Separating Forced and Internal Contributions to Future Northern Hemisphere Storm Track Changes and Associated Precipitation Impacts

Helliesen, Karen E.; Sheng, Bosi; Bollasina, Massimo A.

doi:10.5194/egusphere-2026-3557

Preprints

https://doi.org/10.5194/egusphere-2026-3557

Preprints

26 Jun 2026

| 26 Jun 2026

Separating Forced and Internal Contributions to Future Northern Hemisphere Storm Track Changes and Associated Precipitation Impacts

Karen E. Helliesen, Bosi Sheng, and Massimo A. Bollasina

Abstract. Storm tracks — the preferred pathways of extratropical cyclones in the midlatitudes — are projected to shift poleward, migrate upward in the atmosphere, and change in intensity under anthropogenic climate change. These changes have important consequences for weather and climate variability across the Northern Hemisphere (NH) midlatitudes. Here we investigate future changes in NH storm track strength and position, their driving physical mechanisms, and their impacts on precipitation, using large ensemble (LE) simulations from three CMIP6 models alongside ERA5 reanalysis data. Robust projections are developed under two scenarios (SSP2-4.5 and SSP5-8.5) for both boreal winter (DJF) and summer (JJA) at end-of-century (2070–2100) relative to the present day (1984–2014). The LE approach enables a rigorous characterisation of internal variability and the separation of the forced response from sampling noise. Key results include: (i) a poleward and upward shift of winter storm tracks, driven primarily by tropical upper-tropospheric warming that enhances upper-level baroclinicity, increasing precipitation poleward of ~45° N; (ii) a weakening of summer storm tracks associated with reduced static stability in the upper troposphere, leading to decreased precipitation across the midlatitudes; and (iii) substantial spread among ensemble members, particularly in DJF under SSP2-4.5, highlighting the prominent role of internal variability in shaping projected changes. The contribution of internal variability is reduced under SSP5-8.5 and during JJA, where the externally forced signal dominates. Inter-model differences, linked primarily to differing equilibrium climate sensitivities, emphasise the importance of multi-model LE frameworks for robust climate impact assessment.

Received: 17 Jun 2026 – Discussion started: 26 Jun 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 4971 KB)

Supplement (3722 KB)

Download & links

Karen E. Helliesen, Bosi Sheng, and Massimo A. Bollasina

Status: final response (author comments only)

RC1: 'Comment on egusphere-2026-3557', Anonymous Referee #1, 22 Jul 2026

The authors use three large ensembles of model simulations to examine the projected forced response of mid-latitude storm tracks in the Northern Hemisphere, the role of internal variability, and their associated precipitation impacts. While I find the exercise of quantifying the role of internal variability vs. forced response very important, I feel the authors fell short in this regard for several reasons. First, previous work already examined the forced changes in storm activity, including the underlying physics. Second, the use of the simplistic Eady growth rate to explain the projected changes appears inadequate. Third, the main novelty of this manuscript is the quantification of internal variability, yet the physical mechanism behind its impacts is not analyzed nor discussed. Fourth, the zonal-mean analyses do not provide new insights relative to the spatial change analysis. I elaborate on these points below, along with other major comments.

Major comments:

1. Comparing models with reanalyses. In the first part of the manuscript, the authors evaluate the models' ability to simulate storm climatology from reanalyses. A much more relevant analysis would be to examine how many models capture the spatial historical trends of the reanalyses. I.e., in which percentile, relative to the model spread, the reanalyses lie at each grid point. This would tell us whether models are skillful in capturing multi-decadal changes in the storm tracks. In other words, even if the models are doing poorly at simulating climatology, they could still do a great job of capturing changes in the storm, as the two are not necessarily driven by the same mechanisms. I thus suggest replacing the analysis in Fig. 1 with historical trends rather than climatology. Furthermore, comparing the multimodel or multi-ensemble mean to a single realization (the reanalyses), even when applied to the climatology, could be problematic.

