Ecological Niche Overlap and Prediction of the Potential Distribution of Two Sympatric Ficus (Moraceae) Species in the Indo-Burma Region

Round 1

Reviewer 1 Report

In this study the authors use MaxEnt to model and predict the ecological niche of two coexisting Ficus as well as their most important predictors. They contrast their results using past, current and future conditions to analyze the evolution of those ecological niches along time. After reading the article would like to point out some concerns and suggestions that arose from the paper.

Major comments

-Abstract: I found the section of the abstract exposing the results confusing. I would consider making some modifications. The abstract is the presentation of a paper and a big portion of the potential readers will only read this part. I encourage you to make this section more self explanatory, including all the highlights and information needed to fairly summarize and make your paper understandable.

-Line 66-70: Here it should appear the aim of the study in a clear way. In this manuscript the aim appear blurred between methodological information. I would suggest to clearly highlight the aim of the study and detail the methodology only in the methodology section.

-2.1 section: Which is the criterion you used to select the data? Do you include saplings? How did you spotted the different individuals? Are they under natural conditions (I assume they are, but I would appreciate deeper detail)? What about year of sampling? Have you followed any patter to sample the presences? MaxEnt is a method based on presence-only modeling. Thus, the selection and sampling method needs to be rigorous in order to avoid sample selection bias. More information in the following article: 

Yackulic, C.B., Chandler, R., Zipkin, E.F., Royle, J.A., Nichols, J.D., Campbell Grant, E.H. and Veran, S. (2013), Presence-only modelling using MAXENT: when can we trust the inferences?. Methods Ecol Evol, 4: 236-243. https://0-doi-org.brum.beds.ac.uk/10.1111/2041-210x.12004

General comments

-I don’t understand why you decided to include two current models (one using 10 variables and the other using 15 variables).

-I don’t see the parameter that determines which conditions are unsuitable, marginal or suitable. How do you determine this gradient? Based on a % of suitability that you classified in this 3 groups?

-Line 202-203: I would like to see in the discussion which are the implications of this overlap you mention. Also, the niche theory says that coexisting species tent to differentiate themselves in order to avoid competition. How is this fitting in your story?

-I think the relation between the studied species and its pollinator needs to be better supported and explained, starting at line 226. 

-I think the paper would be benefitted of a deeper explanation on why the studied species are important and why you decided to study those species.

Minor comments

-Line 14: I would remove “Additionally”
-Line 16: Please, describe AUC and LGM. This has not been described before.
-Line 63: What is the usable dataset?
-Line 106: I think you can put the name of the author here and the number of the reference as well.
-Figure 4-5: I would suggest to remove the ocean from the suitable area, since is clearly unsuitable for this species from the ecological perspective.
-Line 121: Use present or current, to be consistent with the terminology used in the table 2.
-Line 137: Why do you talk about global scale here? There is no previous mention about global scale and the paper have a regional scope.
-Line 143-144: Having high habitat suitability and a decrease in pixels seems contradictory to me.
-Line 169-170: % of what? Which is the total and why you selected that total?
-Figure 6: The legend is difficult to read.
-Line 232: Maybe this pollinating wasp also suffers alterations due to climate change that would affect their interactions with Ficus.

Author Response

Reviewer 1

In this study the authors use MaxEnt to model and predict the ecological niche of two coexisting Ficus as well as their most important predictors. They contrast their results using past, current and future conditions to analyze the evolution of those ecological niches along time. After reading the article would like to point out some concerns and suggestions that arose from the paper.

Major comments

-Abstract: I found the section of the abstract exposing the results confusing. I would consider making some modifications. The abstract is the presentation of a paper and a big portion of the potential readers will only read this part. I encourage you to make this section more self explanatory, including all the highlights and information needed to fairly summarize and make your paper understandable.

Response: Thank you for your comment, the abstract has been modified.

-Line 66-70: Here it should appear the aim of the study in a clear way. In this manuscript the aim appear blurred between methodological information. I would suggest to clearly highlight the aim of the study and detail the methodology only in the methodology section.

Response: Thank you very much for the comments, we have clearly highlighted the aim of the study and removed the methodology description to the method section. Please see L81-85.

