Identifying Priority Bird Habitats Through Seasonal Dynamics: An Integrated Habitat Suitability–Risk–Quality Framework

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study presents a novel integrated framework to assess seasonal habitat quality for migratory birds in Wuhan, China, combining MaxEnt, HRA, and HQ models. The topic is timely and addresses critical gaps in conservation planning by incorporating seasonal dynamics. The methodological approach is innovative, and the findings provide actionable insights for urban biodiversity management. However, several aspects require clarification, methodological justification, and deeper contextualization to enhance scientific rigor and readability.

 

1. MaxEnt Model Validation:The manuscript mentions AUC > 0.7 for all models but does not provide species-specific AUC values or uncertainty measures (e.g., standard deviations across cross-validation folds). Include a table or supplementary material detailing AUC scores and variable contributions per species/season. Clarify how spatial autocorrelation was addressed beyond rarefaction. For instance, spatial blocking or independent validation datasets could improve robustness.
2. Habitat Risk Assessment (HRA):The scoring criteria for exposure/consequence (Table 3) lack justification. Explain how thresholds (e.g., 10% overlap for “low” risk) were determined. Were these based on literature, expert opinion, or empirical data? The HRA model uses annual data due to “difficulty in obtaining seasonal land use data.” Acknowledge this limitation and discuss potential biases (e.g., seasonal threats like irrigation in croplands may not be captured).
3. K-means Clustering:While the elbow method is mentioned for selecting K=4, the manuscript does not show the SSE-K plot or statistical validation (e.g., silhouette scores). Provide supplementary figures/analysis to justify cluster number selection. Clarify how clustering results were validated ecologically (e.g., ground truthing or comparison with independent habitat surveys).
4. Spatial Mismatch Analysis:The claim that Cluster 1 (year-round high quality) is “severely underprotected”(Section 4.2) lacks quantitative comparison. Include a statistical test (e.g., chi-square) to compare protection levels across clusters.Discuss potential reasons for mismatches (e.g., historical land-use policies, economic priorities) to contextualize findings.
5. Seasonal Habitat Quality Drivers:The higher suitability of wetlands in autumn/winter is attributed to “hydrological dynamics,” but no data (e.g., water levels, precipitation patterns) are presented to support this. Incorporate hydrological time-series data or references to regional climate studies.
6. Comparison with Existing Literature:The discussion focuses on Wuhan-specific outcomes but lacks broader comparisons. For example, how do the seasonal patterns align with studies in other East Asian flyway cities (e.g., Shanghai, Seoul)?Contrast the proposed framework with other integrated models (e.g., InVEST + circuit theory) to highlight its uniqueness.
7. Practical Conservation Strategies:Recommendations (e.g., “hydrological restoration” in Cluster 1) are not linked to case studies or cost-benefit analyses. Cite examples where similar interventions succeeded/failed in urbanizing regions.

Author Response

General comments

This study presents a novel integrated framework to assess seasonal habitat quality for migratory birds in Wuhan, China, combining MaxEnt, HRA, and HQ models. The topic is timely and addresses critical gaps in conservation planning by incorporating seasonal dynamics. The methodological approach is innovative, and the findings provide actionable insights for urban biodiversity management. However, several aspects require clarification, methodological justification, and deeper contextualization to enhance scientific rigor and readability.

Response: Thank you for your constructive comments and suggestions. The detailed responses are list as follows:

 

Specific comments

1. MaxEnt Model Validation: The manuscript mentions AUC > 0.7 for all models but does not provide species-specific AUC values or uncertainty measures (e.g., standard deviations across cross-validation folds). Include a table or supplementary material detailing AUC scores and variable contributions per species/season. Clarify how spatial autocorrelation was addressed beyond rarefaction. For instance, spatial blocking or independent validation datasets could improve robustness.

Response: Thank you for your professional and constructive comments. First, in the revision, we have added Supplementary Tables S2 and S3. Supplementary Table S2 provides the mean test AUC values and standard deviations from ten-fold cross-validation for each bird species in four seasons. Supplementary Table S3 presents the mean percent contributions of each environmental factors to the habitat suitability models, calculated separately by species and season.

Second, regarding spatial autocorrelation, we applied spatial rarefaction to reduce autocorrelation of bird occurrence points. However, we acknowledge that rarefaction alone may not fully eliminate spatial dependence. As suggested, we have addressed this limitation in the revised Discussion section and suggested that future studies could incorporate spatial blocking or independent validation datasets to improve the spatial generalizability of the model and minimize potential spatial biases.

Please see Tables S2 and S3:

Table S2. Bird species, sample sizes, and test AUC values of the MaxEnt habitat suitability model for spring, summer, autumn, and winter.

Species	Spring		Summer		Autumn		Winter
	N	AUC ± SD	N	AUC ± SD	N	AUC ± SD	N	AUC ± SD
Acridotheres cristatellus	15	0.851 ± 0.13	26	0.855 ± 0.17	49	0.671 ± 0.12	36	0.786 ± 0.10
Alcedo atthis	11	0.909 ± 0.13	13	0.968 ± 0.03	28	0.752 ± 0.21	16	0.694 ± 0.27
Copsychus saularis	20	0.915 ± 0.08	28	0.940 ± 0.06	46	0.815 ± 0.14	16	0.623 ± 0.32
Cyanopica cyanus	32	0.881 ± 0.08	42	0.937 ± 0.07	95	0.817 ± 0.10	52	0.794 ± 0.08
Eophona migratoria	18	0.896 ± 0.10	16	0.932 ± 0.07	34	0.853 ± 0.13	28	0.799 ± 0.11
Gallinula chloropus	18	0.870 ± 0.12	24	0.924 ± 0.14	35	0.725 ± 0.18	29	0.785 ± 0.09
Lanius schach	11	0.743 ± 0.15	14	0.934 ± 0.06	34	0.691 ± 0.15	40	0.832 ± 0.06
Parus minor	16	0.812 ± 0.13	15	0.940 ± 0.09	42	0.836 ± 0.14	21	0.777 ± 0.18
Passer montanus	31	0.873 ± 0.10	36	0.937 ± 0.08	66	0.836 ± 0.11	40	0.739 ± 0.05
Pica serica	23	0.893 ± 0.06	16	0.827 ± 0.21	71	0.754 ± 0.13	50	0.799 ± 0.07
Pterorhinus perspicillatus	31	0.910 ± 0.07	31	0.913 ± 0.11	53	0.817 ± 0.16	24	0.834 ± 0.11
Pycnonotus sinensis	31	0.900 ± 0.10	33	0.927 ± 0.08	73	0.862 ± 0.08	43	0.830 ± 0.11
Spilopelia chinensis	31	0.917 ± 0.05	32	0.919 ± 0.10	70	0.751 ± 0.13	44	0.814 ± 0.10
Spodiopsar sericeus	19	0.869 ± 0.09	18	0.894 ± 0.16	43	0.738 ± 0.15	36	0.777 ± 0.16
Streptopelia orientalis	20	0.866 ± 0.13	27	0.952 ± 0.08	51	0.817 ± 0.21	26	0.711 ± 0.16
Tachybaptus ruficollis	13	0.751 ± 0.24	14	0.813 ± 0.20	57	0.766 ± 0.14	50	0.832 ± 0.07
Turdus mandarinus	33	0.863 ± 0.08	36	0.927 ± 0.10	57	0.803 ± 0.12	47	0.811 ± 0.06

Note: N indicates the sample size; AUC ± SD represents the observed mean and one standard deviation calculated from the replicates of the ten-fold cross-validation procedure.

