The rewilding of Chernobyl: A long term monitoring of passive rewilding after land abandonment using remote sensing

Sammanfattning

In 1986, the nuclear disaster in Chernobyl caused large-scale land abandonment which resulted in one of the largest unintended experiments of passive rewilding, yet no study has combined remote sensing land cover classification using a pre-disturbance baseline with vegetation productivity analysis across the full Chernobyl-Palieski Reserve under an explicit rewilding framework. This thesis quantifies land cover change and vegetation productivity dynamics between the Chernobyl-Palieski Reserve and its surrounding buffer zone between 1986 and 2018. To achieve this, a locally trained Random Forest classification of Landsat imagery is performed and the Enhanced Vegetation Index 2 is used as a proxy for vegetation productivity. Inside the reserve, 89.2% of pixels that were not classified as vegetation in 1986 had transitioned to vegetated cover by 2018 through spontaneous ecological succession, while only 49.8% did so in the surrounding area, with agriculture to shrubland as the dominant transition. Vegetation productivity increased across both zones, with the reserve displaying slightly higher values, consistent with human exclusion as the primary driver. Cross-product comparison with a global land cover product reveals that global products may be insensitive to the transitional shrubland stage through which passive rewilding occurs. Findings from this thesis contribute a large-scale remote sensing case to the emerging evidence base for passive rewilding and highlight limitations of global monitoring products to capture transitional vegetation classes central to ecological succession.