-2.1 section: Which is the criterion you used to select the data? Do you include saplings? How did you spotted the different individuals? Are they under natural conditions (I assume they are, but I would appreciate deeper detail)? What about year of sampling? Have you followed any patter to sample the presences? MaxEnt is a method based on presence-only modeling. Thus, the selection and sampling method needs to be rigorous in order to avoid sample selection bias. More information in the following article: 

Yackulic, C.B., Chandler, R., Zipkin, E.F., Royle, J.A., Nichols, J.D., Campbell Grant, E.H. and Veran, S. (2013), Presence-only modelling using MAXENT: when can we trust the inferences?. Methods Ecol Evol, 4: 236-243. https://0-doi-org.brum.beds.ac.uk/10.1111/2041-210x.12004

Response: Thanks for the reference, we have checked the data for using in MaXent, and have added the sampling information and selecting method in L90-94.

General comments

-I don’t understand why you decided to include two current models (one using 10 variables and the other using 15 variables).

Response: When we predict the distribution in the LGM, Mid-Hol, and future periods, we could only get the bioclimatic variables from the worldclim database. So we used the current model's ten bioclimatic variables to compare with the other period scenario. However, the plant distribution is also affected by some abiotic variables. To see the difference with the model only with bioclimate information, we also include the current model with 15 variables.   

-I don’t see the parameter that determines which conditions are unsuitable, marginal or suitable. How do you determine this gradient? Based on a % of suitability that you classified in this 3 groups?

Response: We used ‘Equate entropy of thresholded and original distributions threshold’ and ‘10 percentile training presence threshold’ which from maxent results to determine this gradient, and we have changed “unsuitable, marginal or suitable” to specific value in Figure 4 & 5.

-Line 202-203: I would like to see in the discussion which are the implications of this overlap you mention. Also, the niche theory says that coexisting species tent to differentiate themselves in order to avoid competition. How is this fitting in your story?

-I think the relation between the studied species and its pollinator needs to be better supported and explained, starting at line 226. 

Response: We have added in L66-75 & L229-233.

-I think the paper would be benefitted of a deeper explanation on why the studied species are important and why you decided to study those species.

Response: We have added the biology of both Ficus species and the special story sharing the pollinating fig wasp in overlapping distribution. Please look L66-75.

Minor comments

-Line 14: I would remove “Additionally”

Response: Deleted. 

-Line 16: Please, describe AUC and LGM. This has not been described before.

Response: The full name has been added.

-Line 63: What is the usable dataset?

Response: Sorry, we made a wrong interpretation here, and deleted this sentence.

-Line 106: I think you can put the name of the author here and the number of the reference as well.

Response: We have added the author and the number of the reference.

-Figure 4-5: I would suggest to remove the ocean from the suitable area, since is clearly unsuitable for this species from the ecological perspective.

Response: Thank you for the comment. That’s because the Sea level in the LGM period (Figure 4A & 5A) was lower than the present, and the continental area was exposed (Bird et al., 2005). However, the geology template of the current is only available to us. That’s why we used the current geology template to project the distribution area in LGM, which showed some area covered by the ocean but actually, there was land exposed during the LGM. And below is the reference for the paleoenvironment and paleo-geology situations.

Bird, M.I.; Taylor, D.; Hunt, C. Palaeoenvironments of insular Southeast Asia during the Last Glacial Period: a savanna corridor in Sundaland? Quat. Sci. Rev. 2005, 24, 2228-2242.

 
-Line 121: Use present or current, to be consistent with the terminology used in the table 2.

Response: Thanks for your suggestion. We have changed “present” to “current” in the text.

 
-Line 137: Why do you talk about global scale here? There is no previous mention about global scale and the paper have a regional scope.

Response: We have changed it to “in the study area”.

-Line 143-144: Having high habitat suitability and a decrease in pixels seems contradictory to me.

Response: We have redescribed this part.

-Line 169-170: % of what? Which is the total and why you selected that total?

Response: Sorry, we have not made it clear here. The %=(suitable area/ study area)×100%. We used these values to compare habitat suitability of F. heterostyla and F. squamosa in this study area.

