Simulation of the potential distribution of rare and endangered Satyrium species in China under climate change

Abstract Satyrium is an endangered and rare genus of plant that has various pharmacodynamic functions. In this study, optimized MaxEnt models were used in analyzing potential geographical distributions under current and future climatic conditions (the 2050s and 2070s) and dominant environmental variables influencing their geographic distribution. The results provided reference for implementation of long‐term conservation and management approaches for the species. The results showed that the area of the total suitable habitat for Satyrium ciliatum (S. ciliatum) in China is 32.51 × 104 km2, the total suitable habitat area for Satyrium nepalense (S. nepalense) in China is 61.76 × 104 km2, and the area of the total suitable habitat for Satyrium yunnanense (S. yunnanense) in China is 89.73 × 104 km2 under current climatic conditions. The potential suitable habitat of Satyrium is mainly distributed in Southwest China. The major environmental variables influencing the geographical distribution of S. ciliatum were isothermality (bio3), temperature seasonality (bio4), and mean temperature of coldest quarter (bio11). Environmental variables such as isothermality (bio3), temperature seasonality (bio4), and precipitation of coldest quarter (bio19) affected the geographical distribution of S. nepalense; and environmental variables such as isothermality (bio3), temperature seasonality (bio4), and lower temperature of coldest month (bio6) affected the geographical distribution of S. yunnanense. The distribution range of Satyrium was extended as global warming increased, showing emissions of greenhouse gases with lower concentration (SSP1‐2.6) and higher concentration (SSP5‐8.5). According to the study, the distribution of suitable habitat will shift with a change to higher elevation areas and higher latitude areas in the future.


| INTRODUC TI ON
Global warming is one of the crucial environmental problems the world faces today (Bayer et al., 2021). The global temperature has risen by about 1°C in the past century, especially in the past 30 years, causing some plant species to move to higher elevation and higher latitude areas, as the Fifth Assessment Report (AR5) of the United Nations Intergovernmental Panel on Climate Change reported (Allen et al., 2014). Climate change in the future will affect the distribution range of species, resulting in a loss of biodiversity and the extinction of endangered species (Bellard et al., 2012). Under climate change conditions, the prediction of suitable habitats of species will be considered important in the future (Kumar et al., 2021).
In research on the geographical distribution range of plants affected by climate change, the species distribution model (SDM) uses distribution data of species and environmental variables in species habitats to predict the realized niche of species (Elith & Leathwick, 2009). These data are combined with environmental data in different periods for simulating the potential distribution areas of species in those periods (Araújo & Peterson, 2012). Among the many algorithms used in modeling species distribution, the maximum entropy approach (MaxEnt) has demonstrated fast modeling, wide use, high accuracy, and stability even with small sample sizes (Merow et al., 2013;Pearson et al., 2007). So it has been the most commonly used species distribution model (Phillips et al., 2006). If the MaxEnt model and ArcGis predict potential species distributions and biodiversity risks, relative strategies can be established to reduce climate change's negative influence on global biodiversity. Satyrium (Orchidaceae) is a rare and endangered plant genus with about 92 species-mostly found in Africa, with five species found in Madagascar alone, and four species found in Asian countries (Mahendran & Bai, 2009). Only three species have been found in China alone (all of which are endemic in China) (Cun, 2005). They are mostly distributed throughout Southwest China. Satyrium is also used for traditional herbal medicinal purposes; for example, Greek medics used Satyrium tubers as aphrodisiacs (Teoh, 2016). Traditional health-care centers in India use the tubers of Satyrium nepalense (S. nepalense) to make energy tonics and cure different types of fevers (Mishra et al., 2018). The number of Satyrium resources has been rapidly diminishing. Its population has a fragmented distribution, and it is in a rare and endangered status because of the value of herbal medicine and the serious deterioration of the ecological environment in recent times (Mahendran & Bai, 2009). Changes in climate will result in changes in the biological phenology period. This in turn will result in changes in the geographical distribution of species and an acceleration in the rate of species extinction. Therefore, it is necessary to adequately understand the changing trends of the geographical distribution of species under climate change conditions and to develop relative protection strategies.  (Figure 1).
It is necessary to analyze environmental variables before they can be used for niche simulation calculations to prevent the multicollinearity of the variables from causing overfitting of the model, because many bioclimatic variables are spatially related (Graham, 2003

| Optimization of model parameters and model building
The feature combination (FC) and regularization multiplier (RM) were adjusted by the ENMeval package in R 4.0.2 software . There are five FCs in the MaxEnt model: linear, quadratic, product, threshold, and hinge (Phillips et al., 2006 (Fielding & Bell, 1997

