贵州喀斯特高原人口分布的自然环境因素 Ⅱ.多元回归分析与地带性研究

Study on the Natural Environmental Factors the Guizhou Karst Plateau： Multivariate Affecting Population Distribution in Regression Analysis and Zonality

以贵州喀斯特高原76个县市（区）农业人口密度、人口密度以及自然地理因素资料，应用多元逐步回归方法，研究贵州省人口分布与其主要影响因子的相关关系，并建立多元回归模型，探讨贵州喀斯特地区的自然地理因素地带性分布的特征，并进一步研究在贵州喀斯特地质条件下人口分布的自然地带性规律。结果表明：在喀斯特地质环境条件下，受经度、纬度以及平均海拔高度之间错综复杂的相互影响关系的作用，各自然地理因素具有一定程度的地带性分布特征，从而导致了贵州高原人口具有地带性分布，表现为人口分布主要集中于西北部的黔中丘原地区。

Based on the data including agricultural population density, population density and physio-geographical factors affecting population distribution from 76 counties, cities and districts in the Guizhou Karst Plateau, in this paper the multivariate stepwi~ regression is used to analyze the relationship between the popula- tion distribution and its main affecting factors in Guizhou Province. A multivariate regression model is developed for lucubrating the zonal distribution features of physiogeographical factors and the zonal distribution laws of population under the geologic conditions in the Guizhou Karst Plateau. The results reveal that there are certain zonal distribution features of physiogeographical factors under the complicated effects and correlations of the conditions of karst geological environment, longitudes, latitudes and average elevation, thus a zonal population distribution occurs in the Guizhou Karst Plateau, that is the hilly areas in central Guizhou Province are densely pop- ulated. The results also reveal that the effect of longitudes on the zonal distribution of population and physiogeographical factors is higher than that of latitudes and average elevation, and there is a close correlation between the longitudes and latitudes, so the zonal distribution of population and physiogeographical factors is mainly affected by the longitudes and latitudes, that is the population density increases with the increase of latitudes and decrease of longitudes. After analyzing the correlations between longitudes, latitudes and average elevation, it is considered that the effect of elevation on the vertical zonal distribution of population and physiogeographical factors can not be ignored.