Blog Archive

Thursday, July 31, 2014

Greg Laden: Volcanoes, Tree Rings, and Climate Models: This is how science works.

by Greg Laden, "Greg Laden's Blog," Science Blogs, July 30, 2014

Mark Your Cosmic Calendar: 775/775

One wonders if anyone felt it. Did Charlemagne feel it as he led his forces across Pagan Saxon Westphalia, knocking down Irminsuls and making everyone pretend to be Christian or else? Did the people of Baghdad, just becoming the world’s largest city, notice anything aside from their own metro-bigness? Did the Abbasid Caliph Muhammad ibn Mansur al-Mahdi have the impression something cosmic was going on that year, other than his own ascendancy to power? Or was it mainly some of the nitrogen molecules in the upper atmosphere that were changed, not forever but for an average of 5,730 years, by the event?
The bent tree like object is said by some to be the, or a, Irminsul, the "pagan" sacred object, destroyed by Charlemagne much as one might destroy a hypothesis, either with, or about, trees.
The bent tree like object is said by some to be the, or a, Irminsul, the “pagan” sacred object, destroyed by Charlemagne, much as one might destroy a hypothesis, either with, or about, trees.
A long time ago, probably in our galaxy but kind of far away, a cosmic event happened that caused the Earth to be bathed in gamma rays in AD 774 or 775. No one seems to have noticed. There is a mention, in 774, of an apparition in the sky that could be related, but talk of apparitions in the sky were more common back then, before they had certified astronomers to check things out. There is chemical and physical evidence, though, of the gamma ray burst. The best evidence is the large scale conversion of stable nitrogen isotopes into unstable carbon-14 isotopes in the upper atmosphere. As you know, radioactive (meaning "unstable") carbon-14 is created continuously but at a somewhat variable rate in the upper atmosphere. Some of that carbon is incorporated, along with regular stable carbon, into living tissues. After the living tissue is created and further biological activity that might retrofit some of the carbon atoms ends (i.e., the thing dies) the ratio of radioactive carbon to stable carbon slowly changes as the radioactive carbon changes back into nitrogen. By measuring the ratio now, we can estimate how many years ago, plus or minus, the originally living thing lived and died.
But it does vary. Solar activity, nuclear testing, other things, can change the amount of carbon-14 that gets produced. And, a cosmic event that happened in 774/775 caused the production of enough carbon-14 to throw off the chronology by hundreds of years. This is seen in the close examination of carbon in the tissue of trees placed in a tree ring chronology. For example:
Screen Shot 2014-07-29 at 2.00.05 PM
Original caption: High-resolution radiocarbon ages, superimposed on annually resolved radiocarbon measurements from Japan and Europe (grey lines and crosses) as well as the IntCal calibration curve based on decadal samples (blue shading), re-sampled at 5-year intervals (light blue crosses). Radiocarbon ages (that is, using 14C, 13C and 12C isotopes) were determined at ETHZ with the MICADAS system.
See the inverted spike there? That is, apparently, gamma rays messing up the radiocarbon chronology. Hold that thought.

