We used data on epiphytic lichen communities in 1215, 0.4-ha plots in the Southwest U.S.A. collected by the Forest Inventory and Analysis (FIA) program to analyze relationships with climate. We sought the climate variables most strongly associated with differences in epiphytic macrolichen communities and described the nature of those relationships, including diversity, community composition, and patterns in individual species. Five lichen community groups were strongly related to temperature and elevation gradients, overall moisture, and summer rain. Lichen abundance was highest in the wettest groups and lowest in the hottest and driest groups. Warm summer monsoonal climates supported the greatest number of species across all plots and within plots. The monsoonal pattern did not occupy a discrete geographic area, but instead formed a gradient, strongest in the southern part of our study area, diminishing to the north and west. In contrast, hot summer monsoonal climates had much lower within-plot richness. Hot, dry climates had the most variation in species composition among plots, but the fewest species within each plot and across all plots. Lichen community gradients had nonlinear relationships with combinations of climate variables rather than strong linear relationships with any single variable, including those derivative climate variables meant to have direct biological relevance. Relationships between air quality and community gradients were weak, potentially overwhelmed by regional climatic variation and complex topographic gradients. Richness of particular functional groups was more strongly related to climate than was overall species richness; functional groups have their own climatic tolerances, owing to the physiological consequences of growth form and photobiont. Presumably species in different functional groups have experienced their own evolutionary tradeoffs, developing peak performance in different climates. On the other hand, overall richness was driven by an even more complex combination of performances relative to climate and was in some functional groups more strongly related to geographic coordinates than to climate variables. Because climatic variables are themselves geographically structured, stronger model fit for geographic coordinates than for climate implies some influence of large-scale historical factors (i.e., factors not clearly expressed in modern climates, such as past climates, vegetation structure, or disturbance regimes).
Vol. 125 • No. 1
Vol. 125 • No. 1
nonmetric multidimensional scaling
nonparametric multiplicative regression