The Indian Ocean Dipole: A Missing Link between El Niño Modokiand Tropical Cyclone Intensity in the North Indian Ocean

Round 1

Reviewer 1 Report

The authors claim a connection between El Nino Modoki and Tropical Cyclone Intensity, which is explained by the Indian Ocean Dipole linking the two together. The evidence they provide to support their hypothesis, however, is insufficient. A key element in their reasoning are Figures 7 - 9. Figure 7 has indeed a local maximum of TCPI occurring 40 years after a local minimum of ENSO. Figure 8, however, displaying the same mode comparison with a different dataset, has no local maximum of TCPI at all after the absolute minimum of ENSO. The authors write that in Figure 9 "A dip in the DMI in 1976 leads the peak in the TCPI in 2006 which is led by DMI in 1989." (lines 349 and 350). So is there a connection of the TCPI peak with the dips or the peaks of DMI? Fig. 7 is more or less identically the same as Fig. 9, and Fig. 10 is the same as Fig. 12. How is this possible with quite different data sets? And if the data sets are very similar, then why are they treated as different? Figures 13 - 15 are missing completely. 

Given the fact that the mode comparisons using dataset-1 and -2 are not consistent with each other and that three essential figures, to which the text refers, are missing, I have to reject the article in its present form.

Author Response

Response to reviewer’s comments

Thank you for your suggestions and the opportunity to revise our paper on ‘Indian Ocean Dipole, a missing link between the El Niño Modoki and the Tropical Cyclone Intensity in the North Indian Ocean’. The suggestions offered have been immensely helpful, and we also appreciate your insightful comments on revising the paper. Please see in blue our response to the reviewer’s comments. All page numbers refer to the modified manuscript with tracked changes.

Reviewer’s comment: So is there a connection of the TCPI peak with the dips or the peaks of DMI? 

Authors’ response: Please see the newly added tables (8-13) and figures (10-12) for detailed findings in the modified manuscript.

Reviewer’s Comment: Fig. 7 is more or less identically the same as Fig. 9, and Fig. 10 is the same as Fig. 12. How is this possible with quite different data sets? And if the data sets are very similar, then why are they treated as different? 

Authors’ response: To facilitate your suggestion, we have incorporated two more datasets and the result remains consistent. Since the two data sets were very similar, we have retained only one of them (Figure 12).

Reviewer’s Comment: Figures 13 - 15 are missing completely. 

Authors’ response: Please see the newly added supplementary figures at the end of the manuscript.

Reviewer’s Comment: Given the fact that the mode comparisons using dataset-1 and -2 are not consistent with each other and that three essential figures, to which the text refers, are missing,

Authors’ response: Please refer to the added figures (10-12) and tables (8-13). The three datasets used results in the same finding. 

 

Author Response File: Author Response.docx

Reviewer 2 Report

This study aims to identify linkages between El Nino Modoki, the Indian Ocean Dipole, and tropical cyclone potential intensity over the Indian Ocean.  It's an interesting and highly relevant topic that certainly warrants investigation.  However, the presentation of the methods, results, and discussion lack clarity.  The authors apply empirical mode decomposition to extract a low frequency (multidecadal) signal from the various datasets examined, but they offer no justification for selecting this mode.  Especially important is that the quasi-periodic low frequency components that they analyze have periods of 30-40 years.  I'd argue that it is not possible to determine lead-lag relationships of such signals over the 64 years of data presented here.  Perhaps the authors can repeat the analysis using a higher frequency (interannual) mode or an average of several modes, which may yield more statistically meaningful results. Given the methodological issues mentioned above and in the comments below, I do not feel that this work is suitable for publication in its present form.

Specific Comments:

Lines 29-30:  There is now a substantial body of research concerning tropical cyclone intensity in the context of climate.  Please provide some citations.  

Introduction: It would be helpful to mention in the introduction that this research specifically focuses on decadal patterns in TC potential intensity.

Lines 53-64:  I appreciate that the authors provide readers with a preview of the results upfront, but this paragraph seems out of place in the introduction.  For example, "IMF" has not yet been defined.  I might consider providing a more generalized summary of the results here, if at all, and moving these lines of text to the results section.

Line 73: Please spell out CPSD since this is its first use within the body of the manuscript.

Lines 80-82: Please indicate the specific reanalysis used and provide a citation (e.g., NCEP/NCAR R1?).

Lines 93-96: Please provide more description of this dataset and any post-processing applied.  Are these data simply average SST anomalies over the Nino4 region, or has other post-processing been applied (I was unable to access the link due to the U.S. government shutdown)?  