Table S3. Mean percent contributions of environmental factors to habitat suitability across all bird species in four seasons.

Environmental factors	Spring	Summer	Autumn	Winter
LUCC	24.56	4.15	2.73	7.95
DtoHR	16.67	15.11	6.76	6.27
DtoLR	18.10	8.16	6.99	4.91
NPP	4.00	21.20	14.55	21.78
DtoU	11.19	17.73	39.49	9.61
DtoF	5.64	15.32	5.16	7.75
DEM	0.20	1.61	0.52	0.76
Slope	3.22	3.63	3.25	2.76
Aspect	4.13	4.77	7.71	11.04
TEMP	0.32	1.36	2.15	9.25
PRE	6.38	1.18	1.62	7.51
NDVI	0.65	0.16	4.44	4.73
DtoW	2.74	4.14	0.87	1.30
DtoB	1.00	0.76	1.67	1.49
DtoC	0.06	0.41	0.37	1.32
DtoBL	1.16	0.32	1.73	1.55

Please see lines 713–719:

Second, although all MaxEnt models performed well with AUC values exceeding 0.7, residual spatial autocorrelation may still influence model performance. Spatial thinning was applied to reduce clustering, but the limited sample size for some species precluded the use of more rigorous spatial validation techniques, such as spatial blocking or independent test datasets. This limitation may have led to a modest over-estimation of predictive accuracy. Future applications should explore more robust spatial validation strategies to enhance model reliability.

 

2. Habitat Risk Assessment (HRA): The scoring criteria for exposure/consequence (Table 3) lack justification. Explain how thresholds (e.g., 10% overlap for “low” risk) were determined. Were these based on literature, expert opinion, or empirical data? The HRA model uses annual data due to “difficulty in obtaining seasonal land use data.” Acknowledge this limitation and discuss potential biases (e.g., seasonal threats like irrigation in croplands may not be captured).

Response: Thank you for your constructive comments and suggestions. First, we have revised the Methods section to provide a more detailed explanation for the threshold values applied in the exposure and consequence criteria. In addition, we have added a “Reference” column in Table 3 to clearly show the sources that informed the threshold values and scoring criteria. Specifically, the table includes seven criteria: spatial overlap thresholds were derived from Samhouri et al. (2012) and Arkema et al. (2014); temporal overlap, intensity, and change in area were based on the HRA model documentation and Caro et al. (2020); connectivity thresholds followed the HRA model guidance and Bastos et al. (2023); water quality was classified according to China’s National Surface Water Environmental Quality Standard (GB 3838–2002);and NPP thresholds were determined with reference to values reported in Li et al. (2018) and Chen et al. (2020). Additionally, we have revised the Discussion section to explicitly acknowledge the limitations and potential biases associated with using annual data in the HRA model.

Please see lines 258–288 and Table 3:

First, habitats and threats were defined by classifying land use data based on the biological and ecological characteristics of the target species (Table S1). Habitat types included forest, wetland, and grassland, while threat types consisted of cropland, urban land, and bare land, all associated with agricultural and industrial activities. Next, a habitat risk assessment framework was established by selecting a set of criteria from both the exposure and consequence dimensions (Table 3). The selected criteria were then graded, scored, rated for data quality, and weighted based on previous research and expert opinions [68–71]. Exposure refers to the degree to which habitats are threatened and includes three criteria: spatial overlap, temporal overlap, and intensity. Spatial overlap quantifies the proportion of habitat area overlapping with threats; temporal overlap indicates the duration of exposure to threats; and intensity reflects the strength of the threats. The classification of these criteria was based on the HRA model documentation and previous studies [68,72,73]. Consequence refers to habitat sensitivity and resilience to threats, represented by four criteria: change in area, connectivity, water quality, and NPP. Change in area reflects habitat sensitivity, indicating the degree of habitat encroachment by threats, with thresholds were defined based on the HRA model documentation and Caro et al. (2020) [67, 73]. Connectivity, water quality, and NPP reflect habitat resilience. Among them, connectivity supports ecological processes, and its classification was determined according to the HRA model documentation and Bastos et al. (2023) [67, 74]. Water quality and NPP indicate habitat health. The classification for water quality was based on China’s National Surface Water Environmental Quality Standard (GB 3838–2002), where Class III water is suitable for fish overwintering and spawning [75]. The classification thresholds for NPP were defined with reference to the values reported by Li et al. (2018) and Chen et al. (2020), which indicated that the mean NPP of forests in Wuhan is approximately 0.5 kgC/m²/a, while non-forest and grassland types are generally below 0.2 kgC/m²/a [76,77]. Subsequently, each criterion was classified into three levels and assigned scores from 1 to 3, with higher scores indicating greater exposure or consequence [78]. Similarly, data quality and criterion weights were categorized into three levels, with scores ranging from 1 to 3, where higher scores denoted lower data accuracy or greater weight [79]. Finally, habitat risk was calculated using the Euclidean distance-based risk function, as de-scribed in Equations (1) – (4) [68,74].

Table 3. Scoring Criteria for Exposure and Consequence Dimensions.

Criteria		Low (1)	Medium (2)	High (3)	Reference
Exposure	Spatial Overlap Area Rating	< 10%	10%–30%	≥30%	Samhouri et al. (2012); Arkema et al. (2014)
	Temporal Overlap Rating	0–4 months per year	4–8 months per year	8–12 months per year	HRA model documentation; Caro et al. (2020)
	Intensity Rating	Low Intensity	Medium Intensity	High Intensity	HRA model documentation; Caro et al. (2020)
Consequence – Sensitivity	Change in Area	Low loss (< 20%)	Medium loss (20%–50%)	High loss (> 50%)	HRA model documentation; Caro et al. (2020)
Consequence – Resilience	Connectivity	Highly (> 100 km)	Medium (10–100 km)	Low (< 10 km)	HRA model documentation; Bastos et al. (2023)
	Water Quality	> 70% of the area with water quality better than Class III	50%–60% of the area with water quality better than Class III	< 50% of the area with water quality better than Class III	China’s National Surface Water Environmental Quality Standard (GB 3838–2002)
	NPP (kgC/m²/a)	High (> 0.5)	Moderate (0.2–0.5)	Low (< 0.2)	Li et al. (2018); Chen et al. (2020)

Please see lines 707–713:

Additionally, due to the absence of seasonally explicit land use data, the HRA model relied on annual inputs. This approach may fail to reflect critical seasonal threats such as irrigation in agricultural areas or tourism pressure in wetlands, thereby limiting the temporal accuracy of risk assessments. Future research should incorporate high-er-resolution, temporally explicit spatial data (e.g., monthly land use or 10 m imagery) to more accurately delineate both temporary and permanent habitats, as well as dynamic urban boundaries.

