Life Cycle Assessment of Oyster Farming in the Po Delta, Northern Italy
Round 1
Reviewer 1 Report
Dear Authors,
There are many flaws in this article, which have prevented me from accepting it for publication.
1.Scenario definition: how can you compare a cradle-to-gate with a gate-to-gate scenario?
2.Inventory data are not reported, except for few figures
3.Why using ecoinvent 2.2 for a 2018 dataset?
4.Ecoindicator 99 is a superseded impact method; why not using ReCiPe? why not following JRCrecommendations
ILCD handbook "Recommendations for Life Cycle Impact Assessment in the European context" http://eplca.jrc.ec.europa.eu/uploads/ILCDI -Recommendation-of-methods-for-LCIA-def.pdf.This article may be useful:
Nisticò, R. (2017). Aquatic-derived biomaterials for a sustainable future: A European opportunity. Resources, 6(4), 65.
I suggest reorganising the materials and methods section towards the 4 LCAphases
Author Response
Dear Authors,
There are many flaws in this article, which have prevented me from accepting it for publication.
1.Scenario definition: how can you compare a cradle-to-gate with a gate-to-gate scenario?
That was a hard point of discussion even among Authors. So, thank you for permitting to explain the point. I personally support the issue that only apparently we are comparing two different systems. We could neglect the contribution of the step seed production in France, because extending the boundaries to the sole transportation is enough to support our conclusion that in-house seed production would be more convenient from the environmental point of view. Including the seed production in France could only have got worse, but not changed the results.
We are comparing two production systems rather than two processes with the same boundaries, with the aim to understand which one should be chosen. Of course, it is worthwhile noting that it is almost impossible to obtain data from French supplier.
As reported in the JRC Recommendations, we are in the case of a comparison of systems that are not fully comparable, or where comparability and equivalence is principally a matter of personal perception.
Concluding, I reckon comparing the two systems in this way is consistent with the LCA and permitted a realistic interpretation of data.
2.Inventory data are not reported, except for few figures
We have now added a more complete inventory data in table 1.
3.Why using ecoinvent 2.2 for a 2018 dataset?
We really apologize for the oversight and a bad typo from previous studies. We have now corrected. We have used ecoinvent 3.3 version
4.Ecoindicator 99 is a superseded impact method; why not using ReCiPe? why not following JRCrecommendations
ILCD handbook "Recommendations for Life Cycle Impact Assessment in the European context" http://eplca.jrc.ec.europa.eu/uploads/ILCDI -Recommendation-of-methods-for-LCIA-def.pdf.
We did not use ReCiPe for the main analysis of oyster production because, as specified in the paper, the unique article published on LCA applied to oyster production had used Ecoindicator 99. We had considered more confident to have the opportunity to compare the results of our study with others.
As general view, I can agree to your comment, in fact we have decided, thanks to your comment and to the comments of the other Reviewer to use ReCiPe, instead of CML, when impact on 1kg of fresh oyster has been calculated in order to compare the results with other species.
Staying on your side, we had to do this (namely calculating impacts with two methods) just because it was very difficult to find other studies in literature that had applied Ecoindicator in their analysis. So firstly we have compared our study with the only one on oysters, then we have to recalculate the impact using a widespread method, in order to compare our study with others on other species.
Anyway, JRC recommendations still includes Ecoindicator among the usable methods.
This article may be useful:
Nisticò, R. (2017). Aquatic-derived biomaterials for a sustainable future: A European opportunity. Resources, 6(4), 65.
Thank you for the alert, we have now added in the introduction data taken from the article and correspondingly cited it in the references list, to complete the general framework of Italian ostriculture.
Please, see the Introduction, at lines 74-76 (text highlighted in yellow, as well as the added references)
I suggest reorganising the materials and methods section towards the 4 LCAphases
We have now changed titles and order within the M&M section, in order to be more fitted to the 4 LCA phases. Thank you for your suggestion. Please see the text at the corresponding lines (text highlighted in yellow).
Author Response File: 
Author Response.docx
Reviewer 2 Report
I thank for the opportunity to review this paper. Generally, the paper is well-written and addresses a relevant aspect of aquaculture farming using LCA and primary data. However, although the paper is worked out well, the clarity of the results and their interpretation is missing, or at it is difficult to follow your argumentation and to understand the main results. The inventory and the data that are used in your article seem to be good, however, I would recommend to run a sensitivity analysis on your model to identify the most sensitive (and most relevant) parameter determining your results. So far you only vary the transport in two scenarios. Particularly, when using ecoinvent 2.2 (the newest is version 3.5!), you should be aware of the quality and age of the data you have used to come to your conclusions or at least reflect on this issue in your discussion (or e.g. you can add additional pedigree matrix approach for your data). Furthermore, you have used the ecoindicator99 (endpoint) and CML midpoint for the LCIA, for which you did not give a reason, why you have used them. There are some newer LCIA methods at hand such as ReCiPe2016 or, it would be good to know if you have followed the ILCD handbook (that compared different LCIA methodologies). I would recommend to focus only on one LCIA methodology otherwise it will get confusing for the reader, why in ecoindicator99 (at aggregated level) your scenarios looked the same but in CML you do see a difference (at midpoint indicator level). There are some typos in the document, which should be corrected after revision. (particularly decimal numbers)
In the following some minor comments:
L24: I would recommend to check the words since LCA analysis is not so commonly used.
