Mycorrhiza-mediated competition between plants and decomposers drives soil carbon storage

– added by Franciska de Vries

This paper, by Colin Averill and colleagues, came out as a Letter in Nature almost two weeks ago. It immediately excited me, as the title suggests that in this paper, the authors are revealing the mechanism through which mycorrhizal fungi increase soil carbon storage. Groundbreaking!

I started to read. What the authors did in this paper was compose a global dataset, consisting of observations of soil organic C, N, and clay content (to a depth of one meter) across a range of vegetation types and biomes. They then assigned values of mean annual temperature (MAT), mean annual precipitation (MAP), and net primary productivity (NPP) to each site, using previously published climate interpolations and satellite-based observations. Finally, they assigned a mycorrhizal status to each of the vegetation types, which could be either arbuscular (AM) or ericoid and ecto-mycorrhizal (EEM).

Mycorrhizal status of each site was assigned based on the dominant vegetation present, and knowledge of its mycorrhizal status. Understorey vegetation (in forests) was ignored, and where no vegetation data was present, they used the vegetation description. So, for example, the vegetation description “grassland” was classified as the AM, whereas “mixed coniferous forest” was classified as EEM.

This dataset was then used to explain soil C storage, using soil N and mycorrhizal status as explanatory variables, while at the same time accounting for variation in climate and other soil properties.

Using mixed effects models, the authors found that in ecosystems dominated by EEM fungi, 1.7 times more C was stored per unit N than in AM ecosystems.

For the first time, this study shows that C global cycling does not only depend on abiotic factors like temperature and moisture, but also on soil mycorrhizal status, highlighting the importance of biotic factors in addition to abiotic factors for soil C storage. This finding supports the results from a modelling study by Orwin et al. (2011), who showed that competition for organic N between EEM and decomposer fungi increases soil C storage.

Averill and colleagues conclude that ‘mycorrhizal functional traits are as important a control over decomposition and soil C storage as are soil chemical properties and the physical protection of soil organic matter’, and that ‘the identity and functional traits of soil microbes exert a control over the terrestrial C cycle’.

So, a pretty exciting paper, but does it really do what it promises? When I read the title, I expected a paper reporting on (a range of) mechanistic experiments to prove that mycorrhizal fungi drive soil C storage. But, rather than a mechanistic study, it is an observational study that uses a powerful data set and an advanced modeling approach to show that there is a relationship between mycorrhizal status and soil C storage. However strong their finding is, and although it holds across a range of ecosystems and biomes, their study does not allow for testing a hypothesis and elucidating a mechanism. As Mark Bradford elegantly says in his News and Views article about the paper: “The authors propose, in line with a previous hypothesis, that these richer carbon stores result from competition for nitrogen between EM fungi and free­living soil microorganisms that feed on organic matter” and: “Pinpointing which mechanism explains Averill and colleagues’ results will require more data and involve challenges common to all large observational data sets, including unobserved variables and spurious cor­ relations”.

I couldn’t agree more.

However powerful the data set in this paper is, there are several issues that would have to be addressed to come up with a conclusive answer of how and whether mycorrhizal fungi drive soil C storage. First of all, important soil properties that can explain both soil C content and mycorrhizal status, such as soil moisture or pH, haven’t been included in the models. Second, certain vegetation types, such as heathland, which is known to be dominated by ericoid mycorrhizal fungi, are missing. Third, the assigned mycorrhizal type might be occurring under a certain vegetation type because of the quality of the C inputs into the soil, which might itself drive soil C storage; something that the authors do acknowledge in the paper. Finally, Mark Bradford calculated, in his News and Views article, that only when the soil contains more that 3 kg N per square meter, C stores in EEM dominated systems exceeds that of AM dominated systems by 1.3 times.

But, despite these nuances that will need addressing in future studies, this is an important study that proposes an important hypothesis, namely that mycorrhizal type is of pivotal importance in driving soil C storage. By formulating this testable hypothesis and identifying a global relationship between soil biota and soil C storage, this study significantly advances the field of plant-soil interactions. It also highlights that disruptions of links between vegetation and mycorrhizal fungi, as a result of global change, might have far-reaching implications for soil C stocks and thus for the climate mitigation potential of soils.

So, I am curious what other people think about this paper! Do you agree or disagree with my views? How do you think we could go about testing the hypotheses proposed in this paper? Or is this enough evidence already? What do you think about the statistical methods used? Feel free to comment – the aim of this journal club is to stimulate discussion – through replies here, but also on Twitter. If you respond on Twitter, please use the hashtag #psejclub.


With thanks to all members of the Soil and Ecosystem Ecology group of The University of Manchester for inspiring discussions.

Full reference: Averill, C., B. L. Turner, and A. C. Finzi. 2014. Mycorrhiza-mediated competition between plants and decomposers drives soil carbon storage. Nature advance online publication.


5 thoughts on “Mycorrhiza-mediated competition between plants and decomposers drives soil carbon storage

  1. colin!

    Super excited to hear about your group discussing the manuscript!

