Code.

Monastery forests in the Ethiopian highlands serve as cultural sanctuaries and critical refugia for Afromontane biodiversity. We conducted floristic inventories (n = 90 nested 20 m × 20 m plots) and demographic censuses (2039 woody stems) across five forest fragments (6.7–82.1 ha) to assess α- and β-diversity, evaluate population viability of key canopy species, and analyse the influence of topography and anthropogenic disturbance on community composition. Coverage-based rarefaction for species richness (Hill q = 0) estimated standardized estimate 32 ± 3.7 to 78 ± 5.2 species ha ⁻ ¹ while Shannon diversity (q = 1) ranged from 3.17 to 3.86 effective species. A Sørensen-based PERMANOVA confirmed significant compositional differentiation among forests (F₄,₈₅ = 3.66, P = 0.001), pair-wise turnover values (β_sim = 0.35–0.62, mean ± SE = 0.48 ± 0.05). Canonical correspondence analysis explained 17.6% of constrained variation, with slope (pseudo-F = 5.9, P = 0.004) and grazing intensity (F = 3.8, P = 0.012) and disturbance (F = 3.1, P = 0.018) emerging as the predominant, non collinear predictors of composition. Olea europaea subsp. cuspidata declined in two fragments (λ ≤ 0.91 yr ⁻ ¹) where seedling: adult ratios were ≤0.15, while Juniperus procera populations (λ = 1.03–1.11 yr ⁻ ¹) were predicted to remain stable to gradually increase by annual, stage-structured matrix models. We recommend a tiered stewardship strategy that combines livestock exclosures, enrichment planting of regeneration-limited taxa and participatory monitoring of vital rates. This study explains how local management modulates β-diversity across rugged terrain and establishes the first population-viability benchmarks for Ethiopia’s church forests by connecting floristic patterns with demographic processes and topographic context.