the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Jul 2025
North Atlantic Oscillation (NAO) in the Paleoclimate Modelling Intercomparison Project (PMIP)
Abstract. The North Atlantic Oscillation (NAO) is one of the main modes of climate variability and the dominant mode of large-scale atmospheric variability in the North Atlantic basin and has large impacts on the European climate, whose future behaviour remains uncertain. Here we assess the NAO response in past and future climates by looking at a comprehensive set of coupled model simulations performed by the Paleoclimate Model Intercomparison Project (PMIP) and the Coupled Model Intercomparison Project (CMIP) for four experiments: the mid-Holocene (6 ka; midHolocene), the Last Glacial Maximum (21 ka; lgm), the last interglacial (127 ka; lig127k) and an idealised future warming scenario with abrupt quadrupled CO2. Although there are various setups across experiments, the midHolocene and lig127k are mainly characterised by altered orbital configurations, inducing variations in the seasonal cycle, and the lgm and abrupt4xCO2 are mainly characterised by various GHG forcing that induces great global temperature change. Our results show that the NAO is sensitive to GHG-forcing-induced temperature changes but not the orbital configurations. NAO weakens in response to cooling and strengthens to warming. The associated teleconnections change consistently with the theory and are sensitive to the change in NAO amplitude. The two orbital experiments do not show a clear change in associated temperature and precipitation. The weakened NAO in the lgm is associated with a cooler and drier northern Europe, while the enhanced NAO in the abrupt4xCO2 causes a warmer and wetter northern Europe as compared to the piControl. No clear relationship is found in the ENSO-NAO teleconnection.
- Preprint
(3570 KB) - Metadata XML
-
Supplement
(133 KB) - BibTeX
- EndNote
Status: open (until 29 Oct 2025)
-
RC1: 'Comment on egusphere-2025-3140', Anonymous Referee #1, 25 Aug 2025
reply
-
AC1: 'Reply on RC1', Anni Zhao, 12 Sep 2025
reply
Huge thanks for reviewing our manuscript. We totally understand the inconvenience caused by the loss of your original comments. No worries at all.
We highly appreciate your pointing out the typos and the need for language improvement. We will revise the manuscript carefully during the revision process.
Thanks for your comments on the usage of common EOFs. The EOF in our manuscript is not for evaluation; it’s the definition of NAO instead. NAO has two commonly used definitions, one is site-based ( based on observations of the difference of normalised SLP anomalies between fixed stations), the other is principal-component based (defined as the leading empirical orthogonal function (EOF) of DJF SLP anomalies over the Northern Hemisphere north of 20°N to 80°N and 90°W to 40°E ). We adopt the latter definition.
Simulations in our study are the monthly mean variables that have been uploaded to the ESGF. We are not able to calculate the number of wet days.
We will enlarge figures 1-3.
In figures 4 and 5, the PI dots vary because we only plot the dots that the models have involved in both the experiment and the piControl in each sub-panel, to provide a more straightforward comparison between the experiment and the piControl. We will add some explanations in the figure captions.
Citation: https://doi.org/10.5194/egusphere-2025-3140-AC1
-
AC1: 'Reply on RC1', Anni Zhao, 12 Sep 2025
reply
-
RC2: 'Comment on egusphere-2025-3140', Quan Liu, 09 Oct 2025
reply
Review of “North Atlantic Oscillation (NAO) in the Paleoclimate Modelling Intercomparison Project (PMIP)” by Zhao, et al.
The authors investigate how the mean state of the European climate and the NAO change in three simulated past climate and one idealised future warming climate. Their results highlight that the NAO is sensitive to GHG-forcing-induced temperature change but not the orbital configurations. They also show consistent changes between the amplitude of the NAO and the precipitation.
A key strength of this paper is the comparison of NAO responses to two distinct types of forcings. The results carry important implications for understanding the NAO response to global warming
However, several issues need clarification. In particular, some statements differ from previous studies without sufficient discussion, there are mismatches between text and figures, and the organization of paragraphs could be improved.
Major points:
- Line 64-68: The mean state change in the NAO index shows spreads among models. However, the reason summarized in the text is not correct, because most models predict a reduced temperature gradient at lower-level as well as an enhanced temperature gradient at upper-level (see Fig. 3 Harvey et al, 2015). Actually, as shown by McKenna et al, 2021, this "large spread" is mainly due to internal variability.
McKenna, C. M., & Maycock, A. C. (2021). Sources of uncertainty in Multimodel Large Ensemble projections of the winter North Atlantic Oscillation. Geophysical Research Letters, 48, e2021GL093258. https://doi.org/10.1029/2021GL093258
- Line 43-48: Changes in the NAO index can be partitioned into changes in its mean state (shift in the NAO index distribution) and changes in its variability (changes in the shape of the distribution), see Liu (2025) and O'Brien and Deser (2023). The last sentence in this paragraph is unclear.Why are these two aspects difficult to separate in paleoclimate reconstructions but “easy to distinguish” in paleoclimate simulations? Further clarification is needed.