Climate Change Is Hard

When volcanoes erupt, they typically spew crap into the air. Some of this material stays in the atmosphere for a while (called aerosols, but not your under-arm deodorant exactly), which will in turn reflect sunlight back out into space prematurely. This causes cooling. It is essential to know how much cooling of the atmosphere happens from aerosols because this is a potentially important factor in global warming. The effect of aerosols caused by volcanoes or industrial activity is an important term in the big giant equation that puts all the different factors together to produce global warming (or cooling). It is important that climate models be able to accurately and realistically incorporate the effects of aerosols. If the science isn’t right on aerosols, climate models may not run true when aerosols are included.
Caldera of Mount Tambora.  When Tambora erupted in 1816 we experienced a year without a summer. Tambora was small compared to many earlier volcanoes which may have produced a few summer-less years in a row.
Caldera of Mount Tambora. When Tambora erupted in 1816, we experienced a year without a summer. Tambora was small compared to many earlier volcanoes which may have produced a few summer-less years in a row.
And indeed there is an apparent problem. When climate models are run and include aerosols, and the results are compared with real life data where we have good proxy-indicators of past climate, the model predictions and the real life measurements don’t line up when aerosols are involved at any significant level. A big volcano goes off, but the proxy record consisting mainly of things like tree rings doesn’t show the level of cooling models predict. This has titillated denialists, as you might imagine, because it shows how the science has it all wrong and the only way to truly understand the climate change is to spend hours in the basement with your spreadsheet and a good internet connection, like Galileo would have done.
In fact, this was an interesting problem that needed to be addressed. The modeling methods had to be wrong, or the paleodata had to be wrong, or something had to be wrong.
In 2012 Michael Mann, Jose Fuentes and Scott Rutherford published a paper in Nature Geoscience proposing a hypothesis to explain this discrepancy. The problem was that when a known volcano went off, the tree ring record in particular tended to show only an anemic result. Volcanoes that were thought to totally mess up the weather seemed to have little effect on trees. This even applied to volcanoes which were very directly observed in recent times, when we know there was an effect because people were putting on sweaters and measuring things with actual thermometers.
Mann et al. proposed that rather than having little effect on tree growth, the volcanoes had a huge effect on tree growth. What was being seen by the dendrochronologists (tree daters, like tree huggers but more serious) as a normal, average growth ring at the time of a volcanic eruption was actually the ring for the next year in line; they were missing, understandably, one or more growth rings. The volcano goes off, the trees don’t grow at all. (The masquerading ring would typically be the year before the missing ring since dendrochronology is done backwards, since we know what year it is now.)
You don’t have to imagine a year in which no tree grows ever anywhere to accept this idea. The trees being used as temperature proxies are more the sensitive type. They respond to temperature changes by growing more or less (warmer vs. cooler). Trees that don’t do this are not chosen for study. This has to do with the species and the setting the tree grows in, combining to make temperature the key limiting factor most years, so that growth ring width reflects temperature more than any other factor. So yeah, when it gets very cool because of a big-ass volcanic eruption, one of those “year with out a summer” deals, the very sensitive trees respond by not growing at all that year. They may have a growth period of a few weeks, but trees don’t simply lay down wood every day they are biologically active. They usually start with leaves, then many move on to reproduction, and once they have finished reproducing, have a cigarette, wash up, whatever, they may lay down wood or roots. (Different species have different patterns.) So, a very short growing season can mean no ring at all. If a really bad nuclear-winter-esque volcano happens, this may go on for a few years. This leads to the growth ring corresponding to the year of the volcano simply not being noticed by the dendrochronologists, with a different year standing in. Over time the record can be thrown off by several years, if there are a few volcanoes and one or more of them affects growth for more than one year.
So two things happen. Years with a very strong cold signal are lost entirely, and the record is quasi-randomly offset by a few years in some but not all tree records (because some will be thrown off, while others are not), so the collective record gets out of alignment. A strong uptick in the signal (the zero growth year) does not contribute to the paleoclimate squiggle of temperature at all, and the other possibly contributing years (after the worst is over) are moved around in relation to each other and average in with less cold years. It’s a mess.
Consider the following made up numbers representing temperature over time. The top table is the hypothetical raw data of tree ring growth in relation to temperature across a very strong cold anomaly as might be caused by a massive volcanic eruption. Depending on the tree, there is one or more years of zero growth. The lower table is the same set of numbers but with the earlier years (top) shifted down to cover the zeros, because that is what would happen if a dendrochronologist was looking at the rings from more recent (bottom) to oldest; there would just be this void and it would be filled with the next data in line.
Screen Shot 2014-07-30 at 7.20.34 PM
Here are the same data graphed showing a clear anomaly in the top chart, but the very clear anomaly utterly disappears because of missing rings and shifting sequences in the lower chart. This is an existential problem for ancient climate events. I squiggle therefore I am.
Screen Shot 2014-07-30 at 7.16.41 PM
Mann et al. proposed adjustments to the record of proxy-indicators of temperature that accounted for missing tree rings at the time of major volcanic events. They made a good case, but it was a bit complicated and relied on some fairly complicated modeling.
Since the publication of Mann et al. there has been quite a bit of back and forth between the climate modelers and the dendrochronologists. I’ve assembled a list of publications and blog posts below. I’ll only very briefly summarize here.
The dendrochronologists had a bit of an academic fit over the idea that they had missed rings. Understandably so. As an archaeologist, I’m partly trained in dendrochronology. There was actually a time when I considered making it my specialty, so I had read all the literature on the topic. I can tell you that missing rings was a serious concern, and taken seriously, and seriously addressed. Seriously, there’s no way modern dendrochronologists would totally miss an entire year’s growth rings. They had ways of dealing with missing rings.
The thing is, it is actually possible to miss rings. Here’s why. The assumption in dendrochronology is that rings can be missed (or for that matter, added) for reasons that allow for correction by cross-dating growth-ring sequences with other trees or even other samples in a single tree. A particular part of a tree can be missing a ring while another is not (especially vertically; the lower part of a tree grows last in many species), or some trees in an area may be missing a ring, but others have that growth ring. This assumption is probably almost always valid; missing rings can be adjusted for by cross-checking across samples. But, if all of the trees of a given species and sampling area have one or more missing rings because of a major volcanic event, that won’t work. But this is not something dendrochronologists are used to.

2 + 2 = 774/775

Eventually Mann and his colleagues put two and two together and realize that the dendrochronologists had a way to test the hypothesis that would not rely on fancy dancy climate-modeling techniques, and that would potentially allow a better calibration of the tree growth-ring record for certain time periods. It was that gamma-ray burst.
That moment in time is a clear marker. Any system involving 14C spanning this time interval should show the spike. Well, what about tree ring records that span both a major volcano and the 774/775 event? If Mann et al. are right, an uncorrected tree-ring record would show a lack of correspondence of any spike at 774/775. But, if missing rings are assumed for sensitive tree records at the time of the volcano, and the tree-ring sequence for those trees shifted, perhaps the records will line up. That would be a test of the hypothesis.
And this is the gist of a letter to Nature from Scott Rutherford and Michael Mann. Very simply put, Mann and his colleagues took this graph, from an earlier paper:
Screen Shot 2014-07-30 at 8.11.52 PM
And changed it to this graphic which shows mainly (see caption) the tree ring sequences that span both the 1258 volcanic eruption, which was a big one, and the 774/775 event.
Screen Shot 2014-07-30 at 8.11.35 PM
This is a gauntlet, being respectfully thrown down. Mann et al. erected a hypothesis that missing tree rings are virtually universal in large parts of the dendrochronological sample for some events, were not accounted for in the tree-ring chronology, and have thus messed up the tree rings as a proxy-indicator for temperature. Various attempts to knock it down have not worked out. Now, Mann has himself provided an excellent way to assail his own idea. It is now up to the tree ring experts to try to knock this hypothesis down. I suspect Charlemagne might have had an easier time knocking down the Irminsul.
I asked Michael Mann how he felt about this latest development in the ongoing saga of the missing (probably) growth rings. He said, “I’m very pleased that we’ve reached some level of reconciliation with our dendroclimatology colleagues: there’s an objective test that is available to determine if there are indeed missing rings in some of the regional chronologies, as we have speculated to be the case. I look forward to seeing the results of those tests. We proposed a hypothesis, other scientists were skeptical of the hypothesis, and now there is a way forward for testing the hypothesis. In the end, a fair amount of good science will have been done, and we will have learned something. This is the way science is supposed to work.”

No comments: