Elymus dahuricus Turcz (E.dahuricusis) is a perennial forage plant of the genus E.dahuricus of the family Gramineae, mostly born in hillside meadows or roadsides, with drought-resistant, cold-resistant, alkali-resistant, sand-resistant growth characteristics, reproduced by seeding. With strong growth force and adaptability, it is usually distributed in Tibet, Hebei, Qinghai, Sichuan, Shaanxi, Gansu, Xinjiang, Inner Mongolia and other regions of China, often used as the main native grass species in ecological restoration projects in China’s plateau areas, and plays an essential role in grassland restoration, grassland improvement, and the establishment of artificial grassland1. In recent years, E.dahuricus has been widely excavated and transplanted due to its wide range of potential applications. However, this has led to the destruction of plant growth patterns and serious depletion of wild germplasm resources. Currently, most of the research on E.dahuricus focuses on propagation and cultivation techniques, however, there is a lack of research on the distribution of its existing germplasm resources and the prediction of its suitable habitats under future climate change scenarios.Therefore, in order to clarify breeding resources, precise distributional surveys of E.dahuricus are necessary to ensure the long-term sustainability of the species and breeding programs2.

Climate is one of the most important environmental factors affecting the geographical distribution of species. Some studies have found that the plateau region has experienced a greater warming compared with other parts of China. Temperature change will directly or indirectly affect the original spatial distribution pattern of species and its related ecological factors, thus changing the distribution area, range and number of species, which will have a greater impact on the structure and function of the original ecosystem3. Due to the interference of climatic factors, the problem of grassland degradation in the plateau area of China is very serious, the productivity of grassland decreases, and the imbalance between grass and livestock is aggravated. For this reason, plants will change their phenology or growth habit accordingly, and select new adaptation areas through range shifts. Up to now, exploring the impacts of climate change on the potential habitats of species has been widely used in the conservation of plant resources and the introduction and cultivation of economic plants, and has become a hot direction in the study of the impacts of global change4. Therefore, predicting suitable habitats for E.dahuricus under climate change scenarios will provide strong theoretical support for the future development and conservation of E.dahuricus germplasm resources.

Species Distribution Models (SDMs), mathematical models that predict the potential geographic distribution of species based on currently known species distribution data and corresponding environmental characteristics, including the Genetic Algorithm Model based on Rule Set Prediction (GARP), Ecological Niche Factor Analysis Model (ENFA), and Maximum Entropy(MaxEnt) Model, have become important tools for exploring species distribution patterns by using environmental information collected from species’ range sites to assess the survival needs of species and then projecting these data onto selected study areas5. The most widely used SDM is MaxEnt, and its reliability in predicting species distributions has been demonstrated in many studies. In addition, MaxEnt has the advantages of lower requirements, higher predictive accuracy and ease of use.In this study, we conducted a comprehensive survey of E.dahuricus in China, selected 180 sites of E.dahuricus occurrence and 38 environmental variables, and predicted the suitable areas for E.dahuricus under future climate scenarios by using the optimized Maxent model and ArcGIS V10.8 software.Our results provide wild resource information and theoretical references for the conservation and rational utilization of E.dahuricus6.

Materials and methods

Species occurrence records and processing

Specimens of E.dahuricus used in this research were obtained from the Digital Herbarium of China (CVH, http://www.cvh.org.cn/), the National Herbarium Resource Centre (NSII, http://www.nsii.org.cn/), the Global Biodiversity Information Facility (GBIF, https://www.gbif.org/), and related literature7.For species distribution points that lack latitude and longitude but have specific locations, the longitude and latitude coordinates of species points are determined by using Baidu coordinate picking system (https://api.map.baidu.com/lbsapi/getpoint/)8. In order to prevent duplicate data within the same range from affecting the accuracy of model prediction, we used the spThin package in R environment to delete a single duplicate data within the highest resolution based on the highest resolution of world climate (http://www.worldclim.org) (2.5arcmin, about 5 km) to ensure that only one distributed data exists within every 5 km9. Finally, 180 distribution record points were finally obtained, and the effective distribution locations are shown in Fig. 1. A schematic representation of the methodology was depicted in Fig. 2.

Environmental data acquisition and processing

Thirty-eight environmental variables reflecting the ecological niches of the species were used in this study, including 19 bioclimatic variables, 15 soil variables, and 3 topographic factors. The bioclimatic variables were obtained from WorldClim version 2.1 (http://www.worldclim.org) and and the spatial resolution was 2.5 arc minutes10,11. The soil variables were obtained from World Soil Database of the Food and Agriculture Organization of the United Nations (http://www.fao.org) and the topographic data were obtained from the Geospatial Data Cloud (http://www.gscloud.cn/)12.

The future climate data are based on the medium-resolution climate system model (BCC-CSM2-MR) of the National (Beijing) Climate Center Climate System Model (NCCCM) in the Coupled Model Intercomparison Project6 (CMIP6), and the BCC-CSM2-MR is more accurate than the other models in CMIP6. Compared with other models in CMIP6, BCC-CSM2-MR has better estimates of temperature, precipitation and atmospheric circulation in China and more reliable simulation results, so we chose this model to participate in the modeling, and it contains four different shared socioeconomic pathways (SSPs)13. In this research, three typical pathways were selected to predict the potential future distribution of E.dahuricus, namely, the global sustainable development pathway SSP126, the medium sustainable development pathway SSP245, and the unbalanced development pathway and conventional development pathway SSP58514.

The effect of environmental variables on the geographic distribution of E.dahuricus was quantitatively assessed based on the 38 environmental factors mentioned above by using the jacknife technique to obtain the significance of the variables, and by using pearson and variance inflation factor (VIF) to test the importance of the correlation and significance of the environmental variables.Under the R environment, Spearman correlation analysis and multiple covariance VIF variance expansion factor analysis were performed to initially screen out environmental variables with correlation less than 0.7 and VIF variance expansion factor value less than 515.

Finally, 12 environmental variables that are more important to the geographical distribution of E.dahuricus were screened out from the 38 environmental variables for modeling, namely, temperature seasonality(bio4), max temperature of warmest month(bio5), min temperature of coldest month(bio6), precipitation of driest month(bio14), precipitation seasonality(bio15), precipitation of driest quarter(bio17), altitude(elev), basic saturation(t_bs), cation exchange capacity of topsoil(t_cec_soil), electrical conductivity of topsoi(t_ece), gravel volume in topsoil(t_gravel), organic carbon content in topsoi(t_oc).

Optimization of the model

In order to avoid overfitting due to the high complexity of models constructed with default parameters, which could result in the predicted distribution of potential habitat for E.dahuricus deviating too much from reality, this study used the ENMeval data package in R 4.1.316.

The Maxent model regularization level consists of two parameters, the regularization multiplier (RM) and the feature combination(FC) optimized by calling the ENMeval package in the R environment.The Maxent model provides five features: linear (L), quadratic(Q), hinge (H), product (P), and threshold (T). In this research, the Maxent software tuned the parameters RM = 1.5, FC = H; in order to optimize the Maxent model, RM was set to 0.5–4. For each 0.5 increase, a total of eight control frequencies were doubled17. At the same time, 48 parameter combinations were calculated using six combinations with one or more features: L; L and Q; H; L, Q and H; L, Q, H and P; and L, Q, H, P and T in permutations. the ENMeval package uses the 48 parameter combinations described above to test the model’s complexity based on the delta AICc value and the 10% test missingness rate, where the lower the value, the more accurately the model predicts18.

Classification of the habitat of E.dahuricus and size statistics

First, the prediction results of Maxent model were transformed into raster data using ArcGIS V10.8 software. The natural breakpoint method was used to categorize the suitable area into four classes: unsuitable area (0 ~ 0.38), low suitable area (0.38 ~ 0.47), middle suitable area (0.47 ~ 0.65) and highly suitable area (0.65 ~ 0.1). On this basis, we used ArcGIS for visualization and a raster calculator to work out the area of each part of the suitable area under different current and future climate scenarios for comparison and analysis19.

