Sunday, October 12, 2014

Tree ring growth vs latewood density: Pat'n Chip get lost in the forest at WUWT

Sou | 3:29 PM Go to the first of 9 comments. Add a comment

Anthony Watts has a "guest essay" by Pat'n Chip (archived here, latest update here), the duo from the CATO Institute. (Part of their job seems to be to spread disinformation about climate science. It's a good lark. Money for jam. They don't have to do much except spend a few minutes every now and then writing stuff and nonsense.)

Today's serving has a woody flavour. They are claiming that a new paper in GRL, by Martin Tingley and colleagues, contradicts a paper by Michael Mann. Michael Mann says, no. That's wrong.

To summarise

As suggested by Steve Bloom in the comments, here are the main points as I see them:
  • Tree ring growth reconstructions do not register the sudden (short-lived) drop in temperature following very large volcanic eruptions of the past millennia. This has implications for estimates of climate sensitivity based on paleoclimatology. (Mann12)
  • Latewood density measures exaggerate the drop in temperature following large-ish recent volcanic eruptions, Krakatau and Novarupta. (Tingley14)
  • The new paper by Martin Tingley et al complements, rather than contradicts, the 2012 paper by Mann et al. 
  • In dendrochronology, density is not the same as tree ring growth and one does not necessarily follow the same pattern as the other.

Initial impressions

Before I discovered Michael Mann's Huff Post article, I looked at the Mann paper and the abstract of the Tingley paper and at first glance it looked to me as if the papers were about different things. Or at least about different volcanic eruptions. I had some questions, though, because both papers were looking at how the sign of a temperature drop after large volcanic eruptions are manifested in tree rings.

So I went to the Tingley paper itself. It's not open access, so unless you can get hold of a copy, you'll have to take what I write at face value. Or you can read what Michael Mann has written about this at Huffington Post.

The impact of very large volcanic eruptions of the past millenia

The paper by Michael E. Mann, Jose D. Fuentes and Scott Rutherford found that while tree rings register most volcanoes, the very big volcanic eruptions that have a sudden large cooling effect, aren't registered in tree rings. The explanation put forward by Mann12, as I understand it, is that they stopped growing because of either or both the sudden spell of extreme cold and/or lack of light, so there possibly weren't tree ring in evidence over that short period following these very large eruptions.

There were only three periods where they noticed this effect. Below I've listed the volcanoes discussed in the Mann et al paper, with the name and estimated VEI in brackets (from Wikipedia). The VEI is discussed here - each successive number represents an order of magnitude difference in ejecta volume.
  • 1258/1259 (Mt Rinjani VEI7)
  • 1452/1453 (Kuwae VEI6)
  • 1809 (Unknown VEI6?) +1815 Tambora VEI7 - double pulse of eruptions.

The Mann et al team made the point that there would need to be more research to see the extent to which any particular individual tree ring series suffers this effect. I imagine it would take a lot of painstaking work to investigate. They also said their finding is very specific and doesn't imply a general problem with dendrochronology. They wrote:
The extent to which the predicted age-model errors are present in actual individual tree-ring series will need to be assessed by dendroclimatologists through the careful reanalysis of the original tree-ring data. It should also be noted that these age-model errors are predicted only for thermally limited treeline environments and for very large eruptions, and do not necessarily imply more general problems in dendrochronology.

Recent large but not so large eruptions in the instrumental era

The Tingley et al paper didn't go back to these extra big explosions. It only went back in time in when there are instruments monitoring temperature. The period they examined was 1850 to 1960, when instrumental records were available but "before the advent of modern divergence". They looked at the following times in particular:
  • 1883 - (Krakatau VEI6)
  • 1912 - (Novarupta VEI6)

They chose those two eruptions because they featured:
  • non-zero Northern Hemisphere sulfate deposits; 
  • cooling in the instrumental record; and
  • observed decreases in growing season solar radiation in the northern latitudes.

Tree ring growth is a very good proxy in general...

