Background & Summary

Mapping species’ spatial patterns and understanding their distribution are crucial for informing species conservation strategies, biodiversity management, ecosystem restoration, and species extinction risk assessment1,2,3. As a guide for global biodiversity conservation, the targets in the Kunming-Montreal Global Biodiversity Framework—aiming to reduce biodiversity loss, restore degraded ecosystems and conserve 30% of land, waters and seas—rely heavily on species distribution data4. As one of the parties to both the Convention on Biological Diversity (CBD) and the Kunming-Montreal Global Biodiversity Framework, China has actively aligned its domestic policies with these international commitments. In 2024, the Chinese government published the “China National Biodiversity Conservation Strategy and Action Plan (2023–2030)”, which calls for the establishment of foundational biodiversity datasets, including species distribution data5.

There are various ways for obtaining species distribution data. Species surveys utilizing monitoring technologies based on environmental DNA (eDNA), passive acoustic monitoring, and visual sensors (e.g., camera-trapping) provide the most accurate information on species distributions6. However, these approaches are extremely costly, making them feasible only for investigating a limited number of species or those within specific regions7,8. Species distribution models (SDMs) are also widely used, typically integrating species occurrence points with environmental variables such as climate and topography to evaluate the likelihood of species presence. Common modeling approaches for constructing SDMs include MaxEnt9,10, AIM11, and machine learning algorithms12. However, limitations in the completeness and accuracy of species occurrence data reduce the reliability of the assessment results produced by these models13,14,15.

Area of Habitat (AOH) is defined as the suitable habitat available to a species within its geographic range16. AOH maps are generated by integrating information on the species’ geographic range, suitable habitat types, and elevation limits17. AOH maps can reduce commission errors for geographic ranges of species, especially not well-known and wide-range species. However, it is notable that for well-known species more accurate assessment methods may exist18. AOH maps are useful for identifying priority conservation areas19,20, assessing conservation gap21, supporting ecosystem restoration22, and evaluating species extinction risks1,23. Moreover, overlaying AOH across species enables the creation of species richness maps, which help identify biodiversity hotspots18.

Previous studies have generated multiple sets of AOH maps for terrestrial animals. For example, Rondinini et al.24 and Ficetola et al.17 generated 300 m resolution AOH maps for global terrestrial mammals and amphibians in 2009, respectively. Lumbierres et al.18 generated 100 m resolution AOH maps for global terrestrial mammals and birds in 2015. Similarly, Mi et al.25 generated 300 m resolution AOH maps for National Key Protected Wildlife (NKPW) in China in 2009. However, these datasets are limited to single-year AOH information, making it impossible to track temporal changes in AOH. The International Union for the Conservation of Nature (IUCN) Red List26 serves as an authoritative assessment of species’ threatened status on a global scale, and existing studies primarily use the IUCN Red List to generate AOH maps17,18,24. However, species on the IUCN Red List might not be fully representative of the threatened status of species in China25. For instance, while the global population of Castor fiber exceeds 639,000 individuals and is classified as Least Concern by IUCN, its Chinese population numbers merely 700 individuals, making it one of China’s rarest aquatic mammals26. Notably, nearly half of the terrestrial species on the NKPW List are classified by the IUCN as Least Concern or Data Deficient, yet these species are considered key protected species in China26,27. The NKPW List is based on rigorous scientific assessments of the threatened status of wildlife in China, incorporating input from diverse stakeholders (e.g., governments, scholars, and the public), and is officially published by the Chinese government and protected by Chinese law, which can be used to summarize the threatened status of species in China25.

The NKPW List contains 988 species classified into two protection level accounting to their degree of preciousness and endangerment, with 235 Class I species representing higher protection priority compared to 753 Class II species. We generate 30 m resolution AOH maps for 720 terrestrial species from the NKPW List (2021 version), covering the years 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2020, and 2022 (Fig. 1). Although the NKPW List contains 988 species (including 869 terrestrial species), we generate AOH maps for 720 terrestrial species sufficient geographic range data are available, excluding marine species and some terrestrial species lacking geographic range data. We used the resulting AOH maps to generate species richness maps for NKPW (Fig. 2). We refined the species distribution data derived from the IUCN Red List by incorporating adjustments to geographic range, elevation, and suitable habitat types (see Methods). These modifications ensured that the generated AOH maps more accurately reflect the actual distribution of species in China.

