Ecological resilience of physical plant–soil feedback to chronic deer herbivory: slow, partial but functional recovery

Ecological resilience to ungulate overbrowsing is an important issue in
forest ecosystem. After chronic herbivory, the recovery rate of understory
vegetation and its related functions can be slow even with decreasing
grazing intensity; thus, detecting elasticity during alternative
successional trajectories is fundamental to understanding state
perturbations. In this context, we focused on physical plant–soil feedback
(functional interactions between plant growths and soil physical
conditions), and evaluated elasticity and recovery processes according to
deer density. The effects of 40-year chronic herbivory by sika deer
(average density 14.7 individuals km-2) on the recovery of understory
plant communities and associated improvements in soil physical properties
in headwater catchments were assessed. Using 8 years of catchment-wide
exclusion (fenced) and reduction (only culled; average 4.3 individuals
km-2) treatments, plot sampling was conducted in 2010 (before treatment)
and 2018 (after treatment). The recovery of vegetation and soil physical
properties were evaluated, and functional plant–soil relationships and
spatial variability were assessed to detect recovery processes during
alternative successional trajectory. Woody species increased only under
the exclusion treatment and the average soil bulk density was lower than
that under reduction treatments. Soil bulk density was negatively
correlated with root biomass in the fenced catchment, and root biomass was
positively associated with woody species richness. Reduced soil bulk
density (~0.5 g cm-3) was observed with greater root biomass and woody
species richness on upper hillslopes in the deer-excluded catchment where
plant coverage was minimal. Successional failure under the reduction
treatment suggested slow recovery with a depressed threshold according to
deer density, indicating a clockwise hysteretic response to deer density.
Unlike plant coverage during the earlier period of overbrowsing, woody
species root development led the recovery of functional physical
plant–soil feedback; however, this was probably limited by the higher soil
erosion rate in riparian areas and an underdeveloped herb layer. Our
results highlight an alternative recovery trajectory of physical
plant–soil feedback driven by an alternative plant element (woody roots)
to degradation trajectory with decreasing plant cover. However, riparian
erosion and herb layer would still suppress recovery. Therefore, recovery
might be slower at landscape scale.