PhD Thesis

Extending the climate envelope: Methodological approaches to integrate ecological prerequisites in species distribution models at large spatial extents

Anja Jaeschke (02/2009-10/2014)

Support: Carl Beierkuhnlein, Michael Hauhs, Christian Laforsch

The following thesis presents potential impacts of climate change on the distribution of protected animal species. Here, different influencing factors like uncertainty in the data basis, dispersal distances, and biotic interactions, as well as their influence on projections of distribution models are analysed. The aim is to amend established climate envelope models by ecological prerequisites and therewith to create a new basis for action for nature conservation. The species considered in this work are protected throughout the European Union and their conservation is a main target, more than ever under the conditions of recent climate change. Distribution changes of animal and plant species, in situ changes of habitats as well as changes in communities and their biotic interactions have to be increasingly expected and can no longer be compensated or mitigated by established management concepts only. There is an increasing amount of literature concerning climate change impacts on organisms and ecosystems. This literature was surveyed to get an impression of the derived knowledge patterns so far and to detect potential knowledge gaps. The analysis reveals large imbalances concerning the spatial distribution of study areas, the studied taxonomic groups and ecosystems as well as the applied methods. Climatic changes are expected to have a relevant influence on the distribution of species. Changes in species’ distributions are already observed and attributed to the recent climate change at least for some species. In order to assess the degree of the awaited distribution changes climate envelope models have been increasingly used in the recent past. They put the spatial distribution of a species in relation to different environmental factors, such as climatic conditions. With their help potential impacts of a changing climate on a species’ distribution can be analysed. The present work uses these climate envelope models to estimate potential range changes of animal species. Beside the pure availability of new climatically suitable areas the accessibility and in situ establishment are main influencing factors concerning the estimation of the future potential distribution of a species. Accessibility is on the one hand determined by the speciesspecific dispersal ability and on the other hand by the permeability of the landscape. The establishment depends also on biotic conditions. Climatic suitability and accessibility of an area are insufficient if the individual cannot discover its essential interaction partner. Odonata are often recognized as good dispersers because of their flight ability. However, having a closer look, their dispersal ability may not be sufficient enough to keep up with the projected climatic changes. This is especially true for damselflies, for which potential suitable areas could develop in the future, but may be not able to reach them on their own or within the next decades because of their small dispersal distance and/or because of the far distance between current and projected areas. Hence, the integration of observed dispersal distances in future projections needs to be given special attention to constrain overall assumptions like ‘unlimited dispersal’ and to receive more realistic projections regarding a species’ dispersal potential. Additionally, biotic interactions need to be increasingly considered in modelling. However, the implementation seems to be problematic. For this reason I developed different approaches to integrate specific interactions in the modelling process and compared these with a model which neglects the interaction. I could show that considering biotic interactions leads to less projected suitable areas and larger potential losses of the target species than a negligence of essential interaction partners and therewith to potentially more realistic results. In the case of habitat types the question “How can we handle complex entities?” arises. To answer this question two principally different modelling approaches were developed: the indirect approach – modelling the distribution of a habitat type using the distribution of its characteristic plant species – and the direct approach – using the distribution of the habitat itself. Both approaches were tested by modelling five grassland habitat types defined by the EU Habitats Directive. Both approaches produce reasonable results, though the indirect approach is at least restricted by the required but actually lacking amount of plant distribution data. Methodological improvements of species distribution models are an essential step to receive more realistic results. However, the knowledge of ecological conditions required by a certain species, i.e. the assumptions about the niche, provides the basis for all models. Ecological demands can differ across large (such as continental) spatial scales and the current knowledge is mainly restricted to a few well-studied species. Hence, also in times of progressing climate change it is worth to focus on monitoring programs and experiments to gain further knowledge on a species` niche. The main focus of this thesis is on the estimation of potential distribution changes of protected animal species caused by climatic changes. It considers thereby not only the assumed relation between climatic conditions and the current distribution, but also integrates further distribution-determining aspects. For this purpose, different approaches were developed and compared. This work contributes to a more comprehensive understanding of the range influencing environmental factors in times of global climate change and therewith to an enhanced basis for actions for nature conservation measurements.