Using the SDMTool package in the R-environment, the center of mass position of suitable habitat for E.dahuricus was calculated for the current and future periods, and the direction of range shift of suitable habitat for E.dahuricus was reflected by the change in centroid position. Geosphere data packet in R language for calculating the centroid migration distance of E.dahuricus under different climate scenarios20.

Construction of the habitat suitability model

In this research, the Maxent model was used to predict the habitat suitability of E.dahuricus in three periods. The processed species distribution point data and environmental variables data were imported into the MaxEnt model to initially establish its species distribution model. The model was constructed by randomly selecting 75% of the sample points as the training dataset, and the remaining 25% as the test dataset to validate the model, with 10 replications21. The weights of the environmental factors were evaluated by Jackknife, and the dominant environmental variables was determined by the contribution rate of each environmental variables22,23,24.

In order to accurately evaluate the reliability of the model, the contribution rate of the environmental variables was analyzed by the Jackknife, and the accuracy of the constructed model was evaluated by using the area under curve (AUC) value of the receiver operating characteristic curve (ROC).The AUC was evaluated between 0.5 ~ 1.0, with AUC < 0.6 indicating a failed prediction; values between 0.6 and 0.7 indicating poor performance; values between 0.7 and 0.8 indicating fair performance; values between 0.8 and 0.9 indicating good performance; and values between 0.9 and 1.0 indicating excellent performance25.

Changes in Spatial pattern of suitable distribution area for E.dahuricus

Spatial units with a probability of species presence ≥ 0.38 were categorized as suitable areas, while areas with a probability of species presence < 0.38 were categorized as unsuitable areas. The potential geographic distribution data of E.dahuricus under current and future climate change scenarios were binarized using the reclassification tool in the ArcGIS 10.8 software and a presence/absence (0,1) matrix for the potential geographic distribution of E.dahuricus was created. Suitable areas are indicated by the value “1” indicating the presence of the species, while unsuitable areas are indicated by the number “0” indicating the absence of the species. Based on this matrix, the changes in the spatial pattern of rose suitability zones under current and future climate scenarios were further analyzed, and three types of changes in the suitability areas were identified: Gain suitable areas, Loss suitable areas, and stable suitable areas. Changes in the spatial pattern of potentially suitable areas under current and future climate change are defined as follows: matrix values 0 → 1 indicate gain suitable areas, 1 → 0 indicates loss suitable areas, and 1 → 1 indicates stable suitable areas26.

Fig. 1
figure 1

Geographical map of the distribution point of E.dahuricus using Arcmap 10.8.0 software (URL: https://www.arcgis.com/home/index.html).

Full size image
Fig. 2
figure 2

Processing methodology in flow diagram.

Full size image

Results

Model optimization and accuracy evaluation

The complexity of MaxEnt has a great impact on the accuracy of predicting species’ suitable areas in the model, and optimization of model parameters can effectively avoid overfitting and significantly improve the prediction accuracy of the model. FC and RM are two key parameters of MaxEnt. In this research, these two key parameters were randomly combined and tested using the ENMeval package based on R software to filter the optimal parameter pairing. Therefore, we chose the parameter settings of RM = 1.5 and FC = LQH for modeling. The extent and complexity of model fit to species distributions was assessed using the lowest increment from the Akaike information criterion (Delta.AICc), the mean difference between the training and test AUC (Avg.diff.auc), and a 10% training omission rate (Mean OR10), and the model’s fit to species distributions was assessed when Delta.AICc was 0 and both Avg.diff.auc and Mean OR10 are both small the optimal parameter combination can be filtered27. The AUC value of simulated training with these parameters is 0.850 (Fig. 3), indicating accurate prediction.

Fig. 3
figure 3

ROC response curve under the Maxent model.