Liu, Quan, et al. "More extreme summertime North Atlantic Oscillation under climate change." Communications Earth & Environment 6.1 (2025): 474.
O'Brien, J. P. and C. Deser, 2023: Quantifying and understanding forced changes to unforced modes of atmospheric circulation variability over the North Pacific in a coupled model large ensemble. J. Climate, 36, 17-35, doi: 10.1175/JCLI-D-22-0101.1.
- Line 196 / Line 215 consistency: The midHolocene and lig127k experiments show weaker meridional temperature gradients than the PiControl. According to Line 64, this should lead to a negative NAO-like mean-state change. Yet Line 215 states that the mean state exhibits a positive NAO-like pattern. Do you have an explanation?
- Line 352: The figure reference is incorrect. Fig. 2a shows temperature changes, not sea-level pressure. Moreover, the text claims “models do not capture a positive NAO-like pattern,” but Figs. 2c and 2f clearly display such a pattern. Please reconcile the text with the figures.
- Figure 4: although the caption notes that “the horizontal locations of dots in a does not have any meaning”, it's easier for readers to understand if the black dots are in the shaded side (right side) of the distribution, as in Figure 5.
- Section 6 title (“remote effects”): I am wondering if “remote effects” is a good title for section 6. It suggests a focus on NAO impacts on remote climates, but much of the section (e.g., Line 349) discusses NAO amplitude changes under different mean-state backgrounds. A more precise title would improve clarity.
- The visibility of the paper could be improved with clearer writing, particularly in the way paragraphs are structured. I noticed that some paragraphs cover multiple topics, and topic sentences are not always clearly stated. For example, in the paragraph starting at line 192, the first sentence is intended as a topic sentence, but it is actually a technical statement that belongs in the methods section (and could be deleted here). Instead, the sentence at line 195 would serve much better as the topic sentence.Another example is the paragraph starting at line 250. It begins with the NAO pattern, then moves on to explained variance, and finally to the magnitude of the NAO index. It would be much clearer if this paragraph were split into separate ones, each focused on a single topic, with a strong topic sentence each. That way, the subsequent discussion of these three aspects (a new paragraph in your manuscript) could be integrated into the corresponding new paragraphs.
Minor points:
- Line 85-88 I suggest to remove the very vague description from “State-of-the-art” to “in different scenarios”.
- Line 175 NAO pattern is a dipole pattern, with negative anomaly at the northern center of action, and positive anomaly at the southern center of action. Therefore, the pattern is always "positive" and the phase is given by the sign of the NAO index. if the EOF gives a "negative" pattern, then both the "eof" and the "pc" should be multiplied by -1. "simulated pattern shows a positive phase" is misleading and should be revised.
- Line 235 to be more concise, “uncertainty in producing the location of the jet stream, arising from both model bias and internal variability”.
- Line 294 I suggest to remove the vague sentence “The NAO is a variation in the atmospheric pressure differences between the Icelandic Low and the Azores High.”
Citation: https://doi.org/10.5194/egusphere-2025-3140-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,317 | 71 | 17 | 1,405 | 30 | 53 | 47 |
- HTML: 1,317
- PDF: 71
- XML: 17
- Total: 1,405
- Supplement: 30
- BibTeX: 53
- EndNote: 47
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
I wrote some comments and saved them inermediately, but they were all lost in the system. Sorry about that, but there were no major issues. Some typoes and the language could be improved. I like the way they maed a synthesis analysis based on different CMIP expeiments and made the most out of available data. One issue to look at is the use of common EOFs for model evaluation and for comparing spatio-temporal covariance structures in the different results (see e.g. Tim Barnet's paper from 1999 https://doi.org/10.1175/1520-0442(1999)012%3C0511:CONSAT%3E2.0.CO;2 and a more recent review from Nature Climate Change http://www.nature.com/nclimate/journal/v7/n10/full/nclimate3393.html).
The analysis is very nice, and it would be interesting to know next whether there is a clearer NAO signal in the number of wet days (or wet-day frequency) than total precipitation. The total preciptation is the product between the number of days, the wet-day frequency and the wet-day mean precipitation. Ther ehave been some indications that the wet-day frequency is more strongly affected by circulation whereas the intensity is influenced by other factors.
Make Fig. 1 & 2 (also 3?) bigger - fill the whole space.
Figs 4 & 5 are difficult to interpret. Why does piControl dots vary?