Interdecadal Oscillation of the Ocean Heat Content as a Contribution to Understanding of Physical Aspects of the Present-Day Climate
Round 1
Reviewer 1 Report
Vladimir Byshev, Anatoly Gusev, Victor Neiman and Alexandra Sidorova: Interdecadal oscillation of the ocean heat content as a contribution to understanding of physical aspects of the present-day climate
This manuscript presents model results from one (INNOM) of the Ocean Circulation models that were compared with observational data during the CORE-II experiment. The model results are used to interpret causes of multidecadal oscillations in the ocean and atmosphere. The presentation is mostly clear. However, there is no justification on how well the model performs in reconstructing present day climate or its variability. This might have been an outcome of the CORE-II experiment, which I am not familiar with. However, this should be mentioned as it is common that different circulation models reproduce differently even with similar forcing. There are at least six references to the authors own publications. I think this is quite many, especially since they appear to be key references for their interpretations of results. There are some language issues with the English text, which make some sentences hard to understand. So the text should be checked by a native English user.
Some details:
Line 24: During the phase of (add during in beginning of sentence)
Line 109: The initial conditions were set at
Line 121: … most pronounced of them…
Line 183 and several others: I think ‘horizons’ should be replaced by e.g. ‘levels’
Line 308: is there a reference to this statement?
Author Response
Response to the Reviewer 1
The authors thank the Reviewer for the helpful recommendation on the manuscript revision.
The Reviewer correctly notes that the results of the World Ocean (WO) circulation reproduced by different model can differ significantly. This can indicate that some models reproduce it better, while other ones do it worse. As an objective assessment of the quality, one can consider correspondence of the numerical modelling results to the WO observational data. A satisfactory model is supposed to reproduce the basic processes including WO climate changes, while the performed measurements are supposed to fit its phase climatic states. Our critical attitude to modelling requires the correct reproduction of the phenomena observed in the WO. The authors paid the most significant attention to this.
1) We performed diagnostic simulations for the Northwestern Pacific zone [21],
2) We involved observation data in this WO region in the scope of one of the greatest hydrophysical experiments in the worldwide oceanographic practice MEGAPOLIGON.
3) We got sure that the model results are consistent with diagnostic simulations and instrumental observation data [8]. We would like to note that the two authors (Byshev and Neiman) were participants of such hydrophysical experiments as POLYGON-70, Polymode, MEGAPOLIGON, Atlantex and others, i.e. they got much experience and practice in the analysis of direct observations in key areas of the WO, containing climatic information on the phase states of the ocean.
The reviewer's comments on the manuscript text (Line 24; Line 109; Line 121; Line 183 and several others) were taken into account, the corrections were made, and the reference was added (Line 308).
As well, the text was significantly modified with respect to English language.
All the corrections were performed with using “trackchanges” TeX package. The submitted file pack contains the two version of PDF-document: one is with history of editing, another (default) is with applied modification.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper entitled “Interdecadal oscillation of the ocean heat content as a contribution to understanding of physical aspects of the present-day climate” by Byshev et al. examined the evolutions of ocean heat content in recent decades. They also discussed the possible reasons accounting for recent ocean heat content variations and attempted to link the ocean heat content variations to the present-day climate. To me, this study is interesting and provides a broader perspective to look at the recent climate change from atmosphere-ocean coupling. However, I find some conclusions in the paper seems to be over-stated and are based on the authors’ speculation on one figure, and this poses a challenge for readers (including me) to follow in some parts of the paper. In addition, I also think the authors need to justify some methods and datasets they used in the paper. Therefore, I recommend major revision for the present manuscript.
1. Could the authors give one figure illustrating the relation between ocean heat content, ocean-atmosphere heat exchange and climate variations (such as dry or humid state) during the recent decades in the introduction?
In addition, the authors declare “the multidecadal oscillation of heat content in ocean is not expressed in the upper mixed layer but manifested in the layer of main thermocline” (L178). If so, how could the multidecadal oscillation of ocean heat content beneath the mixed layer affects the climate (e.g. atmospheric state) with no such signal in the mixed layer?
2. The authors declared around L160 that “The observed two-layer structure of the WO heat content evolution indicates the presence of two different modes and sources of such variability”. To me, I think the conclusion appears to be over-stated. Ocean heat content in the layer of 0-1000m and 1000-5500m both exhibits distinct interdecadal variations (figure 1). These interdecadal variations in different layers could be related and also could be independent. In other words, I personally do not think it is sufficient to conclude it is just one unified mode or two independent modes that controls the multidecadal oscillations in the layer of 0-1000m and 1000-5500m only based on two time series. In addition, the distinct trend shown in the layer of 1000-5500m sterns from the downward penetration of heat content from the upper layer, which I also do not think it is fair to regard the trend as another “independent mode”.
3. L166. The authors declared that “for the period 1948-2007, the heat content of the near-surface 100m layer showed an increasing trend”. Actually, I find it is challenging to see any trend signal in the figure 2. I recommend authors giving the statistical significance value when they calculate trend in all figures.
4. Figures 5-8. I doubt the justification of the method to look at the multidecadal differences by simply calculating the difference between two periods. The anthropogenic influence is inherent in recent decades, and the difference between two periods inevitably comprise the multidecadal oscillations (internal variability) and anthropogenic influence (external forcing).
5. I recommend the authors discussing the uncertainties of the model dataset they used in the study. One way could be comparing the results obtained from model dataset with that obtained from current accessible observational datasets.
6. L365-382. It seems these conclusions are primarily based on the authors’ speculations. If so (even though the proposed mechanism appears to make sense), it should not be put in the conclusions part (and abstract) but only in the discussion part.
Author Response
Response to the Reviewer 2
The authors thank the Reviewer for the helpful recommendation on the manuscript revision.
