1 Introduction
As an important foundation for the survival and development of society, biodiversity has an irreplaceable role in maintaining ecosystem stability
[1]. However, the rapid urbanization and high intensity of human activities have resulted in the increasing loss of biodiversity in urban environments, which has become a common problem worldwide
[2]. High-quality planning and construction of nature reserves has become an effective way to protect biodiversity
[3]. Since 1956, China has established 2,750 nature reserves at various administrative levels, covering a total area of 1.47 million km
2, which accounts for about 15% of the national land cover
[4]. However, in the early stage of reserve establishment, the lack of scientific assessment of conservation priorities has resulted in the overprotection of several species and insufficient protection of some key species in part of protected areas
[5].
Ecological conservation priority area (ECPA) identification is a quantitative technique that designates a priority network of protected areas based on specific conservation objectives, thus providing data support for regional ecological and environmental planning
[6]. In terms of technical operation, foreign scholars began early on to use systematic conservation planning software in biodiversity conservation planning and research. For example, Marco Girardello et al. combined ecological niche modeling and Zonation model to establish important areas for butterfly conservation in Italy and set corresponding conservation priorities
[7]. Juliette Delavenne et al. assessed the prioritization of marine biological reserves in the eastern English Channel by comparing the simulation results of Marxan and Zonation models
[8]. Currently, most of the domestic studies have applied ECPA identification to delineate biological habitats and proposed corresponding conservation strategies based on the biodiversity impact assessment. Jia Tang et al. analyzed the distribution characteristics of nature reserves in Hubei Province from five perspectives, i.e., dominant ecosystems, special ecosystems, endemic ecosystems, ecosystems with high species richness, and special habitats, and then pointed out the protection vacancies that existed in the current priority forest ecosystems
[9]. Fukang Chen et al. took the Guangdong–Hong Kong–Macao Greater Bay Area (the GBA) as the research area and constructed a set of ECPA identification methods applicable to urban agglomerations by integrating ecosystem services and ecological vulnerability indicators
[10]. Focusing on the Minshan Region, Jing Xiao et al. identified the ECPAs for rare and endangered species on the basis of MaxEnt and Zonation modeling results, and proposed a spatial optimization solution
[11]. Using a species distribution and systematic conservation planning model, Jian Zhou et al. recognized the ECPAs for terrestrial vertebrates in Yunnan Province and evaluated the status quo by dissecting their spatial distribution patterns
[12].
In terms of species selection, existing studies tend to consider a single species or taxon as the research object. For instance, Qiqi Luo et al. modeled the spatial distribution pattern and hotspots of waterbird diversity in the GBA using MaxEnt
[13]. Utilizing the same model, Trent D. Penman et al. predicted future distributional changes in snake habitat under a climate warming scenario
[14]. Further, Jiali Zeng et al. identified the suitability distribution of mangrove forests in coastal wetlands
[15]. Nevertheless, this single-taxon perspective often overlooks the holistic features and interrelationships of biodiversity, hindering a comprehensive understanding of the structure, function, and dynamic changes of biomes
[16]. It also hampers the ability to uncover species interactions, patterns of niche differentiation, and the underlying ecological processes, thereby limiting the scientific basis for ecosystem protection and restoration.
Consequently, this research focuses on the terrestrial wildlife in the GBA, including mammals, birds, amphibians, and reptiles. By applying the integrated MaxEnt–Zonation model, it aims to predict the potential distribution pattern of priority terrestrial wildlife that are highly valuable for conservation or severely threatened in the GBA, to identify the prioritized distribution characteristics of the nature reserves and the spatial gaps of existing protected area distribution, thereby proposing an optimization plan for nature reserve construction. The results can contribute to the advancement of biodiversity conservation methods, provide critical support for effective conservation and management of biological resources, and offer referable experiences and methods for sustainable development in other countries or regions.
