A total of 19 bioclimatic and elevation variables were selected, which were downloaded from the WorldClim dataset (http://www.worldclim.org/). Version 2.1 of the WorldClim dataset with a resolution of 30 s was chosen for use in this study, and for the current climate conditions, the average data from 1970 to 2000 were used (Gu et al., 2021). Meanwhile, we selected Model for Interdisciplinary Research on Climate, version 6, to represent future climate scenarios. We chose the data in 2021–2040, 2041–2060, 2061–2080, and 2081–2100 under four shared socioeconomic pathways (ssp): ssp126, ssp245, ssp370, and ssp585 to simulate future climatic scenarios. ssp 126 means lower forcing scenarios (radiative forcing reaches 2.6 W/m² in 2100; the temperature may be lower than 2°C relative to the pre-industrial revolution). ssp245 and ssp370 mean medium forcing scenarios (radiative forcing becomes 4.5 and 7.0 W/m² in 2100). ssp 858 means the highest forcing scenarios (radiative forcing reaches 8.5 W/m² in 2100). According to the model, the temperature may significantly increase at the end of the 21st century, but the precipitation may decrease in arid and semi-arid areas; the trend will become more pronounced as the forcing intensifies. The soil data had a resolution of 30 s and were downloaded from http://soilgrids.org. In addition, we assume that soil conditions do not change in the future climate scenarios and only consider the climatic conditions, as changes in soil properties take place over a long period (Ning et al., 2018) ( Table 1 ).

The strong relation of variables could result in the misinterpretation of the MaxEnt model. Thus, we used ArctoolBox for MaxEnt in Arcgis 10.6 to select variables, and correlation values lower than 0.8 were employed to ensure that the environment is independent (Phillips et al., 2006; Tu et al., 2021). Finally, we obtained 14 environmental variables to build the model for each ephemeral plant ( Table 2 ).

In this study, we have selected six ephemeral plants globally, and each species was typical in the ecosystem, with one having a relatively narrow distribution and another one with a relatively wide distribution ( Figure 1 ). Trigonella arcuate (Fabaceae family) is primarily distributed in Iran, Kazakhstan, Mongolia, China, etc. (narrow distribution), and Tauscheria lasiocarpa (Brassicaceae) is mainly distributed in Kazakhstan, Pakistan, Iran, Mongolia, Kyrgyzstan, Uzbekistan, Turkmenistan, China, etc. (wide distribution). Cold deserts are characterized by cold winters with snowfall, high winter rainfall, and high evaporation over the summer, mainly located in Central Asia. Most plants in cold deserts shed their leaves and have spiny leaves. Anastatica hierochuntica (Brassicaceae family) is distributed in Saudi Arabia, Israel, United Arab Emirates, Palestine, Morocco, Algeria, Libya, Jordan, Egypt, Mauritania, and Oman (widely distributed). Trigonella arabica (Fabaceae family) is mainly distributed in Israel, Jordan, Palestine, United Arab Emirates (narrow distribution). The two typical plants are distributed in hot desert areas with particular survival strategies. In hot deserts, rainfall is usually deficient and concentrated in short bursts between long rainless periods. Evaporation rates regularly exceed rainfall rates. Sometimes rain starts falling and evaporates before hitting the ground. The hot desert is mainly located in Southern Asia and Africa. Plant leaves in hot deserts with small, thick, and covered in a thick outer layer. Crocus alatavicus (Iridaceae) is distributed in Kazakhstan, Uzbekistan, Kyrgyzstan (widely distributed). Gagea filiformis distributed in Kazakhstan, Mongolia, Uzbekistan, Kyrgyzstan, and China (narrow distribution). The two typical ephemeral plants were widely distributed in arid and semi-arid deciduous forest areas. Deciduous forest with sufficient rainfall, but the tall trees shed the sunlight for herbaceous plants.

Species occurrence data were downloaded from the Global Biodiversity Information Facility (GBIF) (https://www.gbif.org) and the Chinese Virtual Herbarium (CVH) (http://www.cvh.ac.cn). We used Arcgis 10.6 to test the cross-correlation of 30 variables and only those variables with a correlation coefficient (r²) < 0.8 were selected (Negrete et al., 2020; Zhao et al., 2022). Each ephemeral plant’s occurrence records were randomly selected from each cell with dimensions of 20 × 20 km (Boria et al., 2014).

