Question: Does the relative importance of stochastic and deterministic factors change during primary succession?
Location: Small depressions (potholes) located on Mount St. Helens, Washington (46°13′51″N, 122°09′10″W, 1290 m).
Methods: Pothole vegetation was described in 1993, 1997 and from 2001 to 2008. Explanatory variables included location and elevation (spatial factors), soil factors and Lupinus lepidus cover from prior years (a fertility surrogate). RDA assessed species-variable relationships. DCA calculated β diversity and within-year heterogeneity. Flexible sorting classified the vegetation. Species composition, richness, cover, H′ and evenness were also calculated.
Results: Vegetation cover increased through 2001, and then fluctuated due to changes in L. lepidus cover. Richness peaked in 2005, after which pioneer species began to decline as persistent evergreen species increased. The six CTs recognized in 2008 were more scattered than were the six different CTs from 2001. DCA demonstrated that woody and rhizomatous species increased as pothole vegetation became less variable. RDA revealed weak spatial relationships in 1993, 1997 and 2001; thereafter, environmental and biological factors became important. The species-explanatory data relationship increased during this study from 10.2% to 36.0%, leaving 64.0% of the variation unexplained.
Conclusions: This is the first temporal study to demonstrate that deterministic control of vegetation development increases during succession. Pothole vegetation has converged somewhat due to deterministic factors, but the initial effects of chance, local disturbances and history remain large and may prevent strong convergence.