L99-101: This sentence is not clear to me, the month are the duration of the one cycle stage?`
L117ff: here you talk about scenarios that haven’t been introduce yet. Please revise the Material and method section. For instance, introduce the scenarios later after describing the main aspects of your study.
L125-126:Doubled with line 100ff
L122+23: check points and comma for numbers throughout the whole text. In English decimal number are: 0.01 and not 0,01.
L153: use abbreviations when you have introduced them earlier, such as FU.
L162: which unit does 110/year have?
Table1: caption should mention the assumed lifespan to understand the numbers listed in the inventory. Maybe it would be good to have a more detailed inventory based on figure 3 processes and add the references accordingly.
Please revise the decimal numbers “.” Vs “,”
Table 3: also here check the decimal “.” Vs “,”
L214: Why and what is surprising? You have only varied the transport, which is not the main contributing process in your system as pointed out later. Which also is why, I would argue that the ecoindicator endpoints impact categories are not adequate for your study.
Figure 4. The graphic is nice and unconventionally designed, but its interpretation is difficult. In order to transport your message, I would recommend to revise it completely, such as a bar plot where you can compare values.
Figure 5. which LCIA method do you refer to? What is the intention of the figure to show no differences between the scenarios?
Figure 6. see comments about decimal numbers (i.e. y axis). What is the unit? Which LCIA method did you use for this graph?
L317. What is a high burden? Do you have quantitative numbers?
Author Response
I thank for the opportunity to review this paper. Generally, the paper is well-written and addresses a relevant aspect of aquaculture farming using LCA and primary data. However, although the paper is worked out well, the clarity of the results and their interpretation is missing, or at it is difficult to follow your argumentation and to understand the main results. The inventory and the data that are used in your article seem to be good, however, I would recommend to run a sensitivity analysis on your model to identify the most sensitive (and most relevant) parameter determining your results. So far you only vary the transport in two scenarios. Particularly, when using ecoinvent 2.2 (the newest is version 3.5!), you should be aware of the quality and age of the data you have used to come to your conclusions or at least reflect on this issue in your discussion (or e.g. you can add additional pedigree matrix approach for your data). Furthermore, you have used the ecoindicator99 (endpoint) and CML midpoint for the LCIA, for which you did not give a reason, why you have used them. There are some newer LCIA methods at hand such as ReCiPe2016 or, it would be good to know if you have followed the ILCD handbook (that compared different LCIA methodologies). I would recommend to focus only on one LCIA methodology otherwise it will get confusing for the reader, why in ecoindicator99 (at aggregated level) your scenarios looked the same but in CML you do see a difference (at midpoint indicator level). There are some typos in the document, which should be corrected after revision. (particularly decimal numbers)
Thank you very much for your suggestion. We overlooked uncertainty analysis in the first version of the paper, even though it is really important for this kind of analysis. Anyway, we have now added the Monte Carlo simulation. Please see the text.
According to ILCD recommendation, and taking into account your suggestion, we recalculate impacts using ReCiPe method, instead of CML initially chosen.
The reason for applying two methods was because the first (ecoindicator) permitted us to compare our results with the unique published on oyster farming (expressed as ecoindicator impact categories), the second to compare our results with other studies on other aquaculture species.
In the following some minor comments:
L24: I would recommend to check the words since LCA analysis is not so commonly used.
We really apologize, it was a bad out of habit! We have now corrected analysis with assessment
L99-101: This sentence is not clear to me, the month are the duration of the one cycle stage?`
Please see the following reply
L117ff: here you talk about scenarios that haven’t been introduce yet. Please revise the Material and method section. For instance, introduce the scenarios later after describing the main aspects of your study.
We really apologize, we have now introduced the scenarios at the end of introduction section, in order to permit to the reader to better understand the experimental development
L125-126: Doubled with line 100ff
We apologize for the oversight; we are now move lines 125-126 immediately after line 100.
L122+23: check points and comma for numbers throughout the whole text. In English decimal number are: 0.01 and not 0,01.
We have now corrected. Please see the text
L153: use abbreviations when you have introduced them earlier, such as FU.
