Optimizing Costs to Collect Local Infauna through Grabs: Effect of Sampling Size and Replication
and Fernando Tuya 4,*
Round 1
Reviewer 1 Report
This study was performed to determine if there is an grab sample size for assessment that is most efficient, where efficiency is defined as the representativeness divided by time to process the sample. I think it is a worthy study because benthic community structure is an important measure of sediment quality and ecosystem health. But the cost of benthic analyses has always been a challenge leading to many studies that did not take enough samples to answer questions properly.
General comments:
Introduction should mention species-area relationships, and its relevance to the study.
Sediment grain size matters and leaving this out of the literature review analysis is a problem. I suspect the statistical relationships would look very different if grain size were accounted for.
Restricting the literature review to Van Veen grabs means that you ignored a lot of methods studies that used other types of grabs and all coring devices. I agree that this restriction is the right approach for comparing studies statistically, but I think this biases your conclusions in the discussion because you are not taking into account all that is known about sampling issues, particularly sediment structure, and sieve size.
Are all the studies listed in table S1 in the reference list, if not, I think the Supplement should also have a Reference list.
The discussion is ok, but it would be nice if there were some more generic recommendations for designing monitoring studies, and/or some discussion of the science (i.e., why the recommendations make sense based on what we know about benthic ecology). At the least, some mention on how the recommendations square with what we know about benthic ecology.
Specific comments:
L14: was the purpose of the lit review to "demonstrate," i.e., confirm something you already knew, or to "discover?" Also, abstracts are better in the third person, so I suggest not using "we" in the abstract. Perhaps start sentence with "A literature review discovered that (1) …"
L17: add something about the site, e.g., muddy or sandy sediments.
L19, 20 23: we.
L24: precision? Probably not the right word.
L27: replace "6.87 l "grab with either G1, G2, or G3, whichever is correct. Also, replace "in conclusion" with something more descriptive, such as "Because of high representativeness and low cost, G? appears to be the optimal sample device to assess infauna in the seabed off Gran Canaria Island".
L36: replace properties and delicate with better words.
L46: Better Google the difference between accuracy and precision. I think you really mean accuracy.
L52: this paragraph needs a mention of the species-are relationship.
L59: Even worse is too few number of locations or treatments to adequately answer questions. Again accuracy, not precision.
L75: rewrite to explain conditioned.
L78: what kind of sediments.
L96-98: punctuation goes inside the closing quotes.
L116: Provide more details on how samples were sieved in the lab. Previous studies have shown more animals are lost in big samples because the washing procedure is inherently more sloppy. May have missed this because you restricted the search to vanveen grabs.
L123: Again, why not sediment type?
L158: I think you should put in some explicit formulas for calculating costs in this paragraph so that this is easy for someone to replicate.
L172: did all the studies use a 1 mm sieve? Needs to be mentioned in the methods. If not, needs some kind of adjustment.
L173: define edf.
L175: The same precision, i.e., number of significant digits, should be provided for all the p values throughout the manuscript. for Table 2, this means use <0.001 or <0.0001, whichever precision you decide to use.
L308: this is the G1 grab, so use that as well.
Author Response
Reviewer 1: (reviewer’s comments in italics)
Introduction should mention species-area relationships, and its relevance to the study.
Reply: we have included the following sentence (Line 61) in this paragraph: “This is because a larger area tends to include a larger number of species [1, 2, 4].”
Sediment grain size matters and leaving this out of the literature review analysis is a problem. I suspect the statistical relationships would look very different if grain size were accounted for.
Reply: we did not include sediment type in our analysis because, contrary to depth, this is a categorical variable (i.e., sandy, muddy, etc), which is hardly to model via statistical GAMs. Moreover, sediment type tends to be correlated with depth, because deep-water samples tend to be muddy relative to shallow-water samples, where sediments tend to be coarser. As a result, despite we acknowledge the relevance of sediment type as a key determinant of infaunal assemblage structure, we decided to maintain this variable out of the scope of this paper. This limitation has been indicated in the discussion – see the following point in this sense.
