Publication details
Modelling fine-resolution plant species richness patterns of grasslands and forests in the Czech Republic
| Authors | |
|---|---|
| Year of publication | 2015 |
| Type | Appeared in Conference without Proceedings |
| MU Faculty or unit | |
| Citation | |
| Description | Species richness patterns have always fascinated ecologists and numerous studies attempted to map, explain and predict species richness across large areas. Such studies usually used inventory or atlas based data with coarse spatial resolution, because fine resolution data were not available. Within large areas, our knowledge of species richness patterns is thus significantly limited to coarse resolution patterns. However, recent development of large databases of vegetation plots provides an opportunity to explore distribution of species richness at very fine resolutions. Here we aim to create maps predicting fine scale species richness of vascular plants in grassland and forest vegetation across the Czech Republic and to examine factors underlying the observed species richness patterns. We used data from the Czech National Phytosociological Database where, 27 002 georeferenced plots of grasslands and 19 764 plots of forests were available. However, data processing showed that only 15 50% of relevés, depending on selection criteria applied, were suitable for modelling. To build predictive models we used Random Forest method which is considered as a very powerful tool for prediction purposes. The modelling of species richness was based on three groups of environmental variables, namely topography & geology, climate and surrounding landscape context. Resulting models explained up to 50% of variability in species richness and residuals showed neither any obvious patterns nor significant positive spatial autocorrelation. When we used our best models to predict species richness of grasslands and forests in 37 760 grid cells, each of them spanning 1.25’ of longitude and 0.75’ of latitude (ca. 1.39 × 1.5 km = 2.09 km2), resulting maps showed meaningful patterns expected based on expert knowledge. |
| Related projects: |