 

3. K-means Clustering: While the elbow method is mentioned for selecting K=4, the manuscript does not show the SSE-K plot or statistical validation (e.g., silhouette scores). Provide supplementary figures/analysis to justify cluster number selection. Clarify how clustering results were validated ecologically (e.g., ground truthing or comparison with independent habitat surveys).

Response: Thank you for your constructive comments and suggestions. We have revised the Methods section to clarify how the optimal number of clusters was determined. The elbow method was initially used to identify a suitable range of K values, followed by silhouette score analysis to confirm the final selection. To support this, we have added Figure S2 in the supplementary materials, which presents both the SSE-K curve and silhouette scores across K values. Furthermore, we agree that ecological validation through ground truthing or independent habitat surveys would offer a more robust assessment of the clustering results. While such data were not available in this study, we conducted a spatial overlay analysis between the identified habitat clusters and existing protected areas in Wuhan, including nature reserves, forest parks, and scenic areas, to assess the ecological relevance of our approach. The analysis revealed a substantial spatial overlap between protected areas and Cluster 1 (46.22 km², 46.51%) as well as Cluster 3 (179.44 km², 46.30%), which represent regions of high habitat quality year-round and in winter, respectively. This spatial alignment suggests that the clustering results are ecologically meaningful and consistent with current conservation priorities in the region.

Please see lines 364–376 and Figure S2:

Then, the optimal number of clusters (K = 4) was determined using the elbow method and silhouette score [86,87]. The sum of squared errors (SSE) was plotted against K values ranging from 2 to 9, and the resulting curve showed an inflection point between K = 3 and 5, indicating a potential elbow. This results was further validated through silhouette analysis, which confirmed that K = 4 was the optimal number of clusters (Fig. S2). Next, the standardized seasonal habitat quality feature data were assigned to the nearest cluster centers using the Euclidean distance metric. Subsequently, to validate the ecological relevance of the clusters and identify existing conservation gaps, a spatial overlay analysis was conducted between the clustering results and the distribution of nature reserves, forest parks, and natural scenic areas in Wuhan. Finally, based on the clustering results and the characteristics of each cluster, corresponding conservation strategies were developed, with habitat patches maintaining high quality across all seasons identified as priority areas for protection.

Figure S2. Determination of optimal number of clusters using the Elbow method and Silhouette Score.

4. Spatial Mismatch Analysis: The claim that Cluster 1 (year-round high quality) is “severely underprotected”(Section 4.2) lacks quantitative comparison. Include a statistical test (e.g., chi-square) to compare protection levels across clusters. Discuss potential reasons for mismatches (e.g., historical land-use policies, economic priorities) to contextualize findings.

Response: Thank you for your constructive comments and suggestions. We fully agree with your recommendation and have revised the manuscript accordingly. Specifically, we conducted an updated spatial overlay analysis between the habitat clusters and existing protected areas, followed by a chi-square test to quantitatively assess differences in protection levels across clusters.

The results revealed significant disparities in protection coverage, confirming that Cluster 1 (representing year-round high-quality habitats) is relatively underprotected compared to other clusters. Although Cluster 1 exhibited the highest protection efficiency (46.51%), it had the smallest protected area (46.22 km²), indicating a substantial spatial mismatch given its ecological significance. To contextualize this finding, we further expanded the Discussion section to explore potential drivers of this mismatch.

Please see lines 614–654:

The current protected area in Wuhan exhibits significant spatial mismatches with habitats of highest ecological value, revealing critical conservation gaps (Fig. 6). The differences in protection levels among the four clusters are statistically significant (χ² = 64.72, df = 3, p < 0.001) (Table 9). Cluster 1, which maintains high-quality habitat throughout the year, has the highest protection rate (46.51%) but the smallest total protected area (46.22 km²), leaving 53.16 km² of ecologically valuable areas unprotected. These areas include key habitats such as Qingshan Wetland, Wuhui Dyke, and the riparian wetlands along the Yangtze River between Yingwuzhou and Baishazhou Bridges. Notably, 60.59% (32.21 km²) of these unprotected areas are located within 5 km of main urban areas, including Baishazhou in Wuchang District and Wugang in Qingshan District, making them particularly vulnerable to industrial infrastructure and intensive human activities. Cluster 3, which consists of habitats with high quality during winter, also exhibits a relatively high protection rate (46.30%) and a much larger total protected area (179.44 km²). Critical wetlands in this cluster, such as Chenhu Wetland in Caidian District, Zhangdu Lake in Xinzhou District, and East Lake in Hongshan District, are formally protected. Cluster 2, characterized by medium habitat quality year-round, contains the largest area under formal protection (187.47 km²), accounting for 22.23% of its total area. Protection areas include Anshan National Wet-land Park in Jiangxia District and Mulan Mountain Nature Reserve in Huangpi District. Cluster 4, which reaches peak habitat quality during migration seasons, has a protection rate of 31.50% (70.21 km² out of 222.89 km²). Unprotected areas in this cluster, such as Bafen Mountain in Jiangxia, Guanlian Lake, and the Tongshun and Maying Rivers in Caidian, continue to face pressure from agricultural activities and infrastructure development.

This spatial mismatch highlights a broader challenge in global conservation efforts [88,104,105]. In rapidly growing cities like Wuhan, these mismatches are often caused by urban sprawl, especially urban planning decisions and urban functional zoning [106]. For example, industrial developments along the Caidian–Hannan border and in western Jiangxia District have been largely driven by economic priorities, often overlooking the ecological roles of these areas, such as supporting seasonal bird migration and sustaining wetland biodiversity. Moreover, Moreover, conservation strategies still tend to focus on large, remote habitat patches, while smaller, high-value habitats within or near urban areas are frequently neglected [92,107]. Although modest in size, these habitats are crucial for maintaining ecological connectivity, particularly as step stones along migratory routes. Yet their scattered distribution and exposure to development pressure have left many without formal protection. Taken together, the effectiveness of our framework in capturing seasonal ecological dynamics and urban complexity enables more accurate identification of habitat clusters and critical conservation gaps. This is essential for balancing ecological integrity with long-term sustainable urban development.

