Predicting Understorey Vegetation Cover from Overstorey Attributes in Two Temperate Mountain Forests

Abstract

It is important to develop a predictive understanding for the environmental controls on understorey vegetation, which harbor most of the plant biodiversity and are the source of food and cover for wildlife. Forest succession models (i. e. gap models) representing overstorey dynamics are not commonly linked to mathematical models of understorey dynamics. This is surprising, given that understorey vegetation clearly responds to changes in the overstorey that result in changing light availability. One difficulty may lie in the coarse representation of light regime captured by most gap models. Linkage of overstorey-understorey models might be facilitated if the diameter structure of simulated stands could be used to drive understorey change, as a proxy for light and other influences. The objective of this study was to determine whether understorey vegetation cover can be adequately predicted by variables derived from overstorey diameter structure alone, or if canopy cover and light availability are important, from additional predictors. Field sampling was conducted at a montane and a subalpine study area in the Swiss Alps. We used regression analysis to assess the relative importance of various overstorey predictors for understorey cover and composition. In the subalpine study area, the relative dominance of graminoids increased with increasing light availability, at the expense of forbs. In the montane study area, forb cover increased sharply with increasing light, while graminoid cover remained at low levels. As a result, the relative dominance of graminoid species declined with increasing light levels. This difference is attributed to the presence of Adenostyles alliariae, a tall, large-leaved forb. The effects of changes in the physical environment on plant community composition were thus mediated by interspecific interactions. This makes it difficult to predict overstorey-induced changes in understorey species composition at the level of functional groups. At both study sites, diameter structure variables were found to provide a reasonable approximation of total understorey cover, cover of the more common species, and species richness. Models of understorey community composition often improved (0–31% increased predictive ability) with inclusion of variables representing the light environment. In the context of gap model development, the great complexity associated with improved representation of light availability must be weighed against the relatively low gain in predictive power that is likely to result. We recommend that efforts to include forest understorey dynamics in gap models begin by considering empirical relationships between understorey patterns and overstorey diameter structure.