Native forest recovery following low‐impact Eucalyptus globulus removal in Galicia: Evidence from a 17‐year chronosequence

The replacement of native forests by exotic tree species planted for forestry threatens biodiversity, ecosystem functioning and fire resilience across many regions. In Galicia, extensive plantations of Eucalyptus globulus have displaced native Atlantic forests, yet the long‐term effectiveness of Eucalyptus removal for ecological restoration remains poorly understood.

Vegetation composition, plant diversity, canopy structure and soil properties were assessed across a 17‐year chronosequence following E. globulus felling using field vegetation surveys, soil physicochemical analyses and soil environmental DNA (eDNA) metabarcoding, with historically managed native forest stands as reference sites. This targets a specific knowledge gap assessing the role of eDNA metabarcoding in monitoring forest recovery trends using above‐ground taxa.

Plant community composition differed significantly among Eucalyptus , intervention and native forest plots. Intervention plots showed compositional change away from Eucalyptus site assemblies and supported a more diverse ground‐flora layer than Eucalyptus plots. However, early successional species, particularly bracken fern ( Pteridium aquilinum ), remained abundant and late‐successional woodland species were underrepresented relative to native reference sites.

Canopy structure showed limited recovery post‐intervention, with non‐linear trajectories indicating slow or uncertain mid‐ to long‐term canopy development after Eucalyptus removal.

Soil properties improved over time after intervention, with increased soil porosity and organic matter content suggesting recovery of soil structure facilitated by low‐impact removal methods. Soil eDNA and field surveys revealed contrasting community signals, indicating methodological differences in detecting restoration progress.

Practical Implication . At this site, complete removal of Eucalyptus globulus was followed by the recovery of plant community composition and key soil properties. Native reference sites in this study reflect historically managed broadleaf woodland rather than mature forest, and the true restoration debt could be greater than chronosequence comparisons suggest. Low‐impact felling methods preserved soil integrity and may have supported regeneration from legacy vegetation. To accelerate succession and woodland recovery, managers could consider complementary actions such as assisted regeneration or targeted planting of late‐successional species; however, targeted research in these areas would provide valuable further guidance. These findings suggest that soil eDNA is best used as a complementary tool rather than a standalone method for monitoring restoration outcomes in this context.