Line 193: It is not clear to me why the fifth IMF is chosen for further analysis.  Given that IMF 5 represents a multidecadal signal, the 64-year dataset used here seems too short to draw statistically meaningful conclusions.

Figure 3:  While I understand that this isn't the primary focus of the study, the authors should note the difference in trends obtained using datasets one and two.  At first glance, these seem to be substantial differences in TCPI trend.  Do other methods for extracting trends yield such differences?  

Line 374: Why was this analysis not repeated using the DMI?

Line 433: "Thus, a high El Nino Modoki could result in a low TCPI season in the North Indian Ocean, at a time lag of about 40 years."  Given that the two climate signals (IOD and ENSO) vary on inter annual timescales, it seems odd that the lag would be so long.  My understanding is that the analysis doesn't exactly look at El Nino Modoki, but rather the low frequency variability in Nino4 SST anomalies.  Finally, given that the datasets only span a 64-year period, the quasi-periodic patterns exhibited by IMF5, which have periods of 30-40 years, do not contain much more than 1 full cycle.  Therefore, it is not clear to me how we can draw conclusions about lead-lag relationships based on only 1 full cycle.  Additional explanation is needed here.

Author Response

Response to reviewer’s comments

Thank you for your suggestions and the opportunity to revise our paper on ‘Indian Ocean Dipole, a missing link between the El Niño Modoki and the Tropical Cyclone Intensity in the North Indian Ocean’. The comments are encouraging and have been immensely helpful, and we also appreciate your insightful comments on revising the paper. Please see in blue our response to your comments. All page numbers refer to the modified manuscript with tracked changes.

Reviewer’s comment: Lines 29-30:  There is now a substantial body of research concerning tropical cyclone intensity in the context of climate.  Please provide some citations.

Authors’ response: We have added more references. Please refer to line 30-37.

Reviewer’s comment: Introduction: It would be helpful to mention in the introduction that this research specifically focuses on decadal patterns in TC potential intensity.

Authors’ response: Suggested changes were made on line 88-89 in the revised manuscript.

Reviewer’s comment: Lines 53-64:  I appreciate that the authors provide readers with a preview of the results upfront, but this paragraph seems out of place in the introduction.  For example, "IMF" has not yet been defined.  I might consider providing a more generalized summary of the results here, if at all, and moving these lines of text to the results section.

Authors’ response: The text has been moved to the results section.

Reviewer’s comment: Line 73: Please spell out CPSD since this is its first use within the body of the manuscript.

Authors’ response: The CPSD method is no longer used in the paper.

Reviewer’s comment: Lines 80-82: Please indicate the specific reanalysis used and provide a citation (e.g., NCEP/NCAR R1?).

Authors’ response: We have specified the reanalysis used and have provided a citation. Please see line number 103-104.

Reviewer’s comment: Lines 93-96: Please provide more description of this dataset and any post-processing applied.  Are these data simply average SST anomalies over the Nino4 region, or has other post-processing been applied (I was unable to access the link due to the U.S. government shutdown)?

Authors’ response: The suggested changes were made on lines numbered 134 through 140.

Reviewer’s comment:   Line 193: It is not clear to me why the fifth IMF is chosen for further analysis.  Given that IMF 5 represents a multidecadal signal, the 64-year dataset used here seems too short to draw statistically meaningful conclusions.

Authors’ response: IMFs are accepted or rejected based on the threshold criterion. We have now added a table delineating the acceptance and rejection values. Please refer to the table (1-7).

 

Reviewer’s comment: Figure 3:  While I understand that this isn't the primary focus of the study, the authors should note the difference in trends obtained using datasets one and two.  At first glance, these seem to be substantial differences in the TCPI trend.  Do other methods for extracting trends yield such differences?

Authors’ response: We extracted the trends using a statistical method and the trends are similar for the significant IMFs (Table 1, 2 and Figure 2, 3). The decreasing trend (Figure (4), Dataset-3, modified manuscript) was not significant and was not retained.

TCPI shows an increasing trend for dataset-1 and 2 (Figure 2 and 3). Both of these are retained during the threshold test. Please see the explanation at the end of the tables 1-3.

 

Reviewer’s comment: Line 374: Why was this analysis not repeated using the DMI?

Authors’ response: This analysis has been repeated with DMI derived from all the three datasets and is shown in figures (6-8) of the modified manuscript.

