We thank the referee for his suggestions made online. We respond in bold and italic below to his comments.

The manuscript by Noirmain et al entitled "Interdisciplinary strategy to survey phytoplankton dynamics of a eutrophic lake under rain forcing: description of the instrumental set-up and first results" is well written and easy to read. The described interdisciplinary strategy provides a very useful tool to decipher the impact of atmospheric processes on lake ecology.

An enormous work has been performed on the atmospheric sciences part but the authors only quickly discuss the ecological implications related to their study. The authors also need to consider further factors related the cyanobacteria biology and ecology. This could for instance apply to Lines 400-405 and Figures 8AB.

A negative effect of rain on organism abundance was mentioned for few taxa. Effect can be seen on Fig. 8B on unicellular group forming cyanobacteria (e.g. Woronichinia, Microcystis, Merismopedia...). Whereas some unicellular individual cyanobacteria (e.g. Synechocystis, Synechococcus, Pseudoanabena...) seem to undergo a positive effect. Cell distribution factor is important because it impacts the growth dynamic and survival of the species.

Moreover microalgae can move and migrate through the water column. Almost the same concentration of cells were fund "before" and "after" RP events in the system (Fig. 8A). The increase concentration at a specific depth implies displacement of the phytoplanktonic community or organismal movement (RP2 towards 1.5 m depth, RP1 towards 3 m). Apparently deeper when the rainfall was more virulent (HIR). This is not described in the results nor correlated to Fig. 8B (diversity) but quickly mentioned in the discussion. Was the water column structure more stable (nutrient, temperature, light, water agitation...) at 1.5 m depth in RP2 and at 3 m in RP1?

The water column does not seem more stable at 1.5 or 3 m after the rain events. No change was found regarding water temperature, conductivity, and oxygen concentrations which were similar up to 3 m deep after RP1 (oxygen concentration=8.72 mg. L-1, conductivity = 116.8 – 117 µS/cm) and after RP2 (oxygen concentration between 77 and 7.9 mg. L-1 and conductivity between 101.2 and 102.48 µS/cm). However, as an abiotic factor, only the water irradiance was different and could explain a slightly different distribution of the phytoplankton abundance at 1.5 m after RP2 and 3 m after RP1. Regarding fig 8B, we did not discuss a potential vertical displacement according to the rain type, convective or stratiform, as other rain events have occurred during RP1 and RP2. We analyzed only three rain events to illustrate the monitoring strategy in this article. However, in a future study, more results and observations will help investigate the potential effect of convective or stratiform rain events on vertical lake stability.

All the cyanobacteria taxa retrieved have different ecological requirements. In the discussion Lm and Lo codons are rapidly mentioned. A principal component analysis can further help to characterize which cyanobacteria taxa was favored under certain rain events (e.g. peculiar nutrient signature).

As suggested by the referee, we performed a supplementary analysis by Multiple factor analysis (MFA) (Fig 9, in the PDF attached) using the abiotic lake factors (water temperature and irradiance and rain amount), phytoplankton taxa (described in Fig 8B), and the lake chemical compositon (Fig 8C & D) to characterize which cyanobacteria taxa could be favored after rain events. Multiple factor analysis (MFA) is an extension of principal component analysis (PCA) tailored to handle multiple data tables that measure sets of variables collected on the same observations. Such analysis was more adapted considering our set of data. The results are presented in the pdf attached.

The MFA analysis confirms a negative correlation between the rain amount and Microcystis, Merismopedia, and Coelomoron abundances, also confirming the significant change of abundance between the “before” and “after” rain period reported for these species in the results (line 403). Nevertheless, as suggested by the referee, we can add this model to the results (if the ninth figure is authorized in the manuscript) as it also shows a positive correlation between the rain amount and the abundance of diatom (Asterionella), colonial microalgae (Elakatothrix), and unicellular picocyanobacteria (Synechocystis) abundances associated with the increase of some inorganic ions (SO42-, Ca2+ and NH4+) and the rain amount. In any case, we will indicate the species that increase following the rain amount in reference to Fig 8B.

