Disturbance is a central driver of forest development and ecosystem processes with variable effects within and across ecosystems. Despite the high levels of variation in disturbance severity often observed in forests following natural and anthropogenic disturbance, studies quantifying disturbance impacts often rely on categorical classifications, thus limiting opportunities to examine potential gradients in ecosystem response to a given disturbance or management regime. Given the potential increases in disturbance severity associated with global change, as well as shifts in management regimes related to procurement of biofuel feedstocks, there is an increasing need to quantitatively describe disturbance severity and associated responses of forest development, soil processes, and structural conditions. This study took advantage of two, replicated large-scale studies of forest biomass harvesting in Populus tremuloides and Pinus bansksiana forests, respectively, to develop and test the utility of a continuous, quantitative index of disturbance severity (DSI) for describing post-harvest response of plant communities and nutrient pools to different levels of biomass removal and legacy retention (i.e., live trees and coarse woody material). There was a high-degree of variability in DSI within categorical treatments associated with different levels of legacy retention and regression models using DSI as a predictor explained a portion of the variation (>50%) for many of the ecosystem- and community-level responses to biomass harvesting examined. Nutrient losses associated with biomass harvesting were positively related to disturbance severity, particularly in P. tremuloides forests, with post-harvest nutrient availability generally declining along the gradient of impacts. Consistent with expectations from ecological theory, species richness and diversity of woody plant communities were greatest at intermediate disturbance severities and regeneration densities of dominant trees species most abundant at highest levels of disturbance. Although categorical benchmarks will continue to be the primary way through which management guidelines are conveyed to practitioners, evaluation of their effectiveness at sustaining ecosystem functioning should be through continuous analyses, such as the DSI approach used in this study, to allow for the identification of minimum benchmarks that ensure a range of desirable outcomes exist across managed landscapes.