Full size image

Potential geographic distribution of E.dahuricus in China

The current distribution area of E.dahuricus was categorized using ArcGIS software to obtain the distribution map under the current climate conditions, and the distribution area is shown in Fig. 4. It is mainly concentrated in the northwest and central regions of China, including Xinjiang, Xizang, Gansu, Qinghai, Ningxia, Inner Mongolia, Shanxi, Hebei, Beijing, Liaoning, Chongqing and other provinces. Among them, the highly suitable areas are mainly located in Xinjiang, Sichuan, Xizang and Ningxia. The simulation and prediction results were basically consistent with the geographic distribution data of E.dahuricus, which further showed that the simulation and prediction results had a certain degree of accuracy. Based on the number of cells occupied by suitable areas, the area occupied by suitable areas can be calculated. As can be seen from Table 1, the suitable area under the current climate conditions is 223.63 million km2, accounting for 23% of the total land area of China, of which the highly suitable area is 395.55 million km2, accounting for 4% of the total land area of China28.

Fig. 4
figure 4

Suitable growing area of E.dahuricus in China under different climate using Arcmap 10.8.0. Software (URL: https://www.arcgis.com/home/index.html).

Full size image
Table 1 Appropriate growing areas of E.dahuricus in China under different Climatic conditions.
Full size table

The influence of main environmental variables on the habitat of E.dahuricus

In order to further elucidate the climatic characteristics of potentially suitable areas for E.dahuricus under the current climatic conditions, the response curves of six environmental factors that have a significant impact on the geographic distribution of E.dahuricus were further investigated29.The 12 environmental variables in the current period were evaluated using the jackknife method in the MaxEnt model, and the percentage contribution (PC) and percentage importance (PI) were calculated for each environmental variable. Environmental variables with greater than 5% influence on the potential geographic distribution of E.dahuricus were bio6, bio15, bio17, t_cec_soil and elev. The overall PC values of the five environmental variables was 83.4%. The PI values greater than 5% were, in order, bio4, bio15,bio17, bio6 and elev (Table 2). At the same time, Fig. 5 shows that variables of regularization training gain (Fig. 5A), test gain(Fig. 5B), and the first two variables of AUC(Fig. 5C) are elev and bio6 suggesting that two variables are more influential than the others.The comprehensive evaluation indicated that the dominant environmental variables affecting the distribution of existing potentially suitable habitat for E.dahuricus are bio17,bio6, elev, bio4, t_cec_soil and bio15.

Fig. 5
figure 5

Results of jackknife evaluations of the predictor variables, using regularized training gain (A), testing gain (B), and AUC on the test data (C) in the MaxEnt model for E.dahuricus.

Full size image
Table 2 Contribution and importance of environmental variables.
Full size table

Prediction of suitable growing area of E.dahuricus under future climate scenarios

Compared with the predicted results of current climate conditions, under future climate conditions (Table 1), the potential distribution area of E.dahuricus in the country shows an increasing trend, which is reflected by the increase in the area of high suitability area; the area of medium suitability area did not change greatly; the area of low suitability area increased greatly in the SSP585-2090 s, and the change was less obvious in other periods, but it was higher than that of the total area of suitability area under the current climate scenario30. The total area of suitable zones further supports the conclusion that temperature has a strong influence on the potential distribution of E.dahuricus.

When compared to current climate conditions, potential distribution area of the E.dahuricus in the future climate scenario changes less, the most obvious change is in the SSP585-2090 s scenario, with a total change area of 39 × 104 km2, and the least obvious change is in the SSP245-2050 s period, with a total change area of only 0.29 × 104 km2 (Table 1). Specifically, the future expansion areas of E.dahuricus are mainly distributed in the edges of the original suitable areas, including the southeastern part of Gansu Province, the northern part of Yunnan Province, the western part of Qinghai Province bordering the Tibet Autonomous Region and Xinjiang Uygur Autonomous Region, and the southern part of Heilongjiang Province, and the new expansion areas are mostly low and high suitable areas. Stabilized areas are mainly distributed in the northeastern region of China, including the northeastern edge of the Inner Mongolia Autonomous Region, the northwestern part of Heilongjiang Province, and the southwestern part of the Tibet Autonomous Region, with the most obvious contraction of suitable areas in the period of SSP126-2090 s, and the contracted areas covering an area of 8.93 × 104 km2. Qinghai Province, most of the Tibet Autonomous Region, and the northeastern part of the Inner Mongolia Autonomous Region31.

