Department of Earth System Science

The montane forests of the eastern part of the Ecuadorian Andes are considered as a hotspot of biodiversity threatened by human activity. Vegetation classification is a central part of ecosystem studies that estimate potential risks and rates of loss, and suggest conservation concepts. Classification can be approached on the basis of floristic or structural data. The present study reviews newer approaches that use a mix of floristic and structural features to characterise vegetation and highlight what ‘structure’ means in each case (especially for the tropics). These mixed approaches are grouped according to the growing degree they consider structure and get independent of plant species identification. Only a few approaches compare floristic classifications with the results of structural investigations and thus develop a tool to describe structure independent of species composition (Webb et al. 1970, Werger and Sprangers 1982, Lux et al. 1994, Le Brocque and Buckney 1997). They all conclude from their data that the investigation of structure leads to comparable results but is much easier to achieve. The results of our structural investigation of a shrub vegetation in higher elevations of the Ecuadorian Andes on plot scale are presented. Although the new investigation system is independent of plant species identification, plant species composition of the 27 plots was registered as a comparison. Shrub vegetation was chosen to ensure a comparatively low number of species that are fairly well documented. Cluster analysis of 134 structural features per stratum led to four groups that were interpreted as three structural units (unit 1, 2, and 3), unit 1 with two sub-units (1a and b). Unit 1 consisted mainly of dwarf shrubs with (1a) herbs, or (1b) shrubs, unit 2 consisted of shrubs only, and unit 3 of shrubs and trees. Geographical distribution of structural units followed a distinct pattern, and was presumably mainly controlled by slope, soil depths and exposure to wind. In the 27 plots, 142 vascular plants from 46 families were found and grouped in two groups. Species composition of group 1 is typical for vegetation above tree line, shrub and grass páramo, that of group 2 for tree line vegetation with ceja andina and upper mountain forest. Comparison of structural units with floristic groups made obvious that the floristic group 1 (páramo) corresponded with the structural unit 1 (dwarf shrub), and the floristic group 2 (ceja) corresponded with the structural unit 2 (shrubs) and unit 3 (shrubs with trees). Our results support the point of view that our structural approach leads to comparable results but is much easier to achieve than a floristic classification. We conclude that if the functioning of the Ecuadorian montane forest ecosystem is to be understood - and this has to be done quickly considering the immense destruction rate - a structure-based classification is the most effective way to describe the vegetation as a basis for further ecosystem research.