Canonical principal component analysis and its application to the study of vegetation-climate relationships in Shanxi, China

Zhang Jintun
1998-01-01
Acta Geographica Sinica
Abstract:Ordination is a group technique of multivariate analysis and becomes more and more important in vegetation studies. Many ordination methods are available in modern ecology, however most of them analyse vegetation composition data only. A new technique combining vegetation data and its environmental data is presented in this paper. Based on Principal Component Analysis (PCA), this method is named Canonical Principal Component Analysis (CPCA). It combines vegetation data and environmental data by the way of multi-regression, similar to that of Canonical Correspondence Analysis (CCA): yj = b0 + b1z1j + b2z2j + ⋯ + bqzqj where yj is the ordination score of the jth entity (sample) ; b0, b1, ⌊, bq are constants calculated by linear regression; zij is the value of the ith environmental variable in the jth entity (i = 1, 2, ⋯, q, q = the number of environmental variables). CPCA is applied to the study of vegetation-climate relationships in Shanxi Province. The outcome illustrates that CPCA describes the relationships between vegetation and climate clearly and is an effective and time-saving methodology. The function of PCA is extended in combination with regression. CPCA has advantages in analysis of vegetation-environmental relationships compared with PCA, and will be widely used in the future. The results of ordination indicate that the geographical distribution of vegetation in Shanxi province is closely related to ecological gradients, i. e. to climatic variations. Among ecological variables, heat and water conditions are the most important to vegetation distribution, particularly in latitudinal distribution in Shanxi. Climatic factors, such as annual temperature, monthly temperature in January, accumulated temperature, annual precipitation, annual sunshine hours, relative humidity etc. play singificant roles in the formation, development and distribution of vegetation in Shanxi Plateau. This relationship is extremely clear in latitudinal variation in Shanxi Plateau. This relationship is extremely clear in latitudinal variation in Shanxi due to its long distance from the north to the south. However, it is not very obvious in longitudinal direction because of its shorter distance from the east to the west and because of its special topographical conditions.
What problem does this paper attempt to address?