Mediterranean landscapes and plant communities relationship

On the basis of an earlier landscape classification of Huelva (Andalusia, Spain) using the Twinspan multivariate classification method, which resulted in eight landscape types, the study sought to test the hypothesis that a relationship may exist between landscape types and plant communities. Samples of serial scrub communities were examined by stratified sampling between landscape types. Samples were classified using phytosociological and Twinspan methods. After merging landscape types and Twinspan-classified community types, the possible association between the two was analysed by means of a two-way contingency table between five landscape units and five plant communities. Findings enabled the null hypothesis, i.e. that both descriptors were independent, to be rejected. Given the high degree of overall agreement between phytosociological and multivariate plant community classifications, inter-community variation was studied via ecological characterization of the syntaxa identified. The results indicated that parent material played a more important role than climate-related variables in accounting for this variation. This conclusion supports the role assigned to each of the two factors in the biogeographical domain, according to the scale of the study.