| Classification of suitable habitat
We imported the MaxEnt model outputs into ArcGIS to convert them into raster data using the conversion tool. Additionally, suitable Satyrium species habitats were determined by a reclassification tool  and placed into four groups: highly suitable habitat (0.60-1.00), moderately suitable habitat (

| Core distributional shifts
Under current and future climate circumstances, the SDM toolbox was used to calculate shifting trends in the area of suitable habitat based on Python, as well as changes in the centroid of the area of suitable habitat (Etherington, 2011). The toolbox is a Python-based GIS application (Brown et al., 2017). Using the toolbox, the researchers calculated shifting trends in areas of suitable habitats and compared the centroids of the regions between current and future climate conditions. The study provides information about the core shifts and distributions of S. ciliatum, S. nepalense, and S. yunnanense.
These species' distributional changes were reduced to a single centroid (central) point, and a vector file was constructed to show the amount and direction of expected change over time. Finally, we examined how the centroid changed with different SDMs to determine whether there were any distribution shifts. When RM = 0.5, FC = LQH, delta.AICc = 0, the model is optimal (Table S1)

| Important environmental variables
The jackknife test results indicated that the top three environmental variables with the largest effect on regularized training gains were as follows: isothermality (bio3), temperature seasonality (bio4), and mean temperature of coldest quarter (bio11) for S. ciliatum; isothermality (bio3), temperature seasonality (bio4), and precipitation of the coldest quarter (bio19) for S. nepalense; isothermality (bio3), temperature seasonality (bio4), and lower temperature of coldest month (bio6) for S. yunnanense when modeling with a single environmental variable ( Figure S1). The contribution rate of environmental variables could also be used to evaluate each environment's importance. The contribution rate of each variable was combined with the distribution of Satyrium species (Table 1)

| Future changes in suitable habitat areas in highly suitable centroid distributions
In regard to future changes in suitable habitat areas, both the gained and lost areas of Satyrium will increase in emission concentrations; however, the gained area will be larger than the lost area ( Figure 6).
According to SSP5-8.5, the number of increased and decreased areas were the highest compared with the current simulation results.
For S. ciliatum, by the 2050s, the area of total suitable habitat would increase to 9.65 × 10 4 km 2 (SSP1-2.6) and 12.63 × 10 4 km 2 (SSP5-8.5), and the gained territory would be in Southern Tibet and Northwest Sichuan. By the 2070s, the areas would increase to 9.56 × 10 4 km 2 (SSP1-2.6) and 26.33 × 10 4 km 2 (SSP5-8.5), and the new territory would be in Southern Tibet, Northwest Sichuan, and Northern Yunnan. The decreased habitat would mainly be distributed in Tibet.