 

Reviewer’s comment: Line 433: "Thus, a high El Nino Modoki could result in a low TCPI season in the North Indian Ocean, at a time lag of about 40 years."  Given that the two climate signals (IOD and ENSO) vary on inter annual timescales, it seems odd that the lag would be so long.  My understanding is that the analysis doesn't exactly look at El Nino Modoki, but rather the low-frequency variability in Nino4 SST anomalies.  Finally, given that the datasets only span a 64-year period, the quasi-periodic patterns exhibited by IMF5, which have periods of 30-40 years, do not contain much more than 1 full cycle.  Therefore, it is not clear to me how we can draw conclusions about lead-lag relationships based on only 1 full cycle.  Additional explanation is needed here.

Authors’ response: We agree that the dataset is not long enough to place our confidence in such a long term cycle. So, considering your suggestion, we have removed the result stating a 40-years lag.

 

Author Response File: Author Response.docx

Reviewer 3 Report

Several editing comments are made along the text (file enclosed), several times a given words is repeated in the same sentence and this should be avoided.

The main comment on the content is about the confidence in the results given the length of the time series, also the CPSD peaks seems to be not clearly resolved, this should be discussed more

Comments for author File: Comments.pdf

Author Response

Response to reviewer’s comments

Thank you for your suggestions and the opportunity to revise our paper on ‘Indian Ocean Dipole, a missing link between the El Niño Modoki and the Tropical Cyclone Intensity in the North Indian Ocean’. The comments are encouraging and have been immensely helpful, and we also appreciate your insightful comments on revising the paper. Please see in blue our response to the reviewer’s comments. All page numbers refer to the modified manuscript with tracked changes.

 

Reviewer’s comment: If you include the atmosphere you should talk about ENSO events

Could be more specific, tropical Pacific Ocean, but do not forget about teleconnections

Authors’ response: Incorporated the suggestions on line 39-46 and added literature related to teleconnections.

 

Reviewer’s comment: This last sentence is a conclusion, should not be here

Authors’ response: The sentence is removed from here.

 

Reviewer’s comment: All this that follows is not introduction, should be in results or discussion

Authors’ response: Line number 53-61 are moved and re-written in the results section

 

Reviewer’s comment: What kind of statistical methods?, should be mentioned

Authors’ response: 5. The dataset was interpolated in time and space by the providers.

Please see lines, 124-128 in the revised manuscript.

 

Reviewer’s comment: What do you mean with ENSO time series, are you using an indexes such as the ONI? this is not clear

Authors’ response: We are using EMI. Please refer to the lines, 134-140 in the revised manuscript.

 

Reviewer’s comment: You have to be consistent with the use of subindexes

Authors’ response: Incorporated your suggestion. Please refer to the equation (1), line number 176.

 

Reviewer’s comment: You  are using the term ENSO signal, I thinks you are referring to SST values, this is only a part of ENSO

Authors’ response: We are using the El Niño Modoki Index in the updated manuscript (Line no.). The index is defined as follows:

 

 

Where, SSTA represents the area averaged SST over central (165°E-140°W, 10°S-10°N), Eastern (110°-70°W,15°S-5°N), and Western (125°-145° E, 10°S-20°N). EMI refers to the El Niño Modoki Index.

The index and the decomposed modes are illustrated in figure (5)

 

Please refer to the equation (4) on line no. 194.

 

Reviewer’s comment Added legends

Authors’ response: We have added legends to the axis in all the figures. Please see figure (2-12)

 

Reviewer’s comment: repeated word

 Authors’ response: Repeated words are removed.

 

Reviewer’s comment: What kind of physical mechanism can explain a relationship with a 40 year time lag? What is the uncertainty on this results given the relatively short time series available (For a 40 year cycle you should have a more that 100 year long data series to resolve this?

 

Authors’ response: We agree that the dataset is not long enough to place our confidence in such a long term cycle. So, considering your suggestion, we have removed the result stating a 40-years lag.

 

 

 

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors have made extensive modifications to their manuscript so that it has improved and is now suitable for publication.

Reviewer 2 Report

The authors made numerous changes that improved the overall quality of the manuscript; however, I still do not find that the data sufficiently support the primary conclusion that the IOD acts as an energy exchange mechanism between El Nino Modoki and TCPI.  Two examples in Tables 12 and 13 are used to illustrate the DMI as a bridge between the EMI and TCPI, which the authors point out in lines 755-759 of the revised manuscript.  Given the long periods of the modes examined, the 60-year dataset does not allow for such conclusions to be drawn.  Perhaps the authors could consider higher frequency IMFs?

Additionally, the authors note that the El Nino Modoki affect on TCPI may be related to the number of TCs that form (e.g., 559 - 567).  Given this, the relationship between El Nino Modoki and TCPI should probably be conditioned on the number of TCs.  Or perhaps, to keep things simpler, the authors could consider a genesis potential index in place of potential intensity.