“Two zones and three centers” distribution and suitable areas shift of an evergreen oak in subtropical China under climate scenarios

Abstract Understanding the impact of climate change on the geographical distribution of species is a fundamental requirement for biodiversity conservation and resource management. Quercus oxyphylla, an evergreen oak endemic to China, plays a crucial role in maintaining the ecological stability in subtropical regions and high economic value attributed to its dark and high‐density heartwood, but the existing resources are close to endangered. Currently, limited knowledge exists regarding its distribution and potential influences of climate change on suitable areas. This study utilized 63 occurrence records and Biomod2 platform, to predict changes in suitable areas for Q. oxyphylla under future climate change. The results revealed that (1) Q. oxyphylla showed a pattern of three disjunctive geographical centers in the eastern subregion of subtropical evergreen broad‐leaved forest region (IVA): Qinling‐Daba Mountains, Nanling Mountains and Wuyi Mountains center. Currently, the highly suitable areas concentrated in two zones divided by the Yangtze River, that is, the northern subtropical evergreen and deciduous broad‐leaved forest zone (IVAii) and the mid‐subtropical evergreen broad‐leaved forest zone (IVAi). (2) The temperature‐related variables, such as annual temperature range (Bio7), the mean diurnal range (Bio2), and annual mean temperature (Bio1), were identified as the key determinants of the distribution pattern. Because of its considerable climatic variations in temperature and water conditions, Q. oxyphylla's habitat displayed a wider climate niche and strong physiological tolerance to climate change. (3) Under future climate scenarios, the suitable area of the species was expected to overall expand with significant regional differences. The suitable area in IVAi was expected to expand significantly northward while that in IVAii was expected to gradually shrink. To address the impact of climate change, it is necessary to develop conservation plans focused around the three distribution centers, implement localized and regional conservation policies, and conduct educational outreach among local people.


| INTRODUC TI ON
Climate change has affected ecosystems worldwide, and its consequences are becoming increasingly evident (Huang et al., 2019;Parmesan, 2006).Global warming has intensified since the Industrial Revolution due to the emission of greenhouse gases.
The IPCC assessment report showed that the global average surface temperature will increase by a minimum of 0.3-1.7°C to a maximum of 2.6-4.8°Cover the 21st century, and the precipitation pattern will also change significantly (Lee et al., 2023;Thackeray et al., 2022).Disruptions in species distributions and ecosystem functions are anticipated as a result of these climatic shifts (Hama & Khwarahm, 2023;Parmesan et al., 2022).
Understanding how different species respond to climate change has become a critical focus for the scientific community (Brown et al., 2016;Chmura et al., 2011).Some species are facing severe threats, including decline, endangerment, or even extinction due to climate change (Peng et al., 2021;Subedi et al., 2024), while others may adapt to climate change due to their strong ecological tolerance, thereby expanding their suitable areas (Bellard et al., 2012).These varied responses are often linked to species' the evolutionary histories and ecological niches, which are closely related to climate sensitivity (Yu et al., 2019).Therefore, understanding how plants respond to climate change and the underlying mechanisms is crucial for biodiversity conservation and addressing climate change.
Species distribution models (SDMs) have emerged as pivotal tools in conservation biology, ecology, and biogeography, as they enable the prediction of a species' geographic distribution based on a suite of environmental variables (Blair et al., 2022;Fois et al., 2018).These models serve as a powerful instrument for deciphering the intricate connections between species and their environmental contexts.Typically, SDMs entail the correlation of a species' presence or abundance with an array of environmental parameters, such as climate, topography, soil composition, and vegetation type (Norberg et al., 2019;Parmesan et al., 2022).Through the analysis of these correlations, SDMs can pinpoint the most influential environmental determinants of a species' distribution and forecast its potential range across various scenarios, including those influenced by climate change (Norberg et al., 2019;Subedi et al., 2024).The insights generated by SDMs are invaluable for conservation initiatives, aiding in the identification of areas of paramount conservation significance and the anticipation of the effects of climate change on species ranges.Among the most frequently employed SDMs are maximum entropy (MaxEnt), generalized linear models (GLMs), random forest (RF), artificial neural network (ANN), one rectilinear envelope similar to BIOCLIM (SRE), and flexible discriminant analysis (FDA) (Subedi et al., 2024;Suicmez & Avci, 2023).Despite their constraints, SDMs have revolutionized our understanding of species distributions and have become an essential tool in the quest to conserve and manage biodiversity in the face of environmental change.Hand.-Mazz. (Handel-Mazzetti, 1929), which is an evergreen tree endemic to subtropical China, belongs to Fagaceae and is included in the List of National Key Protected Wild Plants in China (version 2021).The tree is long-lived and grows in mixed mesophytic forests from 200 to 2900 m above sea level in the subtropical regions of China (Fang et al., 2009;Wu et al., 1999).It was often historically regarded as "miscellaneous wood" and was unknowingly harvested for an extended period, leading to a lack of understanding of this species (Sun et al., 2020;Zhang et al., 1995).In recent years, based on field investigations and wood property studies, we have recognized that this species has favorable ecological functionality, strong adaptability, and high economic value.Its heartwood exhibited a reddishbrown coloration and dense timber.Concurrently, we have also understood that its natural habitat has suffered significant disturbances and degradation due to anthropogenic activities such as cultivation, grazing, deforestation, afforestation, and tourism (Sun et al., 2021).Encroachment, fragmentation, and habitat eradication have resulted in drastic declines in wild populations (Xie et al., 2021).Therefore, mitigating this threat and implementing measures to restore the conditions necessary for habitat and population recovery are imperative.However, the lack of understanding of its distribution patterns and the key environmental factors that regulate its growth presents obstacles to conservation and management efforts.

