Widespread changes to near-surface permafrost in northern ecosystems are occurring through gradual top-down thaw and more abrupt localized thermokarst development. Both thaw types are associated with a loss of ecosystem services, including soil hydrothermal and mechanical stability and long-term carbon storage. Here, we analyzed relationships between the vascular understory, basal moss layer, active layer thickness (ALT), and greenhouse gas fluxes along a thaw gradient from permafrost peat plateau to thaw bog in Interior Alaska. We used ALT to define four distinct stages of thaw: Stable, Early, Intermediate, and Advanced, and we identified key plant taxa that serve as reliable indicators of each stage. Advanced thaw, with a thicker active layer and more developed thermokarst features, was associated with increased abundance of graminoids and Sphagnum mosses but decreased plant species richness and ericoid abundance, as well as a substantial increase in methane emissions. Early thaw, characterized by active layer thickening without thermokarst development, coincided with decreased ericoid cover and plant species richness and an increase in CH4 emissions. Our findings suggest that early stages of thaw, prior to the formation of thermokarst features, are associated with distinct vegetation and soil moisture changes that lead to abrupt increases in methane emissions, which then are perpetuated through ground surface subsidence and collapse scar bog formation. Current modeling of permafrost peatlands will underestimate carbon emissions from thawing permafrost unless these linkages between plant community, nonlinear active layer dynamics, and carbon fluxes of emerging thaw features are integrated into modeling frameworks.

Fungal communities can be used as indicators of various environmental processes in forest ecosystems. The diversity of these communities is linked to aboveground plants and soil properties. We assessed fungal diversity at four Norway spruce sampling sites that were growing on fertile mineral soils (Oxalidosa) in northwestern Latvia. Three sites were managed-a three-year-old clear-cut and fifty- and eighty-five-year-old stands; one site was unmanaged-a naturally regenerated site after wind damage in 1969. For metabarcoding, we used a fungal internal transcribed spacer (ITS2) and high throughput sequencing with the Ion Torrent platform. Our results showed high operational taxonomic unit richness in the samples, with notable variation in community composition between individual plots both within and among sites, with the highest being in managed, middle-aged stands and the lowest in unmanaged. Significant differences in the diversity of soil fungal communities were not detected between the sites. Redundancy analysis indicated that pH, soil organic matter, organic carbon, and nitrogen were the most important soil variables that explained the variation in fungal communities. The unmanaged stand differed notably by community composition. This study highlights the importance of monitoring forest soil environmental parameters and fungal communities to gain a more comprehensive assessment of forestry management regimes.