Comparison of indicator values of forest understory plant species in Western Carpathians (Slovakia) and Vosges Mountains (France)

International audience

anglais. The aim of this article is to make a comprehensive comparative study of plant species behaviour in the Vosges Mountains of Western Europe and the Western Carpathians of Central Europe. We will then look at the consequences of the differences observed on the calibration of environmental variables using the bioindicator values of the plant species. The variables altitude correlated to rainfall and temperature and pH of the A horizon of the soil are used to compare species behaviour. A response curve was established using a logistic model for each species according to these two variables. The response curve is used to determine the indicator values. Most of the species in the two regions react in a significant fashion to one or both of the variables. pH has a more marked effect than altitude in the Vosges Mountains (it affects 69% of species). The reverse is true in the Western Carpathians (pH affects only 28% of the species in this area). Species behaviour varies widely from one region to the other, both for pH and for altitude. The proportion of neutrophilous or subalpine species common to both areas is roughly twice as large in the Vosges Mountains as in the Western Carpathians. These differences in behaviour can be explained by the ecological contexts of each region. The high proportion of neutrophilous and subalpine species in Vosges Mountains is responsible for a shift of predictions of variables with vegetation towards large values of pH and altitude in this land. pH values are better predicted in the Vosges Mountains than in the Western Carpathians. Despite a smaller number of species sensitive to altitude in the Western Carpathians, this variable is better predicted in this land than in the Vosges Mountains. These results show that importance of a variable for vegetation has an influence on the quality of predictions. The wide difference in ecological behaviour of species from one region to the other and the very variable quality of predictions would suggest caution in the use of indicator values and their extrapolation to other regions