-Figure 6: The legend is difficult to read.

Response: Thanks for the comment, we have changed the Figure colors and added the color’s means in the legend.

-Line 232: Maybe this pollinating wasp also suffers alterations due to climate change that would affect their interactions with Ficus.

Response: Pollinating fig wasps with short life cycles are more sensitive to climate change, and will influence the fig-fig wasp mutuslism. We have added the description in L250-254. 

Reviewer 2 Report

The manuscript “Ecological niche overlap and the prediction of the potential distribution of two sympatric Ficus (Moraceae) species in the Indo-Burma region” by Fungjanthuek J. et.al. is devoted to an urgent topic of modeling and predicting plant ranges in a changing climate under past, current and future climates  from LGM to Mid-Hol to 2070.

The authors used ecological niche model (ENM) to evaluate the habitat suitability and to predict the potential habitats of two sympatric species, Ficus squamosa and F. heterostyla. in the Indo-Burma region.

A set of 19 bioclimatic and 5 abiotic variables were chosen to model the habitat of these two ficus species. A Pearson correlation analysis was performed to exclude the highly correlated variables. Ten bioclimatic and 5 abiotic variables were selected to generate model. Then, the maximum entropy model (MaxEnt) was employed to generate the long span ENMs for the two species as a function of time (LGM-Mid-Hol-Current-Future). Finally, the overlap between the sites of two species was evaluated using the indices of equivalence and similarity.

I think the methodology applied was scientifically sound and results obtained were important for the biodiversity issues over the given region.

I have some minor comments after clarifying which I consider the manuscript may be published.

Please clarify:

1. Why only two ficus species were chosen among 800 and their overlap was important to investigate;

2. I would recommend to include all bioclimatic variables in Table 1 or to add excluded variables in Footnotes;

3. Yet, please provide climate descriptions of the region as a whole and if possible of some contrasting sites (South-North; West-East) to show how different climates were, are and will be across the region, especially across the mountains, and through the entire period from LGM  to Future;

4. You used WorldClim data, could you provide some information on how many stations across your region were employed by WorldClim to produce climate layers and how many stations are located there nowadays, specifically at highlands which are covered by the sparce network of weather stations.  

5. Because MaxEnt does not provide bioclimatic ranges for both species, could you provide those in a table?

6. Across mountains, elevation is a significant factor in predicting the distribution of both ficus species (Table 3). As we know, elevation is a proxy of any climatic variable so that the collinearity is high between elevation and a variable. Thus, elevation seems redundant in your analyses. Moreover, elevation has not been and will not be changing for the period LGM – Future.

Please explain what ecological role elevation plays in your model.

7. Technical notes.

a) The authors should improve axis labels for the figures 3, 4 and 5 and make them readable;

b) In Figs 4 and 5, ficus ranges are Unsuitable, Marginal, Suitable – please specify quantitative limits between them;

c) Please explain why you used hundredths of a percent predicting suitable habitats (Section 3.3). Tenths of a percent seem more appropriate.

Author Response

Reviewer 2

The manuscript “Ecological niche overlap and the prediction of the potential distribution of two sympatric Ficus (Moraceae) species in the Indo-Burma region” by Fungjanthuek J. et.al. is devoted to an urgent topic of modeling and predicting plant ranges in a changing climate under past, current and future climates  from LGM to Mid-Hol to 2070.

The authors used ecological niche model (ENM) to evaluate the habitat suitability and to predict the potential habitats of two sympatric species, Ficus squamosa and F. heterostyla. in the Indo-Burma region.

A set of 19 bioclimatic and 5 abiotic variables were chosen to model the habitat of these two ficus species. A Pearson correlation analysis was performed to exclude the highly correlated variables. Ten bioclimatic and 5 abiotic variables were selected to generate model. Then, the maximum entropy model (MaxEnt) was employed to generate the long span ENMs for the two species as a function of time (LGM-Mid-Hol-Current-Future). Finally, the overlap between the sites of two species was evaluated using the indices of equivalence and similarity.

I think the methodology applied was scientifically sound and results obtained were important for the biodiversity issues over the given region.

I have some minor comments after clarifying which I consider the manuscript may be published.

Please clarify:

1. Why only two ficus species were chosen among 800 and their overlap was important to investigate;

Response: Ficus species and their pollinating fig wasps are normally species specific. However, dioecious F. squamosa and F. heterostyla share the pollinating fig wasp, the niche overlap and distribution area are important to understand the coevolutionary mechanism of both partners. We have added relevant description in L66-75.

2. I would recommend to include all bioclimatic variables in Table 1or to add excluded variables in Footnotes;

Response: Thanks. We have included all the variables and added the footnotes, please see the updated table 1.

3. Yet, please provide climate descriptions of the region as a whole and if possible of some contrasting sites (South-North; West-East) to show how different climates were, are and will be across the region, especially across the mountains, and through the entire period from LGM  to Future.

* Our response: I had added the climate descriptions in L49-52.

4. You used WorldClim data, could you provide some information on how many stations across your region were employed by WorldClim to produce climate layers and how many stations are located there nowadays, specifically at highlands which are covered by the sparce network of weather stations.  

Response: It’ s about 400 stations across our study region were employed by WorldClim (Hijmans et al., 2005).

Hijmans, R.J., S.E. Cameron, J.L. Parra, P.G. Jones and A. Jarvis, 2005. Very high-resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25: 1965-1978.

http://www.ncdc.noaa.gov/pub/data/ghcn/v2

5. Because MaxEnt does not provide bioclimatic ranges for both species, could you provide those in a table?

Response: Thanks, we have added a new table 4 for this issue.

6. Across mountains, elevation is a significant factor in predicting the distribution of both ficus species (Table 3). As we know, elevation is a proxy of any climatic variable so that the collinearity is high between elevation and a variable. Thus, elevation seems redundant in your analyses. Moreover, elevation has not been and will not be changing for the period LGM – Future.

Please explain what ecological role elevation plays in your model.

7. Technical notes.
8. a) The authors should improve axis labels for the figures 3, 4 and 5 and make them readable;

Response: We have improved the figures 3, 4 and 5.

1. b) In Figs 4 and 5, ficus ranges are Unsuitable, Marginal, Suitable – please specify quantitativelimits between them;

Response: Thanks, we have changed it. We used ‘Equate entropy of thresholded and original distributions threshold’ and ‘10 percentile training presence threshold’ which from maxent results to determine this gradient, and we have changed “unsuitable, marginal or suitable” to specific value in Figure 4 & 5.

3. c) Please explain why you used hundredths of a percentpredicting suitable habitats (Section 3.3). Tenths of a percent seem more appropriate.

Response: The %=(suitable area/ study area)×100%. We used these values to compare habitat suitability of F. heterostyla and F. squamosa in this study area.

 

Round 2

Reviewer 1 Report

 

Dear authors,

 

From my point of view, you did a great effort improving the manuscript and it look much understandable and appealing now. Still, there are some comments I would like to add:

 

Major comments:

 

I believe some of the lines numeration has been moved in the version I have so it hasn’t been easy to find the modifications you did according to my comments. So maybe I missed something along the text. This being said, I don’t think the methodology explanation of presence selection has been clarified enough, which I assume it might be in 2.1 section. Do you exclude any tree age (saplings)? Are the plants in natural conditions? How are the field surveys distributed? Do they follow any pattern? Making this section clear is very important since MaxEnt is a model using presence-only, otherwise I wouldn’t be so insistent. This is a key point to evaluate the consistency of the research.

 

My previous comment: “-Line 202-203: I would like to see in the discussion which are the implications of this overlap you mention. Also, the niche theory says that coexisting species tent to differentiate themselves in order to avoid competition. How is this fitting in your story?” hasn’t been responded. Did you include something related to it in the manuscript? If you didn’t, why not?

 

Minor coment:

I enjoyed your response to this comment, great job. Is not always about following our suggestoins:

“-Figure 4-5: I would suggest to remove the ocean from the suitable area, since is clearly unsuitable for this species from the ecological perspective.

