Influence of sampling strategy on the relationships between fauna and vegetation structure, plant lifeform and floristics

Abstract

AbstractThe effect of sampling strategy on animal‐habitat relationships was evaluated with data collected from a 50 ha area containing a sequence of tropical vegetation types.Sampling sites were located randomly within defined habitat types (i.e. stratified random sampling) and systematically irrespective of habitat type. At each site the fauna, comprising birds (63 species), reptiles (15 species), amphibia (13 species) and grasshoppers (32 species) were sampled for the abundance of species. Simultaneously, vegetation and related data, comprising vertical structure (52 attributes), ground surface condition (18 attributes), plant lifeform (18 attributes) and the abundance of plant species (200) were recorded.Random and systematic data matrices, comprising sites defined by animal or vegetation attributes, were reduced dimensionally by correspondence analysis. Animal first dimension vectors were then regressed on the first dimension vectors of vegetation structure, lifeform and floristics, respectively.With stratified random sampling, vegetation structure (comprising vertical and ground attributes) and lifeform explained most of the variation in the fauna; floristics were not a significant factor. On the other hand with the systematic data, fioristics explained almost all of the variation in animal abundance and distribution. By removing the ecotonal sites from the systematic data set and recalculating vectors, the animal—vegetation relationships became similar to those generated from the stratified random sampling data.Clearly, the sampling strategy employed in a faunal survey has a major influence on the inventory of species, and on the relative importance of vegetation structure, lifeform and floristics in explaining animal distribution. The presence of ecotones in the systematic data set was highlighted as the key to the difference between the sampling strategies.