Table 9. Protection status of habitat clusters and results of chi-square test.

Cluster	Protected area (km²)	Unprotected area (km²)	Total area (km²)	Protection rate (%)
Cluster 1	46.22	53.16	99.38	46.51
Cluster 2	187.47	655.68	843.15	22.23
Cluster 3	179.44	208.16	387.60	46.30
Cluster 4	70.21	152.68	222.89	31.50
Total	483.34	1069.68	1553.02	31.22

Note: Chi-square test comparing protection status across clusters yielded χ² = 64.72, df = 3, p < 0.001.

 

5. Seasonal Habitat Quality Drivers: The higher suitability of wetlands in autumn/winter is attributed to “hydrological dynamics,” but no data (e.g., water levels, precipitation patterns) are presented to support this. Incorporate hydrological time-series data or references to regional climate studies.

Response: Thank you for your insightful comments. We agree that providing empirical evidence is essential to support the explanation of higher suitability in autumn and winter. In the revised Discussion section, we added hydrological data to support the observed seasonal trends. In 2020, precipitation in Wuhan from June to November exceeded historical averages by 48.9% to 214.2% (Wuhan Water Resources Bulletin, 2020), while groundwater levels remained stable between 21.71 and 23.11 meters (Wang et al., 2025). These conditions likely improved wetland water availability in autumn and winter, contributing to higher habitat suitability for wetland-dependent bird species.

Please see lines 571–579:

For example, in Wuhan, wetland suitability varied from 0.0561 in summer to 0.1854 in autumn, may due to seasonal hydrological conditions. In 2020, precipitation in Wuhan during summer and autumn was significantly higher than the multi-year average, with rainfall from June to November exceeding historical levels by 48.9% to 214.2% [95]. Meanwhile, seasonal groundwater levels remained relatively stable, ranging between 21.71 and 23.11 meters [96]. These findings suggest that abundant rainfall and steady groundwater storage in summer and autumn provided favorable hydrological conditions that helped maintain high wetland habitat suitability in the autumn and winter seasons.

 

6. Comparison with Existing Literature: The discussion focuses on Wuhan-specific outcomes but lacks broader comparisons. For example, how do the seasonal patterns align with studies in other East Asian flyway cities (e.g., Shanghai, Seoul)? Contrast the proposed framework with other integrated models (e.g., InVEST + circuit theory) to highlight its uniqueness.

Response: Thank you for your valuable suggestion. In response, we have revised the Discussion section to compare our findings with studies from other East Asian flyway cities. We have also contrasted our proposed framework with other integrated models such as InVEST combined with circuit theory, highlighting its unique emphasis on seasonal habitat dynamics.

Please see lines 588–612:

By incorporating these refined parameters into the HQ model, we achieved a more accurate assessment of seasonal habitat quality. The results indicate that habitat quality in Wuhan is generally higher in autumn and winter than in spring and summer. This pattern is consistent with local bird field surveys [98,99], suggesting that seasonal habitat conditions influence the spatiotemporal distribution of avian diversity. Similar seasonal patterns have been observed in other East Asian flyway cities. For instance, long-term monitoring in Shanghai’s Dongtan wetlands has revealed significantly larger wintering bird populations than in summer [100]. In South Korea, waterbird species richness peaks during spring and autumn migrations, but declined sharply in winter due to widespread wetland freezing [101,102]. In contrast, Wuhan’s milder winter climate with minimal wetland freezing, allowing high habitat suitability throughout the autumn and winter seasons. These comparisons suggest that, despite regional cli-matic differences, the ecological importance of autumn and winter habitats is broadly consistent across East Asian flyway cities. In addition, our results showed that high-quality habitats were concentrated along a southwest axis (approximately 214° from north), closely aligning with the Yangtze River. This spatial pattern matched migratory bird flight corridors identified through weather radar monitoring [103], further validating the effectiveness of our framework. These findings are consistent with previous studies emphasizing the value of parameter refinement in improving habitat quality assessments [25]. Unlike integrated approaches such as InVEST combined with circuit theory, which primarily focus on migration routes and ecological connectivity [36], our framework emphasizes seasonal habitat conditions. By integrating seasonally differentiated habitat suitability from the MaxEnt model, ecological risk from the HRA model, and refined parameters in the HQ model, our approach provides a reliable and effective tool for evaluating seasonal habitat quality and identifying conservation priorities, particularly in rapidly urbanizing inland regions.

 

7. Practical Conservation Strategies: Recommendations (e.g., “hydrological restoration” in Cluster 1) are not linked to case studies or cost-benefit analyses. Cite examples where similar interventions succeeded/failed in urbanizing regions.

Response: Thank you for your insightful suggestion. We have revised the Discussion section to incorporate real-world case studies and cited examples of similar conservation interventions in both domestic and international contexts. These examples support the feasibility and broader applicability of the proposed strategies.

Please see lines 655–694:

To address these gaps, conservation strategies should be developed based on the characteristics and seasonal dynamics of habitat clusters [106]. First, Cluster 1 should be legally designated as an ecological priority protection zone to prevent urban encroachment and overexploitation. For instance, land reclamation in East Lake should be strictly prohibited, and a 1 km buffer zone should be established to mitigate edge effects. This measure is consistent with China’s Ecological Protection Red Line policy [108–110]. In addition, hydrological restoration, such as dredging blocked waterways in the Houguan Lake, Jinlong Lake, and Suozichang River wetland complexes in Caidian District, should be implemented. This measure has been shown to be effective, as re-storing hydrological connectivity in fragmented wetland systems can significantly im-prove ecological function and biodiversity [111,112]. Second, for Cluster 2, ecological corridors 100–300 meters wide should be constructed along major rivers such as the Han River to connect medium- and high-quality habitats, promoting species movement and genetic exchange [113]. This type of riparian habitat restoration has proven effective in both Beijing, China and the Colorado River Delta, Mexico, demonstrating its broader applicability across urbanizing and degraded river systems [114,115]. These findings underscore the role of riparian corridors in mitigating the negative effects of urbanization on avian communities. Additionally, low-impact agricultural practices, such as reduced pesticide use, should be adopted in adjacent croplands to alleviate long-term ecological pressures. Third, Cluster 3 requires seasonal management [19]. From November to February, activities such as fishing and boating should be restricted in key overwintering habitats such as Chenhu and Fuhe wetlands to minimize disturbances to species like the Baer’s Pochard (Aythya baeri). Meanwhile, restoration efforts should focus on planting native vegetation along degraded wetland edges. These plants not only improve microhabitat conditions but also provide essential shelter and food resources for wintering birds [116], thereby enhancing overall habitat quality and ecological function. Fourth, Cluster 4 should also implement seasonal management strategies [117], but with focus on spring and autumn. Construction and recreational activities should be prohibited during peak migration and breeding periods. This aligns with conservation guidelines by the U.S. Fish and Wildlife Service, which recommend limiting human disturbances, such as construction and noise, during sensitive biological periods, including migration seasons [118,119]. In addition to formally protected areas, Other Effective area-based Conservation Measures (OECMs) offer a complementary approach to biodiversity conservation, particularly in urbanizing regions [120]. OECMs are defined as areas that are not designated as protected areas but are governed and managed in ways that deliver long-term biodiversity ben-benefits [121]. In the context of Wuhan’s habitat clusters, certain urban wetlands, and riparian buffer zones that exhibit ecological value and limited human disturbance could potentially qualify as OECMs [122]. Identifying and integrating these areas into the broader conservation network may enhance ecological connectivity, support seasonal species needs, and promote long-term sustainability goals.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study proposed an integrated framework, “Habitat Suitability–Risk–Quality”, to improve the assessment of seasonal bird habitat quality and to identify priority conservation habitats in urban landscapes, which is quite an interesting study. The study design is well-structured, and the empirical analysis is comprehensive; however, further optimization is needed in the following areas:

