Relative Pollen Productivity Estimates for Mediterranean Plant Taxa: A New Study Region in Turkey

Round 1

Reviewer 1 Report

This is a very welcome and extremely valuable contribution to the field of palaeoecology in that it presents a set of relative pollen productivity estimates (RPPs) for the Mediterranean region. Until very recently, the application of quantitative pollen-based vegetation reconstruction methods such as the Landscape Reconstruction Algorithm (LRA) within the Mediterranean region has been hampered by a lack of RPPs for Mediterranean plant taxa, as RPPs are a key input parameter for these models. This paper therefore represents a significant step forward and opens up a range of exciting new research possibilities for this region. As such, I have no doubt that the paper will be of interest to the palynological research community and that it should be published following some modifications to the manuscript, which I recommend to improve it’s clarity and impact. My comments are listed by line number below.

77: why is it important to have more RPP values for the Mediterranean region? There are a couple of sets already published, so why is this new study needed? I have no doubt that it is needed, but I think this needs to be stated more clearly. For example, is it because there are still some taxa for which we have no RPP estimates? Or perhaps because RPP for the same taxon is sometimes known to vary across it’s geographic range, so perhaps RPPs from SW France are not appropriate for vegetation reconstructions in the eastern Mediterranean?

81: I’m not sure what is meant by the phrase ‘we aim to provide the first data’, since you have already cited two previous works on Mediterranean RPPs. I think your point here could be explained more clearly.

123-4: you refer to monthly average precipitation of 447 mm, but I can’t see this on Figure 2. Is this actually a yearly total you are referring to in the text, rather than a monthly average? The y axes on the graph also appear to be labelled incorrectly – the left-hand axis looks to me like precipitation, and the right-hand one looks like temperature, but they are labelled the other way around – this needs to be checked.

132: you refer to a 50 x 50 km sampling area, implying a square with sides of 50 km length, but Figure 1 shows a 50 km radius. Which is correct?

140-1: you say that no moss samples were present at 3 of the 24 sampling sites. Was it not possible to collect surface soils instead?

148-9: you say that the pollen rain in moss pollsters ‘may’ represent the modern vegetation – this doesn’t sound very confident! Perhaps you could cite some of the literature regarding pollen representation in moss pollsters to help support your argument here, e.g. Mulder and Janssen 1998; Räsänen et al. 2004.

226-7: you say the criterion for a taxon to be selected for ERV analysis was that it had to be represented in both the pollen and vegetation data. Was there a minimum number of sites it needed to be present at, and did you take account of whether it had a good range of values in both the pollen and vegetation datasets?

231: you say here that each of the main taxa were measured in order to calculate their fall speeds, but I got the impression from Table 3 that most of the fall speeds used were previously published values, and the only one you needed to calculate was Olea. Can this be clarified in the text, please?

237: Poaceae is a family name and should therefore not be written in italics. This is a common issue throughout the manuscript and needs to be checked. Only genus and species names need to be written in italics, not family names (typically those ending in -eae).

238-9: there are actually very good reasons not to use Poaceae as a reference taxon, if it can be helped. Because this taxon can contain tens or hundreds of plant species, all with differing ecologies and hence likely different pollen production, and the mix of species is highly likely to vary between regions, it actually makes much more sense to use a single-species reference taxon, as you have done. Most other studies use Poaceae because it’s the only taxon present at all of their sites – not because it is the ideal reference taxon! This should be mentioned here as it is a real strength of your study.

243: can you define what the ‘theoretically optimal shape’ is, please?

Section 3.1: check for and correct instances of family names being written in italics.

251 & 262: why were Juniperus and Fabaceae included in the analysis if they were well-represented in the vegetation but not in the pollen samples?

259: Figure 4 shows Olea and Phillyrea at only 1-2% - I wouldn’t describe these as ‘high percentages’. Do you mean that they were present in many samples, rather than that they were present at high percentages?

263: Asteraceae is not shown on Figure 4 – Centaurea and Anthemis are shown separately.

282-3: this sentence repeats what is said in the previous sentence but one – it isn’t needed.

301-3: it might be worth adding to the caption that the alpha values shown in Table 4 are the RPP estimates. In Table 4, the first set of iterations are labelled GPM – elsewhere in the results and discussion you simply refer to the Prentice distance-weighting method, so I think this should be re-labelled in the table to avoid confusion.

336-7: I’m not sure what other estimations you are referring to here, since it looks from Table 4 as though the RSAP estimate of 102 m is in the middle of the range of values you obtained with other methods, rather than being slightly greater. It is greater than some, but also smaller than others. Or are you referring to published estimates of RSAP? This needs to be clarified.

