Parameterization of biodiversity-productivity relationship and its scale dependency using georeferenced tree-level data

The biodiversity-productivity relationship (BPR) constitutes one of the most fundamental yet challenging topics in ecology. Most described BPRs so far are based on relatively small grain sizes (typically smaller than 1ha), and understanding how the BPRs scale up from individual trees to communities of different sizes, a mission critical for biodiversity conservation and forest management, still remains elusive. Based on a unique 54.12-ha large-scale forest observational site in Northeastern China, where we georeferenced and measured approx. 90,000 trees in two consecutive inventories, we studied BPRs across a cascade of spatial scales using geostatistical analysis and bootstrapping. To explore the underlying mechanisms of the scale dependency of BPR, we further investigated the scale gradient in biodiversity and productivity, respectively, and assessed neighbourhood influences on individual trees. Across all the spatial scales, we found a consistent positive concave-down effect of tree species richness on forest productivity. The elasticity of substitution (theta) represented the degree to which species can substitute for each other in contributing to forest productivity. At the community level, the curvature of BPRs in terms of the elasticity of substitution (theta) declined from around 0.50 and converged at around 0.15, as the grain size increased from 0.01 to 1.00 ha, all else being equal. Similarly, at the individual tree level, neighbourhood diversity coefficient, a linear proxy of theta, declined from around 0.08 and converged at around 0.03, as neighbourhood size increased from 0.01 to 1.00 ha, all else being equal. Synthesis. Our study found a pronounced scale dependency of biodiversity-productivity relationship and quantified, for the first time, how biodiversity-productivity relationship scale up from 0.01 to 1.00 ha. Our findings suggest that biodiversity-ecosystem functioning relationship can be generally scale-dependent and, hence, one must factor in this effect in biodiversity conservation and ecological restoration projects. Our findings also indicate that biodiversity-productivity relationship, predominantly determined by biological processes (i.e. complementarity and diminishing marginal productivity) at a small scale, can be increasingly influenced by statistical (i.e. spatial central limit theorem) and other (such as species-area relationship) processes as the scale increases.