1) In the results analysis section, particularly regarding protected gaps, the boundaries of nature reserves and clustering results are simply overlaid. It is recommended to include the boundaries of scenic spots, forest parks, wetland parks, urban parks, etc., and to clarify the protected gap areas through field surveys;

2) Currently, the assessment of habitat suitability, risk, and habitat quality is conducted at an overall level, lacking analysis of seasonal distribution characteristics of birds across different habitat types, resident birds, or migratory birds, as well as targeted gap analysis;

3) In the discussion section, it is recommended to include a discussion of OECM in the strategies.

4）The entire manuscript, especially the introduction, tends to use Chinese-style English. Further refinement is recommended.

Author Response

General comments

This study proposed an integrated framework, “Habitat Suitability–Risk–Quality”, to improve the assessment of seasonal bird habitat quality and to identify priority conservation habitats in urban landscapes, which is quite an interesting study. The study design is well-structured, and the empirical analysis is comprehensive; however, further optimization is needed in the following areas:

Response: Thank you for your constructive comments and suggestions. The detailed responses are list as follows:

 

Specific comments

1. In the results analysis section, particularly regarding protected gaps, the boundaries of nature reserves and clustering results are simply overlaid. It is recommended to include the boundaries of scenic spots, forest parks, wetland parks, urban parks, etc., and to clarify the protected gap areas through field surveys;

Response: Thank you for your valuable suggestion. In the revised manuscript, we expanded the spatial overlay analysis to include additional categories of protected areas, such as scenic spots, forest parks, and wetland parks. Due to limitations in field access and data availability, we were unable to conduct on-site validation of each gap area. However, this limitation has been acknowledged in the revised Discussion section, and we have proposed field surveys as a key direction for future research.

Please see Section 4.2 “Conservation Gaps and Seasonal Strategies Based on Habitat Clusters” and Figure 6.

Please see lines 732–735:

Fourth, conservation gaps were identified by overlaying habitat clusters with various protected areas (e.g., nature reserves, wetland parks, and forest parks), but lacked field validation. Future research should incorporate on-the-ground surveys to verify these gaps and support the identification of overlooked conservation sites.

 

2. Currently, the assessment of habitat suitability, risk, and habitat quality is conducted at an overall level, lacking analysis of seasonal distribution characteristics of birds across different habitat types, resident birds, or migratory birds, as well as targeted gap analysis;

Response: Thank you for this insightful suggestion. We acknowledge that our current analysis does not distinguish between resident and migratory birds or explore seasonal distribution patterns across different habitat types. This is indeed a limitation. Our primary goal was to develop a generalizable framework that captures the seasonal habitat patterns of most bird species in the study area. We have now clarified this objective in the revised Discussion section and also highlighted the importance of incorporating species-specific ecological traits and seasonal dynamics in future studies.

Please see lines 719–726:

Third, this study employed a multi-species approach based on 17 bird species to identify regional habitat priorities. While this method is effective for revealing general patterns, it does not account for species-specific seasonal responses or ecological preferences. This limitation is particularly relevant for rare or endangered species, whose habitat requirements may differ markedly from those of common species and may be more sensitive to seasonal or habitat changes. Future research should consider more refined, species-level analyses to support targeted conservation strategies for these vulnerable groups.

 

3. In the discussion section, it is recommended to include a discussion of OECM in the strategies.

Response: Thank you for your suggestion. In the revised Discussion section, we have added a paragraph introducing Other Effective area-based Conservation Measures (OECMs) and discussed their potential application in the context of Wuhan’s habitat clusters.

Please see lines 685–694:

In addition to formally protected areas, Other Effective area-based Conservation Measures (OECMs) offer a complementary approach to biodiversity conservation, particularly in urbanizing regions [120]. OECMs are defined as areas that are not designated as protected areas but are governed and managed in ways that deliver long-term biodiversity benefits [121]. In the context of Wuhan’s habitat clusters, certain urban wetlands, and riparian buffer zones that exhibit ecological value and limited human disturbance could potentially qualify as OECMs [122]. Identifying and integrating these areas into the broader conservation network may enhance ecological connectivity, support seasonal species needs, and promote long-term sustainability goals.

 

4. The entire manuscript, especially the introduction, tends to use Chinese-style English. Further refinement is recommended.

Response: Thank you for your suggestion. We have revised the manuscript thoroughly, particularly the Introduction section, to improve the language quality. The revised version has also been reviewed by a native English speaker to ensure clarity and fluency.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This study exhibits use of a novel framework to assess seasonal habitat quality for a set of avian species in Wuhan, China. It demonstrates that conservation strategies should include a consideration of seasonal dynamics and characteristics of various habitats representing natural and land-use areas. The framework and analysis are thorough although admittedly, other reviewers who have knowledge and understanding of the mathematics of the modeling as well as the statistical analysis can better review these processes. What may be missing is consideration of the ecological and biological life parameters of individual chosen species in the habitat suitability evaluation. The bird species were selected by noting those that had the highest occurrence frequencies (and thus were common). Habitat suitability analysis therefore was done with the dataset from all 17 species. There is no mention of habitat requirements dependent on biological and ecological parameters of the species. Habitat suitability for a waterfowl species would be different than for a common species often found in urban areas. A universal habitat suitability prediction was made, yet habitat suitability is also dependent on the species biological factors including ecology, population size, reproductive rate, conservation status, etc.

