Predicting the Habitat Suitability and Distribution of Two Species of Mound-Building Termites in Nigeria Using Bioclimatic and Vegetation Variables
, Azrag. G. A. Abdelmutalab 3,4, Christian W. W. Pirk 1 and Abdullahi A. Yusuf 1
Round 1
Reviewer 1 Report
The authors conducted a study on 3 termite species in NW Nigeria and used modelling to predict current and future distribution. While the idea has merit the execution needs to be improved. This is not a detailed review since the paper has to be improved to allow in-depth review, but I will point out the major problems I had.
The paper is poorly edited. It includes paragraphs that apparently are instructions on how to write results (L97ff) and the latter part of the discussion needs much improvement in English grammar, sentence structure and presenting arguments in a logical fashion. Professional editing services could be used to improve the writing. Literature citations in the text are in an unusual format (check journal requirements). Figure resolution is low, text in Figures is blurry and distorted.
While there is a lot of description of the model in Mat&Met, the sampling description is rather vague. The model is apparently based on 156 termite samples (at least, I assume that that is what “georeferenced points” mean). I assume only presence of these three species was counted, not absence and not population density or species richness (although mentioned in discussion). There was no indication how these points were chosen (random? All termites in an area sampled and identified? Transects?), nor how termites were sampled and identified. While M. subhyalinus had n=126, Odontotermes species (one? many?) only had n=6. Nevertheless, far reaching conclusions about predicted occurrence were also presented for Odontotermes. I question how reliable the model is with that many variables but only 6 samples as input.
No data for current distribution boundaries were given; nevertheless, authors concluded that species will “expand” into other areas. Were those species not already there?
Several times the accuracy of prediction of occurrence was mentioned. Was that compared to actual observations?
The largely unedited Discussion needs major revision concerning language, sentence structure, paragraph content, logical argumentation lines and citation formats. The discussion cannot be properly reviewed in its current state.
The discussion mentions “impacts of the two variables in predicting the current and future climate change on the species richness" [L291]. Species richness was not measured since only 3 species were chosen.
All the claims for “expansion” (eg. L303) need to be checked against the current distribution. The species are probably already there. Perhaps the authors meant an increase in population density, but that was not supported by data merely based on "presence". Conclusions toward “expansion” of Odontotermes species are of limited validity (n=6).
Author Response
REVIEWER No. 1 (diversity-1925533)
Comments and Suggestions for Authors
Comment: The authors conducted a study on 3 termite species in NW Nigeria and used modelling to predict current and future distribution. While the idea has merit the execution needs to be improved. This is not a detailed review since the paper has to be improved to allow in-depth review, but I will point out the major problems I had.
Response: We would like to thank the Reviewer no. 1 for his/her time, encouraging words, and constructive comments that have certainly improved the quality of our manuscript. We have carefully studied all the comments and suggestions, raised and responded to them point-by-point below.
Comment: The paper is poorly edited. It includes paragraphs that apparently are instructions on how to write results (L97ff) and the latter part of the discussion needs much improvement in English grammar, sentence structure and presenting arguments in a logical fashion. Professional editing services could be used to improve the writing. Literature citations in the text are in an unusual format (check journal requirements). Figure resolution is low, text in Figures is blurry and distorted.
Response: We have employed the services of a native English speaker (mentioned in the acknowledgement section) and improved the English grammar/editing of the manuscript. The citations in the main text are updated as suggested so also the references based on the journal requirements and the resolutions of the Figures are improved while the distorted text are now arranged. Example the “title” that appeared on the title of the manuscript is also removed (line 2). We also upload the figures separately.
Comment: While there is a lot of description of the model in Mat & Met, the sampling description is rather vague. The model is apparently based on 156 termite samples (at least, I assume that is what “georeferenced points” mean). I assume only presence of these three species was counted, not absence and not population density or species richness (although mentioned in discussion). There was no indication how these points were chosen (random? All termites in an area sampled and identified? Transects?), nor how termites were sampled and identified. While M. subhyalinus had n=126, Odontotermes species (one? many?) only had n=6. Nevertheless, far reaching conclusions about predicted occurrence were also presented for Odontotermes. I question how reliable the model is with that many variables but only 6 samples as input.
