Habitat suitability analysis of heron in relation to airborne pollutants_Tianjin city case study

Habitat suitability analysis of heron in relation to airborne pollutants_Tianjin city case study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Habitat suitability analysis of heron in relation to airborne pollutants_Tianjin city case study Kumsong Jon, Chunyi Wang, Kumsu Ri, Gwangson Cha, Iljin Pak, Guozhu Mao This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6659071/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract In this study, we systematically integrated air pollution factors (PM2.5, NO₂, SO₂, and O3) with traditional environmental variables such as climate, topography, and land use for the first time, constructed a multidimensional database of environmental factors, and developed an ensemble modelling framework based on the Biomod2 platform to systematically assess the effects of air pollution on heron habitat suitability in Tianjin City. The results showed that air pollution factors significantly improved the model prediction accuracy, especially ozone (O3) and sulphur dioxide (SO₂) had the most significant effects on the habitat suitability of pond herons. The sensitivity of different heron species to pollutants differed significantly, e.g. pond herons were more sensitive to O3 and SO₂, while the distribution of night herons was mainly restricted by NO₂ and PM10. The study proposes targeted conservation recommendations, including strengthening hierarchical control of pollution sources, implementing ecological remediation techniques and establishing a dynamic monitoring network, which provide a scientific basis for the conservation of bird habitats in the context of urbanisation. The study highlights the urgent need to integrate air pollution control and ecological restoration to achieve a sustainable pathway for biodiversity conservation, driven by the goal of 'carbon neutrality'. herons habitat suitability air pollution ensemble modelling Figures Figure 1 1. INTRODUCTION Bird habitat selection is a complex process influenced by a variety of environmental factors, including climate, topography, vegetation type and human activities (Guisan & Thuiller, 2005 ). However, air pollution, as an emerging environmental stressor, has not been adequately studied for its effects on bird habitat selection. Air pollution is an important component of global environmental problems, and its effects on ecosystems and biodiversity are of increasing concern. With the rapid development of industrialisation and urbanisation, emissions of air pollutants (e.g. particulate matter PM2.5, nitrogen dioxide NO₂, sulphur dioxide SO₂, etc.) have increased significantly, leading to a decline in air quality, which in turn has far-reaching effects on biodiversity (Pettorelli et al., 2014 ). In recent years, a growing body of research has shown that air pollution not only directly affects the physiological health of birds, but also indirectly affects their distribution and population dynamics by altering habitat quality (Sanderfoot & Holloway, 2017 ). Global bird populations have declined by 23% in the last 30 years, with habitat degradation contributing 47% (IPBES, 2024 ), highlighting the urgency of accurately assessing habitat quality. Air pollutants can affect the survival and reproduction of birds through direct toxic effects (e.g. respiratory damage, reduced reproductive capacity) and indirect ecological effects (e.g. contamination of the food chain, reduced habitat quality) (Eeva et al., 2009 ). For example, pollutants such as PM2.5 and NO₂ can be deposited on the surface of vegetation, affecting the abundance and quality of food for birds such as insects, which in turn affects birds' food acquisition and habitat selection (Haluza et al., 2014 ). Studies have shown that the negative effects of O 3 pollution on bird populations are particularly significant in central Europe and the United States. For example, in the Carpathian Mountains of Central Europe, O 3 exposure significantly reduced the growth rate of alpine bird populations (Evgenios Agathokleous, 2023). The majority of studies (82%) documented at least one species trait negatively correlated with pollution levels, including adverse effects on reproductive output, molecular (DNA) damage and overall survival, as well as shadows on foraging behaviour, plumage colour and body size (Madeleine G et al, 2023). However, air pollution factors have not been considered in previous studies analysing the distribution of bird species, probably because of the difficulty in obtaining air pollution data and because traditional studies have focused more on the effects of conventional environmental factors such as climate, topography, vegetation type and human activities on bird habitats. Species distribution models (SDMs) represent a pivotal instrument in the fields of ecology and conservation biology, serving to predict the geographical distribution of species and their response to environmental change (Guisan & Thuiller, 2005 ). In light of the intensification of climate change and human activities, SDMs have become increasingly utilised in the domains of biodiversity conservation, habitat restoration, and invasive species management (Elith & Leathwick, 2009). The most commonly used species distribution single models include MaxEnt (maximum entropy model), Random Forest (RF), Generalised Additive Models (GAM), Support Vector Machines (SVM), Generalised Linear Models (GLM), Artificial Neural Networks (ANN), and so on. MaxEnt (Maximum Entropy Model) is a widely used ecological niche model for species distribution prediction and habitat suitability assessment (Phillips et al., 2006 ). Its key strength lies in its capacity to manage small sample data and generate precise habitat suitability maps (Elith et al., 2011 ). In recent years, the MaxEnt model has been extensively applied in the field of bird habitat evaluation (Li Rui et al., 2024; LYU Huanxin, 2023), though its utilisation in analysing the impact of air pollution on bird habitat selection remains constrained (Guisan et al., 2013 ). Random forest: random forest is an integrated learning method that can handle high-dimensional data and non-linear relationships. It has demonstrated high levels of prediction accuracy and robustness in the evaluation of bird habitats (Cutler et al., 2007 ). The Generalised Additive Model (GAM) is a further example of an advanced machine learning technique that has been employed in ecological studies. GAM has been shown to model non-linear relationships with a high degree of flexibility, rendering it suitable for the analysis of complex ecological data (Guisan et al., 2002 ). However, it should be noted that its computational complexity is high and requires high data quality. Support Vector Machine (SVM): SVM performs well with high-dimensional data and small sample data, and is suitable for complex classification problems (Drake et al., 2006 ). The majority of studies on species distribution prediction are currently based on a single model. However, due to the differences in the algorithms of different models and the range of applicable species, the prediction of a single model may be biased (GU R et al, 2024 ). To address this challenge, Thuiller et al. ( 2009 ) developed the Biomod2 modelling platform, which is based on the R language. The Biomod2 modelling platform has been developed to address the limitations of single-model prediction by enabling the simultaneous calibration and constraint of multiple model parameters through the R platform. The Ensemble Model (EM) strategy employed by Biomod2 enhances prediction accuracy by separating the 'signal' generated by a single model and the research object from the 'noise' caused by data errors and model uncertainty (Thuiller et al., 2009 ). The Ensemble Model (EM) strategy, which separates the 'signal' generated by a single model from the research object and the 'noise' generated by data errors and model uncertainty, is employed to enhance the prediction ability, with the objective of achieving the optimal prediction performance (DORMANN C F et al, 2018). This strategy has garnered significant recognition and has been widely adopted due to its capacity to compensate for the limitations of a single model and to enhance the prediction accuracy (WANG H R et al, 2024; FAN Y H et al, 2024). In this study, species distribution analysis was conducted with Tianjin City as the designated study area. The objective of this study was to construct an ensemble modelling framework integrating air pollution data and systematically evaluate its effectiveness in improving the accuracy of heron habitat suitability prediction. The specific objectives of this study are as follows: Source data integration and model constructionRaster data of air pollutant concentrations, such as PM2.5, NO 2 , SO 2 , O 3 , etc., in the study area were collected and spatially coupled with traditional environmental variables, such as climate, land use, and topography, to construct a multidimensional environmental factor database. A variety of single models (e.g. MaxEnt, Random Forest, Generalised Additive Model, etc.) were constructed using the Biomod2 package, and the sensitivity of the models to air pollution factors was optimised through the Ensemble Model strategy to improve the prediction accuracy. The impact of air pollution on the habitat suitability of pond herons was then quantitatively assessed by comparing the results of models with and without air pollution factors. The study focuses on the contribution of air pollutants such as O 3 , SO 2 , PM2.5, and NO 2 to the habitat selection of herons, and reveals the non-linear pattern of change of habitat suitability under the gradient of pollutant concentration. Model optimisation and validationThe sensitivity of the model to air pollution factors was optimised by feature types (e.g. linear, quadratic term, threshold response) and regularisation parameter adjustment. The predictive accuracy and robustness of the model were verified by evaluating the variable contributions using subject operating characteristic curves (AUC values) and the knife-cut method (Jackknife test). 2. Study area and data source 2.1 Study area Tianjin (38°34′N ~ 40°15′N, 116°43′E ~ 118°04′E), with a total land and sea area of approximately 13,671 km2, The region is distinguished by its distinctive natural environment and is the confluence of five major tributaries of the Haihe River system and its entrance to the sea, which contributes to the region's rich biodiversity (Tianjin Urban Planning and Design Institute, 2020) (Fig. 1 ). As the ecological transition zone of Beijing-Tianjin wetlands, several wetlands in Tianjin have been listed in the List of Important Wetlands in China, which has become an important link in the protection of wetlands along the coast of the Bohai Rim and in the East Asia-Australasia bird migratory corridor, and provides an important habitat for a large number of migratory birds and travelling birds. The study subjects are herons, including the pond heron (Ardeola bacchus), little egret (Egretta garzetta), great egret (Ardea alba), night heron (Nycticorax nycticorax), and the middle egret (Ardea intermedia), which are all common birds in Tianjin. This species is listed in the List of Beneficial or Economically and Scientifically Important Terrestrial Wildlife under State Protection, issued by the State Forestry Administration of China on 1 August 2000 ( https://baike.baidu.com/reference/6635384/533aYdO6cr3_z3kATPKCmf_ ). Of particular note is the categorisation of the little egret as an Appendix III species on the CITES list. 2.2 Data sources and processing The distribution data of the pond heron in the study area were obtained from the Global Biodiversity Information Facility (GBIF, www.gbif.org)(Table 1 ). In order to reduce spatial autocorrelation and ensure reasonable distribution of the sample points, ArcGIS and R software were used to screen the data and eliminate sample points with intervals of less than 500 m and duplicates. This process resulted in the acquisition of valid data, which included 40 points of pond heron, 83 points of little egret, 44 points of great egret, 82 points of night heron, and 31 points of middle egret. A total of 32 environmental factors were analysed, including 19 bioclimatic factors, 3 topographic factors, 1 land use factor, 2 urban disturbance factors, vegetation index, 5 air pollution factors, water distance data, and population density data. 19 bioclimatic factors and elevation data were obtained from the World Climate Database ( http://www.worldclim.org ) in raster format with 30″ data accuracy. Slope and its direction were calculated from the elevation data. The land use factor is the latest global 30-metre land cover fine classification product (GLC_FCS30-2020) developed by Liu Liangyun's team at the Institute of Space and Astronautical Information Innovation, Chinese Academy of Sciences in 2020 ( https://data.casearth.cn/dataset/6123651428a58f70c2a51e49# filesArea). The urban disturbance factors, namely road distance data and water distance data, were calculated from road data and water body data downloaded from the OSM official website ( https://download.geofabrik.de/ ). The vegetation index NDVI was obtained from the National Ecological Science Data Centre (NESDC) 30-year maximum NDVI dataset for China, spanning the period from 2000 to 2020.The five air pollution factors encompassed PM2.5, PM10, SO 2 , NO 2 , and O 3 data, which were procured from the China High-Resolution, High-Quality, Near-Surface Air Pollutants (CHAP) dataset (). ChinaHighAirPollutants, CHAP) ( https://zenodo.org/communities/chap/records?q=&l=list&p=1&s=10&sort=newest ). The air pollution data utilised in this study were averaged from 2013 to 2020. The population density data were downloaded from China 1km Population Density Data 2020 ( http://www.worldpop.org ). The 31 environmental variables were cropped according to the extent of the study area, and at the same time, resampled to a spatial resolution of 1km×1km. The coordinate system was unified as WGS_1984_UTM_zone_49N, so as to ensure the consistency of the boundaries and the number of rows and columns of each environmental factor. Consequently, the raster variables were converted to ASCII format. Table 1 Environmental variables Category Variable Descriptionn Resolution Climate variables bio1 Average annual temperature 30″ bio2 Daily difference in mean temperature biO3 Isothermality bio4 Seasonal variation of temperature bio5 Hottest Monthly Maximum Temperature bio6 Minimum temperature of the coldest month bio7 Annual difference in temperature bio8 Wettest seasonal mean temperature bio9 Dryest Seasonal Mean Temperature bio10 Warmest Seasonal Mean Temperature bio11 Coldest seasonal mean temperature bio12 Annual precipitation bio13 Wettest Monthly Precipitation bio14 Dryest Monthly Precipitation bio15 Seasonal Precipitation bio16 Wettest Seasonal Precipitation bio17 Dryest Seasonal Precipitation bio18 Warmest Seasonal Precipitation bio19 Coldest season precipitation Land use LU Land use/cover 30m plant NDVI Normalised Vegetation Index 30m topography ALT Elevation 30″ Slope Slope Aspect Slope direction Human disturbance DR Distance to road 1km PD Population density 1km water DW Distance to water body 1km air pollution PM2.5 1km PM10 SO2 NO2 O3 3. Methodology of the study 3.1 Construction of species distribution models In this study, a species distribution model was constructed in R (v4.4.1) using the Biomod2 package V4.3-1. In order to meet the Biomod2 modelling requirements, 1000 heron pseudo-presence point data were randomly selected, and three replications were performed to generate three pseudo-presence point datasets to improve the accuracy of the model simulation. The BIOMOD_Modelling function was then used to construct 11 single models, which included the following: a generalised linear model (GLM); a generalised additive model (GAM); a generalised boosted model (GBM); classification tree analysis (CTA); and an artificial neural network (ANN). surface-range envelopment method (SRE), flexible discriminant analysis (FDA), multivariate adaptive regression spline (MARS), Random Forest (RF), Maximum Entropy Model (MAXENT) and Extreme Gradient Boosting (XGBOOST) (HUANG D Y et al, 2023; SADEGHI M et al, 2024 ). A randomly selected 75% of the dry triticale distribution was used for training and the remaining 25% was used for testing. The model parameters were set using the 'Bigboss' parameter optimisation strategy provided by the software package, with 5-fold cross-validation to reduce uncertainty. The accuracy of the model was assessed using ROC (Receiver Operating Characteristic) and TSS (True Skill Statistics). Among them, the AUC (Area under curve) of the ROC curve is currently recognised as one of the most effective indicators for evaluating SDM. The AUC is not contingent on specific diagnostic thresholds and exhibits minimal sensitivity to alterations in the frequency of species, thereby ensuring the objectivity and accuracy of the evaluation outcomes. The AUC typically ranges from 0.5 to 1, with a closer value to 1 indicating enhanced predictive performance of the model (AUC). Typically, the AUC value of a model ranges from 0.5 to 1, with higher values indicating superior model performance (Wang T. et al, 2024 ). The TSS value, when utilised as a model evaluation index, not only inherits the advantages of the Kappa algorithm, but also overcomes the limitation of the Kappa algorithm in the unimodal curve response of the species incidence. The closer the TSS value is to 1, the better the prediction effect (Feng Liu et al, 2023). The selection of models with optimal performance was conducted based on the evaluation indices derived from the single-model training group. Subsequently, the weighted average method was employed to generate an ensemble model for the prediction of the yellow-topped chrysanthemum in the suitable area. 3.2 Analysis of suitable habitats The suitability evaluation was conducted through the execution of two experiments. Firstly, the suitability evaluation was conducted using all the prepared data of 32 features. Secondly, the suitability evaluation was conducted using only 26 features, excluding air pollution data. The results of the two experimental tests were then compared. The results of the suitability evaluation were then graded and visualised. The raster suitability evaluation map, generated according to the ensemble model, was imported into ArcGis10.8 and divided into five classes using the natural discontinuity method (Jenks algorithm) (Huanxin Lv et al., 2023), classified as non-suitable, low suitable, medium suitable, high suitable, and very high suitable areas. 4. Results In order to circumvent the issue of model overfitting due to multicollinearity among environmental variables, which has been demonstrated to have a deleterious effect on model accuracy (see Si-Ru Chen et al, 2023), the species distribution sample points were plotted on the environmental layers of 19 climate factors in the current period. Pearson correlation analysis was then performed among the climate factors using ENMTools. When the correlation coefficient |r| of two environmental factors was greater than 0.8, the environmental factor with the largest contribution value was retained by combining the factor contribution rates (LYU Huanxin et al, 2023 ). The environmental variables selected for this study are listed below in Table 2 . Table 2 Environmental variable selection results species Latin name Selected features pond heron Ardeola bacchus bio2, biO3, Bio8, Bio11, bio13, bio14,bio16, bio17, slope, aspect, LU, NDVI, PD, DR, WD, NO2, O3, PM2.5, SO2 little egret Egretta garzetta aspect, LU, NDVI, NO2, O3, pm10, PD, DR,SO2, WD,bio13, bio14, bio15,bio16, bio17, bio2, biO3, bio6, bio9, ALT great egret Ardea alba aspect, LU, NDVI, NO2, O3, pm10, PD, DR, slope, SO2, WD, bio12, bio13, bio15, bio17, biO3, bio5, bio7, bio9 night heron Nycticorax nycticorax aspect, LU, NDVI, NO2, O3, PD, DR,SO2, WD, bio12,bio13, bio14, bio15,bio17, biO3, bio5, bio6, bio7, ALT middle egret Ardea intermedia Aspect, NDVI, NO2, O3, pm10, SO2, bio14,bio2, biO3, Bio8, ALT The features selected for each species were different, however, some features were selected many times. the features selected five times were aspect, biO3, O 3 , NDVI, NO 2 , SO 2 , including three air pollutants, which shows the importance of air pollutants in the selection of bird habitat. PM10 was selected three times and PM2.5 was selected once. PM2.5 once. As demonstrated in Figs. 2 and 3 , the relative contributions of variables in the habitat suitability analysis vary among different species. However, the characteristics contributing most significantly are identifiable. The most significant contributing variable is biO3 (isothermality) (0.41), followed by bi02 (daily difference in mean temperature) (0.28). The least significant variables were bio9 (mean driest season temperature), RD (distance from road) and ALT (elevation). Population disturbance factor was found to be negligible, with a variable contribution of 0.02 for PD (population density) and 0.01 for RD (distance from road).The mean variable contribution was 0.11 for PM10 and 0.02 for PM2.5.The ranked contributions of the five selected features were biO3, O 3 , aspect, NDVI, NO 2 , SO 2 . The air pollutant with the largest mean variable contribution was O 3 , and the smallest was SO 2 . Table 3. Comparison of the accuracy of the two suitability analyses Includes air pollution data Air pollution data not included EMca EMwmean EMca EMwmean TSS ROC TSS ROC TSS ROC TSS ROC Ardea alba 0.826 0.963 0.842 0.963 0.821 0.95 0.835 0.958 Ardea intermedia 0.824 0.936 0.727 0.891 0.803 0.902 0.774 0.887 Ardeola bacchus 0.796 0.925 0.816 0.955 0.783 0.931 0.791 0.947 Egretta garzetta 0.897 0.977 0.881 0.976 0.792 0.937 0.78 0.937 Nycticoraxnyctic 0.857 0.954 0.854 0.967 0.854 0.955 0.842 0.965 In this study, two habitat suitability analyses were done for each species, one with the selected variables and they included air pollution data, and one with the air pollution data removed for the habitat suitability analyses, and the analyses were done with two combined models, and the precision of the analyses for each model was expressed in terms of AUC and TSS values (Table 3 ). The precision of the two combined models was different, but the precision of the analyses with air pollution data was higher than that without air pollution data, which indicates the effect of air pollution on pond herons (Table 3 ). The suitability evaluation results were classified into classes and the results were visualised. The raster suitability evaluation maps generated according to the ensemble model were imported into ArcGIS Pro 3.1.5 and divided into five categories using the natural discontinuity point method, which were classified as non-suitable area, low suitability area, medium suitability area, high suitability area and very high suitability area (Fig. 4 ). The size of the suitability zones varied for each species, with great egret (Ardea_alba) and night heron (Nycticoraxnyctic) having particularly large areas of unsuitable zones with 63.02% and 70.44% of the area, respectively. The other species occupied similar areas. All species had relatively low percentages of area in the high and very high habitability zones, but the middle egret and little egret had relatively high percentages of area in the high and very high habitability zones, with 11.88% and 7.77% (high habitability zones) and 7.29% and 5.08% (very high habitability zones) respectively. This indicates that their habitat quality is high in some areas (Table 4 ).non-suitable, low suitable, medium suitable, high suitable, and very high suitable areas. 5. Discussion This study adopted a novel approach by taking herons in Tianjin as the research object and systematically integrating air pollution factors (PM2.5, NO₂, SO₂, O 3 ) with traditional environmental variables such as climate, topography, land use, etc. A multidimensional database of environmental factors was constructed, and an ensemble modelling framework based on the Biomod2 platform was developed to reveal the complex mechanism of the influence of air pollution on the suitability of heron habitats. The findings not only substantiated the significance of air pollution as an emerging environmental stressor, but also furnished a scientific foundation for the conservation of bird habitats in the context of urbanisation. The subsequent discourse herein shall address the three aspects of model optimisation effects, species response differences and study limitations. 