Sorry for the bad typo. We have now corrected. Please see the text
L162: which unit does 110/year have?
The unit is number of boat runs, as reported at the beginning of the sentence.
Table1: caption should mention the assumed lifespan to understand the numbers listed in the inventory. Maybe it would be good to have a more detailed inventory based on figure 3 processes and add the references accordingly.
The rationale under the inventory is to report the amount of inputs necessary to produce 12 fresh oysters (1 kg), using a mass rather than time criterion. Based on comments of other Reviewer we have now added other inventory data.
Please revise the decimal numbers “.” Vs “,”
Table 3: also here check the decimal “.” Vs “,”
We apologize for the typos, we have now corrected decimal numbers both in the text and tables.
L214: Why and what is surprising? You have only varied the transport, which is not the main contributing process in your system as pointed out later. Which also is why, I would argue that the ecoindicator endpoints impact categories are not adequate for your study.
It was considered quite surprisingly because actually we were comparing transport from France with production in situ, that means the need of more resources in terms of energy, water, nutrients. I sincerely expected that all the nursery and hatchery phase would be more burdening than the sole transport from France.
Figure 4. The graphic is nice and unconventionally designed, but its interpretation is difficult. In order to transport your message, I would recommend to revise it completely, such as a bar plot where you can compare values.
Yes, the graph is nice but probably not so easy to read, especially because the two groups are almost identical. We have accepted your suggestion and change the graph.
Figure 5. which LCIA method do you refer to? What is the intention of the figure to show no differences between the scenarios?
We specified the LCIA to which the figure is refereed. Yes, it was exactly our intention, that graphically can be better appreciate by the reader
Figure 6. see comments about decimal numbers (i.e. y axis). What is the unit? Which LCIA method did you use for this graph?
Thank you for the careful reading. We apologize for the oversight. We have now added the units, corrected the decimal numbers and added the LCIA method which the figure is referred.
L317. What is a high burden? Do you have quantitative numbers?
The comment was referred to the comparison of impact due to transport and the local seed production, so the “high impact” had to be read in that sense. Anyway, thanks to your comment we have delete the high burden. Please see the text.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
Dear Authors,
I acknowledge your effort in replying the comments to the previous version of the article. To strengthen your methodological choices, the replies about system boundaries/comparability, lcia, including relevant related results, should be presented in the main manuscript text. The use of ecoinvent 3.3 should be motivated as well, given the availability of 3.5 for 2018 data.
Author Response
REPLY TO REVIEWER #1
Dear Authors,
I acknowledge your effort in replying the comments to the previous version of the article. To strengthen your methodological choices, the replies about system boundaries/comparability, lcia, including relevant related results, should be presented in the main manuscript text. The use of ecoinvent 3.3 should be motivated as well, given the availability of 3.5 for 2018 data.
Thank you for your further comments, we have now added some more considerations in the text (please see lines 309-316 and 355-359).
Regarding to the use of the latest released version of Ecoinvent, in the meanwhile the 3.6 version has been released the last 12nd September, so we have remade all calculation using that new version.
Reviewer 2 Report
I thank the authors for answering and incorporating most of my comments. Generally, the manuscript is now well structured and good written so that the reader can follow your argumentation. You have added a monte-carlo-simulation (MCS) covering potential uncertainties, which I do appreciate. However, to my mind, it would strengthen the manuscript, if you could discuss your results together with this uncertainty analysis (i.e. maybe an extra subsection including the study limitations). Thus far, you discuss your results within the result section, which is fine for me, and add a MCS on top without discussing it. Also, you have mentioned that some processes (disposal of baskets coated with anti-fouling) are not within the system boundaries, but should be included in a cradle-to-grave analysis. This could be integrated in the conclusion as an outlook on future work such as extending system boundaries (is it required?) or what have you not covered in your analysis but should be included in future studies? Additionally, I would recommend some professional proof reading since some typos and grammar errors are in the whole manuscript. Below you will find some minor comments.
Title.: I would remove the abbreviation in the title since you have this included in the keywords.
Line 18: only EcoIndicator?
Line 24: LCA means life cycle assessment, so LCA assessment is redundant.
Line 74: In 2016, the Italian…
Figure 2 seems to be a copy from an original. Please insert reference.
Table 1. Thank you for revision of the table. Nevertheless, it would contribute to understand your model when you can provide the processes selected from ecoinvent (i.e. foreground and background system) at least in the supporting information for the LCI used.
Figure 4. The main intention of the graph is to compare the results of both scenarios. I would recommend to create one or two graphs only and incorporate the categories accordingly. Since you use relative shares it is easier to compare the results in one graph.
Figure 5. Thank you for your answer. I still don’t understand the value-added to have to graphs showing the exact same thing. I would appreciate a reply and reasoning to my previous comment.