Quercus oxyphylla
In this paper, we aimed to (1) assess the current distribution of Q. oxyphylla and identify the dominant climatic variables influencing it, (2) explore how the suitable area of this species might shift under future climate change scenarios.Given the extensive latitudinal and altitudinal range of this species in subtropical China (Wu et al., 1999), we hypothesized that its distribution will expand in response to future climate change.
implement localized and regional conservation policies, and conduct educational outreach among local people.

| Samples and species occurrence records
Data on the geographical distribution of Q. oxyphylla were mainly acquired from current population and historical collections, representing 220 records from 86 locations.Data were retrieved from the Chinese Virtual Herbarium (CVH, http:// www.cvh.ac.cn/ ), National Specimen Information Infrastructure (NSII, http:// www.nsii.org.cn/ ), Subject Database of China Plant (http:// www.plant.csdb.cn/ ), Plant Photo Bank of China (http:// ppbc.iplant.cn/ ) and relevant published literature (Guo et al., 2019).A taxonomic filtering was applied to all collected data to ensure accuracy and eliminate redundancy.Upon re-examination, we found that certain specimens previously believed to be from Yunnan, western Sichuan, central Guangxi, and eastern Fujian were misidentified (https:// doi.org/ 10. 5061/ dryad.qz612 jmqn), which indicated that previous publications using these records to delineate the geographical distribution of Q. oxyphylla might be inaccurate.After eliminating the misidentified data and incorporating newly occurrence records from our team's fieldwork in recent years, a total of 63 valid occurrence records were obtained encompassing 55 specimen records from 17 herbariums, 4 photography records, 3 field records, and 1 reference record (Figure 1).
The base map used for the analysis was derived from the National Geospatial Information System of China (http:// nfgis.nsdi.gov.cn).
Using ArcGIS 10.8, kernel density analysis was employed to identify the contemporary distribution center (Figure 1).By overlaying the 1:1,000,000 vegetation zoning map of China (Zhang, 2007), which effectively represents species-climate relationships, we were able to ascertain the general correlation between the geographical distribution patterns and climate.All the maps were georeferenced using the WGS84 coordinate system.