2 Study Area
The GBA is located in South China, including Hong Kong Special Administrative Region, Macao Special Administrative Region, and cities of Guangzhou, Shenzhen, Zhuhai, Foshan, Huizhou, Dongguan, Zhongshan, Jiangmen, and Zhaoqing in Guangdong Province, covering 56,000 km
2[17]. It is characterized by high-density urban development, a large population, and a robust economy. Specifically, according to relevant data from the Guangdong Provincial Bureau of Statistics, the population density of the GBA has reached as high as 1,548 people/km
2[18], with a GDP of 13 trillion yuan
[19]. Given its vast territory and the significant topographical variation, i.e., mountains in the north and flat plains in the south, this region has a wide variety of habitat types, including seas, rivers, lakes, wetlands, mountains, and forests. The moderate mean annual temperature (22.9℃), abundant precipitation, and water–heat redistribution facilitated by the Nanling Mountains in the GBA have led to the formation of multiple vegetation types, such as southern subtropical evergreen broadleaf forests and mid-subtropical evergreen broadleaf forests
[20]. These habitats effectively facilitate survival, reproduction, and migration of species, resulting in a rich species diversity, including mammals, birds, amphibians, reptiles, fish, and invertebrates. Among them, there are 77 rare species, encompassing
Sousa chinensis,
Paramesotriton hongkongensis,
Platalea minor,
Chelonia mydas,
Prionailurus bengalensis,
Periophthalmus cantonensis,
Lutra lutra,
Emberiza aureola,
Manis pentadactyla, and
Pelodiscus sinensis, which are flagship species endemic to the GBA. These species exhibit a unique distribution across the region's various ecosystems, forming a complex and diverse ecological network
[21].
However, the rapid urbanization over the past 40 years has led to a dramatic increase in urban population size and density. Simultaneously, extensive areas of natural and agricultural land have been converted into construction land with heightened land fragmentation, causing problems such as resource depletion and degradation of ecological functions and placing substantial pressure on the environment
[22]. Although establishing a biodiversity conservation system and enhancing the overall ecosystem quality have been included as key tasks in the
Outline Development Plan for the Guangdong–Hong Kong–Macao Greater Bay Area, there is still a lack of research on the ECPA identification in the GBA that focuses on multiple species categories. Therefore, it is necessary to carry out a specialized study of the ECPA planning in the GBA to formulate scientific conservation and management strategies that support the holistic conservation of biodiversity in the region.
3 Materials and Methods
This research first predicted the potential distribution ranges of the studied species using the species distribution prediction model, MaxEnt, based on species occurrence data and regional environmental variable data. Then, combined with the spatial conservation planning model, Zonation, the ECPAs of multiple species were identified, which were further utilized to optimize the spatial layout of the nature reserves in the GBA and to rationally delineate the ECPAs.
3.1 Data Sources and Processing
3.1.1 Key Terrestrial Wildlife Species
Within the geographic range of the GBA, this research first screened 457 terrestrial wild vertebrates from The International Union for Conservation of Nature (IUCN) Red List of Threatened Species, The List of Wild Animals Under State Priority Conservation, and The List of Wild Animals Under Priority Conservation of Guangdong Province, which included 28 amphibian species, 64 reptile species, 307 bird species, and 58 mammal species. Subsequently, species occurrence data were obtained through the Global Biodiversity Information Facility and the website of China Bird Report. These records were supplemented with data from the Second National Survey of Terrestrial Wildlife Resources and field sampling data from previous literature. Finally, this research obtained 32 species after manual removal of the species with fewer than five occurrence points, including 7 mammals, 13 birds, 5 amphibians, and 7 reptiles (Table 1).
3.1.2 Environmental Variables
Environmental variables have significant impacts on species distribution
[23]. Based on existing studies and considering the environmental characteristics of the target species' habitats and data availability in the study area, this research selected three categories of environmental variables, namely meteorological variables, topographical variables, and anthropogenic disturbance variables (Table 2). The corresponding sources were: 1) 19 meteorological data for the year 2020 (1 km resolution) were downloaded from WorldClim
[24]; 2) Land use/land cover data for the year of 2020 (30 m resolution) were acquired through the GlobeLand30 dataset
[25]; 3) Digital elevation data (DEM) were accessed through geospatial data cloud (GDEM V3, 30 m resolution); 4) Road data at all levels were obtained from Baidu API, including railroads, highways, expressways, and national highways; 5) Administrative boundary vector ranges for the GBA were derived from the Public Map Service System of Guangdong Province, map No. GS Yue (2023) 1032; 6) Population density data sourced from the 1 km population density dataset of China on WorldPop.
