Model performance and contribution of variables
Ecological modeling yielded an average AUC value of 0.951, while the TSS index was 0.887, classifying the model as very satisfactory. The six bioclimatic variables of annual mean temperature (Bio1), isothermality (Bio3), temperature seasonality (Bio4), precipitation seasonality (Bio15), precipitation of warmest quarter (Bio18), and precipitation of coldest quarter (Bio19) were selected to establish the model (Table 2). Fig. S1 shows the results of the Jackknife test of the variable contribution by Maxent. When used independently, Bio1, Bio3, Bio15, and Bio18 provided very high gains (>0.40), indicating that these four variables contained more useful information than the other variables. Bio4 and Bio19 achieved very low yields when used alone, and did not contain much information. Therefore, Bio1, Bio3, Bio15, and Bio18 were identified as important climatic factors that influence the suitable habitat of C. chinensis.
Response of variables to suitability
The response curves of C. chinensis to the six assessed bioclimatic variables are shown in Fig. S2. As shown in Fig. S2a, when Bio1 is below 5 °C, the probability that C. chinensis exists is extremely low (below 0.5, indicating low probability). With increasing temperature, the probability for C. chinensis to exist gradually increased, and reached the maximum at 22°C with a probability of existence as high as 0.7. When Bio1 ranges from 4°C to 37°C, the survival rate of C. chinensis was high (~0.5). Therefore, the optimum annual mean temperature of C. chinensis ranges from 4°C to 37°C.
As shown in Fig. S2b, when Bio3 ranged from 0 to 45, the survival probability of C. chinensis exceeded 0.5, indicating a benefit for the survival of C. chinensis. Therefore, the optimum isothermality of C. chinensis should remain below 45.
As shown in Fig. S2c, when the temperature seasonality of Bio4 was ~500 or less, the existence probability of C. chinensis was extremely low. Furthermore, from 4000 to 25000, the survival probability of C. chinensis first increased and then decreased, and the survival probability of C. chinensis decreased to above 0.5. Therefore, the optimal temperature seasonality of C. chinensis is 4000-18000.
As shown in Fig. S2d, when Bio15 exceeds 25, the survival probability of C. chinensis rapidly increased and reached a peaked at around 80 (~0.72). From 50 to 130, the survival probability of C. chinensis exceeded 0.5. Therefore, the optimal precipitation seasonality ranges from 50 to 130.
As shown in Fig. S2e, with increasing Bio18, the survival probability of C. chinensis gradually increased and peaked at around 500 mm. Beyond 500 mm, the survival probability of C. chinensis deceased and reached a minimum at around 1400 mm. Furthermore, the optimal precipitation of the warmest quarter ranges from 300 mm to 1000 mm, and from 2500 mm to 3500 mm with the survival probability of C. chinensis exceeding 0.5.
As shown in Fig. S2f, when Bio19 ranged from 0 to 2000 mm, the existence probability of C. chinensis exceeded 0.5. Therefore, the optimal precipitation of C. chinensis in the coldest season ranges from 0 mm to 2000 mm and the survival probability exceeds 0.5.
Model application
Global C. chinensis distribution
The global C. chinensis distribution is shown in Fig. 1a. In Asia, C. chinensis are mainly distributed at latitudes ranging from 20°N to 50°N, which includes central, eastern, and southern China (Fig.1c). C. chinensis also has a small distribution in Japan, India, Afghanistan, Pakistan, Myanmar, Vietnam, Bangladesh, and Turkey as well as minor occurrences in Australia (Fig. 1a, c). However, no distribution was found on Europe, Africa, and America (Fig. 1a). Habitat suitability simulation for three historical periodsThe simulation results of the C. chinensis habitat suitability during three historical periods (last glacial maximum, mid-Holocene, and 1960–1990) are shown in Fig. 2. From the perspective of space, suitable areas for C. chinensis during these three periods concentrated in the central, northern, southern, and eastern parts of China. These areas have a survival probability above 0.5, indicating that C. chinensis in these region benefitted from moderate or relatively high suitability. Compared with the last glacial maximum, the paleoclimatic prediction of the Holocene mid-term CCSM4 climate model indicates that the position in the mid-Holocene changed; moreover, it indicated that the size of the predicted distribution increased. From the mid-Holocene to 1960–1990, the global habitat suitability of C. chinensis gradually decreased, and the area with medium and relatively higher fitness (>0.5) gradually decreased. From the last glacial maximum to the mid-Holocene, the total area with suitability above 0.75 increased by 0.5689 million km2 (i.e., by 25.09%). The area with higher fitness (>0.75) during the mid-Holocene reached 2.8362 million km2, accounting for 1.9% of the global total area. However, from the mid-Holocene to 1960–1990, the total area with suitability above 0.75 decreased by 0.0797 million km2 (i.e., by 2.81%). During the period of 1960-1990, the area with high fitness (>0.75) was 2.7565 million km2, accounting for 1.85% of the global total area. From the last glacial maximum to the mid-Holocene, the total area with suitability of 0.5-1 increased by 0.0875 million km2, while from the mid-Holocene to 1960–1990, the total area with suitability of 0.5-1 decreased by 0.0759 million km2 (Table 3).
Suitable habitat distributions under global warming scenarios
The computed results for the C. chinensis habitat suitability in RCP2.6 and RCP8.5 are shown in Fig. 3 and Fig. 4, respectively. The suitable habitats of C. chinensis decreased in response to climatic warming. In RCP8.5, the total area with intermediate suitability and high suitability for the survival of C. chinensis was less than that of RCP2.6. In RCP2.6, the C. chinensis suitabilities of northern, central, and southern China, North Korea, and the coastal areas of Japan all exceed 0.75. However, the suitabilities of southern Africa, the central and southern parts of North America, and South America ranged between 0.25 and 0.5, while the habitat suitability of the remaining areas was below 0.25. In RCP2.6, the area with suitable habitat was below 0.25 (about 141 million km2), accounting for 94.6% of the global area. Areas where the habitat suitability ranged between 0.25 and 0.5, as well as between 0.5 and 0.75 accounted for 2.0% and 1.5% of the world, respectively, with areas of about 2.9526 million km2 and 2.2519 million km2, respectively. Habitats with suitability exceeding 0.75 accounted for 1.91% of the total area of the world. In RCP8.5, the area suitable for C. chinensis growth between 0.25 and 0.5 was the same as in RCP2.6, and its distribution concentrated in the central and southern parts of North America and South America. Moreover, habitats with suitability above 0.75 were also distributed in Northern China, North Korea, and the coastal areas of Japan. Compared with RCP2.6, for RCP8.5, the areas with high suitability for survival increased by 0.052 million km2; however, areas with intermediate suitability and high suitability for survival decreased by 0.18 million km2. Areas with habitat suitability below 0.25% accounted for 94.8% of the world's total area, with an area of 141 million km2. Compared with PCR2.6, this indicates an increase of 0.3298 million km2. Therefore, in general, habitats suitable for C. chinensis decreased in response to climate warming (Table 4).