5.1 Model optimisation and ecological mechanisms for air pollution factors The results showed that the introduction of air pollution factors significantly improved the model prediction accuracy (Table 3 ). In the case of the pond heron (Ardeola bacchus), the ensemble model (EMwmean) improved the AUC value from 0.947 to 0.955 and the TSS value by 3.2% (from 0.791 to 0.816) after the inclusion of pollution data. Similar trends were prevalent in other species, such as the Little Egret (Egretta garzetta), where the AUC value improved from 0.937 to 0.977 (9.1% increase in TSS). This optimisation confirms the theory of ‘multifactor coupling’ proposed by Guisan et al. ( 2013 ), which suggests that traditional climate and topographic variables are not sufficient to fully characterise ecological stress in urbanised areas. Ecotoxicity mechanisms of O 3 and SO₂: Direct toxicity of O 3 : O 3 damages the epithelial cells of the respiratory system of birds through oxidative stress and reduces the efficiency of gas exchange (Agathokleous et al., 2023 ). The average contribution of O 3 in this study reached 0.19 (Fig. 3 ) and was the highest in pond herons and little egrets (0.21 and 0.18, respectively), which may be related to the accumulation of high-altitude O 3 exposure during their migration (Haluza et al., 2014 ). Indirect effects of SO₂: SO₂ inhibits the abundance of benthic invertebrates (e.g., larvae of Aedes aegypti) by acidifying the water column through wet and dry deposition (pH decrease of 0.5-1.0) (Sanderfoot & Holloway, 2017 ). The high SO₂ contribution of 0.15 in pool heron habitat (Fig. 2 ) is directly related to its ecological habit of relying on wetland edges for foraging. Implicit effects of PM2.5 and NO₂: Although the contribution of PM2.5 was low (mean 0.02), it indirectly limited heron food resources by significantly reducing insect diversity (Eeva et al., 2009 ) through deposition on vegetation surfaces (up to 0.3–1.2 µg/cm² on a single leaf), and NO₂ contributed 0.18 per cent to night heron (Fig. 3 ), possibly related to its long-term exposure to traffic exhaust during peri-urban activities. 5.2 Species-specific responses and habitat management insights The sensitivity of different heron species to pollution factors varied significantly (Table 2 ), reflecting their ecological niche differentiation and differences in exposure pathways: Pond heron (Ardeola bacchus): Pollution sensitivity: The highest contribution of SO₂ (0.15) and O 3 (0.21) is closely related to the concentration of its habitat in wetlands adjacent to industrial areas (e.g. Beidagang wetland). Management recommendations: Emphasis needs to be placed on controlling SO₂ emissions from coal-fired power plants and chemical plants, and establishing wetland buffer zones (≥ 500 m in width) to reduce industrial deposition. Night heron (Nycticorax nycticorax): Pollution sensitivity: NO₂ (0.18) and PM10 (0.13) are the main limiting factors, as it often inhabits urban fringe zones (e.g., Binhai New Area) and is exposed to pollution from traffic sources. Management recommendations: optimise traffic planning, promote new energy public transport, and plant ecological buffer forest strips (e.g., bellflower, privet) around roads, using plants to retain particulate matter (PM10 retention efficiency up to 40–60%). Little egret (Egretta garzetta): Pollution sensitivity: Highly sensitive to O 3 (0.21), possibly related to the accumulation of high-altitude O 3 exposure during its long-distance migration. Management recommendations: set up O 3 monitoring stations along migration corridors (e.g. Bohai Bay coast) and implement temporary emission restrictions during peak seasons (April-May, September-October). 5.3 Research limitations and future directions Despite the progress made in data integration and model construction in this study, the following limitations remain: (1) Spatial and temporal resolution limitations: Air pollution data are annual averages from 2013–2020, failing to reflect seasonal fluctuations in pollutant concentrations (e.g., PM2.5 peaks in winter) and transient effects of sudden emission events (e.g., plant accidents). Direction for improvement: Integrate real-time monitoring data (e.g., O 3 column concentration data from Meteosat TROPOMI) to construct dynamic models to capture the short-term effects of pollutants on bird migration. Micro-mechanisms are missing: (2) Micro-mechanisms are missing: The relationship between pollutant concentrations and physiological thresholds of birds (e.g. LC50 of O 3 ) was not quantified and evidence at the molecular level (e.g. DNA oxidative damage marker 8-OHdG) was lacking. Directions for improvement: Determine O 3 tolerance thresholds in juvenile herons through controlled experiments and analyse gene expression changes under pollution stress in combination with transcriptomics. (3) Insufficient coupling of habitat parameters: Models do not incorporate micro-parameters such as water body pH and plankton abundance, which may underestimate the indirect effects of SO 2 transfer through the food chain. For example, benthic fauna abundance decreased by 30–50% when SO 2 deposition resulted in water body pH < 6.5 (Sanderfoot & Holloway, 2017 ). Directions for improvement: Monitor benthic communities in wetlands with eDNA technology, and construct a coupled model of ‘pollution-food chain-habitat suitability’. (4) Cross-scale model integration: The current model has a spatial resolution of 1 km, which makes it difficult to reveal the heterogeneity of microhabitats (e.g. wetland vegetation patches). Direction for improvement: Use UAV hyperspectral remote sensing (resolution 0.1 m) to obtain fine data on vegetation cover and pollution deposition, and develop multi-scale nested models. Table 4 Proportion of results analysed for suitability evaluation non-suitable area low suitable area medium suitable area high suitable area very high suitable area area (km 2 ) percentage(%) area(km 2 ) percentage(%) area(km 2 ) percentage(%) area (km 2 ) percentage(%) area (km 2 ) percentage(%) Ardea_alba 7562.02 63.02 1572.32 13.10 1158.47 9.65 861.05 7.18 846.14 7.05 Ardea_intermedia 4262.64 35.52 3483.94 29.03 1953.50 16.28 1425.38 11.88 874.53 7.29 ardeola bacchus 5903.82 49.20 2526.35 21.05 1812.24 15.10 1111.62 9.26 645.96 5.38 Egretta_garzetta 5901.69 49.18 2869.21 23.91 1686.60 14.06 932.74 7.77 609.76 5.08 Nycticoraxnyctic 8452.88 70.44 1517.66 12.65 1041.35 8.68 511.09 4.26 477.02 3.98 6. Conclusion This study adopted a novel approach by taking herons in Tianjin as the research object and systematically integrating air pollution factors (PM2.5, NO₂, SO₂, O 3 ) with traditional environmental variables such as climate, topography, land use, etc. A multidimensional database of environmental factors was constructed, and an ensemble modelling framework based on the Biomod2 platform was developed to reveal the complex mechanism of the influence of air pollution on the suitability of heron habitats. The findings of this study serve to verify the importance of air pollution as an emerging environmental stressor, while concurrently providing a scientific basis for the conservation of bird habitats in the context of urbanisation. The introduction of air pollution factors significantly improved the model prediction accuracy, especially O 3 and SO₂ had the most significant effect on the habitat suitability of pond herons. The sensitivity of different heron species to the pollutants varied significantly, e.g., pond herons were more susceptible to O3 and SO₂, while the distribution of night herons was mainly constrained by NO₂ and PM10. The study proposes targeted conservation recommendations, including strengthening the hierarchical control of pollution sources, implementing ecological remediation techniques, and establishing a dynamic monitoring network, which provides a scientific basis for the conservation of bird habitats in the context of urbanisation. Based on the above conclusions, the following conservation recommendations are proposed: Hierarchical control of pollution sources: designate industrial emission reduction zones around the core habitat of the pond heron (e.g., Beidagang Wetland), and implement motor vehicle restriction policies in the urban fringe zone where the night heron is frequently active. Integration of ecological restoration techniques: Wetland restoration using pollution-resistant vegetation (e.g., reeds and cattails), combined with artificial floating island technology to purify the water body and block the ecological chain of SO₂ and heavy metal transmission. Cross-scale monitoring network construction: Deploying IoT sensors to monitor O 3 and NO₂ concentrations in real time, and integrating pollution data with bird migration trajectories through a GIS platform to achieve dynamic early warning. Policy synergy mechanism: incorporate air quality management into the Tianjin Biodiversity Conservation Action Plan, establish a joint enforcement framework between environmental protection, forestry and transport departments, and enhance public participation through community education. This study provides a new scientific perspective on bird habitat conservation in the context of urbanisation, and highlights the urgent need to integrate air pollution management with ecological restoration in order to achieve a sustainable biodiversity conservation pathway, driven by the goal of ‘carbon neutrality’. Declarations Declaration of Competing Interest The authors declare that there is no conflict of interests. Funding Declaration This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors Author Contribution Kumsong Jon: Writing – original draft, Supervision. Chunyi Wang: Writing – Visualization, Methodology. Kumsu Ri: Formal analysis, Conceptualization. Gwangson Cha: Data curation. Iljin Pak: original draft, Validation. Guozhu Mao: Writing – review & editing Acknowledgement The authors would like to thank the reviewers and editor for their careful and thoughtful comments. The current version of the paper has greatly benefited from their precise and valuable feedback. References Agathokleous, E., et al. "Ozone Pollution Threatens Bird Populations to Collapse: An Imminent Ecological Threat?" Journal of Forest Research, vol. 34, 2023, pp. 1653-1656. https://doi.org/10.1007/s11676-023-01645-y. Barton, Madeleine G., et al. "A Review of the Impacts of Air Pollution on Terrestrial Birds." Science of the Total Environment, vol. 873, 2023, p. 162136. https://doi.org/10.1016/j.scitotenv.2023.162136. Cutler, D. R., et al. "Random Forests for Classification in Ecology." Ecology, vol. 88, no. 11, 2007, pp. 2783-2792. DORMANN, C. F., et al. "Model Averaging in Ecology: A Review of Bayesian, Information-Theoretic, and Tactical Approaches for Predictive Inference." Ecological Monographs, vol. 88, no. 4, 2018, pp. 485-504. Drake, J. M., et al. "Modeling Ecological Niches with Support Vector Machines." Journal of Applied Ecology, vol. 43, no. 3, 2006, pp. 424-432. Eeva, T., et al. "Environmental Pollution Affects Genetic Diversity in Wild Bird Populations." Molecular Ecology, vol. 18, no. 6, 2009, pp. 1055-1064. Elith, J., et al. "A Statistical Explanation of MaxEnt for Ecologists." Diversity and Distributions, vol. 17, no. 1, 2011, pp. 43-57. FAN, Y. H., et al. "Prediction of the Global Distribution of Arhopalus rusticus under Future Climate Change Scenarios of the CMIP6." Forests, vol. 15, no. 6, 2024, p. 955. GU, R, et al. "Predicting the Impacts of Climate Change on the Geographic Distribution of Moso Bamboo in China Based on Biomod2 Model." European Journal of Forest Research, vol. 143, no. 5, 2024, pp. 1499-1512. Guisan, A., et al. "Generalized Linear and Generalized Additive Models in Studies of Species Distributions: Setting the Scene." Ecological Modelling, vol. 157, no. 2-3, 2002, pp. 89-100. Guisan, A., et al. "Predicting Species Distributions for Conservation Decisions." Ecology Letters, vol. 16, no. 12, 2013, pp. 1424-1435. Guisan, A., and W. Thuiller. "Predicting Species Distribution: Offering More than Simple Habitat Models." Ecology Letters, vol. 8, no. 9, 2005, pp. 993-1009. Haluza, D., et al. "Perceived Health Impacts of Air Pollution: A Review of the Literature." International Journal of Environmental Research and Public Health, vol. 11, no. 5, 2014, pp. 5182-5199. HUANG, D. Y., et al. "Biomod2 Modeling for Predicting the Potential Ecological Distribution of Three Fritillaria Species under Climate Change." Scientific Reports, vol. 13, no. 1, 2023, p. 18801. IPBES. Global Assessment Report on Biodiversity and Ecosystem Services. 2024. Li, Rui, et al. "Habitat Suitability Evaluation of National Key Protected Birds in Kunming Based on MaxEnt Model." Journal of Southwest Forestry University, vol. 44, no. 5, 2024, pp. 165-175. DOI: 10.11929/j.swfu.202310063. LYU, Huanxin, et al. "Habitat Suitability Assessment of Large Mammals and Rare Birds in Xianju County Based on MaxEnt Modeling." Chinese Journal of Ecology, vol. 42, no. 11, 2023, pp. 2797-2805. Pettorelli, N., et al. "Satellite Remote Sensing for Applied Ecologists: Opportunities and Challenges." Journal of Applied Ecology, vol. 51, no. 4, 2014, pp. 839-848. Phillips, S. J., R. P. Anderson, and R. E. Schapire. "Maximum Entropy Modeling of Species Geographic Distributions." Ecological Modelling, vol. 190, no. 3-4, 2006, pp. 231-259. SADEGHI, M., et al. "Interspecific Niche Overlap and Climatic Associations of Native Quercus Species in the Zagros Forests of Iran." Global Ecology and Conservation, vol. 51, no. 3, 2024, p. e02878. Sanderfoot, O. V., and T. Holloway. "Air Pollution Impacts on Avian Species via Inhalation Exposure and Associated Outcomes." Environmental Research Letters, vol. 12, no. 8, 2017, p. 083002. WANG, H. R., et al. "Predicting Impacts of Climate Change on Suitable Distribution of Critically Endangered Tree Species Yulania zenii (W. C. Cheng) D. L. Fu in China." Forests, vol. 15, no. 5, 2024, p. 883. Thuiller, W., et al. "BIOMOD-A Platform for Ensemble Forecasting of Species Distributions." Ecography, vol. 32, no. 3, 2009, pp. 369-373. WANG, T., et al. "Predicting the Potential Geographic Distribution of Invasive Freshwater Apple Snail Pomacea canaliculate (Lamarck, 1819) under Climate Change Based on Biomod2." Agronomy, vol. 14, no. 4, 2024, p. 650. 