Data preprocessing was conducted using ArcGIS 10.8. First, the national administrative boundary vector data were used as a mask to extract the meteorological variable, topographic variable, and anthropogenic disturbance variable data (30 m resolution) of the GBA in 2020. Second, hydrologic analysis was performed with DEM data, and the Euclidean distance between the corresponding grid and the water system was calculated. Third, the built-up area was extracted from the land use/land cover type data, and the Euclidean distance between the corresponding grid and the grid of built-up area was calculated. Finally, all raster data were converted to ASCII format.
3.2 Research Methods
3.2.1 Operation and Optimization of the MaxEnt Model
The MaxEnt model was used to simulate the maximum entropy distribution probability and to predict the potential distribution of target species based on occurrence data under the assumption of maximum entropy under various conditions
[26]. Specifically, it applies the convex hull method to generate the distribution area of pseudo-occurrence points (background or pseudo-absence points) within the study area, making the result more reasonable and consistent with actual environmental conditions. This approach effectively enhances the model's ability to learn from environmental variables, thereby improving its capacity to accurately simulate species distribution patterns
[27]. To ensure model accuracy and reduce spatial overfitting, ENMTools was employed to consolidate multiple occurrence records of the same species within each 1 km × 1 km grid into a single one, resulting in 1,891 valid species occurrence records. Species with five or more valid occurrence records, along with environmental variables, were imported into the model. Of these, 75% of the data were randomly selected as training data for prediction, while the remaining 25% were used for model validation. Unless otherwise specified, all other parameters were kept at their default settings. Then, the Jackknife method was applied with 10 replicates. The average predicted probability of species occurrence for each grid was obtained. Model performance was assessed using the Receiver Operating Characteristic (ROC) curve, with the Area Under the Curve (AUC) as the accuracy metric. The AUC was interpreted as follows: 0.6 to 0.7 indicates poor performance, 0.7 to 0.8 moderate, 0.8 to 0.9 good, and 0.9 to 1.0 excellent
[28]. After excluding species with AUC values below 0.7, the final predictions of potential distribution areas were retained for the each species.
3.2.2 Relevance Assessment of Environmental Variables
To improve model accuracy and reduce overfitting errors due to variable covariance
[29], this research analyzed the contribution and permutation importance of 28 environmental variables by the Jackknife method and removed those with a contribution rate of less than 1 in the pre-simulation experiments. Meanwhile, Pearson correlation analysis was performed among the environmental variables using ENMTools to calculate the autocorrelation coefficients between the variables. Existing studies commonly exclude less contributing variables with the absolute correlation coefficient smaller than 0.7, thereby ensuring that the environmental variables ultimately modeled for the experiment are reasonable
[29]. However, the environmental variables in this research fluctuated modestly, with the internal correlations showing a relatively stable pattern. Using the 0.7 threshold would lead to the premature exclusion of certain valuable variables, causing the loss of valid information. Therefore, the threshold of 0.8 was applied in this research to identify highly correlated variables.
3.2.3 Operation and Optimization of the Zonation Model
Zonation, Marxan, and C-Plan are current systematic conservation planning tools widely used in research and practice
[6,
30–
31]. Zonation has mostly been used to model large-scale layouts of conservation priority
[32]. Compared with other software, it is able to identify ECPAs for single or multiple species and emphasizes landscape connectivity, maximizing the value of biodiversity conservation
[33]. In this research, Zonation v4 software was used to conduct ECPA planning for key terrestrial wildlife in the GBA. First, using the core area Zonation method, the least valuable grids for species suitability (e.g., poor environment, lack of food resources) were removed while retaining the core distributional area of the studied species. Second, edge removal was employed to eliminate the least valuable grids from the landscape edges to maintain structural connectivity
[34]. During the computation, the value of warp factor determined the number of grids removed every time
[12], a default value of 1 was applied for the warping factor in this research.
