This paper, by Cameron Wagg et al., which was published online early in PNAS last month, describes the results of a very interesting experiment in which the authors manipulated soil biodiversity and measured the effect of these manipulations on a range of ecosystem functions.
More specifically, they created a gradient of reduced soil biodiversity (including a range of faunal and microbial groups) by sieving the soil through a number of decreasing mesh sizes, adding the fraction that passed through the sieve to sterilized soil, while also adding the sterilized fraction that remained on top of the sieve. They then grew plant communities consisting of common grasslands species in the soil for 14 and for 24 weeks, in two separate experiments. At the end of the first experiment, and after 12 and 24 weeks of the second experiment, they measured plant diversity and productivity, carbon sequestration, litter decomposition, nitrogen turnover, N2O emission, phosphorus and nitrogen leaching as ecosystem functions, and fungal and bacterial diversity (by TRFLP), mycorrhizal root colonization (microscopically), and nematode abundance (microscopically).
They then used these data to relate the ecosystem functions measured to the soil biodiversity treatments. In addition, they calculated z-scores for the range of ecosystem functions measured as well as for all groups of organisms quantified, and regressed these against each other to answer the question whether ecosystem multifunctionality is related to soil biodiversity. This approach, of summarising a number of ecosystem processes into one ecosystem multifunctionality index, has been used previously by Maestre et al. (2012).
Their findings are very interesting and will make a lot of soil ecologists very happy: they find that a number of the individual ecosystem functions are reduced with declining biodiversity, but also that ecosystem multifunctionality is positively correlated with overall soil biodiversity.
When taking a closer look at the data, it becomes clear that the reduction in soil biodiversity varies between groups and isn’t linear with the decreasing mesh sizes – mycorrhiza and nematodes drop down sharply after the third ‘dilution’, whereas the other parameters show a more gradual decline. The authors have taken this into account by not only relating ecosystem functioning to the diversity treatments, but also to the abundance and diversity of individual groups. When taking a closer look at this, it becomes clear that the microbial properties measured have a far stronger effect than nematode abundance. In addition, the effect of reduced soil biodiversity on a range of functions is indirect, through effects of plant productivity and diversity.
Of course, it is very easy to criticise aspects of this study. You can question whether bacterial and fungal diversity, microbial biomass, mycorrhizal colonization, and nematode abundance together are a realistic representation of soil biodiversity. For example, why was nematode diversity not assessed? And why not higher trophic levels, such as Collembola and mites? Microbes and nematodes are only a fraction of the soil food web (Fig. 1). With the current analyses, the title ‘Soil microbial diversity and community composition determine ecosystem multifunctionality’ might have been more appropriate.
Also, it would have been interesting to see root biomass in addition to mycorrhizal colonisation – a number of recent papers point to the importance of roots for ecosystem functioning (e.g. Orwin et al. 2010, Grigulis et al. 2013)
A more technical comment relates to the measurement of nitrogen turnover – this was assessed by measuring the uptake of 15N from Lolium multiflorum litter into aboveground L. multiflorum biomass. So, this measurement might be a proxy for L. multiflorum biomass, which decreases with decreasing soil biodiversity, rather than for nitrogen turnover.
On another note, and I would be very interested in other people’s opinion, I am wondering about the value of using an index for ecosystem multifunctionality. True, this averages across ecosystem functions and can therefore inform management to optimize overall ecosystem functioning. However, are the ecosystems that have the greatest average functioning really the most sustainable, and thus, desirable ecosystems? Are all ecosystem functions equally important? There might be trade-offs between different ecosystem functions – for example between crop yield and nitrogen retention, or between decomposition and carbon sequestration. We might want to optimize a certain function in a certain area, of which we already know that it has potential in delivering a certain function, rather than promoting multifunctionality across the board. For example, peatlands store large amounts of carbon because of their low decomposition rates, and agricultural production systems have high yields but low carbon sequestration.
However, in this paper, the multifunctionality index serves the purpose of summarizing overall ecosystem functioning, which shows a strong and positive relationship with soil biodiversity. Done like this, it summarizes a range of measurements that non-specialists might struggle to interpret – thus, it simplifies and reinforces the message of the paper that soil biodiversity determines ecosystem functioning.
Experiments like this require an enormous amount of work, and you simply can’t include everything. It is incredibly difficult to modify soil biodiversity without simultaneously changing soil properties, and the authors of this paper have achieved this by used an elegant method of reducing soil biodiversity. Thus, in contrast to many earlier studies, they were truly able to mechanistically elucidate the role of groups of soil organisms in ecosystem functioning.
This paper adds to the growing body of literature that soil biodiversity plays a crucial role in ecosystem functioning, and highlights the importance of conserving, and promoting, soil biodiversity. That’s what I like to hear!