Although succession has a long history in ecology, rigorous tests of unifying, generalizable principles are rare. A lack of integrated longitudinal studies following major disturbances has constrained the development of successional theory that is consistent, robust, and empirically-based. Studies of community assembly suggest that stochastic, dispersal-dependent, and site-specific patterns are important during succession. However, relatively few studies of succession after disturbance are based on multi-decade, longitudinal observations, or have integrated such observations along a gradient of disturbance. The Mount St. Helens (Washington, USA) volcanic eruption in 1980 created a unique opportunity to test long-held successional theories using a disturbance severity gradient ranging from a “blank slate” to minimal impact. The eruption created primary and secondary successional habitats across a mountain landscape encompassing variation in disturbance history, local topography, resource availability, biological legacies, and proximity to seed sources. We combine data from three long-term (35-year) permanent-plot networks designed to examine plant community structure and diversity across a disturbance gradient that includes 1) primary succession blast zone, 2) secondary succession tree-blowdown zone, and 3) secondary succession tephra-deposit zone (intact forest). We examine temporal trends in species diversity and composition to understand how community stability changes through time across the disturbance severity gradient.
Across the three disturbance zones, species richness increased early in succession (at varying rates) but major components of the community reached a leveling off point at different rates 5-25 years after the eruption. Differences among sites depended on disturbance severity, pre-disturbance history, and microsite characteristics. Species turnover was greatest at the high end of the disturbance gradient (blast zone) characterized by the largest changes in mean rank abundance. These changes were driven by unstable population dynamics of the dominant species (Lupinus lepidus), possibly due to competitive interactions or herbivore pressure. High species turnover in the blowdown zone did not correspond with strong or consistent directional change. Across all three disturbance zones, the community stability metrics (species covariance and asynchrony) that account for species richness showed a significant negative relationship between community stability and variance ratio. This indicates that despite a lack of change in species richness, communities across all disturbance zones have not reached stability in aggregate cover or composition over 35 years. We predict that as woody species increase in the blast and blowdown zones, species turnover and community stability will continue to increase compared to the smaller changes in the less-disturbed tephra communities.