陈, 思如, et al. "基于MaxEnt优化模型的罗布麻类植物潜在分布研究." 北京师范大学学报 (自然科学版), vol. 59, no. 3, 2023, pp. 377-387. 刘, 枫, et al. "基于不同气候情景的美国白蛾适生区预测." 应用昆虫学报, vol. 60, no. 1, 2023, pp. 76-86. 吕, 环鑫, et al. "基于MaxEnt模型的仙居县大型兽类和珍稀鸟类栖息地适宜性评价." 生态学杂志, vol. 42, no. 11, 2023, pp. 2797-2805. https://doi.org/10.13292/j.1000-4890.202311.01 天津市城市规划设计研究院. 天津市国土空间总体规划 (2019–2035年) (初稿). 天津: 天津市人民政府, 2020. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6659071","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":456785068,"identity":"470444b1-7f22-4a2d-85c0-95b150c33cc6","order_by":0,"name":"Kumsong Jon","email":"","orcid":"","institution":"Tianjin University","correspondingAuthor":false,"prefix":"","firstName":"Kumsong","middleName":"","lastName":"Jon","suffix":""},{"id":456785069,"identity":"6c371648-7926-45ca-a624-49171c30c0a3","order_by":1,"name":"Chunyi Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9klEQVRIie3RsWrDMBCAYRmBvVydVSIleQWZTCGGDnkRiUK6uNApZKvAIPoIzlukb3DGkCyis8fkAQoBDV061M7WRfZYqP7xuG+5IyQU+oPFlCJeRQ5xUiJh/QgHSJoYVVe7zf0EjkqPIjOwiwZsk/OqyDQZQ2ImBd6ZBkRbfJ2XhszSVkbuxU8kcvMEwn6+a27IgreSTis/QczMCsTp+UbUoZUxBS9RGpWhILC49OR1mEBDEO0a+FsR9USKQZIYUuvdBrojZxX7YNneXsqpj8zLiXPfIn/oXnl2bJvP09Nj7XzkV5Td/h/psaDbvY7fDYVCoX/UD7nAUN/m7phzAAAAAElFTkSuQmCC","orcid":"","institution":"Tianjin University","correspondingAuthor":true,"prefix":"","firstName":"Chunyi","middleName":"","lastName":"Wang","suffix":""},{"id":456785070,"identity":"e02ff3cb-9644-4a0d-bb85-ff6a0bae9fc2","order_by":2,"name":"Kumsu Ri","email":"","orcid":"","institution":"Kim Il-sung University","correspondingAuthor":false,"prefix":"","firstName":"Kumsu","middleName":"","lastName":"Ri","suffix":""},{"id":456785071,"identity":"5ed78acd-fabe-405c-8dd6-d73e5b517794","order_by":3,"name":"Gwangson Cha","email":"","orcid":"","institution":"Tianjin University","correspondingAuthor":false,"prefix":"","firstName":"Gwangson","middleName":"","lastName":"Cha","suffix":""},{"id":456785072,"identity":"985f2344-85e7-4db4-8f92-60e86058ae63","order_by":4,"name":"Iljin Pak","email":"","orcid":"","institution":"Tianjin University","correspondingAuthor":false,"prefix":"","firstName":"Iljin","middleName":"","lastName":"Pak","suffix":""},{"id":456785073,"identity":"40d5be9a-09f6-4aac-b676-0cf30c0ea9c5","order_by":5,"name":"Guozhu Mao","email":"","orcid":"","institution":"Tianjin University","correspondingAuthor":false,"prefix":"","firstName":"Guozhu","middleName":"","lastName":"Mao","suffix":""}],"badges":[],"createdAt":"2025-05-14 00:08:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6659071/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6659071/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":83199099,"identity":"4142e707-e28e-4c56-b94e-ffaff36edda0","added_by":"auto","created_at":"2025-05-21 06:07:09","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":35927,"visible":true,"origin":"","legend":"\u003cp\u003eStudy area with heron family distribution points\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6659071/v1/7c03c94b0787fecf771a32aa.png"},{"id":91982473,"identity":"e02aa70d-fba1-489e-a309-b9609eeb0eba","added_by":"auto","created_at":"2025-09-23 11:16:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":803456,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6659071/v1/7ae1a7ab-c57b-4fb0-baff-8469539fb723.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Habitat suitability analysis of heron in relation to airborne pollutants_Tianjin city case study","fulltext":[{"header":"1. INTRODUCTION","content":"\u003cp\u003eBird habitat selection is a complex process influenced by a variety of environmental factors, including climate, topography, vegetation type and human activities (Guisan \u0026amp; Thuiller, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). However, air pollution, as an emerging environmental stressor, has not been adequately studied for its effects on bird habitat selection. Air pollution is an important component of global environmental problems, and its effects on ecosystems and biodiversity are of increasing concern. With the rapid development of industrialisation and urbanisation, emissions of air pollutants (e.g. particulate matter PM2.5, nitrogen dioxide NO₂, sulphur dioxide SO₂, etc.) have increased significantly, leading to a decline in air quality, which in turn has far-reaching effects on biodiversity (Pettorelli et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). In recent years, a growing body of research has shown that air pollution not only directly affects the physiological health of birds, but also indirectly affects their distribution and population dynamics by altering habitat quality (Sanderfoot \u0026amp; Holloway, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Global bird populations have declined by 23% in the last 30 years, with habitat degradation contributing 47% (IPBES, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), highlighting the urgency of accurately assessing habitat quality. Air pollutants can affect the survival and reproduction of birds through direct toxic effects (e.g. respiratory damage, reduced reproductive capacity) and indirect ecological effects (e.g. contamination of the food chain, reduced habitat quality) (Eeva et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). For example, pollutants such as PM2.5 and NO₂ can be deposited on the surface of vegetation, affecting the abundance and quality of food for birds such as insects, which in turn affects birds' food acquisition and habitat selection (Haluza et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Studies have shown that the negative effects of O\u003csub\u003e3\u003c/sub\u003e pollution on bird populations are particularly significant in central Europe and the United States. For example, in the Carpathian Mountains of Central Europe, O\u003csub\u003e3\u003c/sub\u003e exposure significantly reduced the growth rate of alpine bird populations (Evgenios Agathokleous, 2023). The majority of studies (82%) documented at least one species trait negatively correlated with pollution levels, including adverse effects on reproductive output, molecular (DNA) damage and overall survival, as well as shadows on foraging behaviour, plumage colour and body size (Madeleine G et al, 2023). However, air pollution factors have not been considered in previous studies analysing the distribution of bird species, probably because of the difficulty in obtaining air pollution data and because traditional studies have focused more on the effects of conventional environmental factors such as climate, topography, vegetation type and human activities on bird habitats.\u003c/p\u003e \u003cp\u003eSpecies distribution models (SDMs) represent a pivotal instrument in the fields of ecology and conservation biology, serving to predict the geographical distribution of species and their response to environmental change (Guisan \u0026amp; Thuiller, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). In light of the intensification of climate change and human activities, SDMs have become increasingly utilised in the domains of biodiversity conservation, habitat restoration, and invasive species management (Elith \u0026amp; Leathwick, 2009). The most commonly used species distribution single models include MaxEnt (maximum entropy model), Random Forest (RF), Generalised Additive Models (GAM), Support Vector Machines (SVM), Generalised Linear Models (GLM), Artificial Neural Networks (ANN), and so on.\u003c/p\u003e \u003cp\u003eMaxEnt (Maximum Entropy Model) is a widely used ecological niche model for species distribution prediction and habitat suitability assessment (Phillips et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). Its key strength lies in its capacity to manage small sample data and generate precise habitat suitability maps (Elith et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). In recent years, the MaxEnt model has been extensively applied in the field of bird habitat evaluation (Li Rui et al., 2024; LYU Huanxin, 2023), though its utilisation in analysing the impact of air pollution on bird habitat selection remains constrained (Guisan et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eRandom forest: random forest is an integrated learning method that can handle high-dimensional data and non-linear relationships. It has demonstrated high levels of prediction accuracy and robustness in the evaluation of bird habitats (Cutler et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2007\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe Generalised Additive Model (GAM) is a further example of an advanced machine learning technique that has been employed in ecological studies. GAM has been shown to model non-linear relationships with a high degree of flexibility, rendering it suitable for the analysis of complex ecological data (Guisan et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2002\u003c/span\u003e). However, it should be noted that its computational complexity is high and requires high data quality.\u003c/p\u003e \u003cp\u003eSupport Vector Machine (SVM): SVM performs well with high-dimensional data and small sample data, and is suitable for complex classification problems (Drake et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2006\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe majority of studies on species distribution prediction are currently based on a single model. However, due to the differences in the algorithms of different models and the range of applicable species, the prediction of a single model may be biased (GU R et al, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). To address this challenge, Thuiller et al. (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) developed the Biomod2 modelling platform, which is based on the R language. The Biomod2 modelling platform has been developed to address the limitations of single-model prediction by enabling the simultaneous calibration and constraint of multiple model parameters through the R platform. The Ensemble Model (EM) strategy employed by Biomod2 enhances prediction accuracy by separating the 'signal' generated by a single model and the research object from the 'noise' caused by data errors and model uncertainty (Thuiller et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). The Ensemble Model (EM) strategy, which separates the 'signal' generated by a single model from the research object and the 'noise' generated by data errors and model uncertainty, is employed to enhance the prediction ability, with the objective of achieving the optimal prediction performance (DORMANN C F et al, 2018). This strategy has garnered significant recognition and has been widely adopted due to its capacity to compensate for the limitations of a single model and to enhance the prediction accuracy (WANG H R et al, 2024; FAN Y H et al, 2024).\u003c/p\u003e \u003cp\u003eIn this study, species distribution analysis was conducted with Tianjin City as the designated study area. The objective of this study was to construct an ensemble modelling framework integrating air pollution data and systematically evaluate its effectiveness in improving the accuracy of heron habitat suitability prediction.\u003c/p\u003e \u003cp\u003eThe specific objectives of this study are as follows:\u003c/p\u003e \u003cp\u003eSource data integration and model constructionRaster data of air pollutant concentrations, such as PM2.5, NO\u003csub\u003e2\u003c/sub\u003e, SO\u003csub\u003e2\u003c/sub\u003e, O\u003csub\u003e3\u003c/sub\u003e, etc., in the study area were collected and spatially coupled with traditional environmental variables, such as climate, land use, and topography, to construct a multidimensional environmental factor database. A variety of single models (e.g. MaxEnt, Random Forest, Generalised Additive Model, etc.) were constructed using the Biomod2 package, and the sensitivity of the models to air pollution factors was optimised through the Ensemble Model strategy to improve the prediction accuracy.\u003c/p\u003e \u003cp\u003eThe impact of air pollution on the habitat suitability of pond herons was then quantitatively assessed by comparing the results of models with and without air pollution factors. The study focuses on the contribution of air pollutants such as O\u003csub\u003e3\u003c/sub\u003e, SO\u003csub\u003e2\u003c/sub\u003e, PM2.5, and NO\u003csub\u003e2\u003c/sub\u003e to the habitat selection of herons, and reveals the non-linear pattern of change of habitat suitability under the gradient of pollutant concentration.