一个多月前，我收到国际刊物对我两篇文章的评审意见。这两篇文章一篇是观测，另一篇是理论。理论部分主要是用我所提出的多尺度相互作用（NMI）模型解释原因。对审稿人的意见，我来回读了十多遍，发现审稿人提的意见都比较中肯，这对于我改进论文有很大帮助。所以这一两个月我忙着修改论文。特别是有一个审稿人的意见比较多，他提出了大多数读者想要提出的问题，是很见地的。通过回答他的问题，我进一步加深了对我的理论模型的理解，也进一步意识到我所提出的理论模型可能是未来这个领域的主要方向（哈哈，在若干年以后。。。，我深信这一点），但要得到国际广泛认可还有很长的路要走。
预计对这审稿人的意见的回复将在30-40页左右。
下面是这个审稿人的审稿意见：
Review of the paper “Decadal Relationship Between European Blocking ......, by D. Luo et al."
In part II of this study,the authors use a barotropic weakly nonlinear multi--‐scale interaction (NMI)model to understand the blocking frequency distribution and investigate the relationships between blocking event distribution, NAO phase, zonal wind strength and storm track intensity, which would help understand the decadal relationship between European Blocking (EB) and NAO. Compared to the full general circulation model, I think the NMI model provides a straightforward way to investigate the roles of NAO phase, background zonal wind and storm track in the life cycle of blocking event.
However, when using such ahighly idealized model to simulate the real atmosphere and target the topic,more efforts are needed to test the robustness of the model results andconvince the readers that the model can simulate the essential dynamics of NAOand European blocking events reasonably well.
For the current version ofthe manuscript, I am concerned that the authors did not provide enough evidence on the robustness of the key results and the similarities between the NMI model simulations and the real NAO, blocking life cycles in the Atlantic--‐European regions. At times, I also feel that part of the manuscript is hard to follow. Isuggest the authors make more efforts presenting the experiments setting and model results more clearly.
The details of my majorconcerns are addressed below:
1) Robustness of the model result
I am concerned that some experiment settings in this study seem too idealized or lack of observationalbasis. An advantage of an idealized model is that one can explore the parameter space more fully to investigate the robustness of the key results. I suggestthe authors at least discuss how the following factors would affect the main conclusions.
--‐ The background zonalwind distribution
I was suspicious of whether the setting of theuniform background zonal wind (as described on page 9 line 16--‐21) is appropriate for simulating the evolution of the NAO and EB event. In thereality, strong jet is observed in the Atlantic region, in which the zonal wind exhibits evident horizontal shear. As revealed in the previous studies, i.e.Thorncroft et al (1993), the wave breaking, which is considered highly related to the daily NAO index, is sensitive to the horizontal shear of the zonal wind.There can be either cyclonic or anti--‐cyclonic wave breaking if the horizontalshear of the zonal wind is varied. Thus, I am concerned that if using more realistic zonal wind distributions in the experiment (i.e. taking into accountof the horizontal shear of the zonal wind), whether the model results and theconclusions made in the paper would be modified.
--‐ Setting of the synoptic eddies
The synoptic eddies are considered playing an important role in driving and maintaining the life cycle of NAO. In the model, as described on page 15 line 8--‐12, the synoptic eddies are setting with a length scale of 1000 km, whose spacial distributions are inthe form as in Figure 5b. I wonder whether the setting of the synoptic eddies also has any observation basis, as for the standing waves in Figures 4 and 5a. I agree with the authors that a typical length scale of the synoptic eddy is in the order of 1000 km. However, some observational studies suggested that themost energetic eddies in the extratropics may lie in the zonal wavenumbers of 4--‐8, which is corresponding to a length scale of the 3000--‐5000 km. Some studies also suggest that synoptic eddies with different zonal wavenumbers mayplay different roles in the low frequency variations as NAO. I am not sure howthe evolutions of NAO and EB depend on the setting of the synoptic eddies.
--‐ The initial conditions
The life cycle of theblocking flow is no doubt dependent on the initial condition. Thus, it isnecessary to show the rational of the setting of initial condition, and whetherit is comparable with the observations. As the variables in the NMI model are all non--‐dimensionalized, it is also hard to compare the values used for themodel parameters at the initial state to that in the real atmosphere. The authors need make this point clear in the paper.
2) Similarities between the NMI model and the real atmospheric flow overAtlantic –European region
In this study, the only waythat the highly idealized NMI model can include geology/topography information and simulate an Atlantic--‐European--‐region--‐like atmospheric flow is to addthe climatological stationary waves (CSW) into the numerical simulations. Ithink the authors should compare Figs. 4 and 5a in detail, and make it clearthat how to connect the zonal position in the model to the real world. Moreexactly, which regions in Figs5a are corresponding to the Atlantic, East andWest European regions, respectively. This would help understand the modelresults in the following context.
As the CSW is primarily topography forced, it should not vary with time and with the phase of NAO or other parameters. Thus, it is very confusing that in different simulations withthe NMI model, the zonal positions of CSW are set differently. This adds the difficulties in connecting the evolutions of the NAO and blocking events in the model to the real atmosphere. As for each experiment, the x--‐ coordinate denotes different positions.
In the paper, the authorsalso show the snapshots of the streamfunctions at every 3 days in the NMI modelto illustrate the evolutions of the blocking flow. As the authors did not show the corresponding snapshot from the reanalysis data, it is hard for me to judge whether the model can simulate the blocking flow reasonably well. In fact, Ithink showing the time evolutions of the NAO and EB indices to compare withthat from the reanalysis data might be more clear and straightforward.Especially for figures 11, 12 14, each of them include 36 subfigures. It is hard for readers to follow every subplot. Given that there are already many figures in the current manuscript, I suggest the authors think of cutting off some of them.
3) Definitions of NAO and blocking indices in NMI model
To compare the NMI modelresults with the observed NAO and EB events, new NAO and blocking indices aredefined to diagnose the model results. The authors need to discuss the analogy between the indices defined for the NMI model and the ones used for reanalysisdata. In addition, I was confused that in different experiments, the NAO andblocking indices are defined differently. For example, on page 20 line 16--‐21and on page 21 line 17--‐20, two different definitions for NAO and EB aredefined. The authors need to explain the rational for the different definitions as well.
Minor comments:
1) In equations 2, there are some variables whose meanings are notelaborated. More specially, variables Ø_p and Ø_w are used in equations 2a 2b and 2c, whose physical meanings are not explained, and these variables are not used anymorein the following context. The authors may think of cutting off these variables.
2) As the authors discussed on page 7 line 10--‐18, the decadal shift of theNAO patterns can also be explained from the viewpoint of wave breaking. Then itis not clear from the authors’ argument that whether the wave breakingdynamics, which is considered highly nonlinear in nature, can be simulated inthe weakly nonlinear model. And how the authors’ model results here is relatedto the previous wave breaking study?
3)In the paper, the authors did a few sets of sensitivity studies with different initial conditions to investigate the rolesof background zonal wind, storm tracks, phase of NAO on the evolutions of EB event. For readers, it is really hard to remember all the settings of the experiments and compare the results of different simulations. To make the paper morereader friendly, the authors may consider listing a table of the experiment setup, in which the settings of the model parameters andthe initial conditions are summarized. I think this will help clarify many details of the numerical experiments, and make model results easier to compare.
4) Page 25line 20: The authors here use the TM index to calculate the blocking intensityof EB event. However, in Part I of this study, the authors use two otherblocking indices denoting the blocking intensity. I don't know why here adifferent index is used and what the differences are between these blockingin dices.
