Disturbance can both initiate and shape patterns of secondary succession by affecting processes of community assembly. Thus, understanding assembly rules is a key element of predicting ecological responses to changing disturbance regimes. We measured the composition and trait characteristics of plant communities early after widespread wildfires in Alaska to assess how variations in disturbance characteristics influenced the relative success of different plant regeneration strategies. We compared patterns of post-fire community composition and abundance of regeneration traits across a range of fire severities within a single pre-fire forest type– black spruce forests of Interior Alaska. Patterns of community composition, as captured by multivariate ordination with nonmetric multidimensional scaling, were primarily related to gradients in fire severity (biomass combustion and residual vegetation) and secondarily to gradients in soil pH and regional climate. This pattern was apparent in both the full dataset (n = 87 sites) and for a reduced subset of sites (n = 49) that minimized the correlation between site moisture and fire severity. Changes in community composition across the fire-severity gradient in Alaska were strongly correlated to variations in plant regeneration strategy and rooting depth. The tight coupling of fire severity with regeneration traits and vegetation composition after fire supports the hypothesis that disturbance characteristics influence patterns of community assembly by affecting the relative success of different regeneration strategies. This study further demonstrated that variations in disturbance characteristics can dominate over environmental constraints in determining early patterns of community assembly. By affecting the success of regeneration traits, changes in fire regime directly shape the outcomes of community assembly, and thus may override the effects of slower environmental change on boreal forest composition.