| Bioclimatic variables
A total of 19 bioclimatic variables (Table 1) were downloaded from the WorldClim database (version 1.4) (http:// www.world clim.org/ ) with a spatial resolution of 30 arcseconds, or approximately 1 km 2 per pixel (Fick & Hijmans, 2017).These variables were used to define the environmental niche of a species, utilizing both present and projected climate data (Hijmans et al., 2005).The time frame for the current period spanned from the 1960-1990s, while the designations of the 2050s and the 2070s refer to the 2041-2060s and 2061-2080s, respectively.The forecasted climate data were aligned with the most recent updates from the 5th IPCC Report.For Q. oxyphylla in China, the Climate System Model (BCC-CSM 1.1) was utilized to estimate future climate impacts based on Representative Concentration Pathway (RCPs) scenarios RCP2.6, and RCP8.5, which represent low-, and high-emission scenarios.These were evaluated during the 2050s and the 2070s (Xie et al., 2021).
In the modeling process, redundant information introduced by strongly correlated variables, which could reduce the reliability and predictability of outcomes, must be excluded (Hundessa  et al., 2018).To avoid overfitting between environmental factors (Gebrewahid et al., 2020), Pearson correlation analysis was conducted on 19 climate factors using the "Hmisc" package in R 4.3.2software (Figure S1).One variable per pair with high correlation coefficient (≥0.8) was removed, resulting in seven climate factors: the annual mean temperature (Bio1), the mean diurnal range (Bio2), the isothermality (Bio3), the temperature annual range (Bio7), the mean temperature of wettest quarter (Bio8), the annual precipitation (Bio12), and the precipitation seasonality (Bio15) (Table 1).

| Modeling species distribution
Data on the occurrence of Q. oxyphylla and seven selected bioclimatic variables during three periods (current, 2050s, and 2070s) were analyzed using a combined modeling approach and utilizes the six common modeling algorithm options (MaxEnt, GLMs, RF, ANN, SRE, and FDA) supported by the Biomod2 platform in R (Thuiller et al., 2024).
To improve species distribution model accuracy, during model simulation, 1000 pseudo-presence point was randomly generated based on the distribution range of each species.A random 75% of sample data was used to train the model, with the remaining 25% used for validation.The weights for presence data and pseudo-presence data were set to be equal.For the six single model algorithms, the models were run with 10 repetitions each, resulting in 60 models.
The model results that met the requirements were then normalized and multiplied by their corresponding weights to be added together.
When integrating the models, the single model weights were automatically allocated based on their TSS values (Subedi et al., 2024).

| Model evaluation
The accuracy of the model prediction was evaluated using the area under the curve of receiver operating characteristic curves (ROC), the true skill statistic (TSS), and Kappa coefficient (Liu et al., 2024).
Higher ROC and TSS values indicated a stronger relationship between the species distribution model and environmental variables, suggesting higher model prediction accuracy (Guo et al., 2015).ROC values range from 0.5 to 1, with values between 0.80 and 0.90 indicating high model simulation accuracy, and values between 0.90 and 1.00 indicating extremely high accuracy (Chen et al., 2022).The TSS evaluation metric ranged from −1 to 1, with values between 0.70 and 0.85 indicating good simulation results, and values between 0.85 and 1.00 suggesting extremely good results (Marmion et al., 2009).
The Kappa statistic was a measure of how the accuracy of the model's predictions compared with those expected by chance.A Kappa value greater than 0.85 indicated that the model's performance is exceptionally good (Araujo et al., 2005).Abbreviations: CV, coefficient of variability; SD, standard deviation.

| Environmental factor importance analysis
Factor importance was used to evaluate the contribution of environmental variables in limiting species' current geographic distribution patterns (Harms et al., 2020).These data relied on a specific pathway for finding the optimal solution, by progressively adjusting the coefficients of individual elements to increase the gain value, then allocating the incremental gain value to the environmental variable that determined the element, and converting it into a percentage before presenting it (Brown, 2014).To analyze the dominant environmental driving variables influencing the suitable distribution of the species under baseline climate, the weight values of each environmental factor from the combined model output by the Biomod2 package were used (Thuiller et al., 2024), and related boxplots were drawn using the ggplot2 package.