Fig. 1
figure 1

Area of Habitat (AOH) map of Black Muntjac (Muntiacus crinifrons). (a) The AOH of the species. (b) AOH changes within the Dexing mine (Asia’s largest open pit copper mine). This species’ habitats are forest and shrubland habitats and has an elevation range of 200–1000 m. Long-term temporal sequences of AOH maps can effectively track dynamic habitat changes, providing insights into how specific threats such as mining activities impact species’ habitats over time. However, we remind users to exercise caution when making temporal comparisons of AOH, as the AOH maps are generated based on land cover rather than cover change. Although GLC_FCS30D employs a continuous change-detection method to produce land cover products, there remain certain errors in temporal stability within the time series.

Full size image
Fig. 2
figure 2

Species richness maps for (a) all species (720 species), (b) class I protected species (195 species) and (c) class II protected species (525 species) in 2020, for example.

Full size image

High-resolution AOH maps with long-term temporal sequences are vital for biodiversity management and conservation strategy development. Our dataset of high-resolution, long-term temporal sequences of AOH maps of NKPW provide a clear depiction of habitat loss processes over time28, elucidating the relationship between the rate of habitat loss and associated threats (Fig. 1)29. These maps can enable the quantification of species restoration benefits, identification of areas with the greatest potential for species restoration, and prioritization areas for targeted restoration efforts11,22. This information serves as a robust scientific foundation for developing effective biodiversity conservation strategies30. Moreover, the field of corporate sustainability reporting (Environmental, Social, and Governance-ESG) also calls for long-term temporal sequences AOH maps to assess the impacts of corporate activities on biodiversity. For instance, the Environmental Sustainability Reporting Standards (ESRS)31, Taskforce on Nature-related Financial Disclosures (TNFD)32, and the Global Reporting Initiative (GRI)33 all emphasize the importance of focusing on habitat changes, particularly nationally threatened species. Therefore, this dataset has significant research value and broad application potential.

Methods

AOH maps are generated by extracting suitable elevation ranges and habitat types from species’ geographic range maps (Fig. 3)16,18. The species’ geographic range maps primarily originate from the IUCN Red List26. Furthermore, we incorporate data from Mi et al.25,34 to refine geographic range maps of mammals, birds, reptiles, and amphibians to better align with species distribution in China. Data sources for each species’ geographic range maps are documented in the species AOH information table35. Species‘ geographic range maps for the IUCN Red List typically include multiple range polygons (extant and probably extant presence; native, reintroduced, and assisted colonization origin; resident seasonality for non-migratory species; and extinct, etc.). All range polygons are retained to generate AOH maps because we aim to capture all habitat information where the species may currently be present or where it may be suitable for restoration in the future. If users require AOH maps for a specific range, they can extract them using the relevant range polygons. Habitat and elevation preference information for each species are primarily obtained from IUCN Red List (version 2024-2)26. Since the IUCN Red List lacks assessment data for some species, we also supplement this information with data from professional biodiversity knowledge platforms, such as eSPECIES (http://www.especies.cn/especies), and relevant literature25. We compile a table summarizing the geographic range, habitat, and elevation preference information and data sources for each species. This table is stored in the database of Science Data Bank35.

Fig. 3
figure 3

Technical flow to produce AOH maps. For example, the Golden Snub-nosed Monkey (Rhinopithecus roxellana) is a forest species with elevation range limits of 1400–2800 m.

Full size image

Since the IUCN Red List does not provide spatially explicit habitat maps, we develop a translation table for terrestrial habitat classes based on the global 30 m land-cover dynamic monitoring product (GLC_FCS30D). The translation table associates each terrestrial habitat type in the IUCN habitat classification scheme with corresponding GLC_FCS30D land cover types at the level-1 habitat class scale. This translation table is also stored in the Science Data Bank35. GLC_FCS30D is the first global fine land cover dynamic product at a 30 m resolution that employs continuous change detection36,37. It uses a refined classification system containing 35 land categories and covers the time span from 1985 to 2022. The product achieves an overall accuracy of 80.88% (±0.27%) for the basic classification system and 73.04% (±0.30%) for the level-1 classification system. Notably, some level-1 land cover types show low accuracy, and users can check the specific accuracy for individual type by combining the translation table and the study of Zhang et al.36. Given the advantages of GLC_FCS30D, including its long-term temporal sequences, high resolution, and high accuracy, we select it as the basis for generating AOH maps. We derive elevation data from NASA’s 30 m Digital Elevation Model38. To ensure AOH map accuracy, we validate results using point prevalence and model prevalence based on the GBIF species point data39. Detailed validation procedures are described in the “Technical Validation” section.