2. The zonal mean analysis. Figs 2-3 nicely discuss the forced and internal variability changes of the strom tracks. Fig. 8 is a very nice addition that should come right after Fig 3. Yet, doing a zonal mean over each basin (Figs. 3-4) does not add any new insights that we have not learned in Fig. 2. In fact, over the Atlantic, such analysis masks the poleward shift over the west Atlantic and the intensification over the East. Since the paper is already too long, and to keep the reader focused on the main take-home messages, I would suggest removing these analyses from Figs. 3-4. Note that also here, comparing the climatology of the reanalysis to the model mean is not the best practice.

3. The underlying mechanisms. The authors use the Eady growth rate to analyze the physics underlying the future forced changes. First, previous studies have already examined some of these changes, and the authors could simply refer to them. For example, the role of reduced meridional temperature gradient in weakening the storms in summer (O’Gorman 2010, PNAS, Lehmann et al., 2014 ERL), and the role of ocean processes, specifically dynamical processes in intensifying the winter Atlantic storm track (Ciasto et al., 2016 J. Climate, Woollings et al., 2012 Nature Geo., Chemke et al., 2022 J. Climate, Hay et al., 2025 J. Climate). Second, the Eady growth fails to capture changes in the storm track. For example, comparing Fig. 5 and Figs. 6-7, the Eady growth rate does not capture the winter poleward shift in the North Pacific or the intensification in the North Atlantic. In summer, it fails to capture the simple wekaning of the storms. I think a much deeper analysis is needed here. But most importantly, rather than revealing the forced changes, the novelty of this paper lies in the role of internal variability, so the authors should focus on how internal variability affects the storm track response. I thus suggest replacing Figs. 6-7, with an analysis of the mechanism underlying the role of internal variability. Third, the authors analyze both the wind shear and the meridional temperature gradient. Those two are the same via thermal wind balance, and only one appears in the Eady growth rate. Thus, it is pointless to analyze both. Fourth, in addition to showing the pattern changes in baroclinicity, you could also show the correlation of baroclinicity with vv across ensemble means.

4. Future scenarios. I appreciate that the authors analyze two different scenarios, but since the results are overall similar, I would put one of those in the SI. The paper is already long enough, and this again would only help the reader follow the main take-home messages better.

5. Missing literature. In the introduction, results, and discussion, the authors do not include previous work that linked strom track changes to precipitation, including Chang et al., 2022 J. Climate. In the introduction, the authors state that the storms have strengthened over the Pacific and Atlantic in recent decades and that models capture that. Recent studies showed that the storms have shifted poleward in the Pacific, and climate models fail to capture this; Chemke and Yuval 2026, Nature.

6. Misuse of baroclinicity. Baroclinic instability, as discussed in the introduction, indicates that both the meridional and vertical structures of buoyancy (or temperature) can affect the available potential energy for storm growth. The authors refer to baroclinicity throughout the manuscript as the meridional temperature gradient, which is incorrect; it refers to both meridional and vertical gradients. Please revise.

7. What is the reason behind choosing the vv metric for storm intensity, rather than EKE or SLP, for example? How do we know if the uu counterpart is not important as well?

8. Throughout the manuscript, the authors use ECS to explain the different behavior of the storms across models. First, how do you estimate ECS? Second, if the storms are driven by temperature gradients, why should they be proportional to changes in the global mean temperature?

Minor comments:

Throughout the manuscript, the authors describe changes in the storm track over the Atlantic in DJF as a poleward shift. This is inaccurate. There is an eastward extension of the storms (intensification over the west Atlantic and Europe), while a slight poleward shift in the west Atlantic.

Citation: https://doi.org/10.5194/egusphere-2026-3557-RC1
RC2: 'Comment on egusphere-2026-3557', Anonymous Referee #2, 24 Jul 2026

The authors investigated the forced response and internal variability of the projected storm track changes in the Northern Hemisphere, utilizing large ensemble simulations. While I believe it is important to understand the future change in the storm tracks, it is difficult for me to recommend publication of the current manuscript. Unfortunately, the current version of the manuscript does not present substantially new results, some of the methodologies are questionable, and the mechanisms are not well explained. Below are my detailed concerns.