Table S1 in Supplemental Materials includes the list of avian species used in the study. It would be helpful to Include not only the common and scientific names but also identify ecological and biological parameters for these species. Habitat assessment should incorporate not only occurrence data, but individual life parameters for the species. Habitat suitability for species in the study may differ for migratory vs non-migratory species, waterfowl requiring water habitats vs passerines commonly found in urban areas, social species living in flocks versus those that live in pairs. For example, the crested myna (Acridotheres cristatellus) is a highly social common passerine with a large range and is non-migratory. It is found in open woodlands, cultivated areas, and urban settings, and has shown the ability to adapt to human-modified landscapes. The little grebe (Tachybaptus ruficollis) is found in open and freshwater coastal waters and is migratory only in areas where the lakes freeze over in winter. Suitable habitat for this species, no matter the season, would be completely different than that of the crested myna.

The manuscript would be strengthened by using examples of case studies of representative avian species to illustrate the habitat suitability in different seasons. This would demonstrate how the framework can be incorporated to inform conservation strategies. This study identifies that seasonal dynamics are an important consideration yet needs the input of ecological and biological parameters for the analysis.

Reference format: When using the citation format (last name of first author {et al. for more than two authors}) in the text, the references should be listed in alphabetical order by last name of first author and without numbers. The format used here is to list the references in order of citation, however there were 18 references that were cited more than once, and the reader had to search through previous entries to find these references. Listing by alphabetical order ensures that the reader can easily find each reference, no matter how many times it is cited in the text.

Please see the comments in the attached document.

Comments for author File: Comments.pdf

Author Response

General comments

This study exhibits use of a novel framework to assess seasonal habitat quality for a set of avian species in Wuhan, China. It demonstrates that conservation strategies should include a consideration of seasonal dynamics and characteristics of various habitats representing natural and land-use areas. The framework and analysis are thorough although admittedly, other reviewers who have knowledge and understanding of the mathematics of the modeling as well as the statistical analysis can better review these processes. What may be missing is consideration of the ecological and biological life parameters of individual chosen species in the habitat suitability evaluation. The bird species were selected by noting those that had the highest occurrence frequencies (and thus were common). Habitat suitability analysis therefore was done with the dataset from all 17 species. There is no mention of habitat requirements dependent on biological and ecological parameters of the species. Habitat suitability for a waterfowl species would be different than for a common species often found in urban areas. A universal habitat suitability prediction was made, yet habitat suitability is also dependent on the species biological factors including ecology, population size, reproductive rate, conservation status, etc.

 

Table S1 in Supplemental Materials includes the list of avian species used in the study. It would be helpful to Include not only the common and scientific names but also identify ecological and biological parameters for these species. Habitat assessment should incorporate not only occurrence data, but individual life parameters for the species. Habitat suitability for species in the study may differ for migratory vs non-migratory species, waterfowl requiring water habitats vs passerines commonly found in urban areas, social species living in flocks versus those that live in pairs. For example, the crested myna (Acridotheres cristatellus) is a highly social common passerine with a large range and is non-migratory. It is found in open woodlands, cultivated areas, and urban settings, and has shown the ability to adapt to human-modified landscapes. The little grebe (Tachybaptus ruficollis) is found in open and freshwater coastal waters and is migratory only in areas where the lakes freeze over in winter. Suitable habitat for this species, no matter the season, would be completely different than that of the crested myna.

 

The manuscript would be strengthened by using examples of case studies of representative avian species to illustrate the habitat suitability in different seasons. This would demonstrate how the framework can be incorporated to inform conservation strategies. This study identifies that seasonal dynamics are an important consideration yet needs the input of ecological and biological parameters for the analysis.

 

Reference format: When using the citation format (last name of first author (Allan et al.)) in the text, the references should be listed in alphabetical order by last name of first author and without numbers. The format used here is to list the references in order of citation, however there were 18 references that were cited more than once, and the reader had to search through previous entries to find these references. Listing by alphabetical order ensures that the reader can easily find each reference, no matter how many times it is cited in the text.

 

Please see the comments in the attached document.

Response: Thank you for your constructive comments and suggestions. The detailed responses are list as follows:

 

Specific comments

1. What may be missing is consideration of the ecological and biological life parameters of individual chosen species in the habitat suitability evaluation. The bird species were selected by noting those that had the highest occurrence frequencies (and thus were common). Habitat suitability analysis therefore was done with the dataset from all 17 species. There is no mention of habitat requirements dependent on biological and ecological parameters of the species. Habitat suitability for a waterfowl species would be different than for a common species often found in urban areas. A universal habitat suitability prediction was made, yet habitat suitability is also dependent on the species biological factors including ecology, population size, reproductive rate, conservation status, etc.

 

Table S1 in Supplemental Materials includes the list of avian species used in the study. It would be helpful to Include not only the common and scientific names but also identify ecological and biological parameters for these species. Habitat assessment should incorporate not only occurrence data, but individual life parameters for the species. Habitat suitability for species in the study may differ for migratory vs non-migratory species, waterfowl requiring water habitats vs passerines commonly found in urban areas, social species living in flocks versus those that live in pairs. For example, the crested myna (Acridotheres cristatellus) is a highly social common passerine with a large range and is non-migratory. It is found in open woodlands, cultivated areas, and urban settings, and has shown the ability to adapt to human-modified landscapes. The little grebe (Tachybaptus ruficollis) is found in open and freshwater coastal waters and is migratory only in areas where the lakes freeze over in winter. Suitable habitat for this species, no matter the season, would be completely different than that of the crested myna.

 

The manuscript would be strengthened by using examples of case studies of representative avian species to illustrate the habitat suitability in different seasons. This would demonstrate how the framework can be incorporated to inform conservation strategies. This study identifies that seasonal dynamics are an important consideration yet needs the input of ecological and biological parameters for the analysis.

Response: Thank you for your valuable comments and suggestions. We have revised Table S1 in the Supplementary Materials to include each bird species name and relevant biological and ecological parameters, including conservation status, migration status, habitat preference, diet, ecological type, and social behavior. These additions aim to provide context for the species used in the modeling process.

Please see Table S1:

Table S1. Bird species selected in this study and their biological and ecological parameters.

