Influence of sample plot size on the data analysis.
| Authors | |
|---|---|
| Year of publication | 2006 |
| Type | Article in Proceedings |
| Conference | In: European Vegetation Survey. 15th Workshop. Vegetation in Agricultural Landscapes and NATURA 2000 |
| MU Faculty or unit | |
| Citation | |
| Field | Botany |
| Keywords | ordinations; data transformation |
| Description | Vegetation samples from plots that are stored in phytosociological databases often differ by one or two orders of magnitude in size, even within a single vegetation type. This fact may influence vegetation analysis, but the extent of this possible influence has not been investigated so far. The effect of sample plot size on ordination results was investigated in forest vegetation in Moravia (Czech Republic), using data from nested plots of 49, 100, 225, 400, 625 and 961 m2. Sampling of different vegetation types over various climatic condition in a larger area of southern Moravia on the one hand and of floristically homogeneous vegetation within a small area on the other hand, provided heterogeneous and homogeneous data sets, which were used for analysis. In both data sets the smallest plot sizes showed the most deviating patterns in ordinations, while samples of larger sizes gave similar results. The transformation of percentage cover also affected the results, more so when samples were collected in homogeneous vegetation. These effects of various plot size on ordination are caused by low species richness and evenness in smaller plots, which are more pronounced in homogeneous data set, and result in a higher beta-diversity of the set of smaller plots. Samples differing in size by one order of magnitude (49 and 961 m2) seem to be inappropriate to combine together in ordination analysis, especially when they were sampled in homogeneous vegetation. |
| Related projects: |