Selective logging is one of the most prevalent land uses of forests worldwide, affecting biodiversity and ecosystem functioning. However, the effect of selective logging on the dual nature of temporal stability, and the scale dependence of this effect, remain to be elucidated. By conducting several decade-long experiments in temperate forest ecosystems, we tested the effects of selective logging on aggregate and compositional stability at multiple spatial scales. As expected, forest ecosystem stability at larger spatial scales was enhanced both by the stability of local scales (i.e. α stability) and asynchronous dynamics among local communities (i.e. spatial asynchrony). We found that the negative effects of selective logging on both facets of forest stability propagated from local to larger spatial scales due to reduced α stability and the biological insurance effects of α diversity. However, both spatial aggregate and compositional asynchrony were not affected by selective logging. Interestingly, despite the selective logging, α diversity still provided biological insurance effects for maintaining aggregate and compositional stability. Our results imply that selective logging may destabilize the aggregate ecosystem functioning and species composition of forest ecosystems at local and larger spatial scales. To our knowledge, this study provides the first evidence of the scale dependence of aggregate and compositional stability of forest ecosystems in response to selective logging. Our findings suggest that forest management should avoid excessive selective logging and strive to protect forest diversity to safeguard the sustainability of the functioning and composition of natural forest ecosystems at multiple spatial scales. Keywords: aggregate stability, compositional stability, diversity, ecosystem functioning, spatial asynchrony, stability