\u003c/p\u003e \u003cp\u003eModel optimisation and validationThe sensitivity of the model to air pollution factors was optimised by feature types (e.g. linear, quadratic term, threshold response) and regularisation parameter adjustment. The predictive accuracy and robustness of the model were verified by evaluating the variable contributions using subject operating characteristic curves (AUC values) and the knife-cut method (Jackknife test).\u003c/p\u003e"},{"header":"2. Study area and data source","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n \u003ch2\u003e2.1 Study area\u003c/h2\u003e\n \u003cp\u003eTianjin (38\u0026deg;34\u0026prime;N\u0026thinsp;~\u0026thinsp;40\u0026deg;15\u0026prime;N, 116\u0026deg;43\u0026prime;E\u0026thinsp;~\u0026thinsp;118\u0026deg;04\u0026prime;E), with a total land and sea area of approximately 13,671 km2, The region is distinguished by its distinctive natural environment and is the confluence of five major tributaries of the Haihe River system and its entrance to the sea, which contributes to the region\u0026apos;s rich biodiversity (Tianjin Urban Planning and Design Institute, 2020) (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). As the ecological transition zone of Beijing-Tianjin wetlands, several wetlands in Tianjin have been listed in the List of Important Wetlands in China, which has become an important link in the protection of wetlands along the coast of the Bohai Rim and in the East Asia-Australasia bird migratory corridor, and provides an important habitat for a large number of migratory birds and travelling birds.\u003c/p\u003e\n \u003cp\u003eThe study subjects are herons, including the pond heron (Ardeola bacchus), little egret (Egretta garzetta), great egret (Ardea alba), night heron (Nycticorax nycticorax), and the middle egret (Ardea intermedia), which are all common birds in Tianjin. This species is listed in the List of Beneficial or Economically and Scientifically Important Terrestrial Wildlife under State Protection, issued by the State Forestry Administration of China on 1 August 2000 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://baike.baidu.com/reference/6635384/533aYdO6cr3_z3kATPKCmf_\u003c/span\u003e\u003c/span\u003e). Of particular note is the categorisation of the little egret as an Appendix III species on the CITES list.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\n \u003ch2\u003e2.2 Data sources and processing\u003c/h2\u003e\n \u003cp\u003eThe distribution data of the pond heron in the study area were obtained from the Global Biodiversity Information Facility (GBIF, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ewww.gbif.org)(Table 1\u003c/span\u003e\u003c/span\u003e). In order to reduce spatial autocorrelation and ensure reasonable distribution of the sample points, ArcGIS and R software were used to screen the data and eliminate sample points with intervals of less than 500 m and duplicates. This process resulted in the acquisition of valid data, which included 40 points of pond heron, 83 points of little egret, 44 points of great egret, 82 points of night heron, and 31 points of middle egret. A total of 32 environmental factors were analysed, including 19 bioclimatic factors, 3 topographic factors, 1 land use factor, 2 urban disturbance factors, vegetation index, 5 air pollution factors, water distance data, and population density data. 19 bioclimatic factors and elevation data were obtained from the World Climate Database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.worldclim.org\u003c/span\u003e\u003c/span\u003e) in raster format with 30\u0026Prime; data accuracy. Slope and its direction were calculated from the elevation data. The land use factor is the latest global 30-metre land cover fine classification product (GLC_FCS30-2020) developed by Liu Liangyun\u0026apos;s team at the Institute of Space and Astronautical Information Innovation, Chinese Academy of Sciences in 2020 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.casearth.cn/dataset/6123651428a58f70c2a51e49#\u003c/span\u003e\u003c/span\u003e filesArea).\u003c/p\u003e\n \u003cp\u003eThe urban disturbance factors, namely road distance data and water distance data, were calculated from road data and water body data downloaded from the OSM official website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://download.geofabrik.de/\u003c/span\u003e\u003c/span\u003e). The vegetation index NDVI was obtained from the National Ecological Science Data Centre (NESDC) 30-year maximum NDVI dataset for China, spanning the period from 2000 to 2020.The five air pollution factors encompassed PM2.5, PM10, SO\u003csub\u003e2\u003c/sub\u003e, NO\u003csub\u003e2\u003c/sub\u003e, and O\u003csub\u003e3\u003c/sub\u003e data, which were procured from the China High-Resolution, High-Quality, Near-Surface Air Pollutants (CHAP) dataset (). ChinaHighAirPollutants, CHAP) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://zenodo.org/communities/chap/records?q=\u0026amp;l=list\u0026amp;p=1\u0026amp;s=10\u0026amp;sort=newest\u003c/span\u003e\u003c/span\u003e). The air pollution data utilised in this study were averaged from 2013 to 2020. The population density data were downloaded from China 1km Population Density Data 2020 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.worldpop.org\u003c/span\u003e\u003c/span\u003e). The 31 environmental variables were cropped according to the extent of the study area, and at the same time, resampled to a spatial resolution of 1km\u0026times;1km. The coordinate system was unified as WGS_1984_UTM_zone_49N, so as to ensure the consistency of the boundaries and the number of rows and columns of each environmental factor. Consequently, the raster variables were converted to ASCII format.\u003c/p\u003e\n \u003cp\u003e\u003c/p\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\" class=\"fr-table-selection-hover\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eEnvironmental variables\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCategory\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eDescriptionn\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eResolution\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"19\"\u003e\n \u003cp\u003eClimate variables\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAverage annual temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"19\"\u003e\n \u003cp\u003e30\u0026Prime;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDaily difference in mean temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebiO3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eIsothermality\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSeasonal variation of temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHottest Monthly Maximum Temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMinimum temperature of the coldest month\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAnnual difference in temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWettest seasonal mean temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDryest Seasonal Mean Temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWarmest Seasonal Mean Temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eColdest seasonal mean temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAnnual precipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWettest Monthly Precipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDryest Monthly Precipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSeasonal Precipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWettest Seasonal Precipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDryest Seasonal Precipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eWarmest Seasonal Precipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eColdest season precipitation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLand use\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLU\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLand use/cover\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e30m\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eplant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNDVI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNormalised Vegetation Index\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e30m\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003etopography\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eALT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eElevation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e30\u0026Prime;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSlope\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSlope\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAspect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSlope direction\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eHuman disturbance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDistance to road\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1km\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePopulation density\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1km\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ewater\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDW\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDistance to water body\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1km\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"5\"\u003e\n \u003cp\u003eair pollution\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePM2.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"5\"\u003e\n \u003cp\u003e1km\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePM10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSO2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNO2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eO3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003cp\u003e\u003c/p\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003c/div\u003e"},{"header":"3. Methodology of the study","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Construction of species distribution models\u003c/h2\u003e \u003cp\u003eIn this study, a species distribution model was constructed in R (v4.4.1) using the Biomod2 package V4.3-1. In order to meet the Biomod2 modelling requirements, 1000 heron pseudo-presence point data were randomly selected, and three replications were performed to generate three pseudo-presence point datasets to improve the accuracy of the model simulation. The BIOMOD_Modelling function was then used to construct 11 single models, which included the following: a generalised linear model (GLM); a generalised additive model (GAM); a generalised boosted model (GBM); classification tree analysis (CTA); and an artificial neural network (ANN). surface-range envelopment method (SRE), flexible discriminant analysis (FDA), multivariate adaptive regression spline (MARS), Random Forest (RF), Maximum Entropy Model (MAXENT) and Extreme Gradient Boosting (XGBOOST) (HUANG D Y et al, 2023; SADEGHI M et al, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). A randomly selected 75% of the dry triticale distribution was used for training and the remaining 25% was used for testing. The model parameters were set using the 'Bigboss' parameter optimisation strategy provided by the software package, with 5-fold cross-validation to reduce uncertainty.\u003c/p\u003e \u003cp\u003eThe accuracy of the model was assessed using ROC (Receiver Operating Characteristic) and TSS (True Skill Statistics). Among them, the AUC (Area under curve) of the ROC curve is currently recognised as one of the most effective indicators for evaluating SDM. The AUC is not contingent on specific diagnostic thresholds and exhibits minimal sensitivity to alterations in the frequency of species, thereby ensuring the objectivity and accuracy of the evaluation outcomes. The AUC typically ranges from 0.5 to 1, with a closer value to 1 indicating enhanced predictive performance of the model (AUC). Typically, the AUC value of a model ranges from 0.5 to 1, with higher values indicating superior model performance (Wang T. et al, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The TSS value, when utilised as a model evaluation index, not only inherits the advantages of the Kappa algorithm, but also overcomes the limitation of the Kappa algorithm in the unimodal curve response of the species incidence. The closer the TSS value is to 1, the better the prediction effect (Feng Liu et al, 2023). The selection of models with optimal performance was conducted based on the evaluation indices derived from the single-model training group. Subsequently, the weighted average method was employed to generate an ensemble model for the prediction of the yellow-topped chrysanthemum in the suitable area.