Abstract
Disturbance and succession have long been of interest in ecology, but how landscape patterns of ecosystem structure and function evolve following large disturbances is poorly understood. After nearly 25 years, lodgepole pine ( Pinus contorta var. latifolia ) forests that regenerated after the 1988 Yellowstone Fires (Wyoming, USA ) offer a prime opportunity to track the fate of disturbance‐created heterogeneity in stand structure and function in a wilderness setting. In 2012, we resampled 72 permanent plots to ask (1) How have postfire stand structure and function changed between 11 and 24 yr postfire, and what variables explain these patterns and changes? (2) How has landscape‐level (among‐stand) variability in postfire stand structure and function changed between 11 and 24 yr postfire? We expected to see evidence of convergence beginning to emerge, but also that initial postfire stem density would still determine trajectories of biomass accumulation. After 24 yr, postfire lodgepole pine density remained very high (mean = 21,738 stems/ha, range = 0–344,067 stems/ha). Stem density increased in most plots between 11 and 24 yr postfire, but declined sharply where 11‐yr‐postfire stem density was >72,000 stems/ha. Stems were small in high‐density stands, but stand‐level lodgepole pine leaf area, foliage biomass, and live aboveground biomass increased over time and with increasing stem density. After 24 yr, mean annual lodgepole pine aboveground net primary production ( ANPP ) was high (mean = 5 Mg·ha −1 ·yr −1 , range = 0–16.5 Mg·ha −1 ·yr −1 ). Among stands, lodgepole pine ANPP increased with stem density, which explained 69% of the variation; another 8% of the variation was explained by environmental covariates. Early patterns of postfire lodgepole pine regeneration, which were contingent on prefire serotiny and fire severity, remained the dominant driver of stand structure and function. We observed mechanisms that would lead to convergence in stem density (structure) over time, but it was landscape variation in functional variables that declined substantially. Stand structure and function have not converged across the burned landscape, but our evidence suggests function will converge sooner than structure.