The Mann et al paper found that in the main, tree rings reflect incidents like volcanic eruptions, except for the really big ones. They went so far as to say that the overall level of agreement is "striking". The authors wrote (my paras):
Given the uncertainties in both the radiative-forcing estimates used to drive the climate models and the temperature reconstruction itself, the overall level of agreement is striking (Fig. 1b). Both simulations capture much of the low-frequency variation in the reconstruction and the residual variability (Supplementary Information) is within the estimated range of internally generated climate noise13.

...but not during extra strong volcanic cooling

They found some notable exceptions:
Yet there is one glaring inconsistency: the response to the three largest tropical eruptions ad 1258/1259, ad 1452/1453, and the 1809+1815 double pulse of eruptions is sharply reduced in the reconstruction.
Both models predict a drop of ~2° C following the 1258/1259 eruption, whereas the reconstruction shows a decrease of only ~ 0.6° C. A similar pattern holds for the two other largest eruptions. Regardless of the predicted cooling, in no case does the reconstruction show more than ~1.2 C cooling relative to the modern base period; that is, there seems to be a ~1° C floor on the cooling recorded by the reconstruction.

Light availability and its impact on tree density

In contrast, the abstract of the Tingley et al paper shows that they were looking at at recent (mostly smaller) volcanic eruptions, Novarupta and Krakatau. They were aiming to assess whether or not the reduced light availability leads to a general underestimation or overestimation of the response. This was a different issue to the what Mann et al found - that over a certain amount of sudden cooling (around 1°C), the tree rings don't register it.

The abstract of Tingley et al makes that quite clear. They didn't look at the volcanic eruptions that were the focus of Mann et al. They looked at the impact of two recent eruptions, Krakatau and Novarupta. The abstract is below:
The fidelity of inferences on volcanic cooling from tree-ring density records has recently come into question, with competing claims that temperature reconstructions based on tree-ring records underestimate cooling due to an increased likelihood of missing rings, or overestimate cooling due to reduced light availability accentuating the response. Here we test these competing hypotheses in the latitudes poleward of 45° N, using the two eruptions occurring between 1850 and 1960 with large-scale Northern Hemisphere climatic effects: Novarupta (1912) and Krakatau (1883). We find that tree-ring densities overestimate post-volcanic cooling with respect to instrumental data (Prob. ≥ 0.99), with larger magnitudes of bias where growth is more limited by light availability (Prob. ≥ 0.95). Using a methodology that allows for direct comparisons with instrumental data, our results confirm that high-latitude tree-ring densities record not only temperature, but also variations in light availability. 

Mann and his colleagues did their work because they were investigating the impact of the largest tropical volcano during the past thousand years, and figured that it should have resulted in a climate cooling of around 2°C. Yet in the tree ring reconstructions of temperature, it didn't show up as such, being "largely absent". In reconstructions using tree-rings and other proxy data the effect was "muted". The paper states:
Here we use a tree-growth model driven by simulated temperature variations to show that the discrepancy between expected and reconstructed temperatures is probably an artefact caused by a reduced sensitivity to cooling in trees that grow near the treeline. This effect is compounded by the secondary effects of chronological errors due to missing growth rings and volcanically induced alterations of diffuse light. We support this conclusion with an assessment of synthetic proxy records created using the simulated temperature variations. Our findings suggest that the evidence from tree rings is consistent with a substantial climate impact of volcanic eruptions in past centuries that is greater than that estimated by tree-ring-based temperature reconstructions.

To summarise, they suggest the reasons for the apparent discrepancy is an "artefact caused by a reduced sensitivity to cooling in trees that grow near the treeline", compounded by secondary effects as follows:
  • chronological errors due to missing growth rings
  • volcanically induced alterations of diffuse light.

Growth vs density

It looks as if both papers have a lot of merit. However there is one other point worth highlighting. What Tingley and his colleagues were looking at was density, they wrote:
Here we further investigate the density of the maximum latewood density response to volcanic cooling in latitudes north of 45° N, and the roles played by climatological spatial patterns of the relative importance of light availability and temperature in limiting plant growth.
On the other hand, as Michael Mann pointed out at Huffington Post, his team was examining growth thickness not latewood density. Now I didn't at first glance see what the difference is, so I looked it up. The growth thickness refers to the width and stem girth as far as I can gather. It's how much growth there is from one year to the next. How the diameter of the tree increases over time. The latewood density is a different parameter. From what I've read, it relates to cell size and cell wall thickness. Both can be used as proxies for various climate parameters.