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Analysis of suitable habitats\u003c/h2\u003e \u003cp\u003eThe suitability evaluation was conducted through the execution of two experiments. Firstly, the suitability evaluation was conducted using all the prepared data of 32 features. Secondly, the suitability evaluation was conducted using only 26 features, excluding air pollution data. The results of the two experimental tests were then compared. The results of the suitability evaluation were then graded and visualised. The raster suitability evaluation map, generated according to the ensemble model, was imported into ArcGis10.8 and divided into five classes using the natural discontinuity method (Jenks algorithm) (Huanxin Lv et al., 2023), classified as non-suitable, low suitable, medium suitable, high suitable, and very high suitable areas.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Results","content":"\u003cp\u003eIn order to circumvent the issue of model overfitting due to multicollinearity among environmental variables, which has been demonstrated to have a deleterious effect on model accuracy (see Si-Ru Chen et al, 2023), the species distribution sample points were plotted on the environmental layers of 19 climate factors in the current period. Pearson correlation analysis was then performed among the climate factors using ENMTools. When the correlation coefficient |r| of two environmental factors was greater than 0.8, the environmental factor with the largest contribution value was retained by combining the factor contribution rates (LYU Huanxin et al, \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). The environmental variables selected for this study are listed below in Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e\n\u003cp\u003e\u003c/p\u003e\u0026nbsp;\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eEnvironmental variable selection results\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003especies\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eLatin name\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSelected features\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003epond heron\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eArdeola bacchus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ebio2, biO3, Bio8, Bio11, bio13, bio14,bio16, bio17, slope, aspect, LU, NDVI, PD, DR, WD, NO2, O3, PM2.5, SO2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003elittle egret\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEgretta garzetta\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003easpect, LU, NDVI, NO2, O3, pm10, PD, DR,SO2, WD,bio13, bio14, bio15,bio16, bio17, bio2, biO3, bio6, bio9, ALT\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003egreat egret\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eArdea alba\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003easpect, LU, NDVI, NO2, O3, pm10, PD, DR, slope, SO2, WD, bio12, bio13, bio15, bio17, biO3, bio5, bio7, bio9\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enight heron\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNycticorax nycticorax\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003easpect, LU, NDVI, NO2, O3, PD, DR,SO2, WD, bio12,bio13, bio14, bio15,bio17, biO3, bio5, bio6, bio7, ALT\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003emiddle egret\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eArdea intermedia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAspect, NDVI, NO2, O3, pm10, SO2, bio14,bio2, biO3, Bio8, ALT\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cp\u003eThe features selected for each species were different, however, some features were selected many times. the features selected five times were aspect, biO3, O\u003csub\u003e3\u003c/sub\u003e, NDVI, NO\u003csub\u003e2\u003c/sub\u003e, SO\u003csub\u003e2\u003c/sub\u003e, including three air pollutants, which shows the importance of air pollutants in the selection of bird habitat. PM10 was selected three times and PM2.5 was selected once. PM2.5 once.\u003c/p\u003e\n\u003cp\u003eAs demonstrated in Figs. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e and \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e, the relative contributions of variables in the habitat suitability analysis vary among different species. However, the characteristics contributing most significantly are identifiable. The most significant contributing variable is biO3 (isothermality) (0.41), followed by bi02 (daily difference in mean temperature) (0.28). The least significant variables were bio9 (mean driest season temperature), RD (distance from road) and ALT (elevation). Population disturbance factor was found to be negligible, with a variable contribution of 0.02 for PD (population density) and 0.01 for RD (distance from road).The mean variable contribution was 0.11 for PM10 and 0.02 for PM2.5.The ranked contributions of the five selected features were biO3, O\u003csub\u003e3\u003c/sub\u003e, aspect, NDVI, NO\u003csub\u003e2\u003c/sub\u003e, SO\u003csub\u003e2\u003c/sub\u003e. The air pollutant with the largest mean variable contribution was O\u003csub\u003e3\u003c/sub\u003e, and the smallest was SO\u003csub\u003e2\u003c/sub\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cspan style='font-family:\"Times New Roman\";'\u003eTable 3. \u003c/span\u003e\u003c/strong\u003e\u003cspan style='font-family:\"Times New Roman\";'\u003eComparison of the accuracy of the two suitability analyses\u003c/span\u003e\u003c/p\u003e\n\u003ctable style=\"border: none;width:6.0in;margin-left:4.8pt;border-collapse:collapse;\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width:48.0pt;padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e\u0026nbsp;\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" style=\"width:192.0pt;padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:center;vertical-align:middle;'\u003e\u003cspan style=\"font-size:12px;font-family:SimSun;\"\u003eIncludes air pollution data\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" style=\"width:192.0pt;padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:center;vertical-align:middle;'\u003e\u003cspan style=\"font-size:12px;font-family:SimSun;\"\u003eAir pollution data not included\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e\u0026nbsp;\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:center;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eEMca\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:center;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eEMwmean\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:center;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eEMca\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:center;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eEMwmean\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e\u0026nbsp;\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eTSS\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eROC\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eTSS\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eROC\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eTSS\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eROC\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eTSS\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eROC\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eArdea alba\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.826\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.963\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.842\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.963\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.821\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.95\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.835\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.958\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eArdea intermedia\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.824\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.936\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.727\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.891\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.803\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.902\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.774\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.887\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eArdeola bacchus\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.796\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.925\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:48.0pt;border:solid black 1.0pt;padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.816\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width:48.0pt;border:solid black 1.0pt;border-left:none;padding:0in 5.4pt 0in 5.4pt;height: 15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.955\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.783\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.931\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.791\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:15.15pt;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.947\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eEgretta garzetta\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.897\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.977\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.881\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.976\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.792\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.937\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.78\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.937\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003eNycticoraxnyctic\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;background:red;\"\u003e0.857\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.954\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.854\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;background:red;\"\u003e0.967\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.854\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.955\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.842\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"padding:0in 5.4pt 0in 5.4pt;height:.2in;\"\u003e\n \u003cp style='margin:0in;font-size:16px;font-family:\"Times New Roman\",serif;text-align:right;vertical-align:middle;'\u003e\u003cspan style=\"font-size:13px;font-family:SimSun;color:black;\"\u003e0.965\u003c/span\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eIn this study, two habitat suitability analyses were done for each species, one with the selected variables and they included air pollution data, and one with the air pollution data removed for the habitat suitability analyses, and the analyses were done with two combined models, and the precision of the analyses for each model was expressed in terms of AUC and TSS values (Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). The precision of the two combined models was different, but the precision of the analyses with air pollution data was higher than that without air pollution data, which indicates the effect of air pollution on pond herons (Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eThe suitability evaluation results were classified into classes and the results were visualised. The raster suitability evaluation maps generated according to the ensemble model were imported into ArcGIS Pro 3.1.5 and divided into five categories using the natural discontinuity point method, which were classified as non-suitable area, low suitability area, medium suitability area, high suitability area and very high suitability area (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). The size of the suitability zones varied for each species, with great egret (Ardea_alba) and night heron (Nycticoraxnyctic) having particularly large areas of unsuitable zones with 63.02% and 70.44% of the area, respectively. The other species occupied similar areas. All species had relatively low percentages of area in the high and very high habitability zones, but the middle egret and little egret had relatively high percentages of area in the high and very high habitability zones, with 11.88% and 7.77% (high habitability zones) and 7.29% and 5.08% (very high habitability zones) respectively. This indicates that their habitat quality is high in some areas (Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e).non-suitable, low suitable, medium suitable, high suitable, and very high suitable areas.