Table 3. Please check decimal numbering (point instead of comma).
Table 4. Please check decimal numbering (point instead of comma).
Table 6. Why have you selected only three impact categories?
L299. Based on the observations, it is not surprising….
L349. RECIPE should be written as ReCiPe
L352. Same here. ReCiPe
L356. I would avoid to name it more sustainable, since you are only focusing on environmental impacts and not on social or economic issues.
L357. … does not require feeding supply
L383. What are your recommendations for future research? Since you are one of the first authors who applied an LCA to Oyster farming it would helpful to know where improvements of your study or of the studied processes can be expected or should be investigated.
Author Response
REPLY TO REVIEWER #2
I thank the authors for answering and incorporating most of my comments. Generally, the manuscript is now well structured and good written so that the reader can follow your argumentation. You have added a monte-carlo-simulation (MCS) covering potential uncertainties, which I do appreciate. However, to my mind, it would strengthen the manuscript, if you could discuss your results together with this uncertainty analysis (i.e. maybe an extra subsection including the study limitations). Thus far, you discuss your results within the result section, which is fine for me, and add a MCS on top without discussing it. Also, you have mentioned that some processes (disposal of baskets coated with anti-fouling) are not within the system boundaries, but should be included in a cradle-to-grave analysis. This could be integrated in the conclusion as an outlook on future work such as extending system boundaries (is it required?) or what have you not covered in your analysis but should be included in future studies? Additionally, I would recommend some professional proof reading since some typos and grammar errors are in the whole manuscript. Below you will find some minor comments.
Thank you again for your careful reading. We tried to strengthen our discussion about MSC and integrate the conclusion section with some consideration bout future developments of research. Please see the text at lines 354-361 and 410-416 respectively and table 4.
Title.: I would remove the abbreviation in the title since you have this included in the keywords.
OK, done. Please see the title.
Line 18: only EcoIndicator?
Of course not, we apologize for the oversight. We have now corrected. Please see line 18
Line 24: LCA means life cycle assessment, so LCA assessment is redundant.
Sorry. Corrected.
Line 74: In 2016, the Italian…
Sorry. Corrected.
Figure 2 seems to be a copy from an original. Please insert reference.
It is not. We have done it using Paint and Powerpoint programmes. The low resolution is probably due to several cut-and-paste from one to the other and viceversa.
Table 1. Thank you for revision of the table. Nevertheless, it would contribute to understand your model when you can provide the processes selected from ecoinvent (i.e. foreground and background system) at least in the supporting information for the LCI used.
We have now added as Supplementary materials the background processes for the two scenarios.
Figure 4. The main intention of the graph is to compare the results of both scenarios. I would recommend to create one or two graphs only and incorporate the categories accordingly. Since you use relative shares it is easier to compare the results in one graph.
Thank you very much for your suggestion, and for emphasizing the need of one graph. I completely agree with you, now I think it is more readable and easier to understand. Please see the new Fig.4
Figure 5. Thank you for your answer. I still don’t understand the value-added to have to graphs showing the exact same thing. I would appreciate a reply and reasoning to my previous comment.
Our aim was to stress the point – that is that the in house production has the same impact of the sole transportation from France and that the actual impacting phase are fattening. But really it could be seen also from the previous figure, so, figure 5 it is quite redundant and has been deleted from the manuscript.
Table 3. Please check decimal numbering (point instead of comma).
Table 4. Please check decimal numbering (point instead of comma).
Oh my God, it is really unforgivable. I’m sorry, now we have corrected the Italian notation.
Table 6. Why have you selected only three impact categories?
Because the three selected (+ water depletion in one case) are represented in all papers regarding LCA and aquaculture, whereas the others are sometimes reported and sometimes not, making difficult to compare different results. Moreover, climate change, eutrophication and acidification are the most important in terms of values, the others usually reach values of 10-5 or more as order of magnitude. We have now added a short explanation of this at line 354-359. Please see the text.
L299. Based on the observations, it is not surprising….
We apologize for the oversight. We have now corrected in the text
L349. RECIPE should be written as ReCiPe
We apologize for the oversight. We have now corrected in the text
L352. Same here. ReCiPe
We apologize for the oversight. We have now corrected in the text
L356. I would avoid to name it more sustainable, since you are only focusing on environmental impacts and not on social or economic issues.
Yes, it could be misunderstood. We have now added “at least from the envinmental point of view”. Please see lines 364-365
L357. … does not require feeding supply
We apologize for the oversight. We have now corrected in the text
L383. What are your recommendations for future research? Since you are one of the first authors who applied an LCA to Oyster farming it would helpful to know where improvements of your study or of the studied processes can be expected or should be investigated.
Thank you for the suggestion. We have now added some recommendation for future research. Please see the text in the section at the end of conclusion section.