Moreover, the differences in phytoplankton concentration and taxa diversity between "before" and "after" calls for emission/deposition fluxes (e.g. Dillon et al 2020 doi:10.1128/AEM.01850-20, Mayol et al 2014 doi: 10.3389/fmicb.2014.00557), very briefly mentioned in discussion. However this is of ecological importance. Retrieved cyanobacteria species from the investigated lakes have been reported from atmospheric samples (e.g. Sharma et al 2007 DOI: 10.1111/j.1529-8817.2007.00373.x). Also microalgae can be emitted after local disturbances such as rain (Tesson et al 2016 doi:10.1128/AEM.03333-15, Wiśniewska et al 2019, 2022) from station Fig 1B2 and be redeposited in local system such as station Fig 1B1 or the lake, therefore it is possible that the retrieved peak of Cr1b with highest phycocyanin value could be partly indigenous.

To ensure a correlation between emission and deposition fluxes in Aydat lake, we should sample airborne microalgae close to Aydat lake's surface. Unfortunately, we did not measure the emission fluxes from the Aydat lake and cannot directly link the phycocyanin detection in the rain (from one rain event, CR1) and the increase of picocyanobacteria abundance at the end of lake sampling. Nonetheless, we discuss deposition fluxes by comparing the level of photosynthetic cells collected in the rainwater with those in the lake, suggesting a very low amount of photosynthetic species introduced in the lake by the wet deposition. Moreover, we also discuss the deposition flux that can impact the diversity in the lake by introducing different species collected by the rain (lines 560-567).

Lines 182-184 - How was the lake sampling performed? Which parameters were investigated? Do the authors refer to Lines 114-116 in situ measurements or were lake water- and/or phytoplankton samples collected? Please describe further.

To provide more details about the lake sampling, we will add this sentence: “In addition to the lake monitoring, we collected lake sampling at three depths, surface, 1.5 and 3 m deep, using a Van Dorn horizontal Bottle Water Sampler (2.2L, PVC) deployed vertically with a weight to take it to the desired depth. Then water was transferred in 15 Liter Jerrican to keep the water temperature stable during transport back to the laboratory (~20 min). “

Line 198 - Why was a 10 µm pore size used? In these filtrates, all microorganisms of a size inferior to 10 µm would be present (including. bacteria, cyanobacteria and other <10 µm eukaryotic microalgae), thus affecting the measurements. Moreover, -20°C storage was applied to the filtrates, conditions under which organismal cell damaging occur, releasing further nutrients in the water. Please explain further.

Line 200 - Which Lugol's iodine solution was used (acidic vs neutral)? What was the final concentration used? How and how long before investigation were the samples stored (temperature, darkness, time)? These are important information for cross studies comparison.

We thank the referee for pointing out an inconsistency with the pore size. The correct size of pores used will be corrected in the new version. The correct sentence is that we filtrated the lake water on a 150 µm Nylon membrane to avoid the presence of zooplankton in lake samples. The filtrate (under 150 µm) was fixed in a neutral Lugol solution (Sigma-Aldrich), and 10 ml of Lugol's iodine stock solution was added to 150 ml of lake filtrated samples for the microscopy. For the nutrient analysis, the lake water was filtrated on 0·2-µm pore-size filters Nylon membrane by rinsing the filtrate previously with 500 ml of ultra-pure Milli-Q water to avoid contamination.

The fixed samples using Lugol were stored in the dark at 4°C and were counted within the year. The lake samples filtrated under 0.2µm were kept at -20°C until the chromatography analysis was performed within the year.

Lines 221-236 - How does the flow cytometer detection method differ from existing protocols? (e.g. Haynes et al 2020, https://www.agilent.com/cs/library/applications/application-analysis-aquatic-plankton-novocyte-5994-2112en-agilent.pdf)

We use conventional flow cytometry measurements with bandpass filters to detect the autofluorescence signals from photosynthetic pigments excited by the 640 nm and 488 nm lasers and collected using FL3, FL2, and FL4 detectors (670 Long Pass,585/42, and 661/16 Band Pass filters respectively) (Dashkova et al., 2016). However, we did not use any fixators like glutaraldehyde because the count of photosynthetic cells can be underestimated (Troussellier et al., 1995). Moreover, the rainwaters were kept at 4°C in the dark (no growth and cell preservation), and we performed the flow cytometry shortly after the rain events (maximum of 48h), ensuring a realistic view of the environmental signature of the samples. The rainwaters, which could not be analyzed after 48h was discarded.