| Suitable range change and centroid migration
A threshold of 0.75 was set to generate a 0-1 raster map for presence/ absence.The number of raster cells representing the presence of each species in different time periods was counted.A bar chart representing the change in species area over different periods was generated.Using the SDMtoolboxs toolkit (Brown, 2014) in ArcGIS v10.8, the centroids of suitable areas were calculated based on all presence/absence distribution maps.Vector files were generated showing the direction and magnitude of change in suitable area centroids between adjacent time periods to represent the direction and distance of species centroid migration under different climate scenarios (Hama & Khwarahm, 2023).

| Model performance
The simulation accuracy of six models differed greatly.By selecting and combining models that exhibited TSS value above 0.7 in simulation results, the stability and accuracy of the overall model's simulation performance were significantly enhanced (Figure 2a).
Among these models, ANN, RF, GLM, MaxEnt, and FDA models demonstrate high simulation accuracy and showed greater consistency in their repeated model analyses.The average TSS and ROC for these models all surpassed the 0.8 threshold.However, most of the results from the SRE did not meet the standards for model integration.

| Importance of bioclimatic variables
The results of the combined model's environmental factor importance showed that bioclimate factors related to temperature were more important than those related to precipitation.Specifically, the annual temperature range (Bio7) was the most crucial climate variable influencing the geographic distribution of Q. oxyphylla, with its impact far exceeding that of other factors (Figure 2b).The next most important factor was the mean diurnal range (Bio2).Among precipitation-related factors, annual precipitation (Bio12) was the most significant, while precipitation seasonality (Bio15) was the least significant (Figure 2b).

| Distribution pattern under the current climate scenario
The sample data of Q. oxyphylla was located between latitudes Basin, Dongting Lake, and Poyang Lake were predicted low suitability regions (Figure 3).
Overall, the predicted suitable areas for Q. oxyphylla were concentrated in the IV Ai and IV Aii vegetation zones of the eastern subregion of the subtropical evergreen broad-leaved forest.
With the Yangtze River as a boundary, the suitable areas in the Qinling-Daba mountains north of the river gradually decreased, while the south of the river significantly increased (Figure 4).As climate warming intensifies over time, the number of grid cells in low and moderate suitability areas would gradually decrease, while the range of high suitability areas significantly would increase (Figure 5a).

| Suitable areas overlap and centroid shift
Under projected future climate scenarios, the suitable range of Q.
oxyphylla would generally expand outward in response to global warming (Figure 6).However, distinct trends emerged under different climate scenarios and across geographical regions.Under the less severe RCP2.6 scenario, the expansion trend showed a slight weakening over time.Conversely, under the more extreme RCP8.5 scenario, the expansion trend intensified slightly (Figures 5b and 6).
Geographically, a clear north-south divide was observed (Figure 6).

| Geographical distribution pattern of Q. oxyphylla
The occurrence records and current highly suitable areas of Q. oxyphylla indicated three disjunctive geographical centers of Qinling-Daba Mountains, Nanling Mountains, and Wuyi Mountains in the eastern subregion of subtropical evergreen broad-leaved forest region (IV A ) (Figures 1 and 3).Interestingly, the mountains where the distribution and suitable areas were located were only in a roughly east-west direction, which coincided with the important geographic dividing region of climate and vegetation zonation in China (Ding, 2013;Zhang, 2007).To the north, the Qinling-Daba Mountains is located on the northern region of the Chinese subtropics (Yang et al., 2018), which documents the northern subtropical evergreen and deciduous broad-leaved forest zone (IV Aii ) (Figure 1).
In the south, the Nanling and Wuyi Mountains stand in the southern part middle subtropical evergreen broad-leaved forest zone (IV Ai ) (Huang et al., 2023).Synthetically, we suggested that the Q. oxyphylla exhibited a disjunctive geographical distribution pattern of "two zones and three centers." The spatial patterns of biological species result from contemporary climate, disturbance, and geological history (Simeone et al., 2016).Q. oxyphylla belongs to the Quercus Group Ilex, also known as Eurasian sclerophyllous oaks, whose distribution displays a more or less continuous range in Eurasia, with higher ecological and taxonomic diversity in the Himalayas and adjacent areas (Denk & Grimm, 2010;Zhou, 1992).On the one hand, Zhou's (1992) research on the origin, evolution, and diffusion of Quercus in China showed that the group of Q. oxyphylla may have originated in the northern section of the Hengduan Mountains and then spread eastward.The Qinling-Daba, Nanling, and Wuyi Mountains were important channels for the west-east migration of plants in Central China (Wang, 1992).
Q. oxyphylla migrated eastward along these two channels after its origin and formed a current pattern.On the other hand, the east-west by frequent crustal movements in the occurrence area (Denk & Grimm, 2009;Simeone et al., 2016).Therefore, we believed that the disjunctive distribution of Q. oxyphylla was possibly linked to the high tectonic activity of past and modern distribution ranges, which led to the absence of Q. oxyphylla in the mountainous (e.g., the Wuling Mountains and Wushan Mountains) and hilly areas among the three centers.Thus, the "two zones and three centers" pattern reflected geographic area disruptions, which might be closely related to its origins, migration, and geological history.to climate change, supporting the "climatic variability hypothesis" (Chan et al., 2016;Wang et al., 2023), which also implied that Q.