Figure 6 shows the distribution of E.dahuricus in China under future climate scenarios. Potentially suitable areas for E.dahuricus in China are mainly located in Sichuan, Gansu, Ningxia and Xinjiang provinces, and the total suitable area for E.dahuricus will increase under the future climate scenario32.Under the SSP126 scenario, the total suitable planting area in the 2050 s and 2090 s is about 212.33 × 104 km2 and 202.19 × 104 km2, respectively, which is 5.05% and 9.59% less than the current period.Under the SSP245 scenario, the total suitable area for E.dahuricus in the 2050 s and 2090 s was about 222.34 × 104 km2 and 243.18 × 104 km2, respectively, a decrease of 0.57% and an increase of 8.74%, respectively, from the current period.Under the SSP585 scenario, the suitable area for E.dahuricus in the 2050 s and 2090 s is about 243.52 × 104 km2 and 262.12 × 104 km2, respectively, which is a decrease of 8.89% and a decrease of 17.21%, respectively, from the current period. In the future climatological scenarios, the unsuitable area will increase compared to the current period, and the low, medium, high, and total suitable areas will increase compared to the current period.In the 2090 s, the total suitable area is the largest in the SSP585 scenario and the smallest in the SSP126 scenario33.

Fig. 6
figure 6

Distribution of suitable habitats for E.dahuricus in China under different future climate change scenarios. using Arcmap 10.8.0 software (URL: https://www.arcgis.com/home/index.html).

Full size image

Prediction of suitable growing area for E.dahuricus in China at different time periods

Figure 7 shows the analysis of changes in the spatial pattern of suitable areas for E.dahuricus under several future climate scenarios. The results show that the suitable area for E.dahuricus under different climate scenarios increases and decreases to varying degrees under future climate change (Table 3). A small portion of the current E.dahuricus suitable area was retained in the next three climate scenarios. The retained area was 223.74 ~ 236.80 × 104km2 and the retention rate was 88.11% ~ 93.26%. The increase area of E.dahuricus under the next three climate scenarios was 8.28 ~ 69.11 × 104 km2 with a growth rate of 3.26% ~ 27.22%. The increase area of E.dahuricus mainly occurs in the northeast of the current suitability zone. The loss area of E.dahuricus was 17.11 × 104 km2 ~ 30.17 × 104 km2 with a loss rate of 6.73% ~ 11.01%. The loss area of E.dahuricus mainly appeared in the southern part of E.dahuricus suitable area34. Comparative analysis of the changes in spatial patterns of potential habitats of E.dahuricus under different future climate change scenarios showed that the suitable habitats of E.dahuricus changed with climate change, and its response to climate change was generally consistent. The changes in spatial patterns of potential habitats under different future climate change scenarios showed that the rate of change in the spatial area of potential habitat area of E.dahuricus was greater in the ssp126 −2090 than in the 2050 s, indicating that the changes in spatial patterns of potential habitats of E.dahuricus were more pronounced in the 209035.

Fig. 7
figure 7

Pattern of spatial changes of E.dahuricus in China under different future climate variability scenarios. using Arcmap 10.8.0 software (URL: https://www.arcgis.com/home/index.html).

