Poznámky k používání průměrných Ellenbergových indikačních hodnot při analýze vegetačních dat
| Title in English | Notes to the use of mean Ellenberg indicator values in vegetation analyses |
|---|---|
| Authors | |
| Year of publication | 2012 |
| Type | Article in Periodical |
| Magazine / Source | Zprávy České botanické společnosti |
| MU Faculty or unit | |
| Citation | |
| Field | Ecology |
| Keywords | analysis of variance; circularity of reasoning; permutation test; unconstrained ordination |
| Description | Ellenberg indicator values (EIV) represent valuable expert information about ecological behaviour of vascular plant species in western and central part of Europe. The mean of EIV for species in vegetation relevé or floristic inventory are considered to be a good proxy of habitat characteristics and relevant surrogate for measured environmental factors, especially when the latter are not available. However, as mean EIV are derived from data about species composition, caution must be applied when using them in analyses with other variables derived from the same data set. Such analyses include, e.g., correlation of mean EIV with ordination axes in unconstrained ordination, analysis of variance of mean EIV between vegetation types or correlation of mean EIV and species richness. As a consequence, such analyses tend to return biased and more optimistic results, especially if their significance was tested using traditional tests, and their interpretation may be therefore inappropriate. In this study, examples using real vegetation data are given and possible remedy is introduced. Due to the biased results, mean EIV are not recommended for use in abovementioned analyses, unless appropriate modified permutation tests are applied. |
| Related projects: |