|
| Aims: Biodiversity is often positively related to the capacity of an ecosystem to provide multiple functions simultaneously (i.e. multifunctionality). However, there is some controversy over whether biodiversity–multifunctionality relationships depend on the number of functions considered. Particularly, investigators have documented contrasting findings that the effects of biodiversity on ecosystem multifunctionality do not change or increase with the number of ecosystem functions. Here, we provide some clarity on this issue by examining the statistical underpinnings of different multifunctionality metrics. Methods: We used simulations and data from a variety of empirical studies conducted across spatial scales (from local to global) and biomes (temperate and alpine grasslands, forests and drylands). We revisited three methods to quantify multifunctionality including the averaging approach, summing approach and threshold-based approach. Important Findings: Biodiversity–multifunctionality relationships either did not change or increased as more functions were considered. These results were best explained by the statistical underpinnings of the averaging and summing multifunctionality metrics. Specifically, by averaging the individual ecosystem functions, the biodiversity–multifunctionality relationships equal the population mean of biodiversity-single function relationships, and thus will not change with the number of functions. Likewise, by summing the individual ecosystem functions, the strength of biodiversity–multifunctionality relationships increases as the number of functions increased. We proposed a scaling standardization method by converting the averaging or summing metrics into a scaling metric, which would make comparisons among different biodiversity studies. In addition, we showed that the range-relevant standardization can be applied to the threshold-based approach by solving for the mathematical artefact of the approach (i.e. the effects of biodiversity may artificially increase with the number of functions considered). Our study highlights different approaches yield different results and that it is essential to develop an understanding of the statistical underpinnings of different approaches. The standardization methods provide a prospective way of comparing biodiversity–multifunctionality relationships across studies. | |
|