Integrating Bioactive Compound Variation and Habitat Suitability to Map the Quality Zoning of Crataegus pinnatifida Bunge Under Human Activity and Climate Change: A Biomod2 Ensemble Modeling Approach

Comments 1: The MM section is pretty complete and well written. I am missing few more detail to help the reader to reproduce the analysis and be more didactic to allow the non-expert to better understand the work. However, with the exception of some cites and minor details I have no concerns with the section.

Response 1: Thanks for your comment on our Materials and Methods section. We revised the Materials and Methods according to your comment.

Comments 2: The results are clear but extense, the figures have minor errors to solve like the font size, more clarity should be preferred. And the supplementary figures are cited however little more description of trends showed in there should be better.

Response 2: Yes, Done. All suggestions regarding language expression, formatting adjustments, and figure revisions have been diligently incorporated. Thank you for your valuable input.

Comments 3: The Discussion can be improved, specially because of the redundance with introduction and the specification of units, statistical uncertainty, and the most important should be declared and disused the bias and limitations of the experimental approximation of the study.

Response 3: Yes, Done. We sincerely thank you for these insightful and constructive suggestions. We have thoroughly revised the Discussion section to address these points, and the changes have significantly improved the manuscript.

Comments 4: L 60-63 The algorithms named here, GLM, GBM, CTA, ANN, SRE, FDA and RF are not referenced the references 12 and 13 are only based on MaxEnt. Can you add a general reference which groups all these algorithms.

Response 4: Yes, Done. Thank you very much for your suggestion. The more appropriate reference has been cited.

(Revised manuscript Line 60)

Comments 5: L 83 Species name in italics.

Response 5: Yes, Done.

Comments 6:L 96-105 Species names in italics.

Response 6: Yes, Done.

Comments 7: L 110 Species name in italics.

Response 7: Yes, Done.

Comments 8: L 128 cite the article instead to mention the url link Fick, S.E. and R.J. Hijmans, 2017. 9 WorldClim 2: new 1km spatial resolution climate surfaces for global land areas. International Journal of Climatology 37 (12): 4302-4315. doi: 10.1002joc. 5086 IF: 2.8 Q3

Response 8: Yes, Done. Thank you for the recommended literature. We have studied it and cited it in our manuscript.

(Revised manuscript Line 125)

Comments 9:  L 132 cite it properly, including the organizations as authors (\(FAO\) and \(IIASA\)), the year of publication (\(2023\)), the title of the database (\(Harmonized\ World\ Soil\ Database\ version\ 2.0\)), and the DOI (\(10.4060/cc3823en\)). For example, the citation could be written as: FAO & IIASA (2023). Harmonized World Soil Database version 2.0. FAO; IIASA. https://doi.org/10.4060/cc3823en

Response 9: Yes, Done. We have properly cited the relevant literature, according to your advice. Thank you.

(Revised manuscript Line 129)

Comments 10: L 149 Specify versions of R and packages used.

Response 10: Yes, Done. Specify versions of R and packages (v 4.2.2) that have been added according to your comment. Thank you.

(Revised manuscript Line 146)

Comments 11:  L 170-171 Why you used this ranges? It is based in some criteria or just for convenience, add a reference or briefly explain why you used intervals and why equal-intervals.

Response 11: Thank you for this question. The model outputs a continuous suitability index (0-1), which is a standard metric in species distribution modeling to represent relative habitat suitability. To classify this continuous output into a map, we employed an equal-interval scheme. This method was chosen for its objectivity, as it allows for a straightforward and unbiased visualization of the entire suitability gradient without being influenced by the specific characteristics of a single model output.

(Revised manuscript Line 168-170)

Comments 12: L 175 The temporal form of the verb is not correct, in MM should not be in past. And even more you didn´t demonstrated since is the first time than it is proposed that the centroids are migrating, you are showing it not demonstrating. Change à2… were visually demonstrated …” by “…can be visually observed (or graphed) …

Response 12: Yes, Done. Thanks.

Comments 13: L 184 Species name in italics.

Response 13: Yes, Done.

Comments 14: L 193 Species name in italics.

Response 14: Yes, Done.

Comments 15: L 203 Is missing a line spacing before subheading in the whole MM section and the are two lines in excess spacing before Results section.

Response 15: Yes, Done.

Comments 16: L 208 In the whole section of Results and conclusions you should write C. pinnatifida in italics.

Response 16: Yes, Done.

Comments 17: L 209 You used Gbif and the virtual herbarium, it is not a rigorous protocol. Detail the protocol or change the sentence, something: like occurrence from Gbif and virtual herbarium were curated and visualized in…

Response 17: Yes, Done. Thank you a lot for your advice.

Comments 18: L 215 The “Whittaker biome classification” is your result? If not add a reference (In this section or MM section).

