生态因子对川渝地区壳斗科植物物种丰富度的影响

The effect of ecological factors on Fagaceae species richness in Sichuan-Chongqing region,China

西南川渝地区是我国壳斗科植物分布最为丰富的地区之一.以川渝地区壳斗科植物为研究对象,以皮尔逊相关系数(r<0.8)以及方差膨胀因子(VIF<10)为准则,选取代表现代气候、生境异质性和历史气候变化的9个生态因子,基于地理加权回归模型(GWR)探究生态因子对壳斗科植物物种丰富度的影响,并与经典的全局最小二乘回归法(OLS)进行比较.结果显示:(1)分布在川渝地区的壳斗科植物有6属共74种,主要以栎属(Quercus)和柯属(Lithocarpus)为主;(2)物种丰富度(SR)表现为南部最高,物种丰富度中心为川西南锦屏山东南部—鲁南山西侧、横断山中部的邛崃山南段—大凉山北段以及四川盆地东南缘大娄山东段,可达到29-35种;(3)GWR模型预测精度优于OLS,其预测结果表明海拔变幅、潜在蒸散量和最暖季降雨量是影响川渝壳斗科植物物种丰富度的主要因子,且影响程度表现出明显的空间差异性.整体来看,GWR模型可为探究物种-生态因子关系的空间异质特征提供一种新的方法,在生物多样性研究中具有较好的应用前景.

The Sichuan-Chongqing region in Southwestern China is one of the regions where Fagaceae plants are the most abundant.This study focused on species richness of Fagaceae in the aforementioned region.We used the Pearson correlation coefficient(r<0.8)and the variance expansion factor(VIF<10)as two criteria to select nine ecological factors that represent the modern climate,historical climate change,and habitat heterogeneity.We explored the effects of these factors on species richness based on a geographically weighted regression(GWR)model.The results showed that(1)there were 74 species from six genera of Fagaceae that were mainly composed of Quercus and Lithocarpus genera in the Sichuan-Chongqing region.(2)Species richness(SR)was the highest in the south,with 29-35 species in the southeast of Jinping Mountains,the west of Lunan Mountains,the middle of Hengduan Mountains,and the eastern part of Dalou Mountains in the southeast of Sichuan Basin.(3)The prediction accuracy of the GWR model is better than ordinary least square(OLS)model.Simulation results of the GWR model showed that the range in elevation,potential evapotranspiration,and precipitation of the warmest quarter were the main factors affecting the SR of Fagaceae;moreover,the effects on SR of the aforementioned factors indicate significant spatial differences.Overall,the GWR model provided a new method to explore the relationship between spatial heterogeneity and species-ecological factors and has good potential applications in biodiversity research.