Abiotic and biotic factors explain independent gradients of plant community composition in ponderosa pine forests

Abstract

General multivariate models have appeal when analyzing ecological systems because they have the potential to shed light on the relative importance of multifaceted factors. We hypothesized that plant community composition is a function of the interactions among three general constructs: abiotic and biotic factors and disturbance history. We sampled vegetation and soils on 75 randomly located 0.05 ha plots across a broad soil gradient within a 110,000 ha ponderosa pine forest landscape in northern Arizona to evaluate this hypothesis. The variation in abundances of 271 understory plant species was reduced to two orthogonal axes using non-metric multidimensional scaling. We evaluated the relationships between ordination results and environmental conditions using two types of structural equation models. ‘Specific’ models provided a detailed perspective of hypothesized relationships among observed variables. A ‘general’ model was evaluated to address the general hypothesis stated above. The general model employed composite variables, which specify the combined effects of multiple factors on a response, in order to evaluate the relative importance of the three theoretical constructs. The results of the final specific model suggested that organic matter and nitrogen explained unique variation in composition, but the ‘total effects’ of organic matter were not significant due to offsetting direct and indirect pathways. Soil texture and pH also constrained where many plant species occurred. Elevation was a strong determinant of species distributions, and its effects could be propagated through a network of processes involving soil mineral and organic properties. Litter depth explained unique variation in community composition, suggesting that plant species differ in their ability to colonize sites with different litter depths independent of the density of pine trees. Composition also differed under varying densities of pine trees since species vary in their tolerance of a competitive overstory. Finally, heavy grazing tended to favor the abundance of non-palatable species in the community. The results of the general model provided new insight into the underlying structure of independent plant community gradients in ponderosa pine forests. The model suggested that abiotic factors generated one gradient in community composition, whereas biotic factors, and to a lesser degree, disturbance history, generated a second independent gradient in community composition. Future evaluations of this composition model across a diversity of ecosystems could determine the generality and utility of this multivariate hypothesis.