Species name	Latin name	Conservation status	Migration Status	Habitat Preference	Diet	Ecological Type	Social Behavior
Crested Myna	Acridotheres cristatellus	LC	Resident	Grassland / Forest	Omnivorous	Terrestrial	Social
Common Kingfisher	Alcedo atthis	LC	Partial migrant	Wetlands / Lakes / Grassland / Forest	Aquatic feeder	Waterfowl	Solitary / Paired
Oriental Magpie-Robin	Copsychus saularis	LC	Resident	Forest / Wetlands	Insectivorous	Terrestrial	Solitary / Paired
Azure-winged Magpie	Cyanopica cyanus	LC	Resident	Forest	Omnivorous	Terrestrial	Social
Yellow-billed Grosbeak	Eophona migratoria	LC	Resident/Partial migrant	Forest	Granivorous	Terrestrial	Social / Solitary / Paired
Common Moorhen	Gallinula chloropus	LC	Resident	Wetlands	Omnivorous–aquatic feeder	Waterfowl	Social / Solitary / Paired
Long-tailed Shrike	Lanius schach	LC	Resident/Partial migrant	Forest / Grassland / Wetlands	Carnivorous	Terrestrial	Solitary / Paired
Japanese tit	Parus minor	LC	Resident	Forest / Grassland	Granivorous–insectivorous	Terrestrial	Social / Solitary / Paired
Eurasian Tree Sparrow	Passer montanus	LC	Resident	Forest / Grassland / Urban areas	Granivorous–insectivorous	Terrestrial	Social
Oriental Magpie	Pica serica	LC	Resident	Forest / Grassland	Omnivorous	Terrestrial	Social
Light-vented Bulbul	Pycnonotus sinensis	LC	Resident	Forest	Omnivorous	Terrestrial	Social
Masked Laughingthrush	Pterorhinus perspicillatus	LC	Resident	Forest / Wetlands	Omnivorous	Terrestrial	Social
Eastern Spotted Dove	Spilopelia chinensis	LC	Resident	Wetlands / Forest	Granivorous	Terrestrial	Social
Red-billed Starling	Spodiopsar sericeus	LC	Resident	Forest	Omnivorous	Terrestrial	Social
Oriental Turtle-dove	Streptopelia orientalis	LC	Resident	Forest / Wetlands	Granivorous	Terrestrial	Solitary / Paired
Little Grebe	Tachybaptus ruficollis	LC	Resident	Wetlands / Lakes	Aquatic feeder	Waterfowl	Solitary / Paired
Chinese Blackbird	Turdus mandarinus	LC	Resident	Forest / Grassland	Omnivorous	Terrestrial	Solitary / Paired

Note: LC indicates species categorized as Least Concern on the IUCN Red List.

We appreciate the opportunity to clarify that habitat suitability was modeled separately for each of the 17 bird species based on their individual occurrence records and environmental factors. This allowed the models to reflect the distinct ecological characteristics of each species. In the final stage, the individual habitat suitability maps were combined into a single composite layer, which was used as an input for the InVEST-HQ model. This approach was intended to identify areas of overall importance for bird conservation at the landscape scale. However, we acknowledge that this method may overlook species-specific needs, particularly for rare or habitat-specialist species. We have addressed this limitation in the revised manuscript under the section “4.3. Limitations and Future Outlook”.

Please see lines 719–726:

Third, this study employed a multi-species approach based on 17 bird species to identify regional habitat priorities. While this method is effective for revealing general patterns, it does not account for species-specific seasonal responses or ecological preferences. This limitation is particularly relevant for rare or endangered species, whose habitat requirements may differ markedly from those of common species and may be more sensitive to seasonal or habitat changes. Future research should consider more refined, species-level analyses to support targeted conservation strategies for these vulnerable groups.

 

2. Reference format: When using the citation format (last name of first author {et al. for more than two authors}) in the text, the references should be listed in alphabetical order by last name of first author and without numbers. The format used here is to list the references in order of citation, however there were 18 references that were cited more than once, and the reader had to search through previous entries to find these references. Listing by alphabetical order ensures that the reader can easily find each reference, no matter how many times it is cited in the text.

Response: Thank you very much for your helpful suggestion. In accordance with the journal’s formatting requirements, we have revised all in-text citations to numerical format (e.g., [1], [1–3] or [1,3]) and reordered the reference list based on the order of appearance in the manuscript. We have also ensured that all references now include full titles, following the ACS citation style.

 

3. Use superscripts (all the same font size) for numbers for all author affiliations. Add periods after numbers. For 3, 4, and 5, second line should be lined up as for 1 and 2.

Response: Thank you for pointing out this issue. We have revised the author affiliations according to the journal’s formatting requirements. Please see lines 4–17 of the revised manuscript.

 

4. Line 23:The framework was applied in Wuhan China, a key migratory stopover along the East Asian–Australasian Flyway.

Response: Thank you for pointing out this issue. We have revised the sentence according to your suggestions.

Please see lines 24–25:

The framework was implemented in Wuhan, China, a critical stopover site along the East Asian–Australasian Flyway.

 

5. Line 30: Wetlandprovided the highest-quality habitats in autumn and winter, grasslands exhibited moderate seasonal 31 quality, and forests showed the least seasonal fluctuation.

Response: Thank you for pointing out this issue. We have revised the sentence according to your suggestions.

Please see lines 31–33:

Wetlands provided the highest-quality habitats in autumn and winter, grasslands exhibited moderate seasonal quality, and forests showed the least seasonal fluctuation.

 

6. Line 46: When using the citation format (last name of first author {et al. for more than two authors}) in the text, the references should be listed in alphabetical order by last name of first author and without numbers. The format used here is to list the references in order of citation, however there were 18 references that were cited more than once, and the reader had to search through previous entries to find these references. Listing by alphabetical order ensures that the reader can easily find each reference, no matter how many times it is cited in the text.

Response: Thank you very much for your helpful suggestion. In accordance with the journal’s formatting requirements, we have revised all in-text citations to use numerical format (e.g., [1], [1–3], or [1,3]) and reordered the reference list to reflect the order of appearance in the manuscript. We have also ensured that all references now include full titles, following the ACS citation style guidelines.

 

7. Line 51: Currently, 40.7 of avian (?) species

Response: Thank you for pointing out this issue. We have revised the sentence according to your suggestions.

Please see lines 53–54:

Currently, 40.7% of all known species worldwide are still threatened, and 21% of natural wetlands have been lost [7,8].

 

8. Line 97: He et al. is listed as #25 and cited on page 2. This is the reason that it is best to alphabetize the references instead of numbering them in order of occurrence;

Line 100: Wang et al. 2023b is listed as #26.

Line 102: Listed as #21.

Line 107: Listed as #7.

Line 698: References should be listed in alphabetical order by last name of first author and without numbers.

Line 707: Journal?

Line 711: Sciences Advances 1(2), DOI: 10.1126/sciadv.1500052

Response: Thank you very much for your helpful suggestion. In accordance with the journal’s formatting requirements, we have revised all in-text citations to use numerical format (e.g., [1], [1–3], or [1,3]) and reordered the reference list to reflect the order of appearance in the manuscript. We have also ensured that all references now include full titles, following the ACS citation style guidelines.

 

9. Line 102: Habitat suitability is also dependent on the species biological factors including ecology, population size, reproductive rate, conservation status, etc.

Response: Thank you for your comment. We agree that biological traits influence habitat suitability. We have revised the sentence according to your suggestions.

Please see lines 97–101:

Yet, habitat suitability for any given species is inherently complex and spatially heterogeneous, influenced not only by environmental factors such as land type, topography, climate, and vegetation abundance, but also by species-specific biological traits including ecological preferences, population size, reproductive rates, and conservation status [43,44].

