Towards a New Way of Understanding the Resilience of Socio-Technical Systems: The Safety Fractal Analysis Method Evaluated

Round 1

Reviewer 1 Report

The manuscript is to validate SAFRAN, a method for accident analysis proposed by the authors, in comparison with other methods, and it will certainly fit the scope of the journal. The authors, however, are advised to brush up the manuscript for publication referring to the following comments.

1. Though the SAFRAN method was briefly described in the introductory chapter, the reviewer could not get an overview of the method at first. It is better to make another chapter to introduce the SAFRAN method separately from the introduction for those who are unfamiliar with the method. In addition to the iterative analysis steps, the accident model, which is given in Subsection 3.1.2 (Line 282-301), and the graphical representation of outcomes, which is given in Subsection 3.1.3 (Line 383-401), of the method can be rearranged into the new chapter.

2. The terms 'model' and 'method' are used exchangeably somewhere in the manuscript, for instance in Line 94 and Figure 1. Clearly define what the model means here and distinguish the both terms.

3. The merits of the SAFRAN method against others by various authors provide the evidence for the conclusion, but just a part of the comparison, 3, 4, and 5 of 32, is given as Appendix B. It is better to include the whole table, not in detail but summarized like a correspondence table or win/loss table.

4. Some English errors and typos.
 Line  13: limit resources -> limited resources
 Line 239: drawn.
 etc.

Author Response

See attached file

Author Response File: Author Response.docx

Reviewer 2 Report

Thank you for an insight in to the development of a novel accident investigation technique. I belive it to be worthy of investigation and presentation but feel that the evidence to conclude the method superior to others is not well structured: I believe it is there but it is very hard to follow. I propose clarifications on this point. 

Major comments:

- Section 3 contains much, perhaps too much explanation but is not overly clear. Please, could you consider summarizing on each point (A1, A2..) or perhaps on a higher aggregate level (a, b..). Things I expected to see are (perhaps in bullet points?). This part is superior to SPARK because...; or this is how resilience is served better than other methods; or users expressed the following advantages. I believe most of it is there but it is a bit of a puzzle finding it. 

- the findings and discussion are blended in chapter 3; please separate a bit more to put evidence in chapter 3 and speculations in chapter 4. 

Minor comments:

- P1: the abstract is unclear as to the objectives of the paper: what do you want to do and did you achieve it?

- P2 & 3 different line spaces 1 & 1,5?

Figure 2 is hard to read: too small letters. Perhaps enlarge a specific section for clarity? 

- Discussion p22, first sentence. After reading all this I’m not so conviced that it is better.

- Lack off appropriate analysis methiods (p22, l 825) I would advise a bit more caution with this commment. If that were true, how did we survive till now?

Line 837 - 841. Is the reduction of variability a resilient solution? Resilient would be to DEAL with variability, not MINIMISING it. Minimising it is actually moving toward traditional barrier management, not away from it. 

 - Even at the conclusion it is not especially clear to me how resilience benefits; I can see it for (traditional) barrier management but not for resilience. 

Line 878 - 880. Same as before, if SMS analysis is poor, how come we don't constantly blow up stuff? I find the current organisation of the paper insfficiently well structured to support this claim. 

Author Response

See attached file

Author Response File: Author Response.docx

Reviewer 3 Report

The presented article deals with application and evaluation of the innovative method SAFRAN, which is determined for analysis of various accidents. The reviewed article fully fits into the conception of this journal with regard to its content as well as from the professional point of view. The structure of article is well-arranged and its elaboration fulfils the defined professional and formal requirements. I also appreciate a fact that the authors applied the SAFRAN method on the real railway-traffic accident, namely on the fatal Grayrigg derailment accident. This aspect makes the article interesting and original. The appendixes A and B are informatively valuable for the readers, as well. The scientific level of the article is sufficient and suitable for publishing in this journal. I have some comments to the article. After their implementation I recommend this article for publishing.

1. I recommend, within the chapter Introduction, to present the method SAFRAN in a little-bit more explanatory way with regard to a better understanding of it for those readers who are not sufficiently familiar with it.

2. It could be interesting for the readers to put into the article some illustrative photo-material relating to the Grayrigg derailment.

3. The Figure 2: “SAFRAN representation of the Grayrigg accident” is too-complex, it is not very clear and thus it loses the informational function. The same scheme is also presented as the Appendix A. Therefore, I recommend to consider a necessity of Figure 2 or a suitable modification of it.

4. In general, it is possible to state that text of the whole article is illustrated using only few figures. I suggest the authors to consider addition of other illustrative figures into the text if any are so appropriate.

Author Response

See attached file

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Correct duplication of 'that' in Line 894.

Reviewer 3 Report

I do not have comments and suggestions for authors

after improvement  of their manuscript and  I reccomend to publish it in the presented form.