364 & 366: you refer to PPEs here – keep terminology consistent with the rest of the paper and use RPP estimates to avoid confusion for readers.

367: the text here says the RPPs shown in Figure 8 were derived using ERV sub-models 1 and 3, but the figure is labelled with sub-models 2 and 3. This needs to be checked and corrected.

370-2: I don’t understand this sentence. It is not possible for ERV sub-model 1 to show lower results ‘in comparison’ to those derived from Prentice, 1/d, and LSM, since these are entirely different things. Do you mean that ERV sub-model 1 consistently gave lower RPPs than the other two sub-models, regardless of the distance-weighting method used?

373-3: ERV model 2 gave similar outcomes to what? ERV model 1? From Table 4, it looks as though model 3 RPPS are actually closer to model 1.

375: ERV sub-model 3 produced lower values – lower than what?

377: you say that ERV sub-model 2 produced lower values for Pinus, but Table 4 shows that it produced the highest values for Pinus using the Prentice and 1/d distance-weighting methods.

377-8: you say that you obtained the ‘most accurate’ results from ERV sub-model 2 and Prentice’s distance-weighting method, but I don’t follow how this was determined. I think it would be helpful to re-write this paragraph to make it more clear how you arrived at this conclusion.

382: replace RPP ‘rates’ with RPP ‘values’. ‘Rates’ doesn’t really make sense in this context.

383: clarify that this creates opportunities for quantitative Holocene vegetation reconstructions in the Mediterranean.

386-93: nice to see comment on the possible influence of the season in which fieldwork was conducted on your results. There has been some work on how season can affect pollen representation in moss pollsters (e.g. Shumilovskikh et al 2021), as well as on vegetation survey and RPP estimates (Farrell et al. 2016). It might be worth citing some of this research here to strengthen your argument.

398-9: I suspect that finding moss pollsters could be a common problem in the Mediterranean region. Could this be overcome, e.g. by using surface soils? Again, there is some published work on this, most recently by Fang et al. (2022).

411: replace RPP ‘rates’ with RPP ‘values’.

432: you refer here to ‘bias’ but it is not clear what you mean, since you have not detailed the results of the study of moss and trap samples in SW Turkey. It is therefore impossible to know what ‘bias’ you are referring to.

444: replace RPP ‘rates’ with RPP ‘values’.

445: rather than ‘low-rate RPPs’, you could just refer here to ‘low pollen production’.

450: replace ‘RPP rates of’ with ‘RPP estimates for’. Could you add deciduous Quercus to this list, since you mentioned earlier that it was an important taxon?

Table 5: this is a nice compilation for RPP estimates for Mediterranean taxa. However, could you please add some explanation as to how the values in the final column (Mediterranean Region) were derived when you have more than two estimates for the same taxon? E.g. how does 6.058 for Pinus in France and 51.625 for Pinus in Turkey become 18.97 for Pinus in the Mediterranean Region?

479: replace RPP ‘rates’ with RPP ‘values’.

498: rather than ‘low-rate RPPs production’, you can just write ‘low RPP’.

502-5: can you attempt to explain here why your RPP estimate for Pinus is so much higher than other published estimates?

524: why do not give your own original RPPs here, rather than those recalculated with reference to Poaceae? I think what is in the conclusion should mirror what is presented in the abstract, i.e. your own original RPP estimates.

520: you state here the need for new RPPs for Cedrus, Olea, and Phillyrea. However, there is a published estimate for Phillyrea from Githumbi et al. (2022), so why is another estimate needed? Is it because geographic differences between SW France and Turkey are likely to lead to different RPPs for the same taxa? I also wonder why you do not mention the herbaceous taxa for which there are still no Mediterranean RPPs here, e.g. Cyperaceae, Centaurea, and Chenopodiaceae.

References:

Fang, Y., Bunting, M.J., Ma, C. and Yang, X. (2022) Are modern pollen assemblages from soils and mosses the same? A comparison of natural pollen traps from subtropical China. Catena 209: 105790

Farrell, M., Bunting, M.J. and Middleton, R. (2016) Replicability of data collected for empirical estimation of relative pollen productivity. Review of Palaeobotany and Palynology 132: 1-13.

Mulder, C. and Janssen, C.R. (1998) Application of Chernobyl Caesium-137 fallout and naturally occurring lead-210 for standardization of time in moss samples: recent pollen-flora relationships in the Allgäuer Alpen, Germany. Review of Palaeobotany and Palynology 103: 23–40.

Räsänen, S., Hicks, S. and Odgaard, B.V. (2004) Pollen deposition in mosses and in a modified ‘Tauber trap’ from Hailuoto, Finland: what exactly do the mosses record? Review of Palaeobotany and Palynology 129: 103–116.

