sep 2022
Summary of the review
This manuscript raises questions related to the physical interpretation of zonal mean jet in the wintertime Southern Hemisphere where the latitudes of jets in the Atlantic and Pacific differ from each other. The authors found from their CMIP6 multi-model analyses that the emergent constraint found for the zonal mean jet between its present-day and future shift does not hold for individual two halves of the hemisphere. Using a toy model, the authors demonstrate that a latitudinal shift of the zonal mean jet does not necessarily reflect a coherent jet shift in different longitudes, by showing that the strengthening/weakening of a zonally localized jet can also influence the zonal mean jet shift as a statistical artifact. Also, the toy model by the authors explains the similarity of the meridional structure of the future change of zonal mean zonal wind among CMIP6 models regardless of the present-day jet profile. I think this journal is appropriate for this article to be published after the following comments are addressed.
Major comments (Not necessarily in the order of importance)
a) Line 74.
The authors diagnose the physical existence of the jet using unfiltered daily data in Figures 2b-c. Instantaneous zonal wind in the southern hemisphere mid-latitudes is sensitive to the synoptic scale cyclones/anticyclones moving eastward due to the background jet stream (Figure 2a). Therefore, the physical meaning of zonal wind between Figure 2a and Figure 2b-c is different; the former is the basic state component, and the latter is the mixture of basic state and eddy components. For the author’s purpose, they should use daily data after applying a low-pass filter of e.g. 8 days to filter out the eddy component.
b) Figure 4c, Figure 5
These figures are based on the toy model where random variables are added to a1 and a2 respectively. Please explain why the values added to a1 and a2 are set independently of each other (i.e. no correlation) in comparison to the CMIP6 results.
c) Figure 5a
Φp of CMIP6 models is less clearly anchored, as the authors also discuss in line 135. There rather seems like two groups of CMIP6 models; one group is non-anchored models and the other is anchored models following the identity line in Figure 5a. The authors should explain what caused the second group. Given the fact that such a case never appeared in the toy model (Figure 5a), I wonder to what extent the toy model is applicable to interpret the anchoring found for CMIP6 models.
d) line 150
While Figures (3) and (A2) conceptually explain that the inter-model difference of the zonal mean jet axis is caused by the Pacific double jet structure, there is no figure showing it is the case in CMIP6. To justify the conceptual argument, please add a panel in Figure 3 which is like the current Figure 3 but for actual CMIP6 models.
Minor comments
a) Equation (1)
Please refine it in a form including subscripts (1,2) to represent the Pacific jet described in Table (1).
- AC1: 'Comment on wcd-2022-42', Philipp Breul, 14 Oct 2022
Status: closed (peer review stopped)
-
RC1: 'Comment on wcd-2022-42', Anonymous Referee #1, 21 Aug 2022
Revisiting the wintertime emergent constraint of the Southern Hemispheric mid-latitude jet response to global warming, by Breul et al
This paper revisits the wintertime emergent constraint on the Southern Hemisphere (SH) jet latitude, which relates the climatological jet position to the future jet shift in CMIP models. Specifically, previous work has argued that this constraint arises primarily as a result of a geometrical effect related to the fact that the wind anomalies in the zonal mean are anchored at the same location regardless of the climatological jet position. It is argued here that the climatological jet latitude differences across models are related to inter-model differences in the relative strength of the single Atlantic jet and the strengths of the two jets in the double jet Pacific structure. The authors argue that the zonal mean jet location is not very physically meaningful given that it is an average over these two distinct jet structures. I think this study is tackling an issue that needed to be resolved and I find the arguments somewhat convincing. I do think that some improvement on the connections between their toy model and the actual CMIP model behavior could strengthen the conclusions considerably. At the moment, the toy model is presented as being able to reproduce the relationships that are found in CMIP, but it might be nice if there were a way to connect the toy model to the behavior of individual models a bit more. I've made some suggestions along these lines below, but overall I think this manuscript is acceptable for publication after minor revisions.
General suggestions:
(1) Improving the linkage between the toy model and the behavior of the individual CMIP models. At the moment, the pieces of evidence for the authors argument are (a)
there is no local connection between the jet position and jet latitude in the two longitudinal sectors separately and (b) the toy model can exhibit similar behavior to the CMIP models when random values are added to the amplitude of a1 and a2. It's not totally clear to me what a1 and a2 represent but I'm assuming it's either the amplitude of the two pacific jets or the amplitude of one of the Pacific jets and the Atlantic jet (see comment below). Anyway, the piece that seems a bit missing is then linking this back to the behavior of CMIP. It seems like it should be possible to then show that there is a relationship between the CMIP zonal mean jet latitude and the amplitude of the relevant jets in a manner that is similar to the toy model. If possible, I suggest the authors investigate whether these aspects can be tied together a bit better e.g., is the climatological latitude of the jet in each model highly correlated with the amplitude of the Pacific southern jet?(2) I think the prior work of Bracegirdle et al 2013 doi:10.1002/jgrd.50153 deserves some discussion. They showed that the emergent constraint holds in the Pacific sector and is kind of there, albeit weaker, in the Atlantic and the Indian ocean sector. The big difference here is probably that you are looking at the winter while Bracegirdle et al used the annual mean, but I think it could still be worth discussing this prior work and why your conclusions differ.