Restricting the literature review to Van Veen grabs means that you ignored a lot of methods studies that used other types of grabs and all coring devices. I agree that this restriction is the right approach for comparing studies statistically, but I think this biases your conclusions in the discussion because you are not taking into account all that is known about sampling issues, particularly sediment structure, and sieve size.
Reply: despite we emphasize with the reviewer, we must restrict our literature review to Van Veen grabs to avoid including confounding factors that can seriously affect the relationships we tested in this study. In any case, to recognize these limitations, we have included in the discussion the following paragraph: “It is important to highlight that our literature review have two limitations. First, the outcomes are limited to studies implemented through collection of infauna by Van Veen grabs. In addition, our study has not considered variation in sediment grain size, as a relevant factor affecting the diversity and abundances of infauna.”
Are all the studies listed in table S1 in the reference list, if not, I think the Supplement should also have a Reference list.
Reply: we have included in the Table S1 a new column with the cite of each considered paper.
The discussion is ok, but it would be nice if there were some more generic recommendations for designing monitoring studies, and/or some discussion of the science (i.e., why the recommendations make sense based on what we know about benthic ecology). At the least, some mention on how the recommendations square with what we know about benthic ecology.
Reply: we have included some general recommendations at the end of the discussion.
Specific comments:
L14: was the purpose of the lit review to "demonstrate," i.e., confirm something you already knew, or to "discover?" Also, abstracts are better in the third person, so I suggest not using "we" in the abstract. Perhaps start sentence with "A literature review discovered that (1) …"
Reply: we have accordingly changed the abstract following these instructions. The passive voice has been always used in this sense.
L17: add something about the site, e.g., muddy or sandy sediments.
Reply: we included “sandy” to clarify this.
L19, 20 23: we.
Reply: the passive voice has been now used.
L24: precision? Probably not the right word.
Reply: accuracy has been used throughout the entire manuscript.
L27: replace "6.87 l "grab with either G1, G2, or G3, whichever is correct. Also, replace "in conclusion" with something more descriptive, such as "Because of high representativeness and low cost, G? appears to be the optimal sample device to assess infauna in the seabed off Gran Canaria Island".
Reply: both suggestions have been accommodated.
L36: replace properties and delicate with better words.
Reply: “delicate” has been replaced by “difficult”.
L46: Better Google the difference between accuracy and precision. I think you really mean accuracy.
Reply: accuracy has been used throughout the entire manuscript.
L52: this paragraph needs a mention of the species-are relationship.
Reply: we have included the following sentence (Line 61) in this paragraph: “This is because a larger area tends to include a larger number of species [1, 2, 4].”
L59: Even worse is too few number of locations or treatments to adequately answer questions. Again accuracy, not precision.
Reply: accuracy has been used throughout the entire manuscript, as indicated above.
L75: rewrite to explain conditioned.
Reply: we have replaced “conditioned” by “affected”.
L78: what kind of sediments.
Reply: “sandy” has been included in this sense.
L96-98: punctuation goes inside the closing quotes.
Reply: corrected at all sites in these sentences.
L116: Provide more details on how samples were sieved in the lab. Previous studies have shown more animals are lost in big samples because the washing procedure is inherently more sloppy. May have missed this because you restricted the search to vanveen grabs.
Reply: The same careful procedures were undertaken irrespective of grab size, which was in turn reflected by varying sorting times between grab sizes. Despite we believe this point should be inherently understood by readers, we have added “carefully” in the sentence to shed light on this.
L123: Again, why not sediment type?
Reply: we did not include sediment type in our analysis because, contrary to depth, this is a categorical variable (i.e., sandy, muddy, etc), which is hardly to model via statistical GAMs. Moreover, sediment type tends to be correlated with depth, because deep-water samples tend to be muddy relative to shallow-water samples. As a result, despite we acknowledge the relevance of sediment type as a key determinant of infaunal assemblage structure, we decided to maintain this variable out of the scope of this paper.