Full size image
Table 3 Variation in the distribution area of E.dahuricus during different periods of time.
Full size table

Changes in the centroid of suitable areas for E.dahuricus under different climate change scenarios

To further explore the response of E.dahuricus to future climate, this study analyzed the displacement of the centroid of potentially suitable habitat for E.dahuricus under different future climate scenarios. The centroid of suitable habitat for E.dahuricus generally tended to move to higher latitudes36. Specifically, the geographic center of potential habitat is currently located in Haibei Tibetan Autonomous Prefecture. In the 2050 s, the centers of E.dahuricus suitable areas under the SSP126 and SSP245 climate scenarios are located in the Haibei Tibetan Autonomous Prefecture and the Haixi Mongolian and Tibetan Autonomous Prefecture. Mongolian and Tibetan Autonomous Prefecture; the center of E.dahuricus suitable area under the SSP585 climate scenario was located in Haibei Tibetan Autonomous Prefecture.2090 s, the center of E.dahuricus suitable area under the SSP126 and SSP585 climate scenarios was located in Haibei Tibetan Autonomous Prefecture. E.dahuricus under SSP126 and SSP585 climate scenarios are centered in Haibei Tibetan Autonomous Prefecture, and the center of E.dahuricus under SSP245 climate scenario is located in Zhangye, Gansu Province.(Fig. 8).

Fig. 8
figure 8

Changes in the geographic distribution of the centroid of suitable areas for E.dahuricus under different climate scenarios. using Arcmap 10.8.0 software (URL: https://www.arcgis.com/home/index.html) [(A: original image; B: enlarged version).].

Full size image

Discussion

This study is based on climate, topography and soil factors and applies the ENMeval data package to optimize the model. This method restricts the background data to the area corresponding to the calibrated position so that the potential geographic distribution area modeled by MaxEnt covers the current distribution points. This approach allows the model parameters to be adjusted to improve the performance of the MaxEnt model, followed by tuning trials by varying the level of regularization, which reduces the complexity of the model, and finally by improving the fit of the predictions to the actual distribution area and by measuring its accuracy through visual inspection of the geographic prediction maps37.The optimized MaxEnt model can effectively reduce the complexity of the model and improve the fit of the prediction results to the actual situation, and predicts the distribution of the species better, and the response curve is obviously smoothed and close to the normal distribution curve, which is in accordance with the Shelford tolerance rule38.

E.dahuricus has a well-developed root system, prefers warmth, is particularly resistant to drought, cold, barrenness, disease and pests, and is well adapted to the soil, and is usually found in Xizang, Hebei, Qinghai, Sichuan, Shaanxi, Gansu, Xinjiang, Inner Mongolia and other places in China.In this study, the current suitable area of E.dahuricus was predicted based on the optimized maximum entropy model, and the results were consistent with the actual distribution, showing that the Maxent model predicts the distribution of E.dahuricus reliably and accurately39.

E.dahuricus, distributed mainly at high altitudes, is ecologically fragile and sensitive to globally changing environmental conditions and human activities. Compared with previous studies, this study introduced topographic and soil factors to predict the potential distribution of E.dahuricus, in addition to using common climatic involvement in modeling, and tested the correlation of the factors to screen out six environmental factors with low correlation and closely related to E.dahuricus40. For example, temperature can directly affect a number of physiological indicators related to photosynthesis, including stomatal conductance, transpiration, and intercellular carbon dioxide concentration, as well as indirectly affect photosynthesis by influencing the decomposition rate of soil organic matter. In addition to temperature, the distribution of E.dahuricus is also related to precipitation of driest quarter (bio17), precipitation seasonality (bio15), cation exchange capacity of topsoil (t_cec_soil) and altitude. Precipitation of driest quarter(bio17) can model the potential distribution of E.dahuricus under extreme drought conditions, which is important for E.dahuricus, which is mainly distributed in environmentally variable highland areas. With the change of altitude, the temperature, water, and nutrients that regulate plant growth will also change. QI found that the genetic variation of germplasm plants was significantly correlated with altitude, and also found that the threshing rate, thousand-seed weight, seed size, yield, and genetic diversity of germplasm plants populations at different altitudes were different41. In addition to the above factors, the distribution of E.dahuricus is also related to species interrelationships, the layout of the water system, soil and many other factors, which have not been addressed in this study, and therefore more factors can be combined to improve the construction of the species distribution model in subsequent studies42. Meanwhile, limited to the existing environmental variables, this study utilized the current topographic, climatic and soil factors to predict the potential distribution area of E.dahuricus under the assumption that anthropogenic disturbances will remain unchanged in the future time period, and the prediction results were somewhat biased43.