1. The forced response of future storm track changes is well known. The result shown in Figure 2 does not appear to be different from that shown in Chang et al. (2012) and many other papers. This lacks novelty to be one of the key results of the paper.

2. The physical mechanism for the forced response of storm tracks is not well explained. The authors argue the increased upper tropospheric temperature gradient explains the poleward shift of the DJF storm tracks. However, diverse mechanisms exist for explaining the underlying mechanism for the poleward shift and cannot be simply connected to the upper tropospheric temperature gradient (Shaw et al. 2019). Generally, the Eady Growth Rate is derived originally from zonal-mean two-layer dynamics, and its interpretation is not convincing when used for longitude sectors and vertically. Previous work such as Lehmann et al. (2014) used tropospheric bulk Eady Growth Rate to explain the storm track changes. Moreover, O’Gorman (2010) noted that Eady Growth Rate fails to explain some aspects of storm track change.

3. Another key result is that the signal-to-noise ratio is greater in summer than winter and is greater for ssp585 than for ssp245 for the storm tracks. This is expected since winter circulation has more variability and the forcing is greater in ssp585. I think this result is too general. However, what can be novel and interesting is the internal variability itself, which has not been studied previously. For example, Fig. 2 shows that certain modes of internal variability can make the summer storm tracks weaken more and less.

4. I found the methodologies used in quantifying the precipitation response questionable. Firstly, I don’t think it is appropriate to refer to the vv term as ETC activity. Meridional wind is maximized between a cyclone and an anticyclone. Moreover, vv term is calculated at the upper troposphere, so it could represent wave activity, not necessarily an ETC. Secondly, Yau and Chang (2020), which is cited in the paper, concluded that 850-hPa EKE best correlates with precipitation events, but an upper-level metric is used here. Other works, such as Hawcroft et al. (2018), used tracking algorithms based on near-surface metrics to quantify future precipitation changes related to ETC changes. In short, previous work has worked on these topics and used more rigorous methodologies.

5. O’Gorman (2010) explains that the storm track intensity is not connected simply to global-mean surface temperature. Additionally, previous work shows that ECS often does not explain the inter-model spread in circulation responses (Grise and Polvani, 2016). Therefore, it is insufficient to explain inter-model differences with ECS differences.

6. There are misuses of references throughout the introduction. Summertime studies should be distinguished from wintertime studies. Future response studies should be distinguished from historical trend studies.

O’Gorman, P. A. (2010). Understanding the varied response of the extratropical storm tracks to climate change. Proceedings of the National Academy of Sciences, 107(45), 19176-19180.
Grise, K. M., & Polvani, L. M. (2016). Is climate sensitivity related to dynamical sensitivity?. Journal of Geophysical Research: Atmospheres, 121(10), 5159-5176.
Shaw, T. A. (2019). Mechanisms of future predicted changes in the zonal mean mid-latitude circulation. Current Climate Change Reports, 5(4), 345-357.

Citation: https://doi.org/10.5194/egusphere-2026-3557-RC2

Karen E. Helliesen, Bosi Sheng, and Massimo A. Bollasina

Supplement

https://doi.org/10.5194/egusphere-2026-3557-supplement

Karen E. Helliesen, Bosi Sheng, and Massimo A. Bollasina

Viewed

Total article views: 64 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
41	18	5	64	6	4	4

HTML: 41
PDF: 18
XML: 5
Total: 64
Supplement: 6
BibTeX: 4
EndNote: 4

Views and downloads (calculated since 26 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	34	15	4	53
Aug 2026	7	3	1	11

Cumulative views and downloads (calculated since 26 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	34	15	4	53
Aug 2026	7	3	1	11

Viewed (geographical distribution)

Total article views: 56 (including HTML, PDF, and XML) Thereof 56 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 11 Aug 2026

Short summary

Midlatitude storms follow preferred pathways known as storm tracks, which shift poleward and change in strength as the climate warms, affecting precipitation across the Northern Hemisphere. Winter changes are less certain due to natural climate fluctuations, while summer changes are more consistent. These findings have implications for regional storm track and rainfall projections and climate adaptation planning.


Total:	0
HTML:	0
PDF:	0
XML:	0