Response: We thank the reviewer for rising this critical concern. Samples in this instance does not refer to individual termite samples but to termite mounds. So, each sample represents an active termite mound with all its inhabitants. We agree with reviewer that 6 points might not be enough for reliable prediction. Hence, we have removed Odontotermes species from the manuscript. As the reviewer mentioned, we used only present data for modelling. The population density or species richness and termite identification was done in separate publications and this paper is follow-up of the previous studies that have been already published (kindly see Aiki et al. (2019), Journal of Thermal Biology: doi: https://doi.org/10.1016/j.jtherbio.2019.102418 and Aiki et al (2020), International Journal of Tropical Insect Science, doi: https://doi.org/10.1007/s42690-020-00330-5. Nevertheless, we have provided some information (line 92 to 119) in the manuscript on the sampling procedure and cited these two papers to provide more information for the reader.
Comment: No data for current distribution boundaries were given; nevertheless, authors concluded that species will “expand” into other areas. Were those species not already there?
Response: Here, we are explaining the results of our model and not the ground truth. The model results are a probability of occurrence and not the actual distribution. It is difficult to conduct a survey for the entire country to assess the distribution due to resource limitation. That is the reason scientist use models to predict the habitat suitability of the species based on known input occurrence data for the species and environmental layers to generate predictions (using the known to predict the unknown). According to our model, some areas are not suitable for the studied termite species under current conditions, however, the suitable areas will increase in future.
Comment: Several times the accuracy of prediction of occurrence was mentioned. Was that compared to actual observations?
Response: We thank the reviewer for this comment. In maxent, the model accuracy is assessed in two ways: 1) sub-sampling method where the input data is splitted into two (70% for training the model and, 30% for validating the model), 2) cross-validation, where each input data point is used to train the model and then for validation. We used the latter-cross-validation method. The advantage of this method that it uses each data point for training and validating the model thus, it minimizes the selection bias, overfitting and provide more accurate results than the sub-sampling method. We have explained this in our manuscript (kindly see section 2.5 on models’ validation). Line 188 – 205.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Higher termites play very important roles in the ecology and environmental sustainability in Africa. This study used the model to predict the distribution of the three mound building termite species in Nigeria. I enjoy reading this manuscript, and suggest the acceptance of this MS after the authors make some minor revisions:
Introduction: It would be valuable to introduce the history and some examples of the Ecological Niche Modeling (ENM). It would help the readers who are unfamiliar with this model. It is also important to state the importance of these termites for Nigeria.
Line 75-84: This information can be placed in the introduction
Figures: All figures are indistinct. It would be important to improve the quality of figures to fit the requirement of publication.
Line 169-171: Please delete this part from the MS
Author Response
REVIEWER No. 2 (diversity-1925533)
Comments and Suggestions for Authors
Higher termites play very important roles in the ecology and environmental sustainability in Africa. This study used the model to predict the distribution of the three mound building termite species in Nigeria. I enjoy reading this manuscript, and suggest the acceptance of this MS after the authors make some minor revisions:
Comment: Introduction: It would be valuable to introduce the history and some examples of the Ecological Niche Modeling (ENM). It would help the readers who are unfamiliar with this model. It is also important to state the importance of these termites for Nigeria.
Response: We thank the reviewer for this comment. We have now introduced ENM as suggested by the reviewer line 67 to 70 and we have discussed about the importance of the termites in the last paragraph of our discussion.
Comment: Line 75-84: This information can be placed in the introduction
Response: Thank you for the observation and suggestion, the information given in line 75 – 84 are details of the study area and Nigeria as a whole. Given that similar description has already been presented in the introduction line 53 – 58.
Comment: Figures: All figures are indistinct. It would be important to improve the quality of figures to fit the requirement of publication.
Response: We have improved the quality of the figures to 500 dpi and uploaded the Figures separately.
Comment: Line 169-171: Please delete this part from the MS
Response: Line 169 – 171 has been deleted.
Author Response File: Author Response.pdf
Reviewer 3 Report
This study investigated the distribution of two termite species, M. subhyalinus and M. bellicosus, under current and predicted future environmental conditions. For this, the authors utilized the Maxent algorithm in R software. I believe this algorithm is appropriate for this study. The authors selected 8 out of the 19 environment variables well. The final model prediction results (Figures 4 and 5) are expected to make a significant contribution to ecosystem preservation plans by predicting changes in Nigeria's ecosystem. I consider this paper to be suitable for publication in this journal.