| Current suitable areas and key bioclimatic variables
oxyphylla might adapt to future climate change due to its wide climate niche.
The results of variable importance analysis verified that the top three bioclimatic variables were temperature factors: Bio7, Bio2, and Bio1, whereas the precipitation factors were relatively less important (Figure 2b).The results were similar to those of other species of Quercus in that temperature-related variables played a key role in controlling their potential biogeographical ranges, such as Q. fabri, Q. mongolica, and Q. phillyreodies (Sun et al., 2020).On the one hand, Q. oxyphylla mainly growed in steep mountainous or hilly areas with high rock exposure, indicating it was a drought-tolerant plant that did not have a high water requirement.In contrast, the leaves of this species were hard and leathery, the lower epidermis and twigs were densely covered with trichomes and had a stronger ability to resist drought than typical mesophytes.Therefore, precipitation was not a key factor restricting the distribution of Q. oxyphylla.

| Potential adaptability of Q. oxyphylla to future climate change
Different species have different mechanisms to adapt to climate change (Dudley et al., 2019).Over time, and with social development, the potentially suitable habitats of species may decrease, remain unchanged, or increase (Feng et al., 2023;Puchałka et al., 2021).Most native plants, including endangered and threatened species, show a trend of shrinking suitable areas, which is closely related to their biological characteristics and inability to shift to suitable habitats (Urban, 2015;Yesuf et al., 2021).Surprisingly, predictions regarding the potentially suitable habitats for Quercus mostly show an increasing trend due to their good tolerance, adapting to a variety of environmental conditions (Gimeno et al., 2009;Sun et al., 2020;Wang et al., 2023).Consistently, in this study, the potentially suitable areas of Q. oxyphylla were predicted to expand with intensifying climate warming (Figures 3, 4 and 6).This might be attributed to their strong adaptability to temperature and drought conditions.An increase in average temperatures would lead to extreme meteorological events, and the frequency and intensity of extreme weather events (e.g., high temperatures and droughts) in most land areas might continue to increase in the future (He, 2022).As a plant with a broad temperature niche and drought tolerance, Q. oxyphylla should be able occurrence and distribution pattern in China.The roman numerals are used for the vegetation region number.I. Boreal Coniferous Forest Region; II.Temperate Coniferous and Broad-leaved Mixed Forest Region; III.Warm Temperate Deciduous Broadleaved Forest Region; IV.Subtropical Evergreen Broad-leaved Forest Region, IV A .Eastern (Humid) Evergreen Broad-leaved Forest Subregion, IV Ai .Northern Subtropical Evergreen and Deciduous Broad-leaved Mixed Forest Zone, IV Aii .Mid-subtropical Evergreen Broad-leaved Forest Zone, IV Aiii .Southern Subtropical Monsoon Evergreen Broad-leaved Forest Zone, IV B .Western (Semi-humid) Evergreen Broad-leaved Forest Subregion; V. Tropical Monsoon Forest and Rainforest Region; VI.Temperate Grassland Region; VII.Temperate Desert Region; VIII.Qinghai-Tibet Plateau Alpine Vegetation Region.
24.5° N and 33.9° N and longitudes 104.7°E and 119.6°E, which showed three disjunctive geographical centers in the eastern subregion of subtropical evergreen broad-leaved forest region (IV A ): (1) Qinling-Daba Mountains center at the junction of Shaanxi, Gansu, and Sichuan provinces; (2) Wuyi Mountains center at the junction of Fujian and Zhejiang provinces; (3) Nanling Mountains center at the junction of Guangxi, Guangdong, and Hunan provinces (Figure 1).Under the current climate scenarios, the potentially suitable areas for Q. oxyphylla laid approximately between latitudes 24.0° N and 35.0°N and longitudes 102.2°E and 120.8°E, far F I G U R E 2 Evaluation results of different models (a) and importance of bioclimatic variables (b).larger than the current actual range (Figure3).The highly suitable areas exhibited a distinct discontinuous distribution pattern, concentrated in two zones: (1) North of the Yangtze River, covering the Qinling-Daba Mountains in the northern subtropical evergreen and deciduous broad-leaved forest zone (IV Aii ).