On the use of observational data in studying biodiversity-productivity relationships in forests

Abstract

The biodiversity-productivity relationship is one of the focus areas in ecological research that is studied primarily through mixed species experiments. Recent efforts in forests, however, increasingly involve the use of observational data, due to the difficulty in establishing long-term, multispecies plantations. Caution is warranted in the observational databased causal relationships between biodiversity and productivity due to the potential confounding effects by environmental variations. In this article, we use a recent forest example to demonstrate how erroneous results could be generated in studying biodiversity-forest productivity relationships when species diversity is highly correlated with environmental variables (multicollinearity). In forestry, erroneous biodiversity-productivity relationships can mislead future research, industry decisions, and policy development. Forest researchers and managers should be aware of the issues associated with collinear data and validate research results with literature reports and professional knowledge. Options to deal with observational data are discussed.