L158: I think you should put in some explicit formulas for calculating costs in this paragraph so that this is easy for someone to replicate.
Reply: Rather than a formula, we have now outlined how the calculations were carried out. In Line 167, we state that: “For each grab type, costs were obtained by multiplying their mean laboratory costs (per sample) by the number of samples necessary to reach a progressive number of species.”
L172: did all the studies use a 1 mm sieve? Needs to be mentioned in the methods. If not, needs some kind of adjustment.
Reply: this is a valid point because studies used either a 0.5 mm or a 1 mm mesh sieve. We then have re-run all GAMS, including “mesh size” as a covariate. This has been included in the methods, where we outline that: “For all GAMs, the mesh size used by each study (mostly 0.5 and 1 mm) was included as a covariate.” Results have numerically changed, and this can be observed in the results section. However, it is important to note that, conceptually, the outcomes of the study are the same.
L173: define edf.
Reply: this is the “estimated degrees of freedom” of GAMs. This has been included in this sentence, at the first mention.
L175: The same precision, i.e., number of significant digits, should be provided for all the p values throughout the manuscript. for Table 2, this means use <0.001 or <0.0001, whichever precision you decide to use.
Reply: This has been corrected, throughout the entire ms, to four decimal digits or via “>” or “<” symbols wherever pertinent.
L308: this is the G1 grab, so use that as well.
Reply: corrected.
Reviewer 2 Report
One major short-coming of the paper was that the papers listed in table S1 were not cited in the Literature Cited section. Considering the literature review was also a major methodological component and result of the paper, those analyzed papers definitely should be cited. (Also, Henkel & Nelson used a 0.1 m2 grab - don't know if this is just a typo in the table or if wrong value was used in analysis - it does look like there is a high effort small grab size dot in Figure 1B)
A likely methodological short-coming, which I can't tell since all the papers aren't cited, is no consideration of the mesh sieve used for sieving. That is likely to affect the abundance and richness far more that the surface areas of the grab. The authors used 1 mm in their local study, but they did not say if they controlled for that in their literature review.
Introduction line 78 - add depth when talking about "shallow water" that can mean different things to different audiences. Yours is really shallow.
Figure 1 - maybe add log () to each x axis. Yes, it says it in the caption, but it was at first startling to see such shallow depths and low effort.
Results line 192 - what is the extra "2" after Table S2? I think typo.
Author Response
Reviewer 2 (reviewer’s comments in italics)
One major short-coming of the paper was that the papers listed in table S1 were not cited in the Literature Cited section. Considering the literature review was also a major methodological component and result of the paper, those analyzed papers definitely should be cited. (Also, Henkel & Nelson used a 0.1 m2 grab - don't know if this is just a typo in the table or if wrong value was used in analysis - it does look like there is a high effort small grab size dot in Figure 1B)
Reply: To accommodate this point, citations of all these papers have now included in Table S1, as a new column. The typo of Henkel and Nelson has been corrected.
A likely methodological short-coming, which I can't tell since all the papers aren't cited, is no consideration of the mesh sieve used for sieving. That is likely to affect the abundance and richness far more that the surface areas of the grab. The authors used 1 mm in their local study, but they did not say if they controlled for that in their literature review.
Reply: this is a valid point because studies used either a 0.5 mm or a 1 mm mesh sieve. We then have re-run all GAMS, including “mesh size” as a covariate. This has been included in the methods, where we outline that: “For all GAMs, the mesh size used by each study (mostly 0.5 and 1 mm) was included as a covariate.” Results have numerically changed, and this can be observed in the results section. However, it is important to note that, conceptually, the outcomes of the study are the same.
Introduction line 78 - add depth when talking about "shallow water" that can mean different things to different audiences. Yours is really shallow.
