Beyond Vegetation Classes Toward a Trait-Based Understanding of Tundra CO₂ Fluxes

Ongoing shifts in tundra vegetation are reshaping Arctic ecosystems, but their impact on carbon exchange remains poorly captured by coarse vegetation classifications. Consequently, it remains unclear how CO₂ fluxes relate to continuous metrics describing vegetation structure. Here, we quantified fine-scale variation in summertime ecosystem respiration (ER20), gross primary productivity (GPP600), and net ecosystem exchange (NEE600) across a heterogeneous tundra landscape and tested how 21 different vegetation structure and functional traits in comparison to vegetation classes explain flux variability. CO₂ fluxes, measured using chamber observations, and vegetation and environmental parameters were collected from multiple plots (n = 211) over multiple field campaigns (2017–2025). CO₂ exchange varied widely among the plots, with GPP600 showing greater variation than ER20 and NEE600 ranging from net CO₂ uptake to minor net CO₂ release. Coarse vegetation classes captured community-level differences in CO₂ exchange, explaining 27%, 34%, and 25% of variation in ER20, GPP600, and NEE600, respectively, with deciduous tall shrub communities showing the highest carbon uptake and respiration, while evergreen dwarf shrubs showed the lowest fluxes. Continuous vegetation properties, however, explained plot-level flux variation more effectively, with the best predictors (e.g. green fraction, maximum vegetation height and leaf nitrogen) accounting for 60% of variation in ER20, 77% of GPP600, and 56% of NEE600. Vegetation variables explained most of the flux variance, with soil temperature and moisture adding ≤4% to marginal R². These findings highlight the importance of continuous vegetation properties for capturing and understanding spatial variation in plot-level CO₂ exchange.