We used the MaxEnt software (3.4.3) (Maxent (amnh.org) to model the six ephemeral plants’ potential distribution (Phillips & Dudík, 2008; West et al., 2016; Abdelaal et al., 2019; Gu et al., 2021; Amiri et al., 2022). We used the receiver operating characteristic to evaluate the MaxEnt model’s performance (West et al., 2016; Abdelaal et al., 2019; Gu et al., 2021). The AUC values range from 0 to 1: AUC >0.9 indicates the highest predictive ability, AUC 0.7–0.9 means good performance, AUC <0.7 indicates that the model is unacceptable (Chen & Chen, 2013; Gu et al., 2021).

This probability was calculated by the occurrence data and randomly selected data with environmental (bioclimatic and soil) variables to generate suitable gradients (Gu et al., 2021; Ye et al., 2021; Amiri et al., 2022). In our study, 75% data were randomly selected to train the model, and 25% data were randomly selected to test each model repeatedly 10 times. A jack-knife test was chosen to evaluate the importance of each variable. In addition, the main environmental variables for each suitability class were extracted from the spatial analysis tool in Arcgis 10.6. (Gu et al., 2021). The probabilities of species distribution were divided into four arbitrary categories with a specific logistic threshold in Arcgis 10.6: unsuitable distribution (0–0.2), lower suitable habitat (0.2–0.4), moderately suitable distribution (0.4–0.6), and highly suitable distribution (0.6–1) (Negrete et al., 2020; Gu et al., 2021; Ye et al., 2021).

To calculate the trends of the potential distribution changes of ephemeral plants in the future, the following equation was used: Ch _dis =Fu _dis - Cu _dis.

Ch _dis means the potential change in distribution between future and current, Fu _dis means the potential distribution under future climate change scenarios, and Cu _dis means the potential distribution in current conditions.

After the model had been built, we obtained the AUC values for T. arcuata, T. lasiocarpa, A. hierochuntica, T._arabica, C. alatavicus, and G. filiformis using the MaxEnt model as 0.977 ± 0.028, 0.985 ± 0.004, 0.984 ± 0.004, 0.995 ± 0.001, 0.977 ± 0.004, and 0.994 ± 0.006, respectively ( Figure 2 ), indicating that all models had excellent performance ( Table 3 ).

Table 7 shows the percentages of different suitable living areas under the current condition. According to the potential suitable areas of T. arcuata, T. lasiocarpa, A. hierochuntica, T._arabica, C. alatavicus, and G. filiformis ( Figure 3 ), the current total potential suitable distribution areas were 448.78 × 10⁴, 1,802.66 × 10⁴, 448.45 × 10⁴, 31.84 × 10⁴, 2,031.45 × 10⁴, and 557.14 × 10⁴ km², respectively. The highest suitable habitat occupied 29.09, 155.32, 46.51, 3.56, 41.50 × 10⁴, and 60.49 × 10⁴ km², while for the moderately suitable habitat, these values were 65.49 × 10⁴, 255.46 × 10⁴, 155.32 × 10⁴, 9.61 × 10⁴, 65.70 × 10⁴, and 112.52 × 10⁴ km² ( Table 7 ).

Moderately suitable distribution means that the specific logistic threshold probability was between 0.4 and 0.6, and the highly suitable distribution was 0.6–1.

The potentially suitable distributions for three types of ephemeral plants were analyzed under 16 different future climatic scenarios (ssp 126, ssp 245, ssp 370, and ssp 585 in 2021–2040, 2041–2060, 2061–2080, and 2081–2100, respectively). The results indicate the changes in the potentially suitable area of six ephemeral plants with a significant difference ( Figures 4 , 5 , 6 , 7 ).

In cold deserts, the total potential suitable area of T. arcuate increased under 11 climatic scenarios, and that of T. lasiocarpa increased under 12 climatic scenarios. The majority of climatic scenarios are under ssp 245, 370, and 585, which means that the distribution of cold desert ephemeral plants will increase under medium and high forces in future climate scenarios. For hot desert plants, the total potential suitable area of A. hierochuntica decreased under nine climatic scenarios (under most lower- and medium-force climate scenarios in 2021–2040 and 2041-2060) and increased in seven (under most high forces climate scenarios in 2061–2080 and 2081–2100). The total potential suitable distribution of T. arabica increased under 10 climatic scenarios, the majority are under high forces scenarios in 2061–2080 and 2081–2100. In deciduous forest ephemeral plants, the total potential suitable distribution of C. alatavicus decreased under 14 climatic scenarios. It indicates that the distribution area would be reduced with the increase in temperature. On the contrary, the total potential suitable distribution of G. filiformis increased under 12 climatic scenarios. The scenarios were under medium and high forces climate scenarios in 2061–2080 and 2081–2100.

