SWC Research @TusnadEcoBear 2024

A species conservation simulation system based on knowledge graph for brown bear movement prediction

Presenter: Albin Ahmeti
Time: Oct 22nd, 13:35

Species conservation, particularly for the brown bear, presents significant challenges. These challenges are typically addressed at local or regional levels through best practices aimed at minimizing human-wildlife conflicts, such as the installation of fences or the removal of problematic bears. In this talk, we model the potential location of identified bears using a knowledge-driven approach based on Knowledge Graphs (KGs). KGs offer a rich data structure that integrates, connects, and harmonizes diverse and disparate data sources, in a way that enables semantic understanding and logical reasoning capabilities.

As part of the Nature FIRST research project, we have developed the Nature First KG, a knowledge graph designed to support biodiversity conservation by linking data on species, habitats, and Natura 2000 sites. Additionally, the KG connects ecosystem basemaps produced within the Nature FIRST—map layers that encode information on habitats, elevation, slope, and other ecological or environmental factors—alongside the Brown Bear Habitat Suitability Index, a weighting scheme that evaluates habitat suitability based on these features. Therefore, the information about the distribution of factors as well as the preference of the bear on the factors is contained in the Nature First KG.

Using this infrastructure, we have built simulations to predict brown bear trajectories. The simulation is based on generating a large number of random points and tracking their future trajectories. The simulation is probabilistic, i.e. the next step in the trajectory of each bear is modelled as an event with multiple outcomes, the probability of each outcome is influenced by the Brown Bear Habitat Suitability Index.