Interpreting co-variation in species richness and productivity in terrestrial vegetation: Making sense of causations and correlations at multiple scales

This commentary examines conventional and recent ideas regarding the interpretation of species richness-productivity relationships (SRPR) in terrestrial vegetation. A new conceptual modelling approach cascading graph diagrams - is used to search for maximum parsimony by distilling and clarifying synthetic linkages between several potential causes of variation and co-variation in these two vegetation attributes at three distinctly different spatial scales: global/continental scale (variation between climatic/geographic regions within a continent, or across latitudes); regional scale (variation between local communities representing different habitat (soil) fertility types or different habitat disturbance levels within a climatic/geographic region); and local community scale (variation between neighbourhood plots within a particular plant community/habitat). In contrast with a number of interpretations in recent literature, the approach developed here emphases that SRPR at each scale in terrestrial vegetation involve a "cascade" of several intermediary causational variables that have not been generally accounted for in previous studies of SRPR. Accordingly, SRPR are expected usually to be correlational, sometimes indirectly causational, but never directly causational, at any scale. Rather than suggesting that causational mechanisms "scale up", the analysis here illustrates that several mechanistic features may be shared across scales and that in some cases, mechanisms may "scale down". This has crucial implications for identifying testable and un-confounded hypotheses for future research and for selecting effective experimental designs and appropriate methods of data analyses for the interpretation of SRPR in natural vegetation.