In addition, considering that the built-up area may significantly affect the classification of ECPAs of the species, this research applied the erase by the mask function in Zonation to exclude the undesirable areas and improve the simulation accuracy. All other parameters were set to their default values. This setup can improve the model's reliability while taking into account the specific characteristics of the study area.
As the conservation value of different species varies, the rarer the species, the higher weight should be given to the model predictions. Therefore, before the identification of multi-species ECPAs, the protection level of each species must be clarified and the corresponding weights must be set
[11,
35]. In this research, weights were set according to the rarity and endangerment levels of different species and their economic and social values (Table 3).
Human activities have a significant impact on ecological land use designation
[36]. The rapid growth of population and economic activities since the 20th century has caused problems such as habitat modification, increased invasive species, climate change, and environmental pollution
[37], which have led to biodiversity crises
[38]. Thus, it is important to consider the potential impacts of anthropogenic activities or land expansion on the ecosystem in ecological studies. Internationally, scholars have included land values or land use fees as variables for model prediction
[39]. In contrast, related studies in China have seldom included human activities in predicting parameters. Moreover, due to the complexity of the situations in each region, it is difficult to accurately calculate the use fee per unit of land. Therefore, to enhance the accuracy and reliability of findings, an innovative approach was adopted in this research. By utilizing the erase by mask approach and taking the distribution of existing built-up area as a variable, the impacts of human activities were rationally incorporated into the identification of the ECPAs.
This research then used R language statistical software to plot the landscape degradation rate–risk of habitat loss curve for different categories of species separately, enabling a comparative analysis of how various species respond to landscape degradation. Meanwhile, key data points were labeled on the curves, such as turning points where the risk of species habitat loss begins to increase significantly when landscape degradation reaches a certain level. Further, the quantitative relationship between landscape degradation and the risk of habitat loss was identified by calculating parameters such as the slope and curvature of the curve. This step served to assess the increasing rate and trend of risk, to identify key factors affecting the risk curve, and to provide a scientific basis for the development of targeted conservation strategies. For example, if the analysis indicates that the risk of habitat loss for a species is most sensitive to a particular indicator (e.g., loss of vegetation cover), then conservation efforts should prioritize monitoring and implementing restoration measures accordingly.
Building on this analysis, the research further identified the critical threshold of remaining habitat proportion for conservation. On the one hand, this threshold ensures that conservation priority areas can effectively protect species distributions, maintain basic biodiversity stability, and prevent severe threats to species survival caused by excessive habitat loss. On the other hand, from the perspective of rational resource use, it enables effective protection of key species and ecosystems while keeping costs manageable and promoting efficient resource allocation.
According to the global 2030 ecosystem conservation target issued by the
Convention on Biological Diversity in 2022, by 2030 at least 30% of terrestrial and inland water and marine and coastal areas should be effectively conserved and managed, and restoration measures should be undertaken for at least 30% of degraded terrestrial and inland water ecosystems and marine and coastal ecosystems
[40–
42]. In line with these targets, this research selected the top 30% grids with the highest conservation value from the Zonation results as the conservation priority areas. To optimize resource allocation and accurately protect ecologically critical areas, this research integrated ecological sensitivity, ecosystem integrity, and threat level of different areas to designate the top 10%, 20%, and 30% conservation value areas as the primary, secondary, and tertiary ECPA, using the reclassification tool in ArcGIS 10.8.
4 Results and Analysis
4.1 Evaluation of MaxEnt Model Simulation
This research calculated the relative contributions of environmental variables to the MaxEnt model. As the results indicated, for mammals, the contribution and permutation importance of altitude, distribution of built-up area, and coefficient of variation of precipitation were more significant. For birds, the contribution and permutation importance of the coefficient of variation of precipitation, precipitation of the coldest quarter, and distance to water body were more prominent. For amphibians, the coefficient of variation of precipitation, precipitation of the driest quarter, and distance to water body were more influential. For reptiles, the contribution and replacement importance of slope and the distribution of built-up area was higher.
