Five questions with Louis Derry
Louis Derry is an Associate Professor in the Department of Earth and Atmospheric Sciences at Cornell University. His research includes biogeochemistry and coupling among climate, tectonics and surface processes. He received a BA in geology from Colorado College in 1981, and a PhD. in Geochemistry from Harvard University in 1990. Geochemical News recently caught up with him via email about the state of his field, the byzantine funding mechanisms of the National Science Foundation, and some near-future research challenges.
1. What is your specialty in the broad field of geochemistry?
I'm interested in lots of things, but most of the ones I actually know something about revolve around low temperature geochemistry and surface processes. I've been interested for a long time in how the Earth's geochemical cycles are coupled to the thermal and tectonic evolution of the Earth, and how all of those are coupled to biological evolution. I have studied these processes on both small spatial and time scales (modern fluxes between rocks, soil, the atmosphere, and plants at the plot scale) to large ones (impact of the Himalaya on geochemical fluxes, or evolution of the Earth's atmospheric O2 levels).
2. What questions are you trying to answer in your current research?
Currently I am thinking a lot about a few things in particular. We're working on understanding metamorphic CO2 fluxes produced by the India-Asia collision. This has me really excited because, at least where we've looked so far they turn out to be large relative to the weathering consumption of CO2. This is an unexpected but fascinating result with potentially far-reaching implications. So we're trying to understand how the fluxes vary along the Himalayan arc, and whether they vary systematically with uplift rate, heat flow or other variables. I'm also working on weathering fluxes from volcanic terranes. We think that they are much more efficient sinks for CO2 than typical cratonic environments, and also more likely to be sensitive to climate. But one problem is that hydrologic fluxes in volcanic terranes are hard to constrain because there are so many small streams rather than a few big river networks, and because there is so much groundwater transport relative to surface water. So there are just some fundamental things we need to understand and quantify better.
We're also trying a different approach to understanding how biogeochemical cycling in the oceans may have worked in the past. Our views of ocean biogeochemistry are so grounded in our understanding of the modern oceans (necessarily), but that makes it very hard to think through the implications of conditions far from today's. We've been working on simulation tools that we can use to understand the behavior of the coupled CNPOS systems in ancient oceans under quite different conditions than we have today, such as low oxygen, or different circulation regimes, or different mechanisms of biogenic particle export and so on. We felt that we really need to rethink the structure of models for this application. The simple ones that people have been using for these sorts of questions aren't quantitatively adequate, and the "sophisticated" ones based on OGCMs are probably inappropriate for many Deep Time problems. This is work "in silico", but we think will be very helpful in posing and testing new hypotheses about biogeochemical dynamics through time.
3. Please give us a short synopsis of your research approach, the types of samples you collect, your analytical methods, and the types of data you work with.
I'm sure that many geochemists would answer this question the same way. Ideally, I like to do some preliminary pondering, and then move to doing some simple calculations or modeling to get a little bit more formal framework of how I think a system or problem behaves. If that appears promising, then we try and design an "experiment" of some kind, often by sampling rocks, soils, stream waters or what have you, and making a set of measurements. The thought process I have just described sounds vague and tentative - and initially, at least, it is. My experience has often been that very good ideas grow out of intuitive thoughts about a larger problem. The hard part is taking those ideas and refining them so that they can be usefully tested. But the real trouble is taking a perfectly good idea and framing it in the format required to write a successful grant proposal. The collective wisdom of the reviewing and panel communities has settled on a very narrow definition of what makes a good proposal, at least in the U.S. In my view this definition is not wrong, but is simply too restrictive. God forbid you should go to NSF with nothing more than a really good idea that you want to work on! Or, what about a proposal with a really interesting problem that you want to explore but without a section on "expected results". The whole concept of "expected results" in a basic research proposal just depresses me - I refuse on principle to write that. And what if you should forget to specify what plastic your bottles are made out of or some other really vital intellectual tidbit? Well, in that case this is obviously not serious science! So, as you can see, I'm not very optimistic about the current system's ability to support really new approaches. I don't blame the funding agencies for this - to quote Walt Kelley and his comic strip Pogo "we have seen the enemy and they is us". We, as proposers, get feedback of this kind, and too often turn around and apply the same ill-considered standards to other people's proposals. I think we as a community can and should do better. It won't by itself change the amount of money NSF or DOE or USDA or whoever has, but it would improve the intellectual climate and eventually, perhaps even the quality of work that was supported. The idea that NSF should fund "transformative" research is nice, but the reviewing and funding apparatus has to really support that before it can become anything other than pretty words. "Incremental" would better describe the current system.
