Top-k comparison between spatial regions with respect to species occurrence

Abstract

Indicative species allow in their abundance the classification of biotopes, a spatially limited area with adapted species. However, because species interact between biotopes, it is also interesting to group biotopes together that are functionally or spatially related to each other, called biotop complexes, leading to implications for nature conservation measures. The classification of biotop complexes poses some challenges, as appropriate methods are still lacking. Building on existing studies from Edler et al. 2016 and Pang et al. 2023 this study presents a new framework that supports the objective identification of indicative species in biotop complexes and enables the searchability of biotop complexes using a localized search engine. This novel approach of using a search engine allows for a finer granularity of resolution, as biotop complexes were not sorted into rigid classes, but individually compared to each other based on a similar indicator species composition, which allow a more precise adaption of nature conservation measures. As a proof of concept, this thesis therefore attempts to answer the question whether biotop complexes with a similar indicator species composition produce similar indicative species in the framework. To support this study, an example dataset was created based on biotop locations in the National Park Kellerwald-Edersee and plant species records from Becker, Frede, and Lehmann 1996 of the national park area. For the implementation, the national park area was divided into cells and in each cell the indicative species were grouped and determined by an indicator species analysis. The indicative species were stored in the search engine and thus allowed the search of cells and therefore biotop complexes by their indicative species. The results showed that the predictions of indicative species by the framework depended on the cell size, as the cell size affected properties such as the biotopes intersecting with the cell or the number of plant species found by the framework. The number of biotop complexes with similar biotop compositions, as well as the similarity of their framework predicted indicative species, also depended on the cell size, while the overall similarity of the indicative species in these biotop complexes tended to be low. Additionally as the creation of the example dataset and the determination of the indicative species show a huge influence on the predictive capabilities of the framework, this thesis offers connection possibilities for future research. This thesis provides a solid foundation upon which future research can be built. The next steps to optimize the various customization options provided by the framework and to determine the optimal cell size to increase the significance of the framework's predictions allowing the framework to support future nature conservation management.