4.2 Potential Habitat Distribution
Figures 1 ~ 4 present the potential habitat distribution for mammals, birds, amphibians, and reptiles, respectively. The red colors on the maps represent areas that are highly suitable for species to inhabit, indicating the potential presence of abundant food sources, suitable climatic conditions, or better ecological conditions. Blue color shows the habitats of lower suitability, where environmental conditions may be less favorable for species survival. The darker the red, the higher the habitat suitability; the darker the blue, the lower the habitat suitability.
4.3 Spatial Conservation Priority Pattern
4.3.1 Habitat Assessment
Raster data representing the potential distribution of 32 key terrestrial wildlife species were input into Zonation by species category. In turn, the spatial conservation priority pattern and the landscape degradation rate–risk of habitat loss curve for the four species categories in the study area were obtained (Fig. 5). Overall, as the landscape degraded, the risks of habitat loss for species in the study area increase. For mammals, the rate of increase in habitat loss risk remains relatively stable when landscape degradation is below 20%; when the degradation reaches 95%, the average risk of habitat loss rises to 40%, after which it accelerates sharply, eventually reaching 100%—indicating an extinction of existing species. For reptiles, the average risk of habitat loss rises to 40% when landscape degradation reaches 95%, followed by a sharp rise to 100%. In contrast, birds and amphibians face a comparatively lower risk of habitat loss. The average risk of habitat loss for birds reaches 25% when the proportion of degraded landscape reaches 95%, and for amphibians, the risk stays below 30% until the degradation reaches 97.5%. Beyond these thresholds, they both experience a phase of sharp increase.
According to the results, when the landscape degradation rate stays below 70%, the proportion of preserved habitat for each category of species remains consistently above 70%. As degradation continues to increase beyond this point, significant changes occur in habitats and the distribution pattern of the species, leading to a decline in conservation effectiveness. Therefore, this research identifies the 70% landscape degradation rate as a critical threshold for achieving effective conservation.
The primary, secondary, and tertiary ECPA results (Fig. 6) show that mammal habitats generally avoid high-density built-up areas and are mainly concentrated in areas with complex and diverse topography, including lowland wetland ecosystems in the Pearl River Estuary, nature reserves along the coastal zone, and mountainous forests and secondary woodlands in the northern GBA. These areas provide rich food resources and less disturbed living environments for mammals, specifically including Dinghu Mountain area in northern Zhaoqing; Luofu Mountain area in northern Huizhou; Futian Mangrove Nature Reserve, Wutong Mountain, and Dapeng Peninsula in Shenzhen; wetlands and mangrove forests close to the Pearl River Estuary in Zhongshan; the coastal areas in western and southern Jiangmen; the coastal areas in the central and southern parts of Zhuhai; the green spaces in northern and southern Macao; and the pristine forests and coastal areas in Lantau Island and UNESCO Global Geopark (Sai Kung) in Hong Kong.
Bird habitats are widely distributed at coastal wetlands and mangrove forests, as well as alluvial plains and paddy fields along the rivers, which supply essential stopover and nesting sites for various migratory birds. Meanwhile, inland open woodlands and grasslands are also important habitats for niche-dependent ground-nesting birds. Nansha Wetland Park in Guangzhou; Lianhua Hill Park and Dapeng Penisula Nature Reserve in Shenzhen; wetland parks in Jinwan District and the mangrove area in Doumen District in Zhuhai; the coastal areas in Haojiang District and Nan'ao Island Nature Reserve in Shantou; Danxia Mountain in Shaoguan; the wetland parks in Xiqiao Mountain and Shunde District in Foshan; Guifeng Mountain National Forest Park in Jiangmen; Dinghu Mountain in Zhaoqing; Xihu Lake and Loufu Mountain area in Huizhou; Wugui Mountain and the wetland areas along the Peral River in Zhongshan; Songshan Lake Eco-Park and the coastal wetlands of Humen Town in Dongguan; the wetlands and rural mountains in Macao; Lantau Island, Meiwo, and Global Geopark (Sai Kung) in Hong Kong are potential bird habitats in the GBA.