Why is the past relevant to the future?

It's worth pointing out the relevance of this to our understanding of what humankind is facing over coming decades and centuries. It is relevant because if the estimate of forcing of recent past events is incorrect, then it will mean that estimates of future warming, based on looking at this history, will also be incorrect.

What Mann12 suggests is that the estimated impact on temperature of very large volcanic eruptions was too low - just looking at reconstructions based on tree ring proxies or a mix of tree ring and other proxies. If the actual drop in temperature was much larger than these reconstructions indicate, then the climate is more sensitive to sudden shocks rather than less sensitive. As Michael Mann wrote at Huffington Post:
...it should be noted that comparisons using the state-of-the-art "CMIP5" climate model simulations employed in the most recent report of the Intergovernmental Panel on Climate Change (IPCC) indicate that paleoreconstructions and model simulations of the past millennium are consistent if and only if the few largest pre-historic volcanic eruptions are masked from the comparison. Otherwise, the model simulations overestimate the amplitude of variation relative to the reconstructions. This provides fairly clear, independent evidence that the tree-ring based paleoreconstructions do indeed show less cooling in response to the few largest pre-historic volcanic eruptions than indicated by state-of-the-art model simulations. Tingley et al fail to confront or address this evidence, which runs contradictory to their thesis.

Equally, the work of Martin Tingley and his team is valuable. It will assist people who use latewood density to estimate past global temperatures, and probably other things that cause changes in trees (light, precipitation maybe). What they found was that in the case of recent large volcanic eruptions, Krakatau and Novarupta, "tree-ring densities overestimate post-volcanic cooling with respect to instrumental data (Prob. 0.99), with larger magnitudes of bias where growth is more limited by light availability (Prob. 0.95)."

Note that Pinatubo would not be relevant here, because it erupted after tree rings diverged from the modern instrumental record (around the 1960s).

Krakatau and Novarupta were an order of magnitude smaller than at least two of the eruptions that Michael Mann and his colleagues wrote about. That plus the fact that the two teams were looking at different features (and effects), means that the two studies are complementary, not contradictory.

Adjusting the proxy reconstructions

To sum up, and taking the two studies together:
  • Very large tropical volcanic eruptions (of VEI7, or equivalent to between 100 km³ and 1000 km³ of ejecta volume), don't register the ensuing cooling in the tree ring growth record.
  • Large volcanic eruptions (VEI6, or equivalent to between 10 km³ and 100 km³ of ejecta volume) exaggerate the effect of cooling in the latewood density record. This is based on Krakatau (tropical) and Novarupta (near-Arctic).

At the regional level, for both Krakatau and Novarupta, the pattern of cooling inferred from examining latewood density is consistent with what is known. Latewood density exaggerated the cooling effect of both in Northern Eurasia, and of Krakatau in Northern Canada, though I couldn't tell from the paper itself by exactly how much the overall cooling effect was exaggerated. It also reads as if there are some regions where it's a bit hard to tell because instrumental records are sparse (eg the Sea of Okhotsk was mentioned).

Martin Tingley and co went into some detail discussing the different character of the two volcanoes. Krakatau was tropical and occurred at the height of the northern hemisphere growing season, where the impact is spread much more widely around the globe. Novarupta was in the high northern latitudes, where the direct impact is largely confined to the northern hemisphere. They were able to explore the nature of the effect in time and space. Their report provides a very clear picture of how the volcanic debris spread throughout the atmosphere and affected weather around the world. It discusses the effect on light penetration and photosynthesis. It's a bit technical but if you are interested, it is a wonderfully detailed account.

It wasn't far to walk

I am a bit surprised that Martin Tingley didn't pop into Michael Mann's office and have a chat while he was preparing his paper. (Maybe he did, I don't know.) You'll also notice that Tingley et al wrote that their finding doesn't necessarily contradict the Mann12 suggestion that extremely large volcanoes could stop tree ring development. They wrote:
In contrast, our results provide no evidence for the missing ring hypothesis of Mann et al. [2012], at least in the spatial domain considered here. The generally cold bias we identify does not necessarily contradict Mann et al. [2012], however, as the cooling bias may overwhelm the influence of any missing rings.

