The capacity of forests to resist structural change and retain material legacies–the biotic and abiotic
resources that persist through disturbance–is crucial to sustaining ecosystem functioning after disturbance.
However, the role of forest structure as both a material legacy and feature supporting carbon (C) cycling
stability following disturbance has not been widely investigated. We used a large-scale disturbance
manipulation to ask whether LiDAR-derived canopy structures as material legacies drive 3-year responses
of NPP to a range of disturbance severity levels. As part of the Forest Resilience Threshold Experiment
(FoRTE) in northern Michigan, USA we simulated phloem-disrupting disturbances at a range of
severities and two disturbance types. We quantified the legacies of forest structure using two approaches:
one measured change in structure and primary production from pre- to post-disturbance and the second
estimated resistance as log transformed ratios of control and treatment values. We found that total
aboveground wood net primary production (ANPPw) remained similar across disturbance severities as
remnant trees rapidly increased rates of primary production. Experiment-wide, disturbance had limited
effects on change in mean structural complexity values; however, high variance underscored large
differences in the magnitude and direction of complexity’s response at the plot-scale. Plot-scale structural
complexity, but not VAI, resistance strongly predicted ANPPw resistance while temporal VAI and
structural complexity changes did not. We conclude that the presence of material legacies in the form of
forest structure may affect primary production stability following disturbance, and that how legacies are
quantified may affect the interpretation of disturbance response.