Response 18: Done. The “Whittaker biome classification” has been added to the section of Materials and Methods.

(Revised manuscript Line 213-214)

Comments 19: L 222 I suppose that hydrothermal is a mistake, you should refer to hydric and thermic conditions, not hydrothermal which are hot water from the volcanic activity. Check and correct it.

Response 19: Thank you, We have substituted the term ‘hydrothermal’ with the more precise descriptors ‘annual temperature’ and ‘annual precipitation’ based on your suggestions.

Comments 20: L 240-241 The word respectively doesn´t make sense here.

Response 20: Yes, Done, thanks a lot! We deleted “respectively”.

Comments 21: L 248 In the panel 2A There is a big proportion of overlapping and is difficult to understand the results, even more since the colors of different algorithms is pretty similar. For example, I cannot distinguish in the figure which is FDA and MAXNET. Did you consider to combine colors and shapes for the panel 2A?

Response 21: Yes, Done. Thank you for pointing out this visualization issue. We have changed panel 2A in Fig. 2.

(Revised manuscript Line245 )

Comments 22:  L 250 Change (A) by (B).

Response 22: Yes, Done. Thank you!

Comments 23: L 263-264 Are you arguing that with human activity the suitable are is higher? In other sentence seems like you are saying the opposite, like in figure 3D were the expansion of the habitat is 1.31 passing from with human to without human.

Response 23: Yes. We sincerely apologize for the improper description in our original manuscript. The section you pointed out has been thoroughly revised to eliminate any potential confusion.

To clarify, there is a conceptual distinction between the ecological niche and habitat suitability. The ecological niche refers to the theoretical multidimensional space of resource use by a species along environmental gradients, while habitat suitability represents the probabilistic projection of how well those niche requirements are met in geographical space. In our results, human activities have led to an expansion in both the suitable habitat and the ecological niche of C. pinnatifida. However, the extent of this expansion differs: the total suitable habitat increased by 4.98%, whereas the ecological niche expanded by 1.31%.

(Revised manuscript Line 267-275)

Comments 24: L 291 Figure 3C doesn´t clarify which one is with and withour human activity. It is WHI with? And OHI without? You should specify it.

Response 24: Yes, Done. The relevant information has been supplemented.

Comments 25: L 384 Panel 5B doesn´t make sense since is only one bar reaching one number, it is better to place the number of square kilometers. I suggest to avoid it, it is not informative, is less informative than the pure number

Response 25: Yes, Done. The figure has been modified according to your requirements.

Comments 26: L 389 The figure 5D What represents? Which this area is? Is a province, is part of a single territory, explain it. And also there are two panels one for each scenario of future climate, add a letter for each panel of each scenario.

Response 26: Yes, Done.

Comments 27: L 396-398 Formally in results section references are not allowed. Move these sentences to Introduction since is not part of your results but this is part of the state of the art..

Response 27: Yes, Done. The revisions have been made in accordance with your suggestions.

Comments 28:L 456-464 Most of this paragraph resembles more introduction than discussion and the final part is quite redundant eith previous paragraphs, I suggest to delete the paragraph or replace the information that is content in the paragraph to a propper place.

Response 28: Yes, Done. We sincerely appreciate this constructive suggestion. The paragraph has been removed as suggested.

Comments 29: L 503 The current distribution is not “align” with your predictions, in fact it is the data to build your models and to stablish your predictions, for sure it mathc what you find. Rewrite the sentence it looks tricky.

Response 29: Yes, we rewrote the sentence according to your suggestion. “The modeled high-quality distribution accurately recaptures the empirical occurrence patterns of C. pinnatifida medicinal materials across China, validating the projection framework.”

Thank you for pointing out the issue.

Comments 30: L 534 Instead of writing the name of the variables write the meaning of what the variables are, for example instead of Bio 4, write temperature seasonality.

Response 30: Done. We write the specific variables instead of the abbreviation.

Comments 31: 34 L 542 You don´t have any patent, avoid it.

Response 31: Thank you for the reminder. The section on “Patent” has been removed.

Comments 1: In the title, add an author citation for the plant's scientific name. Integrating Bioactive Compound Variation and Habitat Suitability to Map the Quality Zoning of Crataegus pinnatifida Under Human Activity and Climate Change: A Biomod2 Ensemble Modeling Approach should be changed to Integrating Bioactive Compound Variation and Habitat Suitability to Map the Quality Zoning of Crataegus pinnatifida Bunge Under Human Activity and Climate Change: A Biomod2 Ensemble Modeling Approach

Response 1: Yes, Done. Thanks for your suggestion.

Comments 2: Line 73: Crataegus pinnatifida, commonly, should be changed to Crataegus pinnatifida Bunge

Response 2: Yes, Done. Thank you.

