Abstract

Topic: Moose and forestry


Cloud-based AI Pipeline for Moose Damage Identification and Mitigation in Swedish and Finnish Forests

Karan Mitra1, Henrique Souza Rossi2, Ari Nikula3, Laura Saggiomo4, Juho Matala3, Pasi Rauto3, Ke Zhang5, Åke Olson5, Matias Hiltunen6, Antti Kukkonen6, Ellinor Lindmark7, Tobias Gramner7

  1. Department of Computer Science, Electrical and Space Engineering, Luleå University of Technology, Campus Skellefteå, Skellefteå, Sweden
  2. Pervasive and Mobile Computing, Department of Computer Science, Electrical and Space Engineering, Luleå University of Technology, Luleå, Sweden
  3. Natural Resources Institute Finland, Rovaniemi, Finland
  4. Department of Wildlife, Fish and Environmental Studies, Swedish University of Agricultural Sciences, Umeå, Sweden
  5. Department of Forest Mycology and Plant Pathology, Swedish University of Agricultural Sciences, Uppsala, Sweden
  6. Lapland University of Applied Sciences, Rovaniemi, Finland
  7. Swedish Forest Agency, Skellefteå, Sweden

Climate change's impact is expected to challenge the forest ecosystem dramatically, especially at higher latitudes. Current moose browsing patterns are already impacting the forests and will likely cause even more damage in the future due to climate change. For instance, forage availability and competition changes may induce moose browsing pressure and browsing damage. This challenge is further exacerbated by different climate mitigation strategies employed in Sweden and Finland. We assert the need for seamless data and knowledge sharing, developing standard moose browsing analysis models, and their integration within climate models, which may assist in efficient moose damage identification and management. Our research aims to build a cloud-based artificial intelligence pipeline for forest risk identification and mitigation. In doing so, an elastic (able to adapt to model and simulation workloads running on clouds) and cost-efficient system will be developed: i) to unify heterogeneous data sources in Sweden and Finland; ii) provide scientists with methods for scalable data management (e.g., upload, download, transform, and sharing); iii) provide an interface to develop moose browsing analysis models, irrespective of data size, type and computation requirements in a seamless and cost-efficient manner. The AI pipeline will enable the scientists to test their hypotheses, for example, by running simulations regarding forest damage; and iv) methods to visualize the results in an immersive way, such as using augmented and virtual reality to assess current and future moose damage scenarios. The visualization system will not only assist the scientists but also help them engage with stakeholders, such as authorities and forest owners, to make an informed assessment regarding possible actions to mitigate damage caused by moose browsing.