We report a different approach to quantifying the autofluorescence to measure specific pigment populations of photosynthetic cells rich in chlorophyll, phycocyanin, or phycoerythrin. An example is showed in the PDF attached (supplementary fig). Indeed, we first select the photosynthetic cells population (total pigments) based on minimal chlorophyll fluorescence on the cytogram by selecting FL3 and SSC, opened on the preselected population without unwanted debris (Cells, plot A & B). Then, to distinguish the photosynthetic cells based on individual pigments autofluorescence, we create a new cytogram by selecting FL3 and FL2 channels taken from the "total pigments" population (plot C). This cytogram was divided into two: a "phycoerythrin rich population" and a not phycoerythrin-rich population. From the "NOT Phycoerythrin population," we create a new cytogram by selecting FL3 and FL4 channels divided into "chlorophyll" and "phycocyanin" populations (plot D). These selections allow us to separate the phycoerythrin population from chlorophyll or phycocyanin.

Line 470 - I disagree with the sentence. First because microalgae encompass both prokaryotic (cyanobacteria) and eukaryotic photosynthetic unicellular organisms. Second because previous studies have investigated the diversity of microalgae in wet depositions including rain. However, the methods used involved capture and growth, not rapid detection based on flow cytometry. The proposed sentence is therefore not proper, please rephrase.

Lines 473-475 - I also disagree with this sentence. One major problem with culturing is that not all organisms can grow in artificial media, therefore applying a selection pressure towards underestimating the environmental biodiversity. Another issue is that all isolated microalgae possess a biome (including bacteria). These bionts can be remove using diverse available methods. However, some microalgae need their bionts to survive. In any cases these should not impede microalgal detection using flow cytometry or microcoscope-based techniques. Please reformulate the sentence.

To clarify lines 470-475, the sentence will be rephrased as below, including new references:

“Moreover, the detection of photosynthetic cells by microscopy in rain samples is often measured on rain cultures after several days (Wiśniewska et al., 2022). However, this methodology seems inappropriate for estimating the natural environmental biodiversity because an unadapted artificial medium for the growth of all microorganisms could apply a selection pressure. Hence, we recommend using flow cytometry or microscopy-based techniques on fresh samples without fixators. Indeed, glutaraldehyde can lead to an underestimation of the count of photosynthetic cells and can alter the intensity of fluorescence according to the cell size (Lepesteur et al., 1993, Troussellier et al., 1995)."

The English language and formulations need to be double checked by a native speaker, several mistakes are present in the text.

A native speaker will check the English mistakes.

Dear Referee 1,

Please find the attached document to have formatted text with the references and the figures together.

Best regards,

Fanny Noirmain

We thank the referee for his suggestions made online and on the pdf. We respond in bold below to his comments (our answers to the major comments of the pdf have been added at the end of the text).

The study title “Interdisciplinary strategy to survey phytoplankton dynamics of a eutrophic lake under rain forcing description of the instrumental set-up and first results” by Noirmain et al. aims to define the fine scale effect of rain and the carried algal particles on a lake physiochemistry and phytoplankton community. They combine methodology from meteorological sciences analyzing cloud structure and origin and raindrop algal cytometry, with characterization of the water column properties and phytoplankton microscopy enumeration, in an innovative approach that aims to reconciliate recent findings of rain algal cell deposition (Dillon et al. 2020, Wisniewska et al. 2022,  both cited thoroughly in the manuscript), with traditional works that explore the relationship between rain events and lake biogeochemistry, such as de Eyto et al. 2016 (DOI: 10.5268/IW-6.4.875). As far known, no efforts have been made to analyze the lake surface and the rain for both chemical composition and photosynthetic organisms.

The study has the potential to provide great insight into rain events effects in lake phytoplankton, short term surface stoichiometry and water column temperature changes. The evaluation of the photosynthetic organisms suspended in rain drops in comparison with the lake phytoplankton can provide great insight into the dispersal rate and mechanisms of the different phytoplankton organisms which is still a poorly understood subject. The inclusion of a phycocyanin channel (also with the chlorophyll channel provides a unique opportunity, together with the real time evaluation of the effects of rain on the physiochemistry of the lake, can provide a great insight on the effects of rain events on the phytoplankton community. Although the title of the article alludes to a presentation of experimental setup and first results, the listed objectives, rationale behind the analyses and discussion aim for a much definite style of work. 

