Jandt, U. (corresponding author, firstname.lastname@example.org) & Bruelheide, H.: (email@example.com) Martin-Luther-University Halle-Wittenberg, Institute of Biology/Geobotany and Botanical Garden, Am Kirchtor 1, 06108 Halle/Saale, Germany von Wehrden, H. (firstname.lastname@example.org): Leuphana University Lüneburg, Centre of Methods & Institute of Ecology, Faculty III, Scharnhorststr. 1, 21335 Lüneburg, Germany
Exploring large vegetation databases to detect temporal trends in species occurrences
Article first published online: 4 AUG 2011
© 2011 International Association for Vegetation Science
Journal of Vegetation Science
Volume 22, Issue 6, pages 957–972, December 2011
How to Cite
Jandt, U., von Wehrden, H. and Bruelheide, H. (2011), Exploring large vegetation databases to detect temporal trends in species occurrences. Journal of Vegetation Science, 22: 957–972. doi: 10.1111/j.1654-1103.2011.01318.x
Co-ordinating Editor: Angelika Schwabe-Kratochwil
- Issue published online: 5 OCT 2011
- Article first published online: 4 AUG 2011
- Received 8 June 2010, Accepted 30 May 2011
- Beech forest;
- Generalized linear model;
- German Biodiversity Exploratories;
- German Vegetation Reference Database (GVRD);
- Semi-dry grassland
Question: Can vegetation relevé databases be used to analyse species losses and gains in specific vegetation types in Germany over time? Does the type of response (increase or decline in relative frequency) conform to observed large-scale environmental trends in the last decades?
Location: Germany. Exploring the German Vegetation Reference Database Halle (GVRD) that was established for forest and grassland vegetation within the framework of German Biodiversity Exploratories.
Methods: Use of generalized linear models (GLMs) for testing changes in temporal frequency of plant taxa in a semi-dry grassland data set (Mesobromion) and a beech forest data set (Fagion). Data were either aggregated by year, decade or by a balanced re-sampling approach. Interpretation of the observed changes was based on species traits.
Results: In both data sets significant temporal changes were observed, although the frequency of the majority of species remained unchanged. In both data sets, species with a temporal increase in frequency had higher Ellenberg N and F indicator values, compared to species that decreased, thus indicating effects of widespread atmospheric nitrogen deposition. In the forest data set, the observed increase in recruitment of deciduous trees pointed to a change in management, while trends in the grassland data set suggested use abandonment, as seen in an increased frequency of woody species.
Conclusion: We demonstrate that vegetation databases represent very valuable resources for analysis of temporal changes in species frequencies. GLMs proved their value in detecting these trends, as also shown by the interpretability of model results with species traits. In contrast, the method of aggregation or re-sampling had little influence on the general outcome of analyses.