We thank the Anonymous Referee #1 for posting a comment on this manuscript. Please find below a response to each remark.

RC1: "General remark: I have wondered about the abbreviation of PELGAS. Maybe it would be a good idea to explain the program either in Methods or Introduction as it is a central part of the data sets."

We thank Referee #1 for the advice.

Zooplankton samples were collected during the successive PELGAS (PELagique GAScogne) integrated surveys carried out over the Bay of Biscay (BoB) French continental shelf, every year in spring from 2004 to 2019 on board the R/V Thalassa. The aim of this survey is to assess small pelagic fish biomass and monitor the pelagic ecosystem to inform ecosystem based fisheries management. Fish data, hydrology, phyto- and zoo-plankton samples and megafauna sightings (marine mammals and seabirds) are concomitantly collected to build long-term spatially resolved time series of the BoB pelagic ecosystem. The PELGAS sampling protocols combines day-time en-route data collection (small pelagic fish and megafauna), with night-time, depth integrated hydrology and plankton sampling at fixed points. Detailed PELGAS survey protocols can be found in Doray et al., 2018c and 2021.

This paragraph will be add to the text at the beginning of the section 2.1 “Sampling”, lines 101-103.

RC1: "I find the presentation of the taxa and OTUs in table 1 and 2 a bit difficult. Usually italics are reserved for genera and species names. I suggest to change that in both tables. Additionally it would be good to add the intruments in the header of table 2."

We understand that italics are not use in a common way here. In the revised manuscript, this font will be suppressed in tables 1 and 2, and taxa and OTUs will be differentiated with different cell colours in both tables. Moreover, the instrument names will be added to the header of table 2.

RC1: I think it would also be good to add at a more prominent place that environmental and other accompanying data are availabel for the same cruise. So far as I noticed this is only mentioned in the concluding remarks. However, for the usability of these datasets environmetnal data are essential and should be highlighted somewhere earlier in the manuscript.

We thank Referee #1 for this interesting remark. Indeed, our zooplankton dataset can be combined with environmental datasets constructed for the same cruise with similar spatio-temporal resolutions which are published in the SEANOE dataportal (Doray et al., 2018a). A comment about the environmental datasets availability will be added at the beginning of the section 2.1 “Sampling” following the PELGAS survey description in the revised manuscript.

RC1: "I have a few orthographic remarks: Line 40: … objects are composed of… / Line 56: Metazoan planktonic organisms…. / Line 77/78: …taxonomic analysts… / Line 86: …these require… / Line 189: quality check of identifications?"

We thank Referee #1 for pointing out these orthographic remarks. It will be corrected in the revised version of the manuscript.

We thank Dr Leo Berline for his suggestions on our manuscript. Please, find below a response to each comment / question:

CC1: I would recommend to clarify/insist on the homogeneity of the two datasets (Zooscan and ZooCAM) in the whole ms.

We appreciate this recommendation. However, the two datasets are not completely homogenous as provided. They are both as taxonomically detailed as our identification skills enabled, and although being intercomparable and interoperable for most of the taxa and most of the oceanographic conditions, there are still some discrepancies between the two instruments, particularly on the abundances estimations (Grandremy et al., 2023b). Yet, those can be easily overcome by aggregating some taxa in one dataset or the other, or both, but the resulting taxonomic resolution would be inferior to that of the provided dataset. We therefore not insist on the homogeneity, rather on the intercomparability and interoperability of the two instruments. We will however take into account your titling suggestion as it actually reflects the combination of instruments already undertaken in a previous paper (Grandremy et al., 2023a).

CC1: The title may state this: replace “from the” with “combining” ZooScan and ZooCAM imaging systems.

Corrected in the revised manuscript.

CC1: Abstract: L40 “are composed”

Corrected in the revised abstract.

CC1: I am not familiar to the meaning of ‘interoperable’. Might be worth clarifying.

Interoperable refers to the FAIR principles (https://www.go-fair.org/fair-principles/). When generating scientific data, it is now recommended to generate them following the FAIR principles, that are: the data should be Findable, Accessible, Interoperable (easy to combine to other data), and Reusable.