I'd like to see longer grants to reasonably well-established investigators so that they could spend less time writing proposals and more doing science. In my personal view I'd like to see five year proposals from senior investigators that outlined a set of really interesting questions and an overall approach, without wasting space on minor details. Do some of my colleagues with a decade or two of excellent scientific productivity really still need to convince me that they can do good lab work and know how to analyze basic geochemical data? I would hope not, but as it stands now, they have to include all that or risk being told that they haven't provided enough "specifics". I'm personally willing to make a bet that, if we give people like Bernhard Peuker-Ehrenbrink or Kate Freeman or Joel Blum (just to pick on them - wink!) funding on a five year basis for a clearly articulated set of great questions, we can safely assume that they'll figure out the field and lab protocols necessary to generate very interesting science. Instead, we require mindlessly detailed 2 year time lines, "expected results", and logistical plans that only a bureaucrat could love. I don't think we, as a community, are better off for it.
Interestingly enough, NSF decided to make such 5 year grants available, but to only a few new investigators. That's a nice concept, but in my experience it doesn't work so well. A depressingly large number of young people waste valuable time writing 5 year plans for expensive proposals that are statistically very unlikely to get funded. Not only do they have to give a detailed science plan, but they also have to fit in some sort of revolutionary education plan in the same 15 pages. That's a tall order for any PI, but especially a new one - unrealistically so in many cases. In my personal view, it makes more sense to have young PIs focus on shorter time frames. This allows them to really focus on well-defined questions on which they can build, and fits with their promotion time scales. We should instead think about longer-term support for people who have demonstrated long term excellence. Others may disagree, but it's time we as a community talked about this.
4. Can you name three fundamental and as yet unanswered questions in your field?
1. How does the long-term carbon cycle feedback really work? The basic idea that has been around for over a century is that weathering rates change as a function of climate on the long time scale, but the details of how this really works are still not clear. We clearly need a means of stabilizing the Earth's climate on long time scales. The very influential models of the 1980's showed quantitatively that this could work with some simple assumptions. A lot of good work since then has unsurprisingly made it clear that "it's not quite that simple". I think the principles are valid, but we have a lot to learn about the mechanisms.
2. How do major orogenic events (the India Asia collision, the Pan-African orogeny...) impact the carbon cycle? They change weathering rates, degassing rates, and impact the organic carbon sub-cycle in multiple ways. We know that the answer to this question is quite different from what many people thought even 15 years ago. But we're still short of having a good quantitative description of how this works.
3. What is going on with the weird and wonderful changes in climate, evolution and biogeochemistry in the Neoproterozoic and Cambrian? The data and chronology have greatly improved over the last decade, both in quantity and quality and the addition of new geochemical tools like MIF sulfur and metal stable isotopes. But we don't have a very synthetic understanding of how and to what extent all the changes we see are linked mechanistically. With dramatic climate swings, major evolutionary milestones, and large amplitude geochemical variations, there's a lot to explain, and modern analogs only get you so far.
5. We know that you are not a prognosticator, but you do know where the field is heading. What, in your opinion, are three important questions that you and your colleagues will be asking in three years?
In three years we'll probably be asking sort of refined versions of the same kinds of questions. Three years is pretty short, and I've learned the hard way that progress just isn't as fast as I always hope it will be. Being a researcher means being an eternal optimist despite good empirical evidence that we should in fact be pessimists. But, overall, I'm excited to see some of the new geochemical tools starting to move from the "exploration" stage to the "application" stage. For a long time I was made quite uneasy by confident predictions that all sorts of new systems were going to be opened up by new instrumentation such as multi-collector ICP-MS, and that this was going to really revolutionize the field. That has proven to be a more difficult road than I think many people understood, but now more and more good work actually addressing interesting large problems is appearing. One other area in my own field that seems ripe for innovation is the link between the geochemistry of organic compounds and metals. At the moment there is need for improved analytical methods as well as a better process-based understanding of how ligands interact with both major elements like Al in soils as well as very minor ones like Fe in seawater, and how these metal-ligand complexes interact with mineral and bacterial surfaces. These are not easy problems, but I believe there is still a lot more to be learned by diving into them.