Finally, the AOH maps for all species are spatially overlaid to generate a species richness map, representing the number of species at each 30 × 30 m grid.

Data Records

AOH maps, species richness maps, and supplementary information tables are stored in Science Data Bank (https://doi.org/10.57760/sciencedb.20221)35. We have generated AOH maps of 720 NKPW species, each with at least one terrestrial-related habitat type for nine years (1985, 1990, 1995, 2000, 2005, 2010, 2015, 2020, and 2022). These maps cover Mammalia (132 species), Aves (370 species), Reptilia (61 species), Amphibia (90 species), Actinopterygii (55 species), Insecta (11 species), and Bivalvia (1 species). AOH maps data are organized by protection level (Class I, Class II) and taxonomic class, and stored in the AOH folder. The map values are set to 1 for the AOH areas and Null for the background areas. Each species’ AOH map is named using its scientific name followed by the year (e.g., Ailuropoda_melanoleuca_2020). The nine AOH maps for each species, corresponding to nine time points, are stored in a single file geodatabase named with the species’ scientific name and Chinese name. Richness maps of all species across the nine time points are stored in the Richness folder. The values in species richness maps represent the number of species at each grid, and the maps are named using the format “Richness_All/Class1/Class2_year”. All maps are stored in file geodatabase format (.gdb), which is accessible and operable using ArcGIS or ArcGIS Pro. The database contains three supplementary information tables: The species AOH information table records the geographic range, habitat, and elevation preference information and data sources for each species; the translation table records the mapping of each terrestrial habitat type of IUCN habitat classification scheme to the land cover type of GLC_FCS30D; and the AOH maps validation table records validation results (Point prevalence and Model prevalence) for each AOH map.

Technical Validation

The accuracy of AOH maps is evaluated using a new method developed by Dahal et al.39. Unlike previous studies that used only point localities and polygons of occurrence to validate AOH maps, this method evaluates the performance of AOH maps by comparing model prevalence with point prevalence at the species level. Model prevalence refers to the proportion of the area of AOH retained within the geographic range, while point prevalence is the proportion of all species points within a species‘ geographic range that are located within the AOH. If point prevalence exceeds model prevalence, the AOH maps perform better than random, indicating that the AOH maps reduce errors associated with geographic range maps and more accurately identify areas where most species occur.

Species occurrence points are downloaded from GBIF40, and only observation points within the species’ geographic range are selected. Points outside the range are excluded due to potential issues, such as misidentification or geolocation errors. A conservative 300 m buffer is applied to all observation points, as suggested by relevant literature39,41, to account for positional errors. This is because recorded locations often represent the observer’s location rather than the exact location of the target species. Given that there are only a few observation points in the GBIF database for some NKPW in China, all observation points within the species’ geographic range are selected.

We obtain all available species occurrence points data for 370 species of NKPW (51.39% of all species) from GBIF and used these data to validate the AOH maps for these species. On average, the AOH maps for all validated species show a mean point prevalence of 0.84 ± 0.20 SD and a mean model prevalence of 0.49 ± 0.27 SD (Table 1), with 97.03% of the species’ AOH maps perform better than random. AOH maps show good performance. However, it should be emphasized that while point prevalence exceeds model prevalence indicating that AOH maps perform relatively better than random, there may be a low point prevalence for a few species (mainly Aves). We provide an AOH maps validation table that details point prevalence and model prevalence for all 370 validated species35. For species with relatively low point prevalence values, we recommend exercising caution when using the AOH maps. Low point prevalence may be due to positional errors between observation locations and species occurrence locations, or the species’ actual ecological niche extends beyond the delineated geographic range42. Additionally, potential errors may still exist due to inaccuracies in species’ habitat and elevation preference information, or the translation table between habitat and land cover18.

Table 1 Mean model and point prevalence for AOH maps of species.
Full size table