Statement 1:
Could the authors give one figure illustrating the relation between ocean heat content, ocean-atmosphere heat exchange and climate variations (such as dry or humid state) during the recent decades in the introduction? In addition, the authors declare “the multidecadal oscillation of heat content in ocean is not expressed in the upper mixed layer but manifested in the layer of main thermocline” (L178). If so, how could the multidecadal oscillation of ocean heat content beneath the mixed layer affects the climate (e.g. atmospheric state) with no such signal in the mixed layer?
Response 1:
The upper active layer (UAL) of the World Ocean (WO) is of a thickness about 1000m (0-1000m depth). The UAL directly interacts with the lower layer of the troposphere. In different regions of the WO, the UAL thickness can vary by hundreds of meters, increasing in the phases of the thermal discharge of the ocean, but decreasing in the phases of heat accumulation. The ocean-air interaction is most active in the cold period of the year, when deep convection is involved (up to 1000m or more, e.g., in the North Atlantic and the Southern Ocean). The upper 100m layer of the ocean (quasi-homogeneous layer – UQL) is a subject of seasonal interaction of the ocean with a layer of the adjacent atmosphere. The UQL significantly experiences phases of thermal accumulation and discharge.
Stastement 2:
The authors declared around L160 that “The observed two-layer structure of the WO heat content evolution indicates the presence of two different modes and sources of such variability”. To me, I think the conclusion appears to be over-stated. Ocean heat content in the layer of 0-1000m and 1000-5500m both exhibits distinct interdecadal variations (figure 1). These interdecadal variations in different layers could be related and also could be independent. In other words, I personally do not think it is sufficient to conclude it is just one unified mode or two independent modes that controls the multidecadal oscillations in the layer of 0-1000m and 1000-5500m only based on two time series. In addition, the distinct trend shown in the layer of 1000-5500m sterns from the downward penetration of heat content from the upper layer, which I also do not think it is fair to regard the trend as another “independent mode”.
Response 2:
One or more series - it depends on what we are considering. If we deal with the WO heat content variability, then one time series is already exclusive information. When using multiple models, you can get multiple series set. There was no such a purpose for us. The only thing we were interested in was the reliability of the simulation results, which, in fact, was emphasized.
Statement 3:
L166. The authors declared that “for the period 1948-2007, the heat content of the near-surface 100m layer showed an increasing trend”. Actually, I find it is challenging to see any trend signal in the figure 2. I recommend authors giving the statistical significance value when they calculate trend in all figures.
Response 3:
We would like to clarify that in this figure, the scale of the y-axis (heat content) is identical for the two different layers (0-50m, 50-100m). For this variant of y-axis scaling, the trend is visually hard to see. We were guided by the principle of comparability by using similar scales. Apparently, it is more important in this case to present the trend more obvious. Therefore, we propose to replace this fugure with a one, which has a different scale along the y-axis for these layers. In this case, the trend is well visualized. The figure has been replaced in the manuscript.
As well, important characteristics concerned with numerical simulations are temporal and spatial resolutions. Since the used spatial resolution was 0.5 degrees in latitude and 1.0 degrees in longitude, therefore, for the temporal resolution of 1 hour, about 17280 profiles or vertical distributions of each parameter are computed in a one-degree square per year, and more than 400000 in a five-degree square. This indicates that the estimates obtained are not random, but depend on the quality and features of the model.
Statement 4:
Figures 5-8. I doubt the justification of the method to look at the multidecadal differences by simply calculating the difference between two periods. The anthropogenic influence is inherent in recent decades, and the difference between two periods inevitably comprise the multidecadal oscillations (internal variability) and anthropogenic influence (external forcing).
Response 4:
The figures 5-8 present the simulated results of estimates of the WO temperature disturbance between the phases of 1950-1974 and 1975-1999, reflecting, in our opinion, mainly the internal thermodynamics of the ocean-atmosphere system. The anthropogenic contribution should probably be traced in a linear trend of characteristics.
Statement 5:
I recommend the authors discussing the uncertainties of the model dataset they used in the study. One way could be comparing the results obtained from model dataset with that obtained from current accessible observational datasets.
Response 5:
Our answer to the first Reviewer is appropriate in this case as well.
The results of the World Ocean (WO) circulation reproduced by different model can differ significantly. This can indicate that some models reproduce it better, while other ones do it worse. As an objective assessment of the quality, one can consider correspondence of the numerical modelling results to the WO observational data. A satisfactory model is supposed to reproduce the basic processes including WO climate changes, while the performed measurements are supposed to fit its phase climatic states. Our critical attitude to modelling requires the correct reproduction of the phenomena observed in the WO. The authors paid the most significant attention to this.
1) We performed diagnostic simulations for the Northwestern Pacific zone [21],
2) We involved observation data in this WO region in the scope of one of the greatest hydrophysical experiments in the worldwide oceanographic practice MEGAPOLIGON.
3) We got sure that the model results are consistent with diagnostic simulations and instrumental observation data [8]. We would like to note that the two authors (Byshev and Neiman) were participants of such hydrophysical experiments as POLYGON-70, Polymode, MEGAPOLIGON, Atlantex and others, i.e. they got much experience and practice in the analysis of direct observations in key areas of the WO, containing climatic information on the phase states of the ocean.
Statement 6:
L365-382. It seems these conclusions are primarily based on the authors’ speculations. If so (even though the proposed mechanism appears to make sense), it should not be put in the conclusions part (and abstract) but only in the discussion part.
Answer 6:
6) We moved the text (lines 365-383) to the end of the Discussion section.
All the corrections were performed with using “trackchanges” TeX package. The submitted file pack contains the two version of PDF-document: one is with history of editing, another (default) is with applied modification.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
I have no further comments.