Shumilovskikh, L.S., Abdulmanova, I.F. and Efimik, E.G. (2021) Does season matter for moss surface sample collection? A case study from Kangur forest-steppe, pre-Urals, Russia. Palynology 45: 191-199.

There are some minor typographical errors and instances of odd phrasing, but the meaning was generally clear throughout the manuscript.

Author Response

Author Response File: Author Response.pdf

Reviewer 2 Report

Based on the analysis of pollen and vegetation data for 21 sites around Gölhisar Lake in southwestern Turkey, the authors have estimated the RSAP and the RPP values of 5 major plant taxa by using ERV model. The results have provided data basis for regional vegetation abundance reconstruction in the Mediterranean area. However, the following questions need to be discussed:

(1) Table 2, in the Pinus brutia forest, the vegetation percentages exceed 100, is this effect the ERV model results?

(2) Figure 6, the log-likelihood values are strange, it should be influenced by the vegetation data beyond 100 m. The authors need to compare the vegetation data from vegetation map and the digital satellite imagery. In addition, the authors suggested that the interval of vegetation data was 1 meter beyond 100 meter in section 2.4. 

(3) Table 4, why the standard deviation close to 0 for most species and models?

(4) Figure 6 and 7, there has no linear relationship between vegetation proportion, pollen proportion and adjusted pollen proportion for Fabaceae, why the authors keep it in the ERV models?

(5) it is recommended to delete section 4.1.

(6) Table 5, please check the integrated RPP data for Fabaceae (0.07), Cupressaceae (2.51), Evergreen Quercus (6.545) and Pinus (18.97), how it was obtained?

(7) Conclusions, I suggest the author move the second paragraph to section 4.4.

Other suggestions:

(1) Please unify the format of all figures.

(2) Figure 2, the legend labels of Y-axis were wrong.

(3) Please check the format of Latin name for all species (family name should not in italics).

(4) Figure 8, there has no results of ERV sub-model 1.

Author Response

Thank you very much for taking the time to review this manuscript. I am grateful your guidance, supporting feedback.Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files. 

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper addresses the relevant issue of the pollen-vegetation relationship at the base of quantitative pollen-based reconstruction of land cover modelling, and provides a set of relative pollen productivity estimates (RPPs) for Mediterranean plant taxa. The complex relationship between vegetation cover and land use in the Mediterranean area could be comprehensively evaluated by the application of the REVEALS model which relies on the estimates of RPPs, but these values are still scarce for Mediterranean plant species. This study presents a set of RPPs for Southwestern Turkey that could be used in REVEALS models for the Mediterranean, thus enabling a better understanding of past (climate and anthropogenic driven) land cover changes in this region.

Based on a robust research design, the paper proposes values of the relevant source area of pollen (RSAP) and RPP estimates for the main taxa of the Mediterranean. The Authors clearly explain the applied methodology, also providing an explanation of some inconsistencies of RSAP results and variations in RPPs – e.g. mainly depending on the openness of the plant landscape. In this regard, the Authors appropriately examine the limitations of the study considering factors that may have led to those results. Also, the discussion is based on well-argued data compared to up-to-date relevant studies.

Overall, the paper is well-written, comprehensive, and represents a stimulating contribution to pollen-based quantitative vegetation reconstructions in the Mediterranean. This study could provide a reference for further research, and I recommend it for publication after corrections to the following (few) minor text editing:

- Cichoridae = Cichorioideae (change also in Fig. 4)

- Anthemidae = Anthemideae

- Plant family and tribe names not in italics.

Author Response

Thank you very much for taking the time to review this manuscript. I am grateful your guidance, supporting feedback.Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files. 

Author Response File: Author Response.pdf

Reviewer 4 Report

This manuscript has systematically studied the relative pollen productivity of a specific area in Türkey. This result is helpful for understanding and reconstructing past climate and vegetation, and the manuscript can be accepted for publishing after revision. The suggestions are as follows:

1. There are many studies on modern pollen in surface soil, including in Turkey. Compared to previous studies, what aspects of pollen studies do you plan to improve in your research? Please add relevant information in the Introduction part.

2. In the Materials and Methods, please provide a detailed explanation of the advantages and disadvantages of the site selection, as well as how to overcome the limitations brought about by this methods.

3. In the Discussion part, please clarify what scientific improvements you have made and what significant progress have you made compared to previous studies on modern pollen in surface soil?

Author Response

Thank you very much for taking the time to review this manuscript. I am grateful your guidance, supporting feedback.Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files. 

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

In the revised manuscript, the author anwsered the main questions and comments of the reviewer, and modyfied the manuscript carefully.

Reviewer 4 Report

all my concerns have been addressed, hence I recommend its acceptance for publication