Minor comments by line number:
l20: It sounds a bit strange to first cite Simpson and Polvani 2016 and Breul et al 2022 in the context of studies that link the jet shift to annular mode timescale and then cite them in the next sentence to say that these studies couldn't find that constraint. Suggest removing Simpson and Polvani 2016 and Breul et al 2022 from the first lot of citations.
l40: What has motivated this choice of 22 models? Why not use all the models?
l46: "chose" --> "chosen"
l48: I think Kidston and Gerber (2010) used this quadratic method to define the jet latitude before Barnes and Polvani (2013) did. Suggest citing them instead.
l55: Simpson et al (2021) might be the more relevant one here since they used CMIP6, while Simpson and Polvani (2016) used CMIP5.
l56: This correlation is also quite a lot higher than was found for CMIP6 by Simpson et al 2021 (they found -0.57). This is probably due to the different models being used, but some motivation for not including some models should probably be given here.l90: I'm not sure that "observed response" is the best phrasing here. It could be mixed up with the climate change signal or the climatological jet latitude in the observations. I think really you're referring to the CMIP behavior here?
l97 and 98: I may have missed it but I don't think you've defined a1 and a2. I'm assuming you're defining the amplitude of the three jets with a1, a2 and a3, but it's not clear to me which jets a1 and a2 correspond to.
l99: Something's not right about this sentence "and the therefore also" is not making sense.
-
RC2: 'Comment on wcd-2022-42', Anonymous Referee #2, 18 Sep 2022
Summary of the review
This manuscript raises questions related to the physical interpretation of zonal mean jet in the wintertime Southern Hemisphere where the latitudes of jets in the Atlantic and Pacific differ from each other. The authors found from their CMIP6 multi-model analyses that the emergent constraint found for the zonal mean jet between its present-day and future shift does not hold for individual two halves of the hemisphere. Using a toy model, the authors demonstrate that a latitudinal shift of the zonal mean jet does not necessarily reflect a coherent jet shift in different longitudes, by showing that the strengthening/weakening of a zonally localized jet can also influence the zonal mean jet shift as a statistical artifact. Also, the toy model by the authors explains the similarity of the meridional structure of the future change of zonal mean zonal wind among CMIP6 models regardless of the present-day jet profile. I think this journal is appropriate for this article to be published after the following comments are addressed.
Major comments (Not necessarily in the order of importance)
a) Line 74.
The authors diagnose the physical existence of the jet using unfiltered daily data in Figures 2b-c. Instantaneous zonal wind in the southern hemisphere mid-latitudes is sensitive to the synoptic scale cyclones/anticyclones moving eastward due to the background jet stream (Figure 2a). Therefore, the physical meaning of zonal wind between Figure 2a and Figure 2b-c is different; the former is the basic state component, and the latter is the mixture of basic state and eddy components. For the author’s purpose, they should use daily data after applying a low-pass filter of e.g. 8 days to filter out the eddy component.
b) Figure 4c, Figure 5
These figures are based on the toy model where random variables are added to a1 and a2 respectively. Please explain why the values added to a1 and a2 are set independently of each other (i.e. no correlation) in comparison to the CMIP6 results.
c) Figure 5a
Φp of CMIP6 models is less clearly anchored, as the authors also discuss in line 135. There rather seems like two groups of CMIP6 models; one group is non-anchored models and the other is anchored models following the identity line in Figure 5a. The authors should explain what caused the second group. Given the fact that such a case never appeared in the toy model (Figure 5a), I wonder to what extent the toy model is applicable to interpret the anchoring found for CMIP6 models.
d) line 150
While Figures (3) and (A2) conceptually explain that the inter-model difference of the zonal mean jet axis is caused by the Pacific double jet structure, there is no figure showing it is the case in CMIP6. To justify the conceptual argument, please add a panel in Figure 3 which is like the current Figure 3 but for actual CMIP6 models.
Minor comments
a) Equation (1)
Please refine it in a form including subscripts (1,2) to represent the Pacific jet described in Table (1).
- AC1: 'Comment on wcd-2022-42', Philipp Breul, 14 Oct 2022
Philipp Breul et al.
Philipp Breul et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 292 | 86 | 11 | 389 | 3 | 1 |
- HTML: 292
- PDF: 86
- XML: 11
- Total: 389
- BibTeX: 3
- EndNote: 1
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1