Black Locust (Robinia pseudacacia L.) in Austria: The Interplay of Climate, Climate Change and Range Expansion KLEINBAUER Ingrid* # , DULLINGER Stefan*, ESSL Franz°, PETERSEIL Johannes°, ENGLISCH Thorsten # * VINCA – Institute for Nature Conservation and Analyses, Vienna, Austria ° UBA - Federal Environment Agency, Vienna, Austria # University of Vienna, Austria contact: [email protected] FIRST The leguminous tree black locust arrived in Europe during the 17th century. By now it has become the most problematic introduced tree species in Austria: threatening silvicultures as well as rare endangered plant communities, in particular species-rich dry and semi- dry, nutrient poor grasslands and thermophilous oak forests. Once established population density increases rapidly due to efficient vegetative reproduction by root suckering. We have to identify regions and habitats at risk of becoming invaded under a warmer climate. THAT‘S WHY NOW In Austria’s pannonical East and Southeast black locust finds perfect conditions for establishment (See Fig. 2). We constructed the environmental envelop for black locust in Austria, using Generalized Linear Models (GLM, McCullagh & Nelder, 1989) to regress several factors (Fig. 3) against presence/absence of black locust. Occurrence data was extracted from the “Mapping the Flora of Austria” database: for each cell (3’ x 5’) in a raster - covering the whole country - the BUT HOW? But: is it really the temperature that limits its spread? And: what happens if climate change forecasts come true and temperatures rise? Fig. 3: Predictors used in the model status of black locust is known (see Fig. 2) In a stepwise backward selection (p<0.05) the following factors were chosen as best predictors (see Fig. 3): • mean April temperature • mean winter precipitation sums • mean number of frost days • land use index • curvatur index © Franz Essl 1 2 3 4 NEXT STEPS To account for differences in recruitment success • in different habitat types + • along a temperature gradient + • under increased Nitrogen availability + • and facing a competitor (native Quercus petraea) … … we started a field experiment in late spring 2006 to test for germination success and survival (see Fig. 6 and 7) The experiment will be finished by the end of 2008. Fig. 6: Experimental Design +N +N +N Forest Bare Soil Grassland locations within each study-site: plots: 1 with and 1 without N added (within each location) 4 different study-sites covering a temperature (T) gradient from about 6 – 10°C Fig. 7: plots in different habitat types forest bare soil grassland 6 Fig. 2: Occurrences of Robinia pseudacacia in Austria (source: Niklfeld, 1998) Austrian Academy of Sciences McCullagh, P. & Nelder, J.A. 1989: Generalized Linear Models. New York, Chapman & Hall Niklfeld, H. 1998: Mapping the flora of Austria and the eastern Alps. - Rev. Valdôt. Hist. Nat. 51, Suppl.: 53-62. Pope, V. D., M. L. Gallani, P. R. Rowntree and R. A. Stratton, 2000: The impact of new physical parametrizations in the Hadley Centre climate model -- HadAM3. Climate Dynamics, 16: 123-146 Jacob et al., 2005: unpublished report within the EU-project prudence - P rediction of R egional scenarios and U ncertainties for D efining EuropeaN C limate change risks and E ffects REFERENCES: FINANCIAL SUPPORT BY: THEN Mean April temperature and mean winter precipitation were recalculated using two different climate change scenarios for the end of the current century: 1) HadAM3 from the IPCC (Pope et al.,2000) and 2) a regionalized model from ETH Zurich (Jacob et al., 2005), CH, which has been downscaled for the whole Alpine region. Please see Fig. 4 and 5 for details in changes and comparison of climate change scenarios. Habitat Suitability Classes - Current Climate Proportion of Area (%) 23,83 8,16 5,29 10,98 51,75 not suitable bad good very good optimal Habitat Suitability Classes - ETH scenario Proportion of Area (%) 15,07 8,25 4,08 6,68 65,93 not suitable bad good very good optimal Habitat Suitability Classes - HadAM3 scenario Proportion of Area (%) 77,65 5,74 4,44 7,49 4,68 not suitable bad good very good optimal Fig. 5: Habitat Suitablility Classes and their area proportion Fig. 4: Distribution of potential habitats for Robinia pseudacacia in Austria 5 The bootstrap-corrected final model’s regression coefficient R² is 0,71 and Somer’s index (Dxy) is 0,89. mean April temperature curvatur index number of frost days land cover index winter precipitation sums mean April temperature curvatur index number of frost days land cover index winter precipitation sums