Comments 3: The ensemble model's credibility is weakened by being built from only two similar algorithms (RF, XGBoost), contradicting the core purpose of combining diverse models to reduce uncertainty.

Response 3: We sincerely thank you for this insightful comment regarding the composition of our ensemble model. We fully agree that. In our study, we initially evaluated 11 candidate species distribution modeling methods. Through this screening process, we found that Random Forest (TSS = 0.9992, ROC = 1) and XGBoost (TSS = 0.9852, ROC = 1) substantially outperformed all other algorithms, which showed notably lower performance (all TSS < 0.8). Given this pronounced performance gap, we made a deliberate, data-driven decision to restrict the ensemble to these two top-performing models.

We acknowledge that this approach represents a trade-off between algorithmic diversity and predictive performance. In future work, we will certainly incorporate your valuable suggestion by exploring strategies to include a wider variety of well-performing models, thereby enhancing the robustness and generalizability of our ensemble framework.

Thank you very much.

Comments 4: The spatial prediction of bioactive compounds relies on simple linear models. The absence of key performance metrics (R², RMSE) and validation makes the quality zonation maps statistically unsubstantiated.

Response 4: We appreciate your constructive feedback on the statistical validation of our quality zonation maps. We acknowledge that our linear regression approach, while statistically significant (P < 0.05), would benefit from additional performance metrics such as R² and RMSE values, as well as cross-validation procedures. This modeling approach is consistent with established methodologies used in comparable studies examining spatial patterns of plant secondary metabolites [1-3]. However, we chose this approach for its biological interpretability and practical applicability given our sample size constraints. We position these maps as preliminary zonation patterns that provide valuable insights into environmental drivers of bioactive compound distribution rather than precise quantitative predictions. Future studies will incorporate more robust validation frameworks and ensemble modeling approaches to enhance predictive accuracy.

Reference

[1] Li X, Wu T, Kang C, et al. Simulation of Pseudostellaria heterophylla distribution in China: assessing habitat suitability and bioactive component abundance under future climate change scenarios plant components. Front Plant Sci. 2024 Dec 4;15:1498229

[2] Guan L, Yang Y, Jiang P, Mou Q, Gou Y, Zhu X, Xu YW, Wang R. Potential distribution of Blumea balsamifera in China using MaxEnt and the ex situ conservation based on its effective components and fresh leaf yield. Environ Sci Pollut Res Int. 2022 Jun;29(29):44003-44019.

[3] Mu Kai-lang, Ran Fei, Feng Yun-qian, et al. Integrated evaluation of habitat suitability and spatial prediction of medicinal quality for Bletilla striata using species distribution models and machine learning approaches. Ind Crop Prod. Volume 236, 2025,121937.

Comments 5: A critical conflict exists where high fruit quality and high leaf quality are predicted in different geographic regions. This fundamental trade-off for cultivators is not discussed or interpreted.

Response 5: We thank you for this valuable comment. We acknowledge that the geographic trade-off between high fruit quality and high leaf quality is a key finding of our study, and we apologize for the omission of its discussion in the original manuscript. In response, we have thoroughly revised the Discussion section.

(Revised manuscript Line 504-510)

Comments 6: The conclusion that human activities "expand" high-suitability areas is questionable without clarifying how the "no influence" scenario was technically modeled (e.g., was the HFI variable set to zero?).

Response 6: Thank you for this important question. To clarify, the "no human influence" scenario was modeled by excluding the HFI variable entirely during model training, not by setting it to zero. The projected output from this no-HFI model was then compared with the model that included HFI. The observed increase in high-suitability area in the scenario with human influence supports our conclusion of habitat expansion.

Comments 7: The study projects future habitat shifts but fails to project how the medicinal quality (citric acid, flavonoids) will change under future climate scenarios, a major gap given the paper's objectives.

Response 7: Thank you. We agree with you that projecting medicinal quality is an important goal.  This, however, requires a different modeling framework (e.g., ecophytochemical models) and data on soil-geochemical futures, which are currently unavailable.

Comments 8: The ecological niche overlap metrics (D and I) are reported but their biological meaning—whether human activity causes a niche shift or expansion—is not sufficiently explained.

Response 8: Thank you a lot for your suggestion. The biological meanings D and I were added to the revised manuscript. Then the influence of human activity on niche shift or expansion of C. pinnatifida was also added to the revised manuscript. (Revised manuscript Line277-285)

Comments 9: The number of occurrence records differs between the Methods (n=331) and Results (Fig. 1A, n=231), creating a discrepancy that must be resolved.

Response 9: Thank you for highlighting this discrepancy. The accurate number of occurrence records used in the analysis is 231. The figure of 331 in the Methods section was an oversight, and we sincerely apologize for this error. We will correct it to 231 throughout the manuscript to ensure consistency.