Modeling the Potential Distribution of Two Species of Shrews (Chodsigoa hypsibia and Anourosorex squamipes) under Climate Change in China

Round 1

Reviewer 1 Report

Reivew of Diversity-1487508

Modeling the potential distribution of Chodsigoa hypsibia and Anorosorex squamipes under climate change in China

Hu et al.

 

Dear Authors,

 

Modeling the potential distribution of species under climate change scenarios is critical work than provide invaluable insights into how these species may be affected climate change. That said, I don’t believe there has been a lot of such models developed for small mammals, especially shrews. It is nice to see that work is being done to consider the effect of climate change on shrews. Thank you for doing this research.

 

Modeling of this sort, using the same 20 covariates used in this study, is fairly common, but to the best of my knowledge not for shrews. The science is valid and in line with similar studies. I found your manuscript was well written. As such, I only have a few minor (but important) suggestions for improvement. These are noted below.

 

Congratulations on a well-prepared manuscript. I look forward to soon seeing it in print.

 

Sincerely,

Detailed Comments:

 

Line 2: The title is unnecessarily awkward because most readers will not know these scientific names. Please rephrase to “Modeling the effect of climate change on the potential distribution of two species of shrews in China”, or something similar.

 

Line 15: Change to “…of two species of shrews (Chodsigoa hypsibia and Anorosorex squamipes)..”

 

Line 22: Rephrase as “These findings demonstrate that different small mammal species respond differently to climate change.”

 

Line 25: Delete this last sentence. It provides little value and is worded awkwardly.

 

Line 26: Spell out the full scientific name of these two species in the keywords. Delete “area”

 

Lines 28-33: The first paragraph provides little value and can be deleted. The second paragraph is a great place to start your paper.

 

Line 46: Replace “modeling” with “approach”. Delete “, an excellent SDM modeling tool,”

 

Line 48: I don’t understand what is meant by “insect disaster warning”. Please clarify or delete.

 

Line 51: Delete “deep”

 

Line 53: Delete “, the smallest mammals in the world,”

 

Line 60: Replace “and/or” with “or”

 

Line 61: Replace “and/or” with “or”

 

Line 62: Provide the English common names for these two species here.

 

Line 63: It is not clear why these two species of shrews were used for this study. Please elaborate and clarify why they were chosen of all the small mammals in China.

 

Line 66: Please elaborate on the global range of these two species and how much of it is in China. This is a key comment. It would be most helpful to modify Figure 1 to show the global range of these species as polygons that extend into adjacent countries.

 

Line 83: Delete “whole”

 

Line 87: Replace “before the remaining ones” with “and those remaining”

 

Line 125: Delete “distributed”, and put the training data before the testing data so that it is in logical order.

 

Line 129: This last sentence is odd. Only two classes of suitablilty is odd, I would have thought a gradient of probabilities would be more useful. Nonetheless, just a comment. No revision suggested.

 

Line 142: There is no information on the number of locations (and date ranges) of the location data used in the models. This is an important gap. Please including at the outset of the Results or in the Methods that specifically notes what data was used for each species – that is, how many records for each species.

 

Line 171: Round this percentages up to whole numbers – e.g., 5%, 5%, and 6%.

 

Line 173: Should this be “decrease” instead of “increase”?

 

Line 176: Figure 3 is great. Please make it larger on the page so it is easier to see.

 

Line 183: I think they also prefer moist forest floors because it aids in maintaining an optimal water balance, which is a challenge for species with a very small body size and high metabolism.

 

Line 202: Should this be “increased” instead of “decreased”

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

You have chosen two shrew species to look at future distribution with MAXENT. I'm not clear on why these were chosen, rather than any other. GBIF shows only 14 records for one and 732 for the other. These data can hardly be regarded as 'comprehensive'. You do not report how many you ended up using. Further, you used 10,000 background points - I presume across all of China? Thus your background to presence ratio may well have affected your results and thus your very high AUC - far too many background points in places the species will never be able to expand to.

I would like to know how many data points were used, and see this distribution. I would also need to see the effect of different background points (reduce the number and the spatial extent). It would then be possible for me to assess the modelling part of this work.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript entitled Modeling the Potential Distribution of Chodsigoa hypsibia and Anorosorex squamipes under Climate Change in China" is quite interesting. However there are some doubts which should be clarified. Thus, I cannot recommend it for publication I its current form. Detailed comments are below

Line 28- why predictions here are used, not the real increase, it refers to the past.

Line 53 – “Shrews, the smallest mammals in the world, have unique physiological and ecological characteristics that make them excellent animals for studying the effects of climate change surface temperatures [18]”  - Well, to support this sentence you cited one study on shrews - but this does not prove that shrews are excellent animals for studying the effects of climate, thus you should explain, why they are excellent.

Line 74 - who obtained? - I suggest a kind of sentence: "A distribution of .. has been studied widely during last thirty years"

Line 96 - how many record of each species have you used in the analysis?

Line 118 - please indicate excluded variables, and r values

Lines 150-157 - I would like to see the fourth ranked variables for both models, and  their contribution

Another thing is, the elevation will not change over time, thus when you predict a range for given species, and when in fact the occupied elevations depend on the climatic conditions, the preferred conditions will be probably shifted, and other elevations would be more suitable. I generally don’t think that elevation should be taken into account or the relation of elevation and climate conditions should be in some way included.

Line 160 - it’s hard to read figure 2, especially axes. Please enlarge the graph.

Line 172 -  there were two climatic scenarios, thus, why there are three results for scenarios? The last two are scenarios and the fist is the current area, this should be clearly indicated.

Line 194  - 500 Celsius?