(2) South of the Yangtze River, including the Nanling Mountains and Wuyi Mountains in the mid-subtropical evergreen broad-leaved forest zone (IV Ai ).Between these two regions, the Wuling Mountains, Xuefeng Mountains, and Wushan Mountains were primarily moderately suitable areas.The hilly areas surrounding the Sichuan South of the Yangtze River, Q. oxyphylla exhibited a significant northward expansion, accompanied by a northeastward shift, spreading from the Yuecheng Mountains to the Wuling Mountains, and from the Xuefeng Mountains to the Wushan Mountains.Northward expansion was also observed in the Wuyi Mountains, reaching the Mulianjiu Mountains and northeastward.In contrast, north of the Yangtze River, the Qinling-Daba Mountains witnessed a less pronounced expansion, with the suitable area shrinking over time and with intensifying global warming.This contraction in the Qinling-Daba Mountains contributed to the overall loss of suitable habitat for Q. oxyphylla, which was particularly notable in the southern and eastern parts of the mountain range, as well as along the southern edge of the Nanling Mountains-Wuyi Mountains.In terms of centroid migration (Figure 7), the centroid of Q. oxyphylla was located in Changde, Hunan province.Under the RCP2.6 scenario, by the 2050s, the centroid was projected to shift 35 km west, and by the 2070s, it was expected to move 135 km southeast to Yiyang, Hunan province.In the extreme climate scenario (RCP8.5), the centroid was anticipated to first migrate 135 km southwest to Jishou by the 2050s, and then relocate 150 km east to Changsha by the 2070s.Overall, under future climate scenarios, Q. oxyphylla primarily exhibited an eastward migration trend.F I G U R E 3 Potential suitable areas of Q. oxyphylla under the current climate condition.Red, green, and blue indicate high, medium, and low suitable areas, respectively.The roman numerals are the same to Figure 1.
Qinling-Daba Mountains and the Nanling Mountains were pushed by the Indian Ocean plate in the Eocene, uplifted by remote effects, and stopped forming in the Oligocene(Xu et al., 2016).The north-south trending Wuling Mountains and Xuefeng Mountains experienced intense uplift movements during the Middle and Late Cenozoic.Recent molecular phylogenetic studies have suggested that the split between the evergreen Groups Cyclobalanopsis and Ilex occurred during the Eocene/Oligocene and that the distribution was constantly affected F I G U R E 4 Projected suitable area for Q. oxyphylla by 2050s and 2070s under the RCP 2.6 (a, b) and RCP 8.5 (c, d) climate scenarios.The roman numerals are the same to Figure 1.
Under current climatic conditions, the distribution areas of Q. oxyphylla is mainly at 200-2900 m altitude between 24° N and 35° N in China (Figure1).These areas have moderate temperatures and sufficient rainfall in the summer and belong to eastern region of the subtropical evergreen broad-leaved forest region (Figure1).Based on the extraction results of climatic factors for each distribution point, in its actual occupied distribution space, Bio1 ranging from 7.8 to 20.8°C, Bio12 from 507 to 1901 mm, Bio7 from 27 to 34°C, and Bio2 from 7.5 to 10.4°C, indicating that Q. oxyphylla had a wide climate niche.The large annual temperature difference and wide altitudinal distribution suggested that the habitat of Q. oxyphylla had large climate variability and has strong physiological tolerance