Amphibians and reptiles share similar living habits, favoring areas with abundant water resources such as streams, lakes, and wetlands, especially those rich in aquatic plants, while also commonly found in hilly and low mountainous forests. These areas can provide suitable thermal conditions and diverse habitats for breeding and foraging. The Qixingyan Scenic Area, Xing Lake, and Dinghu Mountain in Zhaoqing; the wetlands in the Pearl River Estuary area along the western coast and Tai Mountain in the south, and Kaiping watchtowers and ancient villages in the central and eastern inland, and the river and wetland areas of Enping in Jiangmen; Hengqin Island, Jiuzhou Island in Zhuhai, and the mangroves and coastal wetlands in their surroundings; Baiyun Mountain, Dafu Montain forested area, and Huadu Wetland Park in Guangzhou; Huiyang District, Daya Bay, and Luofu Mountain Huizhou; Songshan Lake in southeastern Dongguan; Wutong Mountain and Dapeng New Area in Shenzhen; the country parks in northern Macao, and the wetlands in the south; and the Lantau Island, Global Geopark (Sai Kung), and Hong Kong Wetland Park in Hong Kong are all major potential distribution areas for amphibian and reptile species habitats in the GBA.
4.3.2 Distribution of ECPAs
The overall distribution of ECPAs for of key terrestrial wildlife species in the GBA were delineated by inputting the potential distribution data of the 32 species into the Zonation and assigning appropriate weights to each grid (Fig. 7). The results show that most of the ECPAs distribute in three types of zones with distinctive geographic features, including mountainous forested areas, coastal and seashore areas, and nature reserves on the outskirts of cities. The mountainous forested areas in the northern and central parts of Zhaoqing (e.g., Dinghu Mountain, Qixingyan scenic area) and the Luofu Mountain in Huizhou provide habitats and refuges for diverse terrestrial wildlife. Coastal wetlands and mangroves of the Pearl River Estuary in Jiangmen, Zhongshan, and Zhuhai, as well as the coastal areas from central and western to eastern Shenzhen, are vital for maintaining the ecosystem balance and biodiversity. Songshan Lake Eco-Park in Dongguan, the country park in Macao, and coastal national parks such as Lantau Island and Global Geopark (Sai Kung) in Hong Kong, although situated in urban fringes, offer valuable habitats for wild animals.
By overlaying the overall distribution of ECPAs with the established nature reserves in the GBA (Fig. 8), it is found that most of the ECPAs overlap with the existing nature reserves, but there are still some uncovered conservation gaps.
5 Discussion
5.1 Impacts of Precipitation Seasonality on the Potential Distribution of Key Terrestrial Wildlife Species
According to the analysis results of environmental variables affecting the distribution of key terrestrial wildlife species, this research found that the impact intensity of the meteorological variables on species distribution is stronger than that of topographical variables and anthropogenic disturbance variables. Among them, precipitation seasonality (i.e., Bio16, Bio17, Bio18, Bio19) stands out as the dominant environmental variable shaping the potential distribution of relevant species, reflecting the critical role of precipitation in ecosystem structure and ecological functions
[43].
Precipitation seasonality plays a critical role in ecosystems by affecting water availability, food chain relationships, reproduction, habitat suitability, and species tolerance, and contributes significantly to species habitat distribution. First, an adequate and stable water supply is fundamental to the survival of many organisms. However, seasonal fluctuations in precipitation can directly affect water supply or lead to water shortages during certain periods. This water supply–demand mismatch is a major determinant of species distribution. Second, precipitation seasonality affects the growth and distribution of vegetation
[44–
45], influencing the habitat selection and distribution of the animal species that feed on it. Furthermore, the reproductive activities of many species (e.g., mating, incubation, juvenile development) are closely related to seasonal rainfall. Favorable precipitation patterns can provide optimal breeding conditions, and vice versa can limit the species' distribution. In addition, different precipitation patterns can create diverse habitat types, including wetlands, forests, and grasslands, the suitability of which directly determines the species' habitat distribution. Ultimately, there are significant differences in species' sensitivity to precipitation seasonality. For example, small changes in precipitation may be beyond the adaptive range of species that are highly sensitive to rainfall variability, limiting their habitat distribution.
In summary, future terrestrial wildlife conservation and management efforts should take into account seasonal variations in precipitation and implement timely measures to protect and restore the wildlife habitats and maintain the regional biodiversity and ecological balance. Similarly, future research should also explore in depth the specific impact mechanisms of precipitation seasonality on wildlife population size, distribution range, and migratory behavior to provide a scientific basis for biological conservation and management.