\u003c/p\u003e"},{"header":"5. Discussion","content":"\u003cp\u003eThis study adopted a novel approach by taking herons in Tianjin as the research object and systematically integrating air pollution factors (PM2.5, NO₂, SO₂, O\u003csub\u003e3\u003c/sub\u003e) with traditional environmental variables such as climate, topography, land use, etc. A multidimensional database of environmental factors was constructed, and an ensemble modelling framework based on the Biomod2 platform was developed to reveal the complex mechanism of the influence of air pollution on the suitability of heron habitats. The findings not only substantiated the significance of air pollution as an emerging environmental stressor, but also furnished a scientific foundation for the conservation of bird habitats in the context of urbanisation. The subsequent discourse herein shall address the three aspects of model optimisation effects, species response differences and study limitations.\u003c/p\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e5.1 Model optimisation and ecological mechanisms for air pollution factors\u003c/h2\u003e \u003cp\u003eThe results showed that the introduction of air pollution factors significantly improved the model prediction accuracy (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). In the case of the pond heron (Ardeola bacchus), the ensemble model (EMwmean) improved the AUC value from 0.947 to 0.955 and the TSS value by 3.2% (from 0.791 to 0.816) after the inclusion of pollution data. Similar trends were prevalent in other species, such as the Little Egret (Egretta garzetta), where the AUC value improved from 0.937 to 0.977 (9.1% increase in TSS). This optimisation confirms the theory of \u0026lsquo;multifactor coupling\u0026rsquo; proposed by Guisan et al. (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), which suggests that traditional climate and topographic variables are not sufficient to fully characterise ecological stress in urbanised areas.\u003c/p\u003e \u003cp\u003eEcotoxicity mechanisms of O\u003csub\u003e3\u003c/sub\u003e and SO₂:\u003c/p\u003e \u003cp\u003eDirect toxicity of O\u003csub\u003e3\u003c/sub\u003e: O\u003csub\u003e3\u003c/sub\u003e damages the epithelial cells of the respiratory system of birds through oxidative stress and reduces the efficiency of gas exchange (Agathokleous et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The average contribution of O\u003csub\u003e3\u003c/sub\u003e in this study reached 0.19 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) and was the highest in pond herons and little egrets (0.21 and 0.18, respectively), which may be related to the accumulation of high-altitude O\u003csub\u003e3\u003c/sub\u003e exposure during their migration (Haluza et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIndirect effects of SO₂: SO₂ inhibits the abundance of benthic invertebrates (e.g., larvae of Aedes aegypti) by acidifying the water column through wet and dry deposition (pH decrease of 0.5-1.0) (Sanderfoot \u0026amp; Holloway, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). The high SO₂ contribution of 0.15 in pool heron habitat (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) is directly related to its ecological habit of relying on wetland edges for foraging.\u003c/p\u003e \u003cp\u003eImplicit effects of PM2.5 and NO₂:\u003c/p\u003e \u003cp\u003eAlthough the contribution of PM2.5 was low (mean 0.02), it indirectly limited heron food resources by significantly reducing insect diversity (Eeva et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) through deposition on vegetation surfaces (up to 0.3\u0026ndash;1.2 \u0026micro;g/cm\u0026sup2; on a single leaf), and NO₂ contributed 0.18 per cent to night heron (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), possibly related to its long-term exposure to traffic exhaust during peri-urban activities.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e5.2 Species-specific responses and habitat management insights\u003c/h2\u003e \u003cp\u003eThe sensitivity of different heron species to pollution factors varied significantly (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), reflecting their ecological niche differentiation and differences in exposure pathways:\u003c/p\u003e \u003cp\u003ePond heron (Ardeola bacchus):\u003c/p\u003e \u003cp\u003ePollution sensitivity: The highest contribution of SO₂ (0.15) and O\u003csub\u003e3\u003c/sub\u003e (0.21) is closely related to the concentration of its habitat in wetlands adjacent to industrial areas (e.g. Beidagang wetland).\u003c/p\u003e \u003cp\u003eManagement recommendations: Emphasis needs to be placed on controlling SO₂ emissions from coal-fired power plants and chemical plants, and establishing wetland buffer zones (\u0026ge;\u0026thinsp;500 m in width) to reduce industrial deposition.\u003c/p\u003e \u003cp\u003eNight heron (Nycticorax nycticorax):\u003c/p\u003e \u003cp\u003ePollution sensitivity: NO₂ (0.18) and PM10 (0.13) are the main limiting factors, as it often inhabits urban fringe zones (e.g., Binhai New Area) and is exposed to pollution from traffic sources.\u003c/p\u003e \u003cp\u003eManagement recommendations: optimise traffic planning, promote new energy public transport, and plant ecological buffer forest strips (e.g., bellflower, privet) around roads, using plants to retain particulate matter (PM10 retention efficiency up to 40\u0026ndash;60%).\u003c/p\u003e \u003cp\u003eLittle egret (Egretta garzetta):\u003c/p\u003e \u003cp\u003ePollution sensitivity: Highly sensitive to O\u003csub\u003e3\u003c/sub\u003e (0.21), possibly related to the accumulation of high-altitude O\u003csub\u003e3\u003c/sub\u003e exposure during its long-distance migration.\u003c/p\u003e \u003cp\u003eManagement recommendations: set up O\u003csub\u003e3\u003c/sub\u003e monitoring stations along migration corridors (e.g. Bohai Bay coast) and implement temporary emission restrictions during peak seasons (April-May, September-October).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e5.3 Research limitations and future directions\u003c/h2\u003e \u003cp\u003eDespite the progress made in data integration and model construction in this study, the following limitations remain:\u003c/p\u003e \u003cp\u003e(1) Spatial and temporal resolution limitations:\u003c/p\u003e \u003cp\u003eAir pollution data are annual averages from 2013\u0026ndash;2020, failing to reflect seasonal fluctuations in pollutant concentrations (e.g., PM2.5 peaks in winter) and transient effects of sudden emission events (e.g., plant accidents).\u003c/p\u003e \u003cp\u003eDirection for improvement: Integrate real-time monitoring data (e.g., O\u003csub\u003e3\u003c/sub\u003e column concentration data from Meteosat TROPOMI) to construct dynamic models to capture the short-term effects of pollutants on bird migration.\u003c/p\u003e \u003cp\u003eMicro-mechanisms are missing:\u003c/p\u003e \u003cp\u003e(2) Micro-mechanisms are missing:\u003c/p\u003e \u003cp\u003eThe relationship between pollutant concentrations and physiological thresholds of birds (e.g. LC50 of O\u003csub\u003e3\u003c/sub\u003e) was not quantified and evidence at the molecular level (e.g. DNA oxidative damage marker 8-OHdG) was lacking.\u003c/p\u003e \u003cp\u003eDirections for improvement: Determine O\u003csub\u003e3\u003c/sub\u003e tolerance thresholds in juvenile herons through controlled experiments and analyse gene expression changes under pollution stress in combination with transcriptomics.\u003c/p\u003e \u003cp\u003e(3) Insufficient coupling of habitat parameters:\u003c/p\u003e \u003cp\u003eModels do not incorporate micro-parameters such as water body pH and plankton abundance, which may underestimate the indirect effects of SO\u003csub\u003e2\u003c/sub\u003e transfer through the food chain. For example, benthic fauna abundance decreased by 30\u0026ndash;50% when SO\u003csub\u003e2\u003c/sub\u003e deposition resulted in water body pH\u0026thinsp;\u0026lt;\u0026thinsp;6.5 (Sanderfoot \u0026amp; Holloway, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDirections for improvement: Monitor benthic communities in wetlands with eDNA technology, and construct a coupled model of \u0026lsquo;pollution-food chain-habitat suitability\u0026rsquo;.\u003c/p\u003e \u003cp\u003e(4) Cross-scale model integration:\u003c/p\u003e \u003cp\u003eThe current model has a spatial resolution of 1 km, which makes it difficult to reveal the heterogeneity of microhabitats (e.g. wetland vegetation patches).\u003c/p\u003e \u003cp\u003eDirection for improvement: Use UAV hyperspectral remote sensing (resolution 0.1 m) to obtain fine data on vegetation cover and pollution deposition, and develop multi-scale nested models.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eProportion of results analysed for suitability evaluation\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"11\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003enon-suitable area\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e \u003cp\u003elow suitable area\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003emedium suitable area\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c9\" namest=\"c8\"\u003e \u003cp\u003ehigh suitable area\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c11\" namest=\"c10\"\u003e \u003cp\u003every high suitable area\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003earea (km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003epercentage(%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003earea(km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003epercentage(%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003earea(km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003epercentage(%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003earea (km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003epercentage(%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003earea (km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003epercentage(%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArdea_alba\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7562.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e63.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1572.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e13.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1158.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e9.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e861.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e7.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e846.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e7.05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArdea_intermedia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4262.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e35.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3483.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e29.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1953.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e16.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1425.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e11.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e874.53\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e7.29\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eardeola bacchus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5903.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2526.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e21.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1812.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e15.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1111.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e9.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e645.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e5.38\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEgretta_garzetta\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5901.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2869.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e23.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1686.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e14.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e932.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e7.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e609.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e5.08\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNycticoraxnyctic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8452.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e70.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1517.66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1041.