Ephemeral plants are a particular component of flora that take full advantage of light, water, and temperature to rapidly complete their life cycle in a very short time (Jeffrey, 1992). We use the MaxEnt model to identify the key factors to determine the distribution of ephemeral plants in three types of ecosystems under the current climate scenario. The results show that the significant variables affecting the potential suitable distribution in three ecosystems were inconsistent with our hypothesis.

Compared to the potential distribution change between current and future scenarios modeled by the MaxEnt model, the results show that the distributions of ephemeral plants in the three ecosystems will be changing under most future climate scenarios (SSP126, SSP245, SSP370, and SSP585 in 2021–2040, 2041–2060, 2061–2080, and 2081–2100) (T. arcuate 11 to 16 and T. lasiocarpa 12 to 16).

Studies found that the global arid regions have been expanding significantly due to precipitation decreases and temperature increases over the past 60 years (Feng & Fu, 2013; Huang et al., 2015). The trends will continue in the future, especially in mid-latitude arid and semi-arid regions. Lower precipitation and higher temperature may offer more suitable areas for cold desert plants to survive. However, Chen (2021) found that the arid zones may shrink by the end of the 21st century under RCP8.5 (the highest concentration of greenhouse gas emissions in the future) and RCP4.5 (moderate concentration of greenhouse gas emissions in the future) scenarios, such as in northern China and India. The results were consistent with the finding that the potential distribution area for cold desert ephemeral plants would decrease under most future climate scenarios ( Figures 4 – 7 ). Although shrinkage of the arid area may occur in the cold desert regions of Central Asia, the extent is much lower than the expansion (Chen, 2021). The total aridity region still shows an expanding trend and will likely expand by approximately 10% by the end of this century. It indicates that global climate change could provide more areas for ephemeral plants to survive in the cold desert.

Several studies show that hot desert regions may expand under future climate scenarios, such as in North Africa and Saudi Arabia (Vale & Brito, 2015; Chen, 2021). Our study’s potential distribution area of A. hierochuntica decreases in more than half of future climate scenarios ( Figures 4 – 7 ). It was inconsistent with our hypothesis that the ephemeral plant in hot desert distribution will be expanding in the future. One possibility was that the plants were distributed in North Africa, Egypt, Morocco, Libya, etc. The annual mean precipitation was virtually zero in Libya, Egypt, and Sudan. Still, the Sahara’s rainfall is unreliable and erratic (Vale & Brito, 2015). This uncertainty leads to the prediction results having significant differences. The potential distribution was decreased under nearly half of the future climate scenarios of T. arabica and increased by roughly half of those of A. hierochuntica. It suggests that climate change leads to arid area expansion in the future, which may not significantly affect the distribution of hot desert ephemeral plants.

The deciduous forest ephemeral plant potential distribution area of C. alatavicus was decreased under the majority of future climatic scenarios ( Figures 4 – 7 ), which is consistent with our hypothesis that the distribution of deciduous forest ephemeral plant would be reduced in the future. Zhang and Hu (2009) believed that global warming negatively impacts forest ephemeral plants. If the temperature rises by 1 to 2°C, the potential distribution of early ephemeral plants would be reduced or more fragmented, which may cause some species’ suitable habitats to become reduced or even extinct. Therefore, with the temperature increase in future climate scenarios, the suitable habitats of C. alatavicus would be further compressed.

However, our study found that the potential distribution of G. filiformis was increased in multiple future climate scenarios, which is inconsistent with our hypothesis. The results are due to the main limiting factors between C. alatavicus and G. filiformis being different. The annual mean temperature had the highest contribution for C. alatavicus; the pH had the highest for G. filiformis. The difference in determining factors may indicate the adaptation strategies in species. The potential distribution change was variance in ephemeral plants under the future climate change scenarios: ephemeral plant distribution in cold deserts were increased under the majority of future climate scenarios, but in hot deserts, the distribution was relatively decreased; but in deciduous forest widely-distributed species would decrease, narrow range species would increase. Thus, ephemeral plants with greater adaptive capacity are likely to occupy a wide range of niches, while the niches of those with lower adaptive capacity will shrink further under climate change. This information may be useful for formulating relevant policies to prevent managed species with an expanding potential distribution from becoming invasive species, and for shrinking species, we should minimize damage and protect them from species and in avoid further reductions in species range and their distribution in the future.

In our research, only five soil conditions were input to the MaxEnt model; meanwhile, we also hypothesized that the soil conditions would remain unchanged in the future. Changes in soil conditions should be considered to better understand the distribution of ephemeral plants. Secondly, desert ephemeral plants had only two or three months in early spring to complete the life cycle. However, ephemeral plants may occur in suitable (light, precipitation, temperature) conditions; even in autumn, the seasonal change was not considered. These may explain why the ephemeral plant often occurs in summer and autumn.