Lichen Habitat Detection : Evaluating Lichen Habitats in Northern Sweden Using Remotely Sensed Data and Machine Learning

Sammanfattning

Reindeer herding is a cultural practice carried out by the indigenous Sami people in theBarents region which depends heavily on the availability of lichen as a primary winter foodsource. How lichen react to climate change is an important factor to study to ensure a healthyand thriving Sápmi culture in Sweden. The purpose of this study was to analyse how criticallichen habitats for reindeer husbandry in northern Sweden changes via the usage of remotesensing data in tandem with machine learning. This was conducted by finding satelliteimagery of 2017 and 2024 from the Sentinel-2 satellite covering the reindeer husbandryregions. A Random Forest algorithm was trained and classified the satellite imagery; a changedetection map was produced following this in Google Earth Engine. Results were assessedwith a confusion matrix and four statistical accuracy metrics, which all indicate that theclassifier has a high accuracy and performs consistently. This study finds indications ofdiminishing habitats where lichen thrives, primarily in, but not limited to, the northern parts ofthe study area with possible consequence of reducing ideal grazing grounds for reindeer andexerting pressure on remaining lichen habitats.