Multivariate analysis of phenological data
Fischer, Hagen S.
Phenological changes over the year influence the species cover and consequently the data of the relevés. This may have an undesired effect on the results. Due to differences in climatic conditions between years and in different regions, the date of the year is not an adequate variable to describe the date of sampling. The present paper presents a new multivariate analysis of symphenological relevés to describe the influence of phenology on floristic structure of the vegetation and for bioclimatological analysis. The aim of the investigation was to analyse the influence of the cutting date and fertilizer input on the species composition and cover of hay meadows. A total of 850 relevés was taken in Bavaria. The cover of all species was measured quantitatively and the phenological state was determined for all the species. Two new multivariate methods for analysing phenological data are described. Phenological sequencing is a heuristic algorithm that yields in a rank order describing the phenological development of the vegetation within a vegetation period. The phenological index is based on probability functions and results in a continuous description of vegetation's phenology. The phenological index can be used in canonical analysis as a covariable to eliminate undesired effects of the phenological development during the sampling period on the results of multivariate analysis. Furthermore the phenological index can be used to characterise the bioclimatological conditions of different regions or sites. Finally the phenological probability function can be used to generate a rank order of the phenological development of species. The phenological index has a significant influence on the floristic composition of the vegetation. The correlation between the date of the year and the phenological index showed a high significant relationship. The regression lines for the regression between date and phenological index for the years 1995 and 1996 in the district Ansbach (in North Bavaria) are shown. The regional climate of several Bavarian districts is characterised in terms of their average phenological development.