5.2 Optimization of Spatial Layout
In comparison with the existing nature reserves, this research identified the conservation gaps for key terrestrial wildlife in the study area, which are concentrated in the built-up areas with intense human activities in the eastern and southern GBA (Fig. 9). The three major cities, namely Shenzhen, Dongguan, and Hong Kong, have a denser distribution and higher prioritization of ECPAs. However, the available land for further urban development is limited. Thus, how to properly deal with the relationship between urbanization and ecological conservation has become a common development challenge for these cities. To avoid the encroachment of built-up area into ecologically valuable areas, it is essential to define land use boundaries rationally and formulate and refine regulations on conservation areas. In practice, a proportion of dividends from economic development can be invested in ecological protection and restoration, to coordinate the relationship between economic development and ecological protection, as well as to realize their mutual development. Additionally, innovative technologies such as vertical greening can be implemented to create more habitats within the built-up areas to protect and enhance biodiversity, realizing the sustainable urban development and objectives of ecological conservation and promoting the harmonious coexistence of human beings and nature.
5.3 Reasonableness and Limitations in Incorporating Anthropogenic Disturbance Variables
By using the erase by mask approach and treating the distribution of built-up area as an environmental variable, this research effectively eliminated existing development areas and improved the accuracy of conservation area calculations, thereby rationally incorporating the impacts of human activities into the process of identifying ECPAs. However, this method also presents certain limitations. First, as built-up areas are primarily used for residential and industrial purposes, their internal environments are highly influenced by human activities, making it difficult to maintain original natural ecosystem functions. Therefore, removing them helps improve the ecological integrity and effectiveness of conservation area calculations, contributing more to achieving the goal of ecological conservation. However, it is somewhat reductive to assume that built-up areas are unsuitable as habitats. In fact, some species may have gradually adapted to the developed environment. Arbitrarily excluding these areas may overlook their internal ecological potential and result in the omission of potential conservation areas.
6 Conclusions and Prospects
Taking the GBA as the study area and 32 key terrestrial wildlife species as the research object, this research successfully predicted the ECPAs in this region by integrating the species distribution prediction model MaxEnt and the multi-species conservation planning model Zonation. The conclusions drawn are as follows. 1) Precipitation seasonality has a significant impact on the potential distribution of the species, and the suitable areas are mainly concentrated in areas with abundant precipitation and strong water retention capacity. 2) The ECPAs are mainly located in the northern and central mountainous and forested areas of Zhaoqing, the coastal areas of Jiangmen, the central to southern coastal areas of Zhuhai, the central northeastern parts to coastal areas of Zhongshan, the central to coastal areas of Huizhou, the southeastern Dongguan, the central, western, and coastal areas of Shenzhen, the northern and southern Macao, the coastal area of Hong Kong. 3) The predicted ECPAs overlap with most of the established nature reserves, but there are still gap areas in the eastern and southern coasts of the GBA.
The results of this research can provide valuable references for advancing ecological conservation in other regions. Nevertheless, this research is only a conceptual planning attempt, and several factors need to be considered when developing concrete planning objectives. First, the species occurrence data should be collected and surveyed in a more detailed and comprehensive manner. The studied species only covered key terrestrial wildlife and has not yet included all protected species in the study area. Meanwhile, the occurrence data used are based on partially available public sources and are mostly about planned conservation areas or where observers have presented. More infrared camera monitoring or field surveys are needed, particularly in areas beyond existing reserves and in less accessible regions to further clarify species distribution information. Second, dynamic factors such as future urban expansion and climate change should be incorporated into the planning of nature reserves. The impact of population growth and rapid urbanization on urban ecology cannot be ignored. In addition, in the context of global warming, environmental variables (e.g., precipitation pattern, temperature) will continuously change. As a coastal region, the GBA is vulnerable to the impacts of climate-related risks (e.g., sea level rise, storm surge). Therefore, the combined effects of these factors must be considered in future planning to ensure the sustainability and adaptability of nature reserve designation.
PDF
(11877KB)