 

10. Line 109: It is applied to Wuhan, a key stopover along the East Asian–Australasian Flyway, to identify priority habitats with distinct seasonal characteristics.

Response: Thank you for your comment. We have revised the sentence according to your suggestions.

Please see lines 114–117:

The framework was implemented in Wuhan, Hubei, China, a key stopover along the East Asian–Australasian Flyway, one of the world’s major migratory bird routes that supports the annual movement of millions of waterbirds, to identify priority habitats with distinct seasonal characteristics.

 

11. Line 134: Reference?

Response: Thank you for your suggestion. We have added the data source in the revised manuscript, referencing the Main Data Bulletin of the Third National Land Survey of Hubei Province (2021).

Please see lines 131–134:

Its ecosystems are primarily composed of freshwater and forest ecosystems, including 1,380.42 km² of water bodies and 2,308.86 km² of forested land, together accounting for 43.38% of the total area (Department of Natural Resources of Hubei Province, 2021).

 

12. Line 140: listed as Critically Endangered on the IUCN Red List of Threatened Species

Response: Thank you for the suggestion. We have revised the sentence as advised.

Please see lines 137–139:

This includes the Baer’s Pochard (Aythya baeri), listed as Critically Endangered on the IUCN Red List of Threatened Species, highlighting the ecological importance of this region.

 

13. Line 153: The Study Area in Wuhan, Hubei, China.

Response: Thank you for your suggestion. We have revised the figure caption to “Figure 1. The Study Area in Wuhan, Hubei, China” as recommended.

Please see line 207 and Figure 1.

 

14. Line 214: The framework for identifying priority bird habitats through an Integrated Habitat Suitability-Risk-Quality Assessment. There should be another box in the Bird habitat suitability box incorporating ecological and biological parameters for the study species.

Response: Thank you for your constructive suggestion. In response, we have revised the framework diagram by adding a sub-box under “Bird Habitat Suitability” to explicitly incorporate the ecological and biological parameters of the study species.

Please see line 214 and Figure 2:

 

Figure 2. The Framework for Identifying Priority Bird Habitats through an Integrated Habitat Suitability–Risk–Quality Assessment.

15. Line 227: Samples of what - occurrence data?

Line 229: Clarify - what was considered a sample? Were there more samples for certain species than for others?

Response: Thank you for your comment. We have clarified that “samples” refer to species-specific occurrence records. Differences in sample sizes across species and seasons are detailed in Table S2.

Please see lines 222–226:

Each species was required to have at least five occurrence records per season to meet the model’s minimum data requirement [59]. The final sample sizes, defined as species-specific bird occurrence records, were 373 (spring), 421 (summer), 904 (autumn), and 598 (winter) for the 17 bird species (see Table S2).

 

16. Line 242: The bird species were selected by noting those that had the highest occurrence frequencies. Habitat suitability analysis therefore was done with the dataset from all 17 species. There is no mention of habitat requirements dependent on biological and ecological parameters of the species. All were lumped together. Habitat suitability for a waterfowl species would be different than for a common species often found in urban areas. I don't see how a universal habitat suitability prediction can be made.

Response: Thank you for your valuable comment. We fully agree that habitat suitability is strongly influenced by species-specific ecological and biological traits. In response, we have revised the Methods section to clarify that habitat suitability modeling was conducted separately for each of the 17 bird species, with MaxEnt models independently run for each species and each season, in order to reflect differences in their habitat preferences and responses to environmental variables. Furthermore, the final composite habitat suitability map was generated by averaging the species-level predictions at the grid cell scale. This method enabled the identification of habitat hotspots while retaining ecological distinctions among species.

Please see lines 220–251:

The input data consisted of bird occurrence data and environmental variables. First, bird occurrence data were spatially rarefied by species for each season at a 100 m resolution using the SDMs tool to reduce spatial autocorrelation and overfitting [57,58]. Each species was required to have at least five occurrence records per season to meet the model’s minimum data requirement [59]. The final sample sizes, defined as species-specific bird occurrence records, were 373 (spring), 421 (summer), 904 (autumn), and 598 (winter) for the 17 bird species (see Table S2). Importantly, habitat suitability modeling was conducted separately for each species to preserve ecological differences in habitat use. Species-specific MaxEnt models were run for each season, accounting for variations in habitat preferences, behaviors, and ecological traits. In addition, 16 environmental variables related to land use, topography, climate, vegetation, and distance factors were selected based on previous studies and expert opinions, and are listed in Table 2 [60–62]. Among these, precipitation, temperature, NDVI, and distance to urban areas were the variables that reflected seasonal dynamics. To mitigate multicollinearity, Pearson correlation analysis was performed, ensuring no strong correlations among variables (|r| < 0.85; see Fig. S1).

To ensure model stability and result reliability, several parameters were configured [63,64]. First, ten-fold cross-validation was applied, and the area under the receiver operating characteristic curve (AUC) was calculated to evaluate the model performance [65]. Based on Swets’ classification for AUC [66], model accuracy is categorized as follows: low (0.5–0.7), good (0.7–0.9), and excellent (>0.9). In this study, the AUC values for habitat suitability predictions across all bird species and seasons exceeded 0.7, indicating that the model exhibited good predictive and explanatory ability (see Table S2). Second, bird habitat suitability was output in logistic format, representing species presence probability on a scale from 0 to 1, with higher values indicating greater habitat suitability [54].

To obtain a composite habitat suitability map that reflects the importance of habitats for multiple species, the seasonal suitability values of all species were averaged for each grid cell. This method facilitated the identification of habitat hotspots while maintaining interspecific distinctions through species-specific modeling. Finally, habitat suitability for forest, wetland, and grassland was extracted and categorized into six classes using the natural breaks method. These classification results served as the foundation for subsequent habitat quality modeling.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

the authors did good jobs for revision. I am pleased to accept it.

Reviewer 3 Report

Comments and Suggestions for Authors

This study exhibits use of a novel framework to assess seasonal habitat quality for a set of avian species in Wuhan, China. It demonstrates that conservation strategies should include a consideration of seasonal dynamics and characteristics of various habitats representing natural and land-use areas. The authors did a very fine job of addressing all the reviewer's comments, including mention that biological and ecological parameters are important and further fine-tuned analysis should be done. Table S1 in Supplemental Materials was updated with additional biological and ecological parameters. Table S1 was noted in the response report, however the updated table was not included in the Download Manuscript section (only the original file of Supplemental Materials was included). Please be sure that the updated Supplemental Materials file is used for the publication. 

This is a very well written manuscript to propose a novel framework for integrated habitat suitability, risk, and quality analysis to identify priority bird habitats, considering seasonal dynamics and pressures at urban edges. Results can inform conservation planning and strategies for avian species and offers a tool to contribute to overall biodiversity conservation.