Unfortunately, the measurements provided seem disconnected. The water sampling times are too far away to assign any causality of the chemical and community changes to rain events and not wind induced or upstream inputs (which seems the case given the steady decrease of temperature in CR1). Mesocosm experiments could have helped isolate the rain effects from the basin and wind effects.

"The suggestion to use a mesocosm is interesting and would help isolate rain to wind effects, but we do not have this material at our disposal and cannot perform this experiment. For this article, we focus on a case study of three rain events to illustrate a monitoring strategy of the atmosphere and lake. As we analyzed the rain impact on the lake stratification at a fine scale in real-time, we were able to report a causality between the rain intensity and the lake temperature. To better show the wind effect on the lake temperature, as suggested by the reviewer, we added a correlation between the water temperature and the wind during the rain events. So, during HIR, associated with low wind speed, we report an immediate decrease in the diurnal surface water temperature. The correlation between the water temperature at the lake surface and the wind was positive but not significant (r=0.8, p-value = 0.3333). In contrast, the correlation with the rain intensity was significantly negative (r=-0.7, p-value=1.2.10-06), suggesting the predominant water temperature decrease due to the rain amount at the surface.

On the other hand, during CR1 associated with a higher wind speed, the correlation between the water temperature and the rain intensity was lower (r=-0.24, p-value=1.3.10-03), whereas those obtained with the wind speed was significantly negative (r=-0.63, p-value = 2.1.10-03) until 3 m deep. Moreover, as the wind was higher during RP2, it could be argued that it accelerated the decrease of the lake temperature by mixing the water column during this period and weakening the lake stratification strength. This analysis and correlations with the wind could be added to the manuscript to show better the relative contribution of wind and rain in water temperature as the wind events.

Although we sampled the lake as soon as possible, we agree that a more frequent sampling would increase the understanding of the causality between biochemical rain and lake biochemical composition. Nevertheless, for this article, it was not the principal aim as we wanted to illustrate a strategy with high monitoring. Indeed, we present only three rain events as an example for the case study that does not allow us to connect precisely the rain and lake biochemical composition".

The lake phytoplankton and the rain photosynthetic cells are measured with two different methods that are hard to reconciliate. The rain cytometry does not seem to fix the cells with glutaraldehyde like Dillon et al. 2020, so the cells present in rain might be over or underestimated by growth or death in the rain collector chamber.

Overall, this work provides the first attempt to measure the rain effect on a lake on the fine scale. It can be improved by a bigger connection between the variables and timescales used to measure them and the questions set in the introduction. To properly answer said questions I suggest performing cytometry on the lake surface (and/or microscopy on rain samples), isolate the effects from rain and watershed inputs using mesocosm enclosures, and shortening the sampling time after the rain event. Although the main missing link is the disconnection of methods for phytoplankton analysis for the lake and for the rain, with comparative 2d cytograms of known cultures/species or lake samples, the two datasets can be made compatible. Regarding the manuscript structure, the introduction and the discussion need higher cohesiveness and streamlining, and I suggest that it undergoes severe rewriting. The results’ figures could also benefit from trimming down to the ones specifically pertaining to the questions set.

The cytometry method used to detect the photosynthetic microorganisms in the rainwater was developed to cover the size of particles ranging from 0 to 30 µm. This methodology is well adapted for species in rainwater, as their size commonly found is under 30 µm in the atmosphere. On the contrary, the microscopy-based method is not well adapted due to the very low number of photosynthetic cells in rainwater. Moreover, in the literature, we found that some authors let the cells grow for 30 days before estimating the diversity, which leads to underestimating the diversity as all species cannot grow in an artificial medium. So we decided not to use this method.

On the contrary, the photosynthetic lake species can be composed of colonial or filamentous forms larger than 30 µm. Therefore, the flow cytometry is inappropriate and could lead to underestimating the concentration of the lake's colonial or filamentous species. It is why we estimated the lake phytoplankton abundance by microscopy.

However, we also measured the phytoplankton by flow cytometry, using the same recorded parameters developed for the rainwater. An example is presented on the plots of the PDF attached, showing a comparison between rain and lake samples measured by flow cytometry. The flow cytometry based-method detects small cyanobacteria (like Synechococcus and Synechocystis), which grow especially after RP2 (plot B, PDF attached). On the contrary, the microscopy counts show that many cyanobacteria were present before RP1 and RP2 (plot D, PDF attached), which were not detected by the flow cytometry based-method, corresponding to colonial cyanobacteria (Microcystis, Merismopedia...) (Fig 8B).