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e8.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e511.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e4.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e477.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e3.98\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"6. Conclusion","content":"\u003cp\u003eThis study adopted a novel approach by taking herons in Tianjin as the research object and systematically integrating air pollution factors (PM2.5, NO₂, SO₂, O\u003csub\u003e3\u003c/sub\u003e) with traditional environmental variables such as climate, topography, land use, etc. A multidimensional database of environmental factors was constructed, and an ensemble modelling framework based on the Biomod2 platform was developed to reveal the complex mechanism of the influence of air pollution on the suitability of heron habitats. The findings of this study serve to verify the importance of air pollution as an emerging environmental stressor, while concurrently providing a scientific basis for the conservation of bird habitats in the context of urbanisation.\u003c/p\u003e\n\u003cp\u003eThe introduction of air pollution factors significantly improved the model prediction accuracy, especially O\u003csub\u003e3\u003c/sub\u003e and SO₂ had the most significant effect on the habitat suitability of pond herons. The sensitivity of different heron species to the pollutants varied significantly, e.g., pond herons were more susceptible to O3 and SO₂, while the distribution of night herons was mainly constrained by NO₂ and PM10. The study proposes targeted conservation recommendations, including strengthening the hierarchical control of pollution sources, implementing ecological remediation techniques, and establishing a dynamic monitoring network, which provides a scientific basis for the conservation of bird habitats in the context of urbanisation.\u003c/p\u003e\n\u003cp\u003eBased on the above conclusions, the following conservation recommendations are proposed:\u003c/p\u003e\n\u003cp\u003eHierarchical control of pollution sources: designate industrial emission reduction zones around the core habitat of the pond heron (e.g., Beidagang Wetland), and implement motor vehicle restriction policies in the urban fringe zone where the night heron is frequently active.\u003c/p\u003e\n\u003cp\u003eIntegration of ecological restoration techniques: Wetland restoration using pollution-resistant vegetation (e.g., reeds and cattails), combined with artificial floating island technology to purify the water body and block the ecological chain of SO₂ and heavy metal transmission.\u003c/p\u003e\n\u003cp\u003eCross-scale monitoring network construction: Deploying IoT sensors to monitor O\u003csub\u003e3\u003c/sub\u003e and NO₂ concentrations in real time, and integrating pollution data with bird migration trajectories through a GIS platform to achieve dynamic early warning.\u003c/p\u003e\n\u003cp\u003ePolicy synergy mechanism: incorporate air quality management into the Tianjin Biodiversity Conservation Action Plan, establish a joint enforcement framework between environmental protection, forestry and transport departments, and enhance public participation through community education.\u003c/p\u003e\n\u003cp\u003eThis study provides a new scientific perspective on bird habitat conservation in the context of urbanisation, and highlights the urgent need to integrate air pollution management with ecological restoration in order to achieve a sustainable biodiversity conservation pathway, driven by the goal of \u0026lsquo;carbon neutrality\u0026rsquo;.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eDeclaration of Competing Interest\u003c/h2\u003e\n\u003cp\u003eThe authors declare that there is no conflict of interests.\u003c/p\u003e\n\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eDeclaration\u003c/p\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eKumsong Jon: Writing \u0026ndash; original draft, Supervision. Chunyi Wang: Writing \u0026ndash; Visualization, Methodology. Kumsu Ri: Formal analysis, Conceptualization. Gwangson Cha: Data curation. Iljin Pak: original draft, Validation. Guozhu Mao: Writing \u0026ndash; review \u0026amp; editing\u003c/p\u003e\n\u003ch2\u003eAcknowledgement\u003c/h2\u003e\n\u003cp\u003eThe authors would like to thank the reviewers and editor for their careful and thoughtful comments. The current version of the paper has greatly benefited from their precise and valuable feedback.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAgathokleous, E., et al. \u0026quot;Ozone Pollution Threatens Bird Populations to Collapse: An Imminent Ecological Threat?\u0026quot; Journal of Forest Research, vol. 34, 2023, pp. 1653-1656. https://doi.org/10.1007/s11676-023-01645-y.\u003c/li\u003e\n\u003cli\u003eBarton, Madeleine G., et al. \u0026quot;A Review of the Impacts of Air Pollution on Terrestrial Birds.\u0026quot; Science of the Total Environment, vol. 873, 2023, p. 162136. https://doi.org/10.1016/j.scitotenv.2023.162136.\u003c/li\u003e\n\u003cli\u003eCutler, D. R., et al. \u0026quot;Random Forests for Classification in Ecology.\u0026quot; Ecology, vol. 88, no. 11, 2007, pp. 2783-2792.\u003c/li\u003e\n\u003cli\u003eDORMANN, C. F., et al. \u0026quot;Model Averaging in Ecology: A Review of Bayesian, Information-Theoretic, and Tactical Approaches for Predictive Inference.\u0026quot; Ecological Monographs, vol. 88, no. 4, 2018, pp. 485-504.\u003c/li\u003e\n\u003cli\u003eDrake, J. M., et al. \u0026quot;Modeling Ecological Niches with Support Vector Machines.\u0026quot; Journal of Applied Ecology, vol. 43, no. 3, 2006, pp. 424-432.\u003c/li\u003e\n\u003cli\u003eEeva, T., et al. \u0026quot;Environmental Pollution Affects Genetic Diversity in Wild Bird Populations.\u0026quot; Molecular Ecology, vol. 18, no. 6, 2009, pp. 1055-1064.\u003c/li\u003e\n\u003cli\u003eElith, J., et al. \u0026quot;A Statistical Explanation of MaxEnt for Ecologists.\u0026quot; Diversity and Distributions, vol. 17, no. 1, 2011, pp. 43-57.\u003c/li\u003e\n\u003cli\u003eFAN, Y. H., et al. \u0026quot;Prediction of the Global Distribution of Arhopalus rusticus under Future Climate Change Scenarios of the CMIP6.\u0026quot; Forests, vol. 15, no. 6, 2024, p. 955.\u003c/li\u003e\n\u003cli\u003eGU, R, et al. \u0026quot;Predicting the Impacts of Climate Change on the Geographic Distribution of Moso Bamboo in China Based on Biomod2 Model.\u0026quot; European Journal of Forest Research, vol. 143, no. 5, 2024, pp. 1499-1512.\u003c/li\u003e\n\u003cli\u003eGuisan, A., et al. \u0026quot;Generalized Linear and Generalized Additive Models in Studies of Species Distributions: Setting the Scene.\u0026quot; Ecological Modelling, vol. 157, no. 2-3, 2002, pp. 89-100.\u003c/li\u003e\n\u003cli\u003eGuisan, A., et al. \u0026quot;Predicting Species Distributions for Conservation Decisions.\u0026quot; Ecology Letters, vol. 16, no. 12, 2013, pp. 1424-1435.\u003c/li\u003e\n\u003cli\u003eGuisan, A., and W. Thuiller. \u0026quot;Predicting Species Distribution: Offering More than Simple Habitat Models.\u0026quot; Ecology Letters, vol. 8, no. 9, 2005, pp. 993-1009.\u003c/li\u003e\n\u003cli\u003eHaluza, D., et al. \u0026quot;Perceived Health Impacts of Air Pollution: A Review of the Literature.\u0026quot; International Journal of Environmental Research and Public Health, vol. 11, no. 5, 2014, pp. 5182-5199.\u003c/li\u003e\n\u003cli\u003eHUANG, D. Y., et al. \u0026quot;Biomod2 Modeling for Predicting the Potential Ecological Distribution of Three Fritillaria Species under Climate Change.\u0026quot; Scientific Reports, vol. 13, no. 1, 2023, p. 18801.\u003c/li\u003e\n\u003cli\u003eIPBES. Global Assessment Report on Biodiversity and Ecosystem Services. 2024.\u003c/li\u003e\n\u003cli\u003eLi, Rui, et al. \u0026quot;Habitat Suitability Evaluation of National Key Protected Birds in Kunming Based on MaxEnt Model.\u0026quot; Journal of Southwest Forestry University, vol. 44, no. 5, 2024, pp. 165-175. DOI: 10.11929/j.swfu.202310063.\u003c/li\u003e\n\u003cli\u003eLYU, Huanxin, et al. \u0026quot;Habitat Suitability Assessment of Large Mammals and Rare Birds in Xianju County Based on MaxEnt Modeling.\u0026quot; Chinese Journal of Ecology, vol. 42, no. 11, 2023, pp. 2797-2805.\u003c/li\u003e\n\u003cli\u003ePettorelli, N., et al. \u0026quot;Satellite Remote Sensing for Applied Ecologists: Opportunities and Challenges.\u0026quot; Journal of Applied Ecology, vol. 51, no. 4, 2014, pp. 839-848.\u003c/li\u003e\n\u003cli\u003ePhillips, S. J., R. P. Anderson, and R. E. Schapire. \u0026quot;Maximum Entropy Modeling of Species Geographic Distributions.\u0026quot; Ecological Modelling, vol. 190, no. 3-4, 2006, pp. 231-259.\u003c/li\u003e\n\u003cli\u003eSADEGHI, M., et al. \u0026quot;Interspecific Niche Overlap and Climatic Associations of Native Quercus Species in the Zagros Forests of Iran.\u0026quot; Global Ecology and Conservation, vol. 51, no. 3, 2024, p. e02878.\u003c/li\u003e\n\u003cli\u003eSanderfoot, O. V., and T. Holloway. \u0026quot;Air Pollution Impacts on Avian Species via Inhalation Exposure and Associated Outcomes.\u0026quot; Environmental Research Letters, vol. 12, no. 8, 2017, p. 083002.\u003c/li\u003e\n\u003cli\u003eWANG, H. R., et al. \u0026quot;Predicting Impacts of Climate Change on Suitable Distribution of Critically Endangered Tree Species Yulania zenii (W. C. Cheng) D. L. Fu in China.\u0026quot; Forests, vol. 15, no. 5, 2024, p. 883.\u003c/li\u003e\n\u003cli\u003eThuiller, W., et al. \u0026quot;BIOMOD-A Platform for Ensemble Forecasting of Species Distributions.\u0026quot; Ecography, vol. 32, no. 3, 2009, pp. 369-373.\u003c/li\u003e\n\u003cli\u003eWANG, T., et al. \u0026quot;Predicting the Potential Geographic Distribution of Invasive Freshwater Apple Snail Pomacea canaliculate (Lamarck, 1819) under Climate Change Based on Biomod2.\u0026quot; Agronomy, vol. 14, no. 4, 2024, p. 650.\u003c/li\u003e\n\u003cli\u003e陈, 思如, et al. \u0026quot;基于MaxEnt优化模型的罗布麻类植物潜在分布研究.\u0026quot; 北京师范大学学报 (自然科学版), vol. 59, no. 3, 2023, pp. 377-387.\u003c/li\u003e\n\u003cli\u003e刘, 枫, et al. \u0026quot;基于不同气候情景的美国白蛾适生区预测.\u0026quot; 应用昆虫学报, vol. 60, no. 1, 2023, pp. 76-86.\u003c/li\u003e\n\u003cli\u003e吕, 环鑫, et al. \u0026quot;基于MaxEnt模型的仙居县大型兽类和珍稀鸟类栖息地适宜性评价.\u0026quot; 生态学杂志, vol. 42, no. 11, 2023, pp. 2797-2805. https://doi.org/10.13292/j.1000-4890.202311.01\u003c/li\u003e\n\u003cli\u003e天津市城市规划设计研究院. 天津市国土空间总体规划 (2019\u0026ndash;2035年) (初稿). 天津: 天津市人民政府, 2020.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"herons, habitat suitability, air pollution, ensemble modelling","lastPublishedDoi":"10.21203/rs.3.rs-6659071/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6659071/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn this study, we systematically integrated air pollution factors (PM2.5, NO₂, SO₂, and O3) with traditional environmental variables such as climate, topography, and land use for the first time, constructed a multidimensional database of environmental factors, and developed an ensemble modelling framework based on the Biomod2 platform to systematically assess the effects of air pollution on heron habitat suitability in Tianjin City. The results showed that air pollution factors significantly improved the model prediction accuracy, especially ozone (O3) and sulphur dioxide (SO₂) had the most significant effects on the habitat suitability of pond herons. The sensitivity of different heron species to pollutants differed significantly, e.g. pond herons were more sensitive to O3 and SO₂, while the distribution of night herons was mainly restricted by NO₂ and PM10. The study proposes targeted conservation recommendations, including strengthening hierarchical control of pollution sources, implementing ecological remediation techniques and establishing a dynamic monitoring network, which provide a scientific basis for the conservation of bird habitats in the context of urbanisation. The study highlights the urgent need to integrate air pollution control and ecological restoration to achieve a sustainable pathway for biodiversity conservation, driven by the goal of 'carbon neutrality'.\u003c/p\u003e","manuscriptTitle":"Habitat suitability analysis of heron in relation to airborne pollutants_Tianjin city case study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-21 06:07:05","doi":"10.21203/rs.3.rs-6659071/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"43771149-3b9e-4e95-b399-9d8858d58f57","owner":[],"postedDate":"May 21st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-09-23T11:08:34+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-21 06:07:05","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6659071","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6659071","identity":"rs-6659071","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}