Global pattern and local variation in species–area relationships

摘要: Aim  We conducted a meta-analysis of species–area relationships (SARs) by combining several data sets and important covariates such as types islands, taxonomic groups, latitude spatial extent, in hierarchical model framework to study global pattern local variation SARs its consequences for prediction. Location  One thousand nine hundred eighteen islands from 94 SAR studies around the world. Methods  developed generalization power-law model, HSARX which allows: (1) inclusion multiple focal parameters (intercept, slope, within-study variance), (2) use effect modifiers based on collection studies, (3) modelling between- variability. Results  The was average had wide confidence intervals. slope 0.228 with 90% limits 0.059 0.412. intercept, variability showed great heterogeneity result interaction modifying covariates. Confidence intervals these were narrower when other addition area accounted for, thus increasing accuracy predictions species richness. significant latitude, taxa island type indicated that ‘typical’ latitudinal diversity gradient can be reversed isolated systems. Main conclusions  relationship underlying provides good fit non-nested across vastly different scales taking into account allows researchers explore complex interactions among variables, explicitly scale dependence, make robust levels (island, study, global) associated prediction From perspective, it is not but matters.