Thoughts on NutNetSubmitted by drupaladmin on 20 October 2011.
Nice write-up in Science this week on NutNet ("Nutrient Network"), a huge (68 sites, 12 countries, 6 continents) experiment looking at the effects of nutrient enrichment and mammalian herbivore removal on grassland communities. This ongoing experiment was started six years ago by a group of grad students and postdocs (several of whom I'm proud to count as friends) who met at NCEAS. NutNet has already yielded an important Science paper (Adler et al. 2011) and a number of other publications, with many more in the pipeline (some of which I hope get submitted to Oikos!)
A few off-the-cuff thoughts:
- What a bang for the buck! NutNet is supported by one, count 'em, one NSF grant for $322,000 USD. A good chunk of that money goes towards supporting the data-sharing (e.g., paying the salary of a postdoc to run the centralized database). The experiment itself is simple and cheap to set up and sample, and so many of the NutNet sites are paid for with small "in kind" contributions from the NutNet members. It's a very different model than, say, NEON, which is going to collect many different kinds of data, using much fancier instrumentation--and which is going to cost $434 million USD and will only cover one country. I'm not saying my American colleagues should dump NEON and replace it with, um...(does long division in head)...150-odd NutNets. For one thing it would be naive to assume that all the money going to NEON would otherwise have gone to projects initiated by individual ecologists. For another thing, I don't know that there are 150-odd questions in ecology ripe for a NutNet-type approach. But it does serve as a reminder that high-impact science need not be expensive science.
- In some ways, NutNet is very new school, NCEAS-influenced science: highly collaborative, based on data sharing, relying on the internet to facilitate said collaboration and data sharing. But in some ways, it's very old school: it's all about collecting new data rather than meta-analysis of existing data, in order to address a question we were already interested in rather than a question we only thought to ask because of the data we happened to have. Indeed, NutNet was born out of frustration with the limitations of existing data, because previous nutrient addition and herbivore removal experiments all used different methods, thereby making meta-analyses difficult to interpret. NutNet also is old school in that the new data can be collected simply and cheaply, without using advanced technology. No remote sensing, no genomics, not even any computationally-intensive stats--just fertilizer, fences, and generalized linear models. It's almost like citizen science--except the "citizens" are professional ecologists!
- I am now officially kicking myself for not joining NutNet. I actually thought about it briefly when it was first starting up, but decided I needed to focus on getting my own research program off the ground. In retrospect, that was probably a mistake. Not for mercenary reasons like "Geez, I could've had my name on a Science paper without doing much work or spending much money", but just because it would've been fun to collaborate with some good friends, and because it would've taught me a new way of working (although I have subsequently started some collaborative lines of work).
So, what other questions in ecology could be usefully addressed via a NutNet-type approach?