CC1:

• L76: decent spatio-temporal scale: ‘decent’ is not a scientific word.
• L91: ‘open’ = freely available.
• L113: “...laboratory with the zooscan in 2019.” I suggest to remove ‘in 2019’ which is useless but confusing with the sampling dates.
• L138: I would add “net collection” to insist on the fact that it is the same sample that can be analysed in both ways.
• L151-153: “ break free from ..” unclear sentence. I guess authors mean to control the maximum number of object.
• L176: ZooCAM sofware?
• L184: “that were used”
• L191-192: “To be used together” I guess “combined” is the proper word.

All these specific comments have been corrected in the revised manuscript.

CC1: L190-200: for 2016, how were the samples splitted between Zooscan and ZooCAM?

A comparison study was done to ensure the datasets originating from each instrument are homogenous and can be combined for ecological studies (Grandremy et al., 2023b). All the zooplankton samples from year 2016 (61 sampling stations over the whole BoB continental shelf) were imaged with both instruments. This comment has been added to the text to clarify this point, in the section 2.6.

CC1: L201 comparable: can you better quantify (correlation?) 

Corrected in the revised manuscript.

CC1: Fig 2, 3, 4: add reference for the gridding procedure.

The gridding procedure is presented in Petitgas et al. (2009) and Petitgas et al. (2014). See also Doray et al. (2018c) and Grandremy et al. (2023a) for application examples. These references have been added in the figures’ legends.

CC1: L252 Table 1, 2. Why not sorting according to abundance? It is more meaningful than taxonomy.

We agree with Dr Berline. Both tables have been sorted according to the counts. In Table 1, the sorting was done according the counts calculated from the ZooCAM data.

We thank Referee #2 for his/ her review and all the comments posted. Please, find below our answers to each comment.

RC2: The fact that presenting and providing the full dataset (both on the form of raw data and in the form of images) is a really large contribution to make important and precious dataset as FAIR as possible and the way the dataset is presented contributes to a clear understanding of the protocols operated, but also on the analysis conducted to obtain the final raw dataset. The fact that both datasets have been cross calibrated to provide basis on their Interoperability (the ‘I’ in FAIR) is also one really nice and important contribution (but is published and extensively inspected in https://doi.org/10.1002/lom3.10577 , maybe the citation and link to this latter could be made more visible), and increase its robusteness.

We thank Referee #2 for the positive comment. At the time we wrote this data paper, the publication presenting the interoperability of both instruments was under review. It is now published and the citation of this publication has been updated with the DOI in the present manuscript, in the section 'Reference'.

RC2: In addition to the provided “raw” data (hosted in the platform SEANOE, and freely available to anyone; therefore, making it FA(i)R), I would strongly recommend to provide the “pre-analyzed” data (ie. the ones used to currently produce the figures 2,3, and 4 (abundance, size and shanon index, maybe separated by instrument types, since here it appears unclear which instrument is used for the year where both were used for intercalibration). Such additional data availability will make the dataset even more “Reusable” for users without having to recompute everything from raw data.

The figures 2, 3 and 4 in the manuscript illustrate a few examples of the zooplankton community’s descriptors that can be derived from these datasets, but this list is not exhaustive. Biomass, biovolumes, size spectra can also be calculated from the data. The use of these datasets will depend on the scientific question of each study in which they will be included. The main interest of providing raw individual data is to allow users to aggregate the data at the taxonomic resolution they require for their specific scientific needs but also to use the data at the individual level as provided and / or to define functional traits based on imagery or taxonomy. Also, the figures 2, 3 and 4 show the data spatially smoothed over a grid with a cell size set at 0.3° in latitude and longitude. Providing raw individual data at each sampling stations will allow users to work at the spatial resolution needed (sampling station or spatial smoothing over grids with different cell sizes or averaged over the whole Bay of Biscay). Therefore, we think that providing the raw individual data at each sampling station, as presented here, is the easiest way to make these datasets reusable.

RC2: The fact that the raw images are also provided is also a very important feature, although providing 6Go of dataset at once will certainly limit the number of people which will actually access to this massive dataset (except for needs of machine learning techniques). In addition to this, I would provide a way to interactively access and inspect the imaging dataset without actually downloading it. Maybe providing the full list of ecotaxa projects (currently accessible from what I see, e.g. https://ecotaxa.obs-vlfr.fr/prj/2952 they seems to be separated by instrument and years of sampling) and allowing readers to get a preview and inspect the dataset could be a really nice addition, maybe as an additional table in the manuscript.

