Abstract
Climate-driven tree mortality impacts forest ecosystems, yet studies rarely integrate above- and below-ground processes. Consequently, the temporal dynamics of rhizosphere microbial communities following mortality and their link to vegetation changes remain unclear. This study examined coupled shifts in species composition and rhizosphere microbial communities following multiple-tree mortality events in Pinus densiflora stands. We employed a chronosequence approach by reconstructing tree mortality timing and analyzing temporal changes in above-ground species composition alongside rhizosphere microbial diversity. Random Forest (RF) traced microbial community transitions from healthy to declining and dead trees, identifying core genera indicative of tree health, followed by an analysis of co-occurrence network dynamics along the health-decline-mortality sequence. As a result of this study: (1) Despite increased light from canopy decay, understory changes remained subtle. However, 15 years post-mortality, snag collapse expanded gaps, promoting active P. densiflora regeneration and signaling potential recovery. (2) Rhizosphere microbial α-diversity in healthy trees was lower or similar compared to declining and dead trees. (3) RF analysis identified the abundance of specific microbial groups-notably Oidiodendron spp. (Ascomycota) and Umbelopsis spp. (Mucoromycota)-as a key indicator of healthy P. densiflora forests. This finding suggests that specific functional microbial composition (key microbial groups associated with tree health), rather than overall diversity, is a critical factor for tree health, which is supported by our α-diversity results showing that healthy forests had lower diversity than declined forests. (4) In fungal communities, compared to healthy P. densiflora stands, declining and recently dead stands (3 years post-mortality) formed denser and more complex networks (more nodes and edges). Notably, this complexity was characterized by significantly lower modularity within the largest connected component, despite exhibiting similar or higher whole-network modularity. In contrast, networks at the oldest mortality sites (~15 years) became smaller and simpler, with similar modularity and resembling healthy stands at the whole-network level. By examining temporal above- and below-ground changes, this study provides insights into forest regeneration processes, emphasizing above-ground-below-ground interactive dynamics and highlighting that the abundance of specific indicator microbial groups, rather than overall microbial diversity, plays a more critical role in assessing tree health.