This study found that E.dahuricus is distributed in many parts of the country, which is closely related to its excellent resistance to stress and drought, among which the highly suitable areas are mainly concentrated in the western part of Sichuan Province, the southern part of the Xizang Autonomous Region of China, the area in the northeastern part of Qinghai Province bordering Gansu Province, and the area in the southeastern part of Qinghai Province bordering Sichuan Province, which is in line with the current distribution of E.dahuricus, which is concentrated in the distribution points of the region. Under future climate scenarios, the highly suitable area will migrate northward, the area of the highly suitable area for E.dahuricus in Qinghai Province will increase, and the rest of the area will be more stable, and the utilization and protection of E.dahuricus should mainly focus on the highly suitable area44. With the global climate continuously warming, the total potential distribution area of E.dahuricus in China shows an increasing trend, which is similar to the results of Guo for the distribution prediction of E.dahuricus in Northwest Sichuan45. The increase in temperature satisfies the heat demand of plants to a certain extent, thus favoring the growth and development of plants. WANG found that warming treatment significantly promoted the growth of E.dahuricus46. JIANG found that warming was favorable to the increase of E.dahuricus height, leaf area index, and aboveground and belowground biomass, and at the same time, the normal growth of the plant could only take place within a certain temperature range, within which the plant growth accelerated with the increase of temperature, but after exceeding a certain range, the excessively high temperatures were detrimental to the growth of the plant, and this also explains why the area of suitable habitat peaked in the SSP126 scenario in 2041–2060 and then declined in the SSP245 and SSP585 scenarios47. Meanwhile, with the warming of the climate, the center of mass of E.dahuricus migrated, and the direction and distance of migration varied in different scenarios in different periods, but the general direction of migration was to the southwest first, and then to the northeast. This is similar to the results predicted by Chen for 100 species distributed across Europe, northeastern North America, and Oceania.

Conclusion

In the present study, the default parameters of the model were optimized to obtain the best combination of model features with FC = LQHPT and an adjustment multiplier of 1.5. The mean value of the area under the ROC curve (AUC) was 0.850, which indicated that the Maxent model had a high prediction accuracy of the potentially suitable growing areas for E.dahuricus. The results showed that the distribution of E.dahuricus was influenced significantly by six environmental factors: temperature seasonality(bio4),min temperature of coldest month(bio6),precipitation of driest quarter(bio17),and precipitation seasonality(bio15),cation exchange capacity of topsoil(t_cec_soil) and elev.The cumulative contribution of these factors to E.dahuricus distribution was 88.3%. According to this study, E.dahuricus is most suitable to grow in Xinjiang, Xizang, Gansu, Qinghai, Ningxia, Inner Mongolia, Shanxi, Hebei, Beijing, Liaoning, Chongqing and other provinces.Meanwhile, under the current climatic scenario, the suitable habitat area of E.dahuricus is about 223.63 × 104km2, accounting for 23% of China’s total area. Under the future climate change scenario, the suitable area will be further expanded with the increase of temperature, especially the increase of temperature in the high suitable area. Climate change will also cause the center of mass of E.dahuricus to migrate to higher latitudes, and the direction and distance of migration will change under different climatic conditions. Therefore, the results of this study have laid a solid foundation for the rational and effective utilization and conservation of E.dahuricus.

However, this study only predicted the effects of climate, soil and topography on E.dahuricus, and explored the distribution of E.dahuricus in China and its changes in suitable habitats, which has a certain limitation. In the future, we can also collect more data on the distribution of E.dahuricus in other regions or even worldwide, taking into account the effects of species interactions, human activities, and other bioecological factors, in order to conduct a more comprehensive study on the suitable environmental conditions conducive to the growth of E.dahuricus.