Although the sonication could help dissociate the colonial species, it could also damage cyanobacteria or flagellate cells, which will not be detected by flow cytometry. Moreover, chlorophyll a is not the best proxy to assess environmental diversity as the increase in chlorophyll a can be due to a higher proportion of large species (Visser et al., 1996; Felip & Catalan, 2000).

Due to these technical issues and more clarity in this article, we select only the appropriate methods to detect the rain's photosynthetic cells by flow cytometry and the lake phytoplankton by microscopy.

The air mass analysis does not provide additional support to the questions set out in the introduction apart from a brief mention in the discussion about how CR rain was from the lower cloud system and not the higher one of marine origin. They discuss solar radiation and rain effects on the water column thermal structure, while avoiding wind effects, which is usually the dominant factor determining short term mixing changes (as clearly seen in the CR event).

In our view, it is essential to maintain the air mass analysis as it could explain one part of the dynamics of photosynthetic cells. However, with only the three rain events presented here, it is complicated to conclude about the origin of photosynthetic cells according to the air mass origin (from the sea or the continental, results are controversial on the topic), and more rain events analysis is necessary. Nevertheless, we aim to illustrate a strategy to investigate the potential link, and it seems crucial to keep such analyses in our article. Thus, we will improve the introduction to present this information more clearly and to improve the link between the questions set in the introduction and the discussion referring to the dynamic of photosynthetic cells in the rain. Likewise, we will add more bibliography citations related to the dynamic of these cells and will also add the following sentence in the introduction:

“We illustrate a strategy to monitor the dynamic of photosynthetic cells in the rain by characterizing the cloud and rain physical properties, the meteorological variables, and the air mass origin.”

The discussion will also benefit in a cloud type-oriented organization, with sections for HIR and CR going each from the cloud source, the rain physiochemistry up to the lake chemistry changes and finally phytoplankton changes, instead of partitioning into their methodological counterparts. The text contains minor errors and will greatly benefit by proofreading by a native English speaker.

The cloud-type-oriented organization could be interesting. However, with only three rain events analyzed, we think it is premature at this stage as we did not perform a climatology study with a higher number of rain events. Nevertheless, this will be performed in fore coming studies.

With the current state of this work, it might be worth considering partitioning the work in two:

Analyzing the rain drops cytometry (and/or microscopy) and its relationship with the different sources and characteristics of the rain clouds, with additional 2d cytograms of representative samples of the lake or cultures to known in greater detail the composition of said cells. Optimally, cytometry of the lake surface before and after the event should be performed, but the timing should be precise to avoid changes due to algal migration.

It cannot be done that way because the culture will not give us more detail about the cells found in the rainwater (see the answer on the cytometry method above). Therefore, even if we used the culture of known species to adjust cytometry parameters, we have only information about the pigment types and the cell densities, which do not allow species identification. Moreover, 2d cytograms of known cultures/species or lake samples could not be compared with rain samples as species in culture do not represent all those found in the atmosphere (some of them are uncultivable). In any case, the flow cytometry could not give greater detail on the lake composition of said cells as we were limited in a size range to focus on the bioaerosols found in the rain and not those found in the lake.

Report on the lake phytoplankton and physiochemistry changes before and after different precipitation events but including a further discussion on wind effects and watershed inputs until a decoupling is achieved with mesocosm deployments.

Because we did not have a mesocosm at our disposal for comparisons, we cannot argue about the wind effects and watershed inputs.

Ions reported should be in the context of cyanobacterial biomass changes, i.e. NH4+, NO3-, and PO4-, and a point should be made of the viability of the rain droplets milieu as a growth media for the airborne cyanobacteria, while the other ones (that are used as fingerprints to identify the origin of the clouds) can be just mentioned in the text or in supplementary figures.

The atmospheric conditions are stressful for these species (UV, temperature variation, osmotic chock…). However, as mentioned in the literature, they could be viable, so maybe the nutrient composition favors their viability. We did not point out the viability of the rain droplets media because the rainwater is a temporary environment for the cells. Indeed, the photosynthetic cells can be scavenged in the cloud "long-range transport in the air mass" or washed out below the cloud from the air column. It is similar to ions. Hence, we did not know if their origin came from the cloud or the air and how long they stayed in this environment.

