Vulnerability of the Maritime Network to Tropical Cyclones in the Northwest Pacific and the Northern Indian Ocean

Round 1

Reviewer 1 Report

Well revised.

Author Response

Thank you for your comments!

Reviewer 2 Report

The manuscript is well-written and the topic is interesting. However it can be improved by addressing the following:

The impact of tropical cyclones, especially in terms of losses/damages and consequences to the industry should be made more apparent in Section 1.  Statistical evidence on expansion of maritime trade and frequency of cyclones should be presented when there are mentions of these in the first paragraph. Section 1, lines 65-67: how will vulnerability be studied? Would be good to briefly introduce this here. Parts of Section 1 are actually a review of previous work done. I suggest that these be carved out to be under a new Section named Literature Review.    There are errors in the figure numbers in the main text. Please correct these. Section 3.2, lines 222-227: Is it possible to present the results in a table form, instead of in a summarised manner? The real life scenario that sees the "deletion" of a series of ports would be cyclones that affect many different ports at the same time. We do not see this often. How would your study remain meaningful regardless of this? When one port is down, diversion of cargo and ships occur. Has this been taken into account in your methodology or implications? Also, when one port is removed from the network, there would be other remaining ports that become more important in the network as cargo and ships will be diverted to these remaining ports. Does your methodology address this changes in importance every time a port is removed?  Section 4 should also discuss about the implications of the results from different perspectives of the industry stakeholders.  Section 5, line 346, "convenience and shipping costs": these were not illustrated by your results. 

Author Response

Thank you for you comments! Please check the attached files for the detail responses.

Author Response File: Author Response.pdf

Reviewer 3 Report

This article deals with an interesting topic that is, how a spatial network and its nodes become affected by disruptions, in this case storms and cyclones. The main problems of the article is that it deals with "potential" disruptions, not actual disruptions. That could work, but given the spatial spread of potential risk, it fails to prove the precise effects of storms on the maritime network (liner shipping). Perhaps, a different categorization of storms should be used. Detailed comments can be found below.

 

abstract
- the structure of maritime networks: better announce the topic by dealing with circulation, traffic in general?
- findings are mainly regional-based, what about wider implications for network studies and maritime studies?

introduction
- northwest pacific: do you mean northeast asia?
- the maritime network = nodes and routes; yes when seen as a graph, but more thought should be given to this approach as it is not the only possible one
- fixed route: only for liner shipping
- the literature review should be better structured, as it is mixing maritime network studies and cyclones definitions

materials and methods
- it should be discussed what would be the influence of year gaps between cyclone data and shipping data
- 4 grades to classify climatic events: is this a conventional method?
- what is "usage degree"?
- why using the deletion/removal method if only part of port activity is affected? this would need more insights from qualitative case studies

results
- obviously different: yes, but how?
- it is more useful to map the inverse of the clustering coefficient to see the main hubs, keeping zero values unchanged
- why not the betweenness centrality; and how is the ASPL calculated at each port as it is a global measure?
- multihub axial radiative network: impossible to understand
- fig 2 does not show a power law, as the exponent falls under 1; it should be between 2 and 3 according to Barabasi
- pole-axial structure: see above, better use understandable concepts
- the ports require 2 transit steps: hard to follow, needs further explanation
- figure 4: better use a single color and a grayscale from dark red to light yellow to better see the variation
- same for figure 5, and give numbers to each class
- indexes: indices
- table 3: this table does not support the analysis; perhaps it should be complemented by correlations?
- table 4 should be mapped instead to better see the results

Author Response

Thank you for you comments! Please check the attached files for the detail responses.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors have adequately addressed the comments. 

Author Response

Thank for you comments! The latest manuscript has been polished up by a native speaker.

Reviewer 3 Report

I do not think that the authors did a sufficient work. I therefore insist more about the necessity to carefully listen to the critiques and address them one by one. Often the responses consisted in "we have corrected this line xxx", but there is only limited proof that the authors took a sufficient amount of time to rethink their approach and mistake. Another example is when authors refuse to change a figure without a strong argument. Lastly, it is not "academic" to say that because the existing literature says something, it just should be followed. In the case of scale-free exponents, you should not listen to existing maritime studies but to founders of (complex) network analysis, like Barabasi. Replying that you want to follow existing studies is not respecting the reviewer's judgement. Basically I found that all of the changes were not respecting the reviewer's judgement.

Author Response

Thank you for your comments again. We apologize that our changes did not respect your judgement. We have added the requested information in the latest manuscript. Moreover, we have changed Fig 1 and revised the analysis of the port degree distribution. We hope our manuscript earns your approval.

Author Response File: Author Response.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Review comments

 

It is an interesting research to analyze the network robustness under the scenarios of tropical cyclones.

The followings are comment to improve the quality of paper if it will be revised.

 

Title The looks very long. I suggest authors to rephrase the title as more compact one.

 

Introduction The contribution of the paper is not clear. Is the main contribution is to analyze the vulnerability of maritime networks under a new dimension of natural hazards? If the authors focused on driving useful policy implications, the discussion and conclusion parts are weak. 

 

2.1 Data Authors need to describe the characteristics in more detail. Is the ship schedule data yearly one or for certain period of time? If we assume the tropical cyclones give the impact during certain period of time, using the yearly data may not be a good idea.

 

88-89 How to build the network when there is multiple ship schedules between ports? Did you dichotomize the weight?

 

135 and equation (5) How authors handle the shortest path when two ports are unreachable due to cyclones?

 

2.2.2. Complex network method Section title is not clear. Do you mean the measures for network robustness? In addition, why authors do not consider the size of giant component which is commonly used for robustness analysis?

 

2.2.3 Analysis of network vulnerability Authors need to justify the node removal scenario. Current node removal scenario by removing the ports with high hazard looks fine. However, if we assume that the tropical cyclones usually occurred in specific geographical region, authors may need to consider other scenario by removing ports in specific region(s).

 

Table 2. Degree means the number of connections (binary variable) or the schedule frequency?

 

212 The argument is not clear. What is the relationship between the hazard levels and the network measures?

 

Table 4. Why the average shortest path lengths of some countries are decreased?

 

Discussions and conclusions I strongly suggest the authors to enrich discussion and conclusion. Since the paper does not focus on the methodological improvement, the authors need to add more policy implications from the results. Need to discuss limitations and future research areas too.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Section 2.2.2: I wonder the definition that each port is connected in this paper. Only when two ports are directly connected without any stops between them, is it regarded as they are connected? Or all of (more than three) ports that one liner service call at are defined as connected each other? I guess the authors define it as the former meaning, but I think the connectivity in the latter meaning is more important.

 

What kind of significant negative impacts are expected if a cyclone attack the container port? In most cases, it can only last in the very short term (a few days at most). Do the liner shipping companies or terminal operators consider that it is significant risk for their business? Did the authors interview with them? I think some ports (e.g. those in the Philippines) are more vulnerable for natural hazards due to lacking of infrastructure, but for example, most Japanese ports already have some countermeasures against such risks. In that sense, the authors’ conclusion that the network metrics in the Philippine ports are the highest is interesting for me.

 

How the authors chose the hazardous ports to delete for each interval of 10%? The graphs shown in Figure 5 are heavily depending on their order.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

see attachment

Comments for author File: Comments.pdf

Reviewer 2 Report

I understand the authors' responses to my comments, but those to my second and third comments should be reflected in the manuscript.