Reply: we have accordingly included the depth (2-3 m) of out study site in this sentence.
Figure 1 - maybe add log () to each x axis. Yes, it says it in the caption, but it was at first startling to see such shallow depths and low effort.
Reply: we prefer to maintain Fig. 1 as it stands, to avoid adding more text into the figure. This is information is already in the caption.
Results line 192 - what is the extra "2" after Table S2? I think typo.
Reply: corrected.
Reviewer 3 Report
While not an entirely unique area of study this is a useful contribution to the field. However it is not clear to me whether or not the authors consider the depth of grab penetration into the sediment significant factor in sampling efficiency.
Sediment particle size measurement is only touched upon briefly and how this is sampled from the grab is relevant.
I feel the authors by restricting their search to Van Veen grabs may have missed a number of relevant studies, admittedly some which may be in the grey literature.
Perhaps consider the relevant issues associated with temporal sampling as well as spatial studies.
Marine Pollution Monitoring and Management Group (2003). UK National Marine Monitoring Programme Green Book. Marine Pollution Monitoring and Management Group
[www.marlab.ac.uk/FRS.Web/Delivery/Information_resources/information_resources_view_documents
.aspx?resourceId=18479]
Humber Benthic Field Methods Workshop, Hull University 1997. Collection and processing of
macrobenthic samples from soft sediments; a best practice review. Environmental Agency R and D
Technical Report E1-135/TR: 1-viii, 1-128.
Riddle, M. J. (1989a). Precision of the mean and the design of benthos sampling programmes:
caution advised. Marine Biology, Berlin 103: 225-230.
Riddle, M. J. (1989b) Bite profiles of some benthic grab samplers. Estuarine, Coastal and Shelf
Science 29: 285-292.
TI think conclusions and because of the nature of the paper, recommendations should be clearer
Author Response
Reviewer 3 (reviewer’s comments in italics)
While not an entirely unique area of study this is a useful contribution to the field. However, it is not clear to me whether or not the authors consider the depth of grab penetration into the sediment significant factor in sampling efficiency.
Reply: we assume this penetration to be similar between the three Van Veen grab types. To shed light on this, we have included in methods a sentence (Line 118).
Sediment particle size measurement is only touched upon briefly and how this is sampled from the grab is relevant.
Reply: initially, we did not include sediment type in our meta-analysis (literature review) because, contrary to depth, this is a categorical variable (i.e., sandy, muddy, etc), which is hardly to model via statistical GAMs. Moreover, sediment type tends to be correlated with depth, because deep-water samples tend to be muddy relative to shallow-water samples. As a result, despite we acknowledge the relevance of sediment type as a key determinant of infaunal assemblage structure, we decided to maintain this variable out of the scope of this paper. This limitation has been indicated in the discussion – see the following point in this sense.
I feel the authors by restricting their search to Van Veen grabs may have missed a number of relevant studies, admittedly some which may be in the grey literature.
Reply: we totally emphasize with this view. However, we must restrict our literature review to Van Veen grabs to avoid including confounding factors that can seriously affect the relationships we tested, by means of GAMs, in this study. In any case, to recognize these limitations, we have included in the discussion the following paragraph: “It is important to highlight that our literature review have two limitations. First, the outcomes are limited to studies implemented through collection of infauna by Van Veen grabs. In addition, our study has not considered variation in sediment grain size, as a relevant factor affecting the diversity and abundances of infauna.”
Perhaps consider the relevant issues associated with temporal sampling as well as spatial studies.
Reply: because the discussion is already long, particularly after the inclusion of new paragraphs, we prefer not to include more considerations.
I think conclusions and because of the nature of the paper, recommendations should be clearer
Reply: we have rewritten some considerations in this sense, i.e. a last paragraph has been included.
Round 2
Reviewer 3 Report
As the authors have noted the comments made earlier the paper can approved for publication. Although in future they should consider the point about penetration depth of grab more fully. Nonetheless this is a useful study.