Specific comments included inline in the annotated pdf are enumerated here:

PDF comment for line 183: This 15 minute drought break seems very arbitrary and specifically to obtain 2 rain periods from what should be a single long one. Better support and justification is needed for this choices.

We agree that the two rain periods could be viewed as belonging to a larger rain event. However, our goal is to study the possible high temporal links between rain and its origin and characteristics and its effect our lake water composition. Thus, it seems important to separate the different phases within a rain event as distinct periods, not only as a function of dynamics but also composition.

In our example, it appears that the two separate periods actually show slightly different air mass origins, thus reinforcing our choice to select a drought period equivalent to the best interval to ensure a new bottle change between samples.

Line 202: Justify this normalization procedure.

The abundances of species (Fig8 A) were transformed into relative abundances to counter the heterogeneity in the number.

Fig 8A: Phytoplankton abundance should be avoided when talking about the whole community since it includes disparate taxa like Synechocystis (2 µm unicellular) and Microcystis (500+ µm colonies), biovolume/biomass should be reported instead.

The biovolume/biomass is particularly interesting in the food web study, which is not the case here. So, we prefer to keep the dynamics of phytoplankton in cellular concentration.

Line 545: concentration shouldn't be compared, given that the direct mass of rain is very small in comparison to the whole water column (accounting for the watershed is a different issue).

Line 548, we reported that the wet deposition did not influence the lake's chemical composition, certainly due to the small mass of rain compared to the lake (similar to the photosynthetic cells brought by wet deposition), very low in contrast to phytoplankton cells line 563).

Line 563: A discussion about phytoplankton dspersion and biogeography will greately enhance this section.

We did not have the relevant information about species diversity in the rain samples. Thus, we cannot discuss the correlation between photosynthetic cells in the rain and the lake. Nevertheless, we will improve the introduction with citations describing the species encountered in the atmospheric compartments, especially in the rain.

Line 573: If they were "pulled down by mixing", then the charophytes and chlorophytes would have also decreased. Seems like these groups' buoyancy mechanisms created a downward migration. Lm and Lo codon have low surface/volume ratio on an individual basis, and not high as listed here. See: Reynolds 1994 (DOI: 10.1007/BF00007405)

We thank the referee for pointing out an inconsistency with the surface-to-volume ratio, which is low for Lm and Lo species. Therefore, we will rephrase the sentences as follow:

"Microcystis, Merispomedia, and Coelomoron belonging to Lm and Lo codons have a low tolerance to mixing. Therefore, their decrease after the rain events could be explained by a dilution caused by their transport through the entire mixed depth. However, when temporary stratification was back after the rain, a secondary thermocline nearer the surface could help these species to quickly regulate their vertical position by buoyancy mechanisms to benefit from optimal conditions (nutrient and water irradiance)."

In order to improve the discussion on wind effects, we will add sentences to describe the species increasing after the rains in link the rain and wind events:

On the contrary, after the rain events, we reported a shift in species composition from colonial picocyanobacteria towards unicellular picocyanobacteria (Synechococcus and Synechocystis) and also the increase of diatoms (Asterionella and Melosira) and microalgae (Elakatothrix). Synechococcus, Synechocystis, and Asterionella have a high surface-to-volume ratio and can quickly grow in the mixed layer depths if there is no light-limitation, as is the case during the lake campaign (the euphotic zone, 13 m deep, exceeds those of the thermocline, 7 m). Indeed, Reynolds (1994) reported that optimal division rates of these species could lead to doubling their abundance per day when there is no light limitation, suggesting they can quickly develop to become dominant after the rain events.

On the other hand, we also reported the presence of larger unicellular algae, Closterium, present only after RP2, and the coenobial green algae Elakatothrix, which increased after the rain despite its low growth rate. Hence, as RP2 was characterized by stratiform rain events and high wind speed, we suggest that the wind increased diffusivity, allowing green algae to stay longer in the water column where no light-limitation was reported. Indeed, the mixing could favor larger and heavier species to stay suspended and their development, as has already been reported in the literature (Blottiere et al., 2014).

Dear Referee 2,

Please find the attached document to have formatted text with the references and the figures together.

Best regards,



Fanny Noirmain