Seasonal not annual rainfall determines grassland biomass response to carbon dioxide

After a brief hiatus, journal club is back and this time we’re discussing a paper by Hovenden et al from Nature in May exploring the interaction between carbon dioxide and rainfall on plant biomass.

At this point it seems that we ecologists have a reasonably good idea of what effect many environmental variables – like water, temperature, and carbon dioxide – have on certain ecosystem parameters, though always with a few caveats and exceptions thrown in to keep it interesting.  However, our understanding of how these variables interact and the effects of these interactions, especially over various temporal and spatial scales, is still pretty woeful.

For example, we know that plants need water to grow, and when there isn’t enough water they stop growing – very straightforward. Plants also need CO2 to grow, and in general higher CO2 levels lead to higher plant biomass.  This is because increased CO2 allows for higher rates of photosynthesis and greater water use efficiency.  Due to the greater water use efficiency, we also expect that the effect of elevated CO2 (eCO2) on plant growth will be greater when water scarce.  Basically, with higher CO2­, plants can photosynthesize more per unit available water, so will be able to grow more before the water runs out compared to plants grown at lower CO2 levels.

As with so many things in ecology, what we predict is exactly what we see … except when we don’t.  If the relationship between water availability, eCO2 and plant biomass is so straightforward, biomass responses to eCO2 would always be positive and we would see the strongest responses in the driest years.  I bet you can see where this is going…

TasFACE ring

Photo credit: TasFACE website

Hovenden et al looked at data from a nine year FACE experiment in Tasmania (TasFACE) and found that the eCO2 effect was far from consistent across years.  Some years there was no discernable eCO2 effect on biomass, some years it was positive (like we’d expect) and one year it was actually strongly negative; and these responses were not correlated with annual rainfall or soil water availability.

Instead, Hovenden et al found that the biomass responses to eCO2 were strongly correlated with seasonal rainfall variability.  Higher rainfall in the summer resulted in a positive effect of eCO­2 on biomass, as we would expect.  Summer rain at the site tends to come in short, sharp bursts, so the increased water use efficiency would allow the plants to maintain growth for longer between rain events.  However, increased rain during the spring and autumn were correlated with a negative effect of eCO2 on biomass.  During these cooler, wetter periods plants don’t grow as much and it is likely that increased rain would leach nutrients from the soil.  This was supported by a strong negative relationship between spring rain and soil nitrogen availability.

It seems probable that such a relationship between seasonal rainfall and eCO2 effects on biomass could be seen throughout temperate and seasonally wet systems, and that this could have big implications for global carbon models.  It also highlights the importance of looking beyond plants to fully understand the mechanisms that drive responses to climate change.

I would love to see similar analyses of other FACE datasets to see if these trends are replicated in other systems.  It’s an important finding, but opens up lots of other interesting questions: How does vegetation type or soil type effect the relationship between seasonal rainfall and eCO2 effects on biomass?  Does seasonal temperature variability affect the relationship significantly?  What about increased nitrogen pollution or fertilisation – would increased nitrogen deposition overturn the negative relationship between high spring/autumn rain and the eCO2 effect on biomass?

As always, we’d love to hear what you think about the paper.  Is it the best paper you’ve ever read or do you think it contains some fundamental flaw? Does it raise interesting questions or link well with something else you’ve read recently?  Would you use similar methods or could you propose a better protocol?  Let us know in the comments or on twitter with hashtag #psejclub!

Finally, don’t forget about our joint meeting with the Plant Environmental Physiology group coming up in October.  All the details, including links for registration and abstract submission, are available here.  It’s going to great!

2 thoughts on “Seasonal not annual rainfall determines grassland biomass response to carbon dioxide

  1. Relena Ribbons (@relenaribbons)

    Thank you for posting this Sarah. I agree that it would be really interesting for other FACE sites to look at similar analyses. This is an interesting study and I was glad to see that the authors looked into potential mechanisms for observed patterns.

    For me, this paper brings up an interesting point that goes back to talking about increased CO2 and climate change. There are a lot of different ways to approach talking with non-scientists about the subject, but I still find myself offering an overly complicated response. The effects of climate change are not necessarily directly aligned with how systems are functioning under present climatic conditions. This makes climate change research a really pressing issue (I know I’m preaching to the choir), and exciting because some seemingly straightforward predictions (such as those presented in this paper) may not hold true under future climate conditions. I don’t mean to hijack the discussion about how to talk about climate change with non-scientists… this paper just got me thinking about it.

  2. Sarah Post author

    You’re right that how we talk about climate change is really important, and often really difficult. In this case, interactions between environmental variable mean that ecological responses aren’t always straightforward. With multiple variables changing at the same time it’s incredibly difficult to predict where the balance will fall in the future with regard to any particular ecological response or across different temporal and spatial scales. To deal with this, it seems we need two things – better understanding of interactions and better communication skills. More studies explicitly looking at interactions could help us to understand responses, mechanisms and thresholds, while analyses of existing data from large experiments (hint, hint, FACE data-holders) could help to identify which effects are general trends and which are more locally variable. For better communication skills, there are loads of courses available (Imperial offers 5 under the ‘public engagement’ heading), but there’s also no substitute for practice – talk to people, blog, help with public engagement events and projects (shameless project promotion… You’ve hit on one of my favoured topics and I’m always happy to discuss science communication and public engagement, but I’ll try not to derail the thread here too far away from the paper!


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 )

Connecting to %s