Analysis of Relationships Between Productivity and Diversity Using Experimental Results From the Long-Term Ecological Research Network - Robert B. Waide and Michael R. Willig

We convened a working group from September 10-16, 1996, in order to synthesize data available from the 18 Long-Term Ecological Research (LTER) sites and the literature to search for general patterns in the relationship between productivity and diversity of both producers and consumers. Twenty-two participants took part in this meeting, and two follow-up meetings were attended by a group of seven of the original participants. Our analyses included original data from LTER sites and comprehensive searches of the literature. In addition to analysis and synthesis of these original data, we used mathematical modeling to develop and test new analytical techniques and to propose new extensions of ecological theory.

The initial product from the working groups is a series of seven papers that have been submitted as a Special feature to Ecology. Waide and Willig (submitted) summarize the results of the working group and point out new research directions. Mittelbach et al. (submitted) provide a synthetic overview of published works (over 100 publications) addressing the relationship between species richness and productivity within the context of current theory. They conclude that considerable variation in the form of the relationship (positive linear, negative linear, unimodal, and random) exists regardless of taxonomic perspective (plants versus animals) or frame of reference (within communities, between communities, and between biomes). More specifically, a positive monotonic relationship was pervasive in studies that comprise data from two or more biomes, whereas the relationship was most commonly unimodal when data derive from two or more communities within a biome. No consistent relationship occurred when data were limited to replicate sites within communities. In contrast, Scheiner et al. (submitted) adopt a definition of scale that includes sampling unit, grain, focus, and extent. Based on exponential, power, and logistic functions, they develop a quantitative protocol for examining the scale-dependence of the relationship between species richness and productivity. From simulation analyses, they conclude that the logistic function is a more accurate reflection of the species area relationship over a range of species density values and species-abundance distributions. Using the frameworks established by Mittelbach et al. (submitted) and Scheiner et al. (submitted), Gross et al. (submitted) assess the relationship between species diversity (i.e., species density and species richness) and net annual primary productivity in low stature vascular plants communities at six sites in the LTER network. They document the scale-dependence of the relationship, and interpret their results in light of the pattern accumulation hypothesis. When the focus was on fields, observed patterns depended on the extent of the analysis (i.e., at best weakly negative and usually non-significant considering variation within community types, but consistently unimodal considering variation among fields that spans community types within grasslands or North America). When the focus was on community types or plots, relationships usually were non-significant. Dodson et al. (submitted) synthesized information on the interrelationships between species richness, area, and productivity for 33 well-studied lakes. Analyses of lacustrine phytoplankton, rotifers, cladocerans, copepods, macrophytes, and fish uniformly revealed a unimodal response by richness to productivity when controlling for lake area.. They further examined the effects of short- and long-term whole-lake fertilization on the relationships between productivity and species richness, and found them to be variable and taxon-specific. Gough et al. (submitted) focus on the effects of experimentally increased productivity on species density and species richness in predominantly herbaceous communities at seven LTER sites. They used a meta-analytical approach based on the ratio of mean response in fertilized plots to the mean response in control plots, to compare effects among communities and among LTER sites. In general, nutrient enrichment decreases species density by a constant amount regardless of initial levels of productivity, suggesting that the unimodal pattern is unlikely to exist at these scales, in part because of the absence of the opportunity for an evolutionary response by the biota. Moore and de Ruiter (submitted) develop a trophic dynamics model that predicts a modal relationship between species richness and productivity available to a system. Richness is constrained at low productivities because of a lack of sufficient energy to support higher trophic levels. At high levels of productivity, richness is low due to instabilities associated with oscillatory dynamics. At intermediate productivities the potential for a higher number of stable assemblages leads to increased diversity.