| DISCUSS ION
The Satyrium plant species are important resources, but they are endangered and rare (Mishra & Saklani, 2012). MaxEnt models were applied to predict current and future potential distribution areas of S. ciliatum, S. nepalense, and S. yunnanense. According to the four different future scenarios, the suitable habitat area for Satyrium in China will be extended. These results are consistent with the results of a study on the Houttuynia cordata Thunb (Ceercao) habitat suitability; the potential suitable area in the high-emission scenario was greater than that in the low-emission scenario in the study . The increased precipitation in China in the SSP5-8.5 emission scenario was higher than in the low concentration emission scenario (Yue et al., 2021). This shows the increased precipitation in the high concentration emission scenario can solve the limitations of precipitation on distribution of species, and the increased precipitation in the low concentration emission scenario cannot reduce or solve the limitations of precipitation on the distribution of species. This may also be the reason the most suitable habitat area and the increasing area of the Satyrium species are the largest in the high concentration emission scenario, particularly the SSP5-8.5 scenario. The prediction shows that the upcoming future will be the most suitable for the growth of the Satyrium species. It shows that, generally, the suitable habitat is stable and the species will not face extinction due to climate change. This study shows the new areas for species distribution. The findings are in agreement with those of previous studies, indicating that in some regions, habitat suitability of plant species improves under climate change conditions (Feng et al., 2020;Li et al., 2019). The distribution of species under future climatic scenarios predicts that the area of species covering 20% of the earth's surface will face the risk of extinction, and about 15%-37% of the species will be endangered (Thomas et al., 2004). One of the reasons for the endangered status of Satyrium is that most Satyrium species are hermaphroditic. Some species rely on birds and other animals for long-distance transmission, the species population is small, and the quality of seed development is low because of long-distance pollination (Johnson et al., 2011). The suitable habitat area for Satyrium in China will move to higher altitude areas in the northwest side of China in the 2050s and 2070s, where the humidity and habitat are suitable for the growth of the Satyrium species. The centroid of the distribution of endangered plant Semiliquidambar cathayensis (Hamamelidaceae) will move northward (Ye et al., 2020). The distribution of suitable habitat of Paeonia delavayi will shift to higher elevations . With the global warming, more plants tend to migrate to high altitude and high latitude areas (Sekercioglu et al., 2008 (Deblauwe et al., 2016;Wang et al., 2021). In regard to S. nepalense, precipitation is one of the environment variables influencing its geographic distribution. Satyrium is a perennial terrestrial orchid distributed in Southwest China (Deng et al., 2019). Terrestrial orchids are typically cold tolerant, live in low-temperature environments, and are highly distributed in areas with large amounts of annual rainfall (Phillips et al., 2011;Poff et al., 2016). More studies have shown that the dominant variables restricting geographical distribution of plants are energy supply, plenty of water, and cold tolerance (Zhou & Wang, 2000). Based on the prediction of the MaxEnt model, the area of the current potential suitable habitat of Satyrium is mainly located in Southwest China, which has a typical subtropical monsoon climate (Luo et al., 2021). Previous studies have indicated that precipitation is the main variable affecting plant growth, regeneration, nutrient cycles, and community productivity in different habitats (Miranda et al., 2011). Climate change can determine geographical distribution of species, and geographical distribution of species can respond to changes in climate (Warren et al., 2021). Global warming has changed the structures of terrestrial ecosystems, which have changed the habitats and geographical distribution functions of species in turn (Ghini et al., 2012). For instance, climate, topography, soil, human disturbance, and spatial constraints are significant to the distribution of many different spatial scales (Eiserhardt et al., 2011;Parisien & Moritz, 2009). The environmental data used in the study included bioclimatic variables and topographic variables in WorldClim. We did not study soil type, land use, human activities, biological interaction, and other factors that influence the distribution of Satyrium. The more factors that are included, the more accurate the prediction will be. Therefore, other factors affecting the potential suitable habitat of Satyrium should be studied in the future, with different niche prediction models predicting areas of potential suitable habitats. This will make the prediction of results more reliable.
Generally, the evaluation index of prediction model accuracy is the ROC curve, which has the following advantage: the AUC value is not affected by the threshold and can be used for the comparison of various models (Fielding & Bell, 1997). However, further research shows that it is inaccurate to evaluate accuracy only by AUC value because it cannot reflect wrong spatial distribution information, and calculate the omission rate and error rate by a single method (Lobo et al., 2008;Peterson et al., 2008). Nowadays, the information measurement index AUC is used to select the ideal setting parameters of MaxEnt (Muscarella et al., 2014); the method of combining omission rate and AUC value is used to obtain the best parameters (Radosavljevic & Anderson, 2014); TSS value is used to evaluate accuracy (Bedia et al., 2011); and multiple parameters are set to select the optimal result (Elith et al., 2010;Warren & Seifert, 2011). TSS can accurately correct the overall accuracy of the model without being affected by the size of the species validation data set and without depending on the model threshold, therefore, AUC and TSS values are used as evaluation indexes in this prediction (Bedia et al., 2011).

| CON CLUS ION
In the study, the potential suitable Satyrium habitat in China was predicted based on the MaxEnt model optimized by the ENMeval package. Temperature was the common variable in the three species. In regard to S. nepalense, precipitation is one of environment variables influencing its geographic distribution. The potential geographical distribution of Satyrium under current climatic conditions is primarily on the southwest side of China. Satyrium's highly suitable habitat area, moderately suitable habitat area, and total suitable habitat area were predicted to increase under future climate change scenarios. The highly suitable habitat area of Satyrium showed a tendency to shift to areas with higher elevation. In regard to future changes in suitable habitat areas, both the gained and lost areas of Satyrium will increase in emission concentrations; however, the gained area will be larger than the lost area. These results can provide theoretical guidance for protecting endangered plants in China and offer further reference for biodiversity protection in China.

This study was supported by the Key Research and Development
Program of Zhejiang Province (grant number 2019C02024).

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest for the publication of this study.

This article has earned Open Data, Open Materials and Preregistered
Research Design badges. Data, materials and the preregistered design and analysis plan are available at Dryad Digital Repository (https://doi.org/10.5061/dryad.k98sf 7m6v).

DATA AVA I L A B I L I T Y S TAT E M E N T
All authors agreed to deposit data from this manuscript in a public repository. Data were submitted to Dryad (the DOI number is https://doi.org/10.5061/dryad.k98sf7m6v).