REPEATABILITY OF THE FRENCH HIGHER VEGETATION TYPES ACCORDING

Abstract

Higher vegetation types are generally determined by successive approximations and defined by a common consent. Instead, they might be statistically determined and repeated, according to a numerical method called ‘socio-ecology’. This method deals only with floristical data, but gives them an ecological meaning by a previous calibration of the relations between plants, computed as ecological indices. It is applied to a pair of two homologous samples, each having 2.000 relevés and coming from the 60.000 relevés stored in the French data bank ‘Sophy’. Each sample covers the main ecological gradients of the bank, it defines a hierarchy of vegetation types and it explains half the peculiarity of a type with only 10 to 30 discriminant plants, out of the 5.000 plants observed in the relevés. Results : 1) The discriminant plants may characterize the vegetation types, including the higher ones, in a coherent and readable form. 2) In the two independent classifications, having different structures, the same vegetation types are repeated. They are the reciprocal nearest types, in the socio-ecological space. Though the two classifications have no one relevé in common, the repeated types have nearly the same discriminant plants. 3) At the highest level, two clear-cut main types show the difference between light and shadow. The same herbaceous discriminant plants, for a type, and the ligneous or sciaphilous ones, for the other, have similar fidelities and constancies in the two classifications. 4) Such a numerical agreement, instead of common consent, appears again in the sub-types, which remind the classical ones, but which are repeatable.