关键词: 白刺花, 最大熵模型, ArcGIS, 潜在分布区, 环境变量

Abstract: Sophora davidii is a unique legume shrub species in China, which is of great significance for water and soil conservation in drylands due to its strong drought tolerance. In order to understand the change of potential distribution area of S. davidii in China under climate change scenarios, 211 geographic distribution data of S. davidii and nine climate factors were collected. The maximum entropy model was used to simulate S. davidii distribution in the last interglacial period, the last glacial maximum (MIROC, CCSM), current, and future (2050s, 2080s). The reliability of the model was evaluated by receiver operating characteristic curve, jackknife method, environmental contribution rate, and permutation importance. The changes of the threshold of main environmental variables and the range of suitable distribution area of S. davidii were analyzed. The results showed that the AUC values of the five periods were greater than 0.95, indicating well performance of this model. The most suitable distribution areas of S. davidii were mainly concentrated in Southwest, Northwest, North China, and Central South China. The most suitable distribution areas of S. davidii decreased in Southwest during the last interglacial period and moved to Hebei and Shandong provinces. During the last glacial maximum, the most suitable distribution areas in Northwest China decreased and moved to the southwest. In 2050, the most suitable distribution areas shifted from Yunnan Province to Sichuan Province. In 2080, the most suitable distribution areas in Yunnan Province decreased. Temperature and precipitation were the main environmental factors affecting S. davidii distribution. Under the scenario of climate change, the suitable area of S. davidii decreased to 17.78% of present area. Compared with the last glacial maximum, the suitable area of S. davidii was reduced. Compared with the last interglacial period, current suitable areas slightly increased. Compared with current climate, the suitable areas will increase by 12.07% in 2050 and 1.67% in 2080.

Key words: Sophora davidii, maximum entropy model, ArcGIS, potential distribution area, environmental variable.