Discussion in general - I rather expected a paragraph comparing the climate effects on the small mammals, this would be much more interesting than explanation of the climatic conditions effect on these species. Besides, I don’t see a clear conclusions.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

The Authors present a study on the effects of climate change on the distribution of two species of small mammals. The approach is not original, as there are a huge amount of studies analysing the same research question. However, the methodology is not correctly used, as I explain below in detail:

1. Table 2: Why did the Authors provided AUC and variable contributions for the future projections? I am supposing that the Authors modelled directly the future, which is completely wrong. Future models (projections) can only be obtained by projecting the current to the future. Maxent has an option to project the current model to several future scenarios. Future projections cannot be modelled directly as future presences are not available. You cannot validate a future projection with current presences. A future projection can only be validated with future presences, which are obviously not available. For the same reason, it is not possible to obtain variable contributions. Remember, that the projection assumes that the relationship between the species and the environment will not change in the future. If not, it will be not possible to project the current model to the future. Consequently, please delete Figure 2 and the associated results in the main text.  Please note that this can only be by java script, as Maxent GUI option only allows to define one scenario. I recommend reading Sillero et al 2021 for a step-by-step modelling guidelines. What it is correct to do is count how many presences will have a high suitability index in the future scenarios.

2. The worldclim variables from the 5th IPCC report has been substituted by the variables from the 6th IPCC report:

https://worldclim.org/data/cmip6/cmip6_clim10m.html

I recommend repeating the models with the updated variables and the new future scenarios: the four Shared Socio-economic Pathways (SSPs): 126, 245, 370 and 585.

3. L107: Removing ambiguous or duplicated records do not reduce spatial auto-correlation, but the level of point clustering. Spatial auto-correlation always refer to a specific variable, not only to the position of records. Please Sillero & Barbosa 2021 and Sillero et al 2021 for a detailed explanation. In fact, after checking Figure 1, it seems that A. squamipes might have clustering problems. It is the clustering level of the records at the Southwest part of China real, or a product of uneven sampling effort? In that case, further filtering should be needed.

4. Why the regularisation parameter was set to 1? Modelling only two species allows the Authors to search for the best regularisation parameter.

5. Why did the Authors defined a size of background points? The last version of Maxent chooses automatically the best size for the study area. I recommend not modifying the size of background points.

6. I recommend calculating null models following Raes & ter Steege (2007) methodology for a better evaluation of Maxent. Please, note that Maxent is a background-presence algorithm, so it does not use absences to the fit the model. Therefore, absences should not be used to validate Maxent models. As we do not have a better method than using the AUC, it is recommendable to calculate null models for a better evaluation.

7. Figure 3: It is better to present the continuous models, not the binary ones, as the former provide more information.

Minor changes
 
L48: Relating instead of combining.

L52-55: Maxent has been used to model all types of species and other types of point processes. Please, provide more references and not only related to invasive or endangered species.

L88-L91: I suggest formulating the objectives as research questions, avoiding to say whether the objective is interesting.

L89: What do the Authors mean with real coordinate records? I suggest to delete that part of the sentence, which is not very relevant. If not, 'real' should be changed by 'correct'.

L97: Explain why do the Authors think the C. hypsibia is more vulnerable to climate change.

Table 1: Please, indicate the selected variables. 

L145: What version of Maxent was used?

L167: Please, the SD values of the AUC results.

 

Neftalí Sillero

 

References
- Raes, N, and H ter Steege. 2007. “A Null-Model for Significance Testing of Presence-Only Species Distribution Models.” Ecography 30 (5): 727–736.
- Sillero, Neftalí, and A Márcia Barbosa. 2021. “Common Mistakes in Ecological Niche Models.” International Journal of Geographical Information Science 35 (2). Taylor & Francis: 213–226. doi:10.1080/13658816.2020.1798968.
- Sillero, Neftalí, Salvador Arenas-Castro, Urtzi Enriquez‐Urzelai, Cândida Gomes Vale, Diana Sousa-Guedes, Fernando Martínez-Freiría, Raimundo Real, and A.Márcia Barbosa. 2021. “Want to Model a Species Niche? A Step-by-Step Guideline on Correlative Ecological Niche Modelling.” Ecological Modelling 456 (April): 109671. doi:10.1016/j.ecolmodel.2021.109671.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I remain concerned that the model has been over-fitted. You had only 30 samples for  C. hypsibia (23 training and 7 testing according to your methods), along with 5,000 background points (many of which could be in areas where C. hypsibia is actually present), along with 20 explanatory variables.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript entitled The manuscript entitled Modeling the Potential Distribution of Chodsigoa hypsibia and Anorosorex squamipes under Climate Change in China" has been generally improved according to the comments. The Authors responded to my comments in a sufficient extent. however I still have some comments to the improved manuscript, still, perhaps, due to my laconic comments, some of my suggestions were misunderstood. Below are listed detailed comments.

 

2.3.Authors indicate, that they have excluded the elevation from the factors, thus it should be clearly indicated in methods, especially as they accessed this data (elevation).

108-109 – this is really small sample size, and should also be discussed

140 - please add the information which variables were finally used in modeling, this is important for the model evaluation.

209-213 - how current scenario can be showed as increased?, increased comparing to what? If I understand in a proper way, this is a percent of suitable areas as indicated on the fig4

– also the fig 4 capture, not area of the species distribution, but the suitable area or area of suitable habitats

L 270- authors respond, that they follow my suggestion regarding climate effects on the small mammals in the discussion. This however was a misunderstanding. I expected a general comparison of climate effects on small mammals, not only the two studied species. This would make the discussion more interesting an valuable for readers.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report

I have only minor comments - see attached pdf.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf