Anthony’s Science Blog

58 days until I submit my dissertation…

…to my committee that is. They will probably have me revise some things at that point, but 58 days is the big huge deadline that I am working toward.

I’m on track for my 58 days. My first chapter has been accepted for publication in Arctic, Antarctic, and Alpine Research. My second chapter is also completed and is being reviewed by my advisor. I plan to incorporate his suggestions and submit that paper sometime this month to the journal Ecosystems. But there is of course a lot still to do. I am working on my third chapter right now, which I will to send to my advisor as soon as it is done. The good news is that I am not teaching and can thus focus solely on this third chapter.

After that I move to Ohio. That’s right: Holy Toledo. I just accepted a postdoctoral position in Toledo with Dr. Mike Weintraub, whom I know from his stint as a postdoc here at CU. Mike, currently a faculty member at the University of Toledo, was recently funded by NSF to examine the effects of altered growing season length on C and N cycling in the arctic. The project is three years in length, based at the Arctic LTER at Toolik Lake, so I’m psyched to be going to Alaska this summer and to be working with Mike, who is a great guy.

Mad Scientist

From an article in this week’s Nature about South Korean scientist Woo Suk Hwang who has been on trial for fabricating his research…

In August, at the last of 43 judicial hearings on the affair, prosecutors sought four years in prison for Hwang. It was the end of an epic 40-month trial that saw the chief judge change three times, heard testimonies from more than 60 witnesses and included some bizarre moments — such as when Hwang claimed he could not account for some of his funding because he had used it to pay Russian mafia for access to the DNA of a frozen mammoth that he hoped to clone.

Wow.

Summer field work

I got married!!!

Pretty awesome. We had a great wedding on September 5th and an equally great honeymoon to Hawaii. Also, Jackie knows me pretty well (of course) and so she made me this awesome nitrogen cake!

But before wedding madness took hold, I had a great summer field season on Niwot Ridge. It was my last field season of my dissertation and I did a vegetation survey of my study area, which was pretty fun. We measured hundreds of quadrats. Here are Riley, Andrea, and Carly with our beloved quadrat:

I can’t wait to put all the data together and analyze it with my nitrogen data. However, I have a bunch of other manuscript and postdoc related work I have to get to before I can do that. Here’s a little friend that we came across! Thanks to Carly for these pictures.

We got pretty good at identifying the plants this summer, even the grasses and sedges, though the Poa spp. are still basically impossible to tell apart, even with the extremely generous help of expert Colorado botanist Nan Lederer. Even for those though, we have a good set of morphospecies which I will go with.

Thanks again to the awesome field crew!!

Ecological Society of America Meeting 2009

I’ve seen lots of good talks this year and have had lots of fun conversations with fellow ecologists that I’ve gotten to know over the years. This evening I will be presenting a poster, which is the same as the poster I presented earlier this year at the Front Range Student Ecology Symposium. I actually kind of wish that I had an updated version of my poster since I have made a lot of progress on analyzing this data since I made the poster, but it should still be fun to chat with everyone.

I’ve seen some good talks this week, including lots of nitrogen talks (of course) and other good global change and biogeochem ones as well.  I especially enjoyed Dan Liptzin’s talk on rapid changes in oxygen levels in tropical soils. He showed us how the oxygen levels (and inversely, CO₂ levels) in soils track the tropical rains in Puerto Rican rainforests. He also showed, however, that within a very similar area, overall oxygen levels were high, low, or variable with no easy way to guess which spots would be which. As always, more research is called for.

The conference is being held in Albuquerque this year. It’s been fun poking around, though I have to say that my general impression is that there are a lot of sketchy characters here. I saw some sketcher harrassing a conference attendee at the bus stop yesterday and my hand was on my cell phone in case the situation worsened (it did not). Also, two nights ago when I approached the house I’m staying at, some random dude hopped up off the porch swing and took off into the night! On the plus side though, I love the little adobe-style houses and I had some awesome tacos yesterday. I unfortunately don’t have any pictures of the city. Gotta remember to bring my camera when I come to these things!

2009 Green Lakes Valley Snow Survey

I didn’t do my own snow survey this year but I had a lot of fun participating in the annual Green Lakes Valley snow survey for the Niwot Ridge LTER. Here’s me with my snow-measuring pole on snow-covered Green Lake 4.

All of these pictures are courtesy of my snow-measuring buddy Jordan Parman, grad student in geography. The hole in the picture formed when Jordan fell through the snow. The lake is of course solid ice below the snow, but there was this weird empty spot running in a sort of snow fracture across the center of the lake. That was pretty cool. Here’s a better picture of the fracture.

Probably the hardest part of the survey was snowshoeing at 11,500-12,000 feet all day long. Here’s me on a slope:

Probably the coolest part was the fantastic scenery in the glacially carved Green Lakes Valley. Here’s Green Lake 3 and Kiowa Peak.

Thanks to the organizers Kurt and Casey for doing a great job and to Liz for cooking. And thanks to all participants for good company. Oh, and also Kathy Clegg of INSTAAR, who baked the large number of delicious pies we ate, which is pretty damn awesome of her. THANKS!

Quantifying spatial heterogeneity

I’ve thought a lot about quantifying spatial heterogeneity over the course of my dissertation. To demonstrate different types of spatial heterogeneity, I made the graphic below. I originally included this graphic in a draft of a paper that’s part of my dissertation. I’m nixing it for the next draft of the paper since the reviewers complained bitterly, one of whom said that “These generalized maps can be found in standard textbooks on geostatistics.”

While the reviewer is correct that maps similar to all parts of this figure can be found in various locations, I don’t think it’s really true that they have been presented together in this way before (post if you find it; I of course won’t rule out its existence elsewhere). But anyway, for me this formulation helps in classifying the different kinds of spatial heterogeneity that can be observed. Maybe it will be useful for others too.

Changes in three different types of heterogeneity are shown: overall variation, the magnitude of spatial autocorrelation, and the scale of spatial autocorrelation. Three corresponding metrics can be used for each of these types of variation: the coefficient of variation (CV) as a metric of overall variation, the relative magnitudes of the modeled variogram nugget and sill as a metric of the magnitude of spatial autocorrelation, and the variogram range as a metric of the scale of spatial autocorrelation. The second two types might not make any sense if you are not familiar with geostatistics. If you are curious, I recommend Model Based Geostatistics by Diggle and Ribeiro (2007).

In the figure, heterogeneity increases from left to right. Only one type of heterogeneity is modified for each arrow. In the three simulated grids on the left, the coefficient of variation increases. In the two upper grids, the scale of spatial autocorrelation is reduced while leaving the magnitude constant; vice versa for the two lower grids. Grids are simulated 50 x 50 Gaussian random fields with the same random number seed, which ensures that differences are due only to parameter choices.

Climate Change Seminar

This semester, I’ve been attending a fun seminar on the effects of climate change on ecosystems run by Professor Nichole Barger. We have covered a range of fun topics from species distribution models (which I presented with my classmate Brian) to climate change and invasive species to the societal effects of climate change on mangrove ecosystems.

For our final class, we are read a couple articles discussing institute of climate change science and its future directions. Specifically, these articles were dissenting on the scientific consensus that we need bigger better modeling efforts to understand climate change, an interesting perspective. These papers brought up some interesting points.

The first one we read was from the journal Eos and argued that the cost of new supercomputers to run climate models was not worth it because of the irreducible complexities that the authors argue would render these models useless. Also, they said, we don’t need more climate models because we already understand the range of possible risks with climate change. They cite that such a supercomputing facility could cost up to $1 billion, which indeed is a lot of money. In general, I was pretty unconvinced by their argument. While I agree that there may in fact be better ways to spend a billion science dollars, I think it would only be for opportunity cost, not because the models are doomed to fail. On the contrary, I think this sounds like a worthwhile project.

I think the authors of the paper were missing a fundamental fact about the sociology of progress in science. Frequently, it is the means not the end that are important for a research endeavor. The serendipitous discoveries, unexpected results, useful technological spinoffs, and the collaborative spirit that can develop behind a common research goal are all way more important for the success of science as an institution than whether these supercomputers can actually produce spot-on climate predictions. To quote author Mike Hulme (from our other reading), “[These large modeling endeavors] have a heuristic rather than predictive function.”

This was one of many interesting issues we discussed in the seminar. I felt like we had a great group in this class and it was fun to hear the perspectives of my fellow grad students as well as several non-scientists who participated in our seminar. Cheers to all participants and to Nichole for organizing.

Biogeochemistry Reading Group: Gordon Bonan

On March 3, Dr. Gordon Bonan visited us from the terrestrial sciences section an NCAR and we had a great discussion about modeling the effects of ecosystems on climate. Gordon did his undergrad as well as his Ph.D. at UVA and then eventually moved on to a postdoc at NCAR where he has been ever since.

Gordon got interested the interaction between ecosystems and global climate in 1988 when tropical deforestation catching headlines. He was fascinated by the huge impact of the rainforest on regional and global climate, which is why he eventually came to NCAR to work with a researcher named Bob Dickinson who studied the topic. Although he started on tropical rainforests, he eventually also became interested in how northern forests and other ecosystems affect climate.

Overall, we had a great discussion with Gordon and a couple of points really stuck with me from our chat with him.

First, we had a fun discussion with Gordon about public interpretation of his research. In a section of his paper that we read, he explains that deforestation in the boreal forests would actually help mitigate climate change due to the greater snow exposure, which reflects solar energy away from the Earth. We asked him about this and he explained that this conclusion comes from an earlier study he did in which he and his colleagues simulated the elimination of the boreal forest to see what the effects on climate would be.

We asked him if this study was ever used as a talking point by climate change skeptics. He told us that indeed, his study was seized upon by none other than Rush Limbaugh. Gordon explained to us how Rush worked his usual cherry picking shenanigans by using this study as a counterpoint to the environmental movement at the expense of broader context. Later, when we were discussing the many technological climate mitigation schemes that have been suggested recently, Gordon explained that “there is a lot of fodder for the Rush Limbaughs out there.”

My other favorite part of the discussion was when we talked about the inclusion of element cycling into the next generation of climate models. While the current models are basically based only on atmospheric physics, modelers have recently begun to incorporate biogeochemical cycling of carbon and even nitrogen.

Gordon said that because of the unknown parameters involved with such modeling, he thinks that this may lead to less overall certainty in the global temperature predictions in the next IPCC report. This would be a reversal of the trend in IPCC reports where scientists have voiced more and more confidence in their predictions over time. It will be interesting to see if this indeed occurs because it could have important implications for biogeochemistry research and would also be pretty interesting from a sociology of science perspective.

Showing data underlying bar graphs

A few years ago, I posted on Edward Tufte’s message board advocating that people show the data that underlie their bar graphs. Here’s the thread. My post included this graphic meant to demonstrate the double standard for continuous data where the data is clearly shown on a scatterplot,  vs. categorical data that is most often portrayed using “dynamite plots.”

The basic argument is that showing the underlying data (points for a small sample, or boxplots for a larger one) allows readers to see the range, distribution, central tendency and its uncertainty, sample size, and outliers. On the other hand, the dynamite plot shows only the central tendency and its supposed uncertainty by way of the standard error, which has its own questionable assumptions and other issues.

Probably aided by the endorsement that the graphic originally got from Edward Tufte, I’ve now seen it picked up in a couple other places by statisticians who are likely trying to make the same point about the double standard. Here’s a slide show by Dr. Frank Harrell that includes the graphic and here’s a handout on data analysis made by Dr. Rafe Donahue.

I should take this opportunity to point out that I am certainly not the originator of the idea that bar graphs hide information instead of showing them. For example, Dr. William Magnusson has authored a number of critiques along these lines, such as this one in the Bulletin of the Ecological Society of America (may require subscription), in which his Figure 2 makes a similar point but without the discussion of the blatant double standard between continous and categorical x-axis variables.

It’s been fun to see this graphic float around and I hope even more people use it! Even better, maybe people will stop hiding their data in bar graphs.

Front Range Student Ecology Symposium

I had a great time yesterday presenting at the Front Range Student Ecology Symposium at CSU Fort Collins. It was an impressively well organized conference and I had a lot of really nice conversations with CSU grad students and others about my poster, which is here:

This poster was a continuation of a previous poster I made a couple years ago. Together, they summarize most of my findings on hotspots of N cycling on Niwot Ridge, which are admittedly a little meager, but for better or worse, there they are!

I also saw a cool presentation by one of the conference organizers, Aaron Berdanier, in which he talked about understanding water limitation in the alpine. I met Aaron last summer when he was working on Niwot Ridge. His idea was to try to figure out whether water limitation of plants was being caused by direct impacts on stomatal conductance or by indirect impacts on microbial nutrient mineralization.

Both of these things could be occurring at different times and it would be neat to try to separate them in field experiments. Aaron presented some nice data on δ¹³C of Polygonum vivipara, a common alpine forb, showing that their carbon isotopes didn’t vary much in different moisture conditions. This is pretty interesting, since it is somewhat contradictory to a map I made of soil carbon isotopes in alpine and subalpine meadows in which soil δ¹³C correlates fairly well with moisture regimes across the landscape.

There are many things that could be going on here and it would be cool to figure it all out.