    Franciska- you’re right on that we do not present a mechanistic test of this hypothesis. However, we felt that we had strong grounds to looking for a global relationship between soil C storage and mycorrhizal type based on published experiments suggesting an inhibitory effect of EM fungi on soil C decomposition, by means of competition for N with free-living decomposers. I would offer Lindahl et al. 2010 ISME, McGuire et al. 2010 Oecologia, and Gadgil and Gadgil 1971 Nature as experimental support for this mechanism. Furthermore, there is evidence that EM fungi rely more heavily on organic N sources and AM fungi, and we have theoretical reasons to believe that increased reliance on organic N should lead to N limitation of saprotrophic decomposers, slowed decomposition, and hence increases in soil C storage.

    So, although we do not explicitly test the mechanism of action, we felt the literature had set the stage by providing the mechanistic experiments and careful modeling analysis that supported this hypothesis. We aimed to test the generality of this hypothesis in nature.

    It might be worth noting that I am running manipulative field experiments to carefully test this hypothesis along a gradient of EM fungal abundance. So, hopefully we’ll have more data to challenge this hypothesis with soon.

    Again, stoked your group was interested in the manuscript and took the time to discuss it.

    Colin Averill

  2. Pablo García-Palacios

    First of all congratulations to Franciska and all the people involved in the launching of this exciting journal discussion club!

    And congratulations to Colin Averill and colleagues for such an interesting paper.

    In my opinion, the approach followed in this study is still valid to address the question of whether ecosystems dominated by EEM fungi storage more soil C than those dominated by AM fungi. It uses a global dataset of 227 sites encompassing 4 different biomes, and the model selection procedure seems to have evaluated major potential confounding factors driving soil C storage at large spatial scales such as plant productivity, climate and texture. Of course, it´s always difficult to find the definitive mechanism with such an observational approach, but I think the authors do show quite supportive evidence. It´s true that a manipulative test is still lacking to confirm the hypothesized mechanism (N competition between EEM fungi and free-living microbes), but the authors show an interesting global scale pattern.

    This manipulative test would complement the previous modelling results of Orwin et al. (2011) and would provide experimental support to the study of Colin Averill and colleagues. With this I want to say that a manipulative test would allow testing the former mechanisms but would lack the generality of biomes and spatial scales provided by this study. Complementary approaches, complementary results. But I totally understand Franciska’s point, maybe promoted by a title and abstract written in a way that makes you think this study actually tested the specific mechanisms?

    I also wonder why the authors did not directly assess the contribution of litter chemistry differences among EEM and AM plants. Global databases of leaf traits, such as TRY, could have been used to test the influence of the dominant plant species chemical composition, at least for some well-studied traits such as leaf/litter SLA or N content.

    Again, congrats for this nice initiative coming from the Plants-Soils-Ecosystems special interest group.

    Pablo García-Palacios

  3. Michael Van Nuland

    Hi All,

    Congratulations Colin and to your colleagues on this paper, my group has been informally discussing it for the past few days! And thank you Franciska for creating this excellent club to work through these ideas!

    I agree with Franciska, this paper is a nice and important step in thinking about the global consequences of plant-soil linkages. At times it may rely a little too heavily on previous mechanistic studies for my taste, but for the most part do try and include most (but not all, like Franciska already mentioned) confounding factors that are known to influence soil C dynamics.

    My major concern relates to what both Franciska and Pablo tangentially brought up, namely that there is little consideration of variation in plants (or plant-soil interactions) beyond the broad categorizations of vegetation type and AM/EM. Vegetation types harbor various plant communities (i.e., understory veg like Franciska highlighted), all with different niche spaces and functional traits that influence their conditioning of the belowground environment (i.e., Pablo’s comment on leaf traits and litter chem). I understand that it is virtually impossible to include every potentially important factor into the models and see what gets spit out on the other end, but there are a range of studies illustrating that there are ecosystem consequences of “finer scale” (read: not global, but still relatively large) variation in how plants and soil communities interact based on genetically controlled plant traits and microbial community dynamics. In short, while it’s a thought-provoking exercise to evaluate global patterns, there is much variation in plant-soil interactions that these models do not incorporate. Following up with Pablo’s suggestion of adding leaf traits into the mix might start to get at this and could yield some interesting patterns.

    One minor point, I believe the WorldClim database uses Annual Precip (not MAP, but it is unclear if the authors calculated mean values for this variable).

    Again, I greatly enjoyed reading this paper and have had many interesting conversations about its core ideas and results. Good luck with your future work and I’m excited to see what comes of your field experiments!

    Michael Van Nuland

  4. Relena Ribbons (@RelenaRibbons)

    Everyone has brought up great points already. It is always challenging to try and take observational studies and coerce them into a global mechanistic model. This study makes a lot of great steps in the right direction, and also serves to point out where additional data is needed to refine the model (for example the inclusion of understory vegetation or additional vegetation cover types that have distinct EM/AM fungi relationships). Global syntheses will naturally not always hold at local scales, but they do provide insights into general relationships, and in this case spur an increased research effort to challenge/validate the hypothesis. In this case I think the model is a great starting point, but I feel that the mechanistic understanding needs to be more deeply vetted, and explicitly tested with field and laboratory tests.

    I am that concerned some readers might conclude that only mycorrhiza-mediated interactions are controlling carbon in all environments and that other plant-soil interactions always play a smaller role. It is comforting that scientists are trained to scrutinize data and contextualize the results rather than just take everything at face value.

  5. Pingback: #psejclub discusses: plants, fungi, competition and carbon storage | Plants-Soils-Ecosystems

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s