Predicting the ecosystem consequences of biodiversity loss at the landscape level

Abstract

The experimental and theoretical BEF research spurred by the growth of global change ecology led to a robust set of results that strongly suggest that withintrophic level changes in diversity are associated with changes in the magnitude and variability of ecosystem processes (Loreau et al. 2002; Loreau and Hector 2001). To be sure, many questions remain unanswered. To what extent are experimental results due to facilitation, niche complementarity, or sampling? Which components of biodiversity (is it functional diversity, the identity of species, or dominance) are the most important in determining the impacts of biodiversity loss? Compared to abiotic controls, do biodiversity controls over ecosystem functions matter? Is stability (low variability in ecosystem function in high diversity treatments) due to insurance (Yachi and Loreau 1999), statistical averaging (Doak et al. 1998), or other causes? These questions are not readily addressable by the small-scale experimental approaches that have typified BEF research, but large-scale experiments are impractical. Perhaps answers to these questions may come using observational studies and the BioMERGE frameworks. Lessons learned from the controversies surrounding early phases of BEF research warn against indiscriminate employment of the BioMERGE framework. The BioMERGE framework has been partially and successfully employed by Solan et al. (2004) to predict ecosystem function (specifically, biogenic mixing) in a marine (estuarine) ecosystem. Similar studies are under way for North American and European grasslands and New World tropical rainforests. Clearly, BEF is scaling up. Perhaps the next step should be to combine a traditional combinatorial BEF experiment with a BioMERGE study and compare their findings.