Spatial analysis of remnant tree effects in a secondary Abies-Betula forest on the eastern edge of the Qinghai–Tibetan Plateau, China

•Juveniles of shade-tolerant species show spatial attractions to remnant trees.•Juveniles of shade-intolerant species show spatial repulsions to remnant trees.•Our results are in part consistent with the Janzen–Connell hypothesis.•Juveniles distribute within a particular range of annuli around remnant trees.•Remnant trees play a significant role for the restoration of Abies-Betula forests.

Remnant trees, remaining after large-scale disturbance of forests, have been shown to exhibit noticeable ecological effects on the recovery of ecosystem functions. Although it is well known that remnant trees may serve as dispersal foci, their effects on the spatial pattern of later stages of dominant tree species in natural secondary forest have not been fully appreciated. To explore the spatial influences of remnant trees on later stages of dominant tree species of a secondary Abies-Betula forest on the eastern edge of the Qinghai–Tibetan Plateau, we mapped all the trees in a 200 m × 200 m plot. We used spatial point pattern analysis including O-ring statistics to analyze the spatial associations between old-growth remnant trees and two dominant tree species of later stages. Abies faxoniana and Betula albo-sinensis, comparing different size-classes at different spatial scales. Our results showed that saplings, small trees, and medium trees of the shade-tolerant species, A. faxoniana, showed significantly positive associations to the old-growth remnant trees of both dominant species, suggesting that remnant trees provided favourable sites for the regeneration of shade-tolerant species. In contrast, small, and medium trees of the shade-intolerant species, B. albo-sinensis, showed significantly negative associations to old-growth remnant trees of both dominant species, suggesting that the remnant trees hinder their regeneration. For old-growth remnant trees, the O12(r) function value (density) of their seedlings, saplings, small trees, and medium trees was calculated at increasingly greater concentric scales (with radius from 0 to 100 m from the remnant trees). Tree density first increased rapidly from 0 to 10 m or 20 m radius (distance from remnant tree) and then decreased continually away from remnant trees, consistent in part with the population recruitment curve proposed by the Janzen–Connell model. We found that saplings, small trees, and medium trees of A. faxoniana tended to distribute within a particular range of annuli around the remnant trees of A. faxoniana, with the densities of the three largest size-classes reaching a maximum within 3–6 m of the remnant tree. The densities beyond the parent remnant trees continued to show an increasingly expanding zone of attraction around the remnant tree. In effect, the small and medium A. faxoniana trees coalesced around remnant trees to create increasingly enlarged canopy cover areas, which served simultaneously as protected zones for A. faxoniana regeneration and extended exclusion zones for B. albo-sinensis and similar shade intolerant species.