Glacier foreland succession and the fading effect of terrain age

Abstract

AbstractQuestionsWhat are the most important factors explaining present‐day variation in species composition in a glacier foreland? Does the rate of species compositional change in glacier forelands decelerate through primary succession? How do data set properties and analytic methods influence our understanding of glacier foreland successional dynamics?LocationNigardsbreen glacier foreland, western Norway.MethodsWe sampled the species composition and recorded 21 explanatory variables in 74 plots, distributed on eight classes of terrain age (dated moraines). Gradients in species composition found using global non‐metric multidimensional scaling (GNMDS) ordination were interpreted with split‐plot generalized linear models. Yearly succession rates were calculated from plot positions along a vector of maximum compositional change related to terrain age in the interpreted GNMDS ordination.ResultsWe interpreted the main gradient in species composition as being related to a complex gradient with soil moisture and soil nutrients as primary constituents. Terrain age contributed to the second most important gradient. Succession rates were nonlinear with time. Terrain age explained variation in species composition only when plots from the two youngest terrain age classes were retained in the data set.ConclusionsIn contrast to the majority of studies of glacier foreland successions performed so far, we find that terrain age is not the principal factor that explains present‐day variation in species composition. Instead, local environmental variables are the main determinants of species composition. This result emphasizes the importance of taking environmental gradients into account when variation in glacier foreland vegetation is studied. The limited importance of terrain age in our study is interpreted as likely due to this glacier foreland being situated below the tree line, the relatively long distance between the bulk of the studied foreland and the glacier snout, and inclusion of few plots from young terrain in our data set. The non‐linearity of succession rates with time implies that a linear time‐since‐deglaciation variable is inappropriate for constrained ordination of glacier foreland vegetation.