Abstract
As ecosystems face unprecedented change and habitat loss, pursuing comprehensive and resilient habitat restoration will be integral to protecting and maintaining natural areas and the services they provide. Microbiomes offer an important avenue for improving restoration efforts as they are integral to ecosystem health and functioning. Despite microbiomes' importance, unresolved knowledge gaps hinder their inclusion in restoration efforts. Here, we address two critical gaps in understanding microbial roles in restoration—fungal microbiomes' importance in “reconstructive” restoration efforts and how management and restoration decisions interactively impact fungal communities and their cascading effects on trees. We combined field surveys, microbiome sequencing, and greenhouse experiments to determine how reconstructing an iconic landscape feature—tree islands—in the highly imperiled Everglades impacts fungal microbiomes and fungal effects on native tree species compared with their natural counterparts under different proposed hydrological management regimes. Constructed islands used in this research were built from peat soil and limestone collected from deep sloughs and levees nearby the restoration sites in 2003, providing 18 years for microbiome assembly on constructed islands. We found that while fungal microbiomes from natural and constructed tree islands exhibited similar diversity and richness, they differed significantly in community composition. These compositional differences arose mainly from changes to which fungal taxa were present on the islands rather than changes in relative abundances. Surprisingly, ~50% of fungal hub taxa (putative keystone fungi) from natural islands were missing on constructed islands, suggesting that differences in community composition of constructed island could be important for microbiome stability and function. The differences in fungal composition between natural and constructed islands had important consequences for tree growth. Specifically, these compositional differences interacted with hydrological regime (treatments simulating management strategies) to affect woody growth across the four tree species in our experiment. Taken together, our results demonstrate that reconstructing a landscape feature without consideration of microbiomes can result in diverging fungal communities that are likely to interact with management decisions leading to meaningful consequences for foundational primary producers. Our results recommend cooperation between restoration practitioners and ecologists to evaluate opportunities for active management and restoration of microbiomes during future reconstructive restoration.