F
Projected suitable area changes for Q. oxyphylla by 2050s and 2070s under the RCP 2.6 and RCP 8.5 climate scenarios.(a) indicates the percentage of different suitable grid cells out of the total predicted grid cells.(b) indicates the percentage of grid cell changes relative to current climate suitable area.to adapt to future climatic conditions in eastern area of Chinese subtropics.Simultaneously, results revealed exhibits distinct adaptability in different geographical regions and climate conditions.Under the backdrop of a warming climate, the high suitable area for the species would increase (Figures4 and 5a), indicating warming might be conducive to the reproduction and survival of the species.Additionally, there were significant regional differences in the suitable habitat of Q. oxyphylla (Figures4 and 6).In the Qinling-Daba Mountain region north of the Yangtze River, the suitable habitat area gradually shrinks under climate change, which indicated that climate change could lead to be more vulnerable to threats of the species in the region.Conversely, the region from the Nanling-Wuyi Mountains region south of the Yangtze River would experience a significant northward expansion.This suggested that the mid-subtropical evergreen broad-leaved forest zone (IV Aii ) was a stable habitat for Q. oxyphylla, and future conservation efforts could focus on this area.These findings highlighted the complex and region-specific impacts of climate change on species distribution, emphasizing the vulnerability of certain regions (Qinling-Daba mountains) to habitat loss under future climate scenarios (Figure6).Furthermore, based on our research findings, developing species-specific conservation action plans is critical for sustainable conservation of Q. oxyphylla.Moreover, to mitigate the risk of local extinctions driven by climate change, a pressing task is to identify suitable areas and establish new populations with high genetic F I G U R E 6 Projected suitable area change for Q. oxyphylla by 2050s and 2070s under the RCP 2.6 (a, b) and RCP 8.5 (c, d) climate scenarios.F I G U R E 7 Shifts of centroids of Q. oxyphylla under different climate scenarios.diversity through translocation conservation efforts.Due to the differing responses of northern and southern populations to climate change, habitat conservation for Q. oxyphylla demands a focused approach incorporating sustainable local and regional policies, adaptive management strategies, and habitat restoration.Based on our investigations of modern distribution and field surveys, most Q. oxyphylla populations are found near villages outside protected areas, primarily in "Fengshui" forests or secondary forests.Thus, a priority lies in educating local communities about the importance of the Sharp-leaved Oak, its habitat, and the impacts of climate change.Furthermore, given the valuable timber resources and broad climatic adaptability of the species, afforestation efforts in climatically suitable regions can contribute positively to the sustainable use of forests under changing climate.5 | CON CLUS IONSAs one of the few evergreen oaks in subtropical regions, Q. oxyphylla exhibited a unique disjunctive distribution pattern of "two zones and three centers" and a wide altitudinal distribution, indicating its broad climatic niche and ecological tolerance.The climate-related variables (Bio7, Bio2, and Bio1) were the key climatic factors that determined the distribution pattern, which might also be related to the evolution and migration routes of Quercus since the Tethys Sea.As global climate warming intensifies, the distribution of Q. oxyphylla would exhibit regional sensitivity and vulnerability.The suitable area in north of the Yangtze River will gradually shrinking.Meanwhile, the suitable area in the south of the Yangtze River would expand northward on a large scale and show an eastward migration trend.These findings provided a scientific basis for developing effective conservation strategies and sustainable resource utilization of the species under climate change.Nonetheless, this study only focused on the species' climatic niche rather than its actual niche.Other factors like species interactions, geographic barriers, and disturbances, might also influence the species' distribution within suitable areas.Consequently, further researches would be needed into the ecological mechanisms of Q. oxyphylla population dynamics and habitat changes to provide a scientific foundation for adaptive management.
Descriptive statistics of 19 bioclimatic variables for 63 occurrence sites of Q. oxyphylla.
Note: Bold indicates variables used for modeling.