We thank Referee #2 for the idea of providing a way for the users to check the imaging dataset without necessarily downloading it. Each dataset corresponds to a project on the platform Ecotaxa that gather the data from all the years for each instrument (2004-2016 for the ZooScan and 2016-2019 for the ZooCAM). We now provide the names of both of these projects to allow the readers to browse and inspect the individual vignettes on Ecotaxa, in the section 4. On the other hand, we would like to clarify that the 6 Go dataset available through SEANOE constitute the individual images only that can be downloaded separately from the numerical data. We assume that these sets of images will be useful in the machine-learning field where computers are powerful enough to handle these very large datasets.

RC2: Although convincing in GrandRemy et al 2023, the intercomparison of instruments is certainly not fully perfect. And from table 3 only, it seems that some taxonomic groups do exhibit some discrepancies in between the two datasets (clearly overseen by one instrument). This is the case for Eucalanidae, Foraminifera, Isopoda, Mysida, Pontelidae, Rhincalanidae, Sapphiridae at least. However, this intercomparison is made complicated since data provided in table 1 are “raw counts” and total counts per instruments are not equal. Eventually providing percentages would also allow a better feel on this, and maybe a more complete disclaimer on potential discrepancies at fine taxonomical resolution between the two datasets could be needed in the section 2.6… and could also be served by table 2 (e.g. rhizaria in zooscan been identified under the label harosa in zoocam).

Reviewer 2 is right, the intercomparison is not perfect, as already discussed in a comment posted in the Discussion by Dr Leo Berline.

The Tables 1 and 2 have been modified to show the percentage calculated for each taxa and OMGs. The discrepancies showed by some taxonomic groups between instruments originates from the differences in sample preparation protocols before the image acquisition, the imaging techniques and quality, and whether the samples were imaged live or fixed. For example, the mineralized protists (here, Rhizaria) dissolve in formalin and are considered underestimated in preserved seawater samples (Biard et al., 2016). Also, the random orientation of objects in the ZooCAM flow cell leads to a loss of taxonomic identification accuracy due to the difficulty to spot the specific features needed for the identification (Colas et al., 2018; Grandremy et al., 2023b). This is particularly acute for copepods, where the ZooScan seems to provide better identification capabilities to experts, as the organisms are imaged in a lateral view most of the time whereas the ZooCAM often images them in a non-lateral, randomly-oriented view, preventing the visualisation of specific features. A detailed discussion about how to explain the discrepancies between the ZooScan and the ZooCAM can be found in Grandremy et al. (2023b). This paragraph has been added at the end of the section 2.6.

RC2: One modification that should also been conducted is the probably very incorrect usage of the term OTU to defines the different taxonomic groups used here. While it may correspond to the initial definition of OTU (operational grouping of closely related individuals…. Here by morphological properties), this latter is currently too much used by genomic approaches to not bring to confusions. Additionally, here some “OTU” are used to define non biological entities (therefore not having any “taxonomy”, such as bubbles or detritus), making even the initial definition of OTU non pertinent. More importantly with the initial definition of what is called “taxa”, those are indeed the real OTU (following the initial definition). All this to show that the usage of the term OTU is definitively confusing and I would strongly advice to not use this term and use a more adequate one (morphological groups, “Operational Morphological groups” or any more pertinent terms).

We understand that the term OTU is confusing. It has been replaced by the term “Operational Morphological Groups” (OMG) as this category combined aggregated morphological groups, non-adult life stages of taxa and non-living objects groups (e.g.: detritus, artefacts). The change has been made throughout the manuscript, mostly in the section 3.1 and in Tables 1 and 2.

RC2: Table 2 seems inverted in the columns: the title state that we are inspecting Zooscan and Zoocam data while the rest of the caption seems to indicate that the table is oriented on showing Zoocam (left) and Zooscan (right) (therefore inverted into order). Since the two sides of the table have no labels, this makes the interpretation somewhat confusing. I would advise to keep the same order (and add labels to the table).

We thank Referee #2 for pointing out the confusion on Table 2. As stated in the table’s caption, taxa and OMGs appearing exclusively in the ZooCAM dataset are listed in the left column while those appearing exclusively in the ZooScan dataset are listed in the right column, following the same order as in Table 1. The title of Table 2 caption have been modified to “ZooCAM and ZooScan not common taxa and OMG.” and the instrument names have been added to the table’s header for more clarity.

RC2: Table 3: ephyra not ephira

Corrected in the Table 3.