Response: Thank you for the comment. That’s because the Sea level in the LGM period (Figure 4A & 5A) was lower than the present, and the continental area was exposed (Bird et al., 2005). However, the geology template of the current is only available to us. That’s why we used the current geology template to project the distribution area in LGM, which showed some area covered by the ocean but actually, there was land exposed during the LGM. And below is the reference for the paleoenvironment and paleo-geology situations.

Bird, M.I.; Taylor, D.; Hunt, C. Palaeoenvironments of insular Southeast Asia during the Last Glacial Period: a savanna corridor in Sundaland? Quat. Sci. Rev. 2005, 24, 2228-2242.

Author Response

Reviewer #1

Dear editor

We thank you for the invitation to revise and resubmit our manuscript (#Forests-1727937). We have substantially improved the manuscript referring to one reviewer’s comments. We also have invited a professional editor with native English speaker to polish the English again.

 

Comments

Dear authors,

From my point of view, you did a great effort improving the manuscript and it look much understandable and appealing now. Still, there are some comments I would like to add:

R: Thank you so much!

Major comments:

I believe some of the lines numeration has been moved in the version I have so it hasn’t been easy to find the modifications you did according to my comments. So maybe I missed something along the text. This being said, I don’t think the methodology explanation of presence selection has been clarified enough, which I assume it might be in 2.1 section. Do you exclude any tree age (saplings)? Are the plants in natural conditions? How are the field surveys distributed? Do they follow any pattern? Making this section clear is very important since MaxEnt is a model using presence-only, otherwise I wouldn’t be so insistent. This is a key point to evaluate the consistency of the research.

Response: Thank you for your comments. We have added some description in detail in 2.1 section.

“Ficus squamosa Roxb. (Subgenus Sycomorus, Section Sycocarpus) is distributed in Nepal, Bhutan, Sildcim, north-east India, Myanmar, Laos and China (Yunnan), while F. heterostyla Merr. (Subgenus Sycomorus, Section Hemicardia) is distributed in Thailand, Vietnam, Laos, and Cambodia [37]. Both Ficus species show overlapping distribution in the Xishuangbanna region of China and share an undescribed pollinating fig wasp (Ceratosolen sp.) [28]. From 2012 to 2019, we conducted several field surveys and sample collections in China, Laos, Myanmar, Vietnam, Cambodia and Thailand (98.5° E–109.5° E and 9.5° N–26.5° N). We investigated the distribution regions of both species in the Flora of Thailand [37] and Flora of China [38], as well as other potential distribution regions. We only recorded the occurrences of mature individuals that found in natural conditions. We also downloaded the distribution data of both species from Global Biodiversity Information Facility (GBIF) website. All data obtained were exported to ArcGIS v10.5, with and duplicate presence data within 1 km removed. Finally, the distribution data of the two sympatric species included 55 records of F. squamosa and 67 records of F. heterostyla (Figure 1 & Table S1) . We retained these records to match climate variables for further model analysis.”

A Supplemantary Table S1 was added, and also updated the Figure 1 with data sources.

 

My previous comment: “-Line 202-203: I would like to see in the discussion which are the implications of this overlap you mention. Also, the niche theory says that coexisting species tent to differentiate themselves in order to avoid competition. How is this fitting in your story?” hasn’t been responded. Did you include something related to it in the manuscript? If you didn’t, why not?

Minor coment:

I enjoyed your response to this comment, great job. Is not always about following our suggestoins:

“-Figure 4-5: I would suggest to remove the ocean from the suitable area, since is clearly unsuitable for this species from the ecological perspective.

Response: Thank you for the comment. That’s because the Sea level in the LGM period (Figure 4A & 5A) was lower than the present, and the continental area was exposed (Bird et al., 2005). However, the geology template of the current is only available to us. That’s why we used the current geology template to project the distribution area in LGM, which showed some area covered by the ocean but actually, there was land exposed during the LGM. And below is the reference for the paleoenvironment and paleo-geology situations.

Bird, M.I.; Taylor, D.; Hunt, C. Palaeoenvironments of insular Southeast Asia during the Last Glacial Period: a savanna corridor in Sundaland? Quat. Sci. Rev. 2005, 24, 2228-2242.

Response: Thank you so much for your enjoying the response.

Author Response File: Author Response.docx