Which does acknowledge that the Mann12 hypothesis about tree ring suppression may be correct. The thing is, the Mann et al paper didn't mention missing tree rings in the context of lesser volcanic explosions like Krakatau and Novarupta. Their paper was about much larger eruptions. They found that for very large volcanic eruptions, the expected temperature drops of 2°C or more, just didn't show up in the tree ring reconstructions. The tree ring growth didn't appear to register below around 1°C.

Cut them some slack

To cut Martin Tingley and colleagues some further slack, if they missed the words "use a tree-growth model", they'd have had to check the references cited by Mann12 to see whether Mann12 was looking at reconstructions based on growth or density. It wouldn't have taken long to check. The third sentence of Mann12 showed which studies they were referring to:
This effect, however, is largely absent from tree-ring reconstructions of temperature6–8, and is muted in reconstructions that employ a mix of tree-rings and other proxy data9,10.

The first reference (6) was to Esper et al (2002), which looked at "centuries-long tree ring widths", not latewood densities. The second was D'Arrigo et al (2006), which looked at "tree-ring width (and limited density [Luckman and Wilson, 2005]) data". The third was Briffa et al (2008), which looked at "ring-width measurements", not latewood density. The first mixed reconstruction referred to the IPCC AR4 WG1 report, in which Chapter 6 is Paleoclimate. The second was Mann et al (2008), and the Supplemental Information indicates the tree proxies were tree ring growth not density.

Scientists: Please avoid giving deniers talking points

I'm also a bit surprised that scientists (and journal editors) continue to give ammunition to deniers. Perhaps journal editors and reviewers could add a "denier-alert" to their checklist when reviewing papers for publication. They could check the wording used, and make sure it doesn't trigger denier attacks. I don't expect that was the intention of Martin Tingley, Alexander Stine or Peter Huybers, who are all solid scientists in their own right (if a bit careless in this case).

From the WUWT comments

What you'll have noticed is that for the second time in a month, WUWT-ers are embracing dendrochronology. They used to deride it mercilessly. Also, of course, they'll take any chance they can to employ the Serengeti Strategy.

NeilC decides that Tingley14 is "right" and Mann12 is "wrong" and writes:
October 10, 2014 at 4:20 am
Is there anything about climate science when compared to reality that MM and the rest have got right?
angech urges Anthony Watts to support defamation (again)
October 10, 2014 at 4:47 am
Anthony, should you put a link to Steyn in humor and satire or political climate. Sorry if there is one there and I missed it. It would be appropriate for any one interested in tree rings, circuses etc

Stephen Richards thinks that refining the proxy analysis means that all dendrochronololgy is "wrong". He's wrong.
October 10, 2014 at 5:08 am
Oh you mean that treemometers are not treemometers. Well, well, who’d a thunk it.

markstoval joins in the lynch mob and writes:
October 10, 2014 at 5:17 am
To quote some climate scientist or the other, “We can’t find Mikey Mann’s intelligence and it is a travesty that we can not.”

Wondering Willis Eschenbach thinks volcanoes are "hilarious" and mistakenly thinks that Michael Mann overestimates volcanic effects, when Mann12 demonstrated the complete opposite, that tree ring reconstructions underestimate the effect of very large volcanoes:
October 10, 2014 at 6:08 am
Hilarious … see my many posts on volcanoes, and in particular on Mann’s overestimation of volcanic effects.

Scott Scarborough doesn't realise that geologists have figured out exactly when volcanic eruptions took place in the last 1,000 years, and writes:
October 10, 2014 at 6:30 am
Mann’s contention makes no sense. How could tree rings know a cooling is from Volcanos or just a plain cooling of the planet? Is he contending that increased aerosols in the air cause trees to grow faster? If tree rings don’t respond correctly to a cooling how can you be sure they respond correctly to a warming?

Pamela Gray sits back in her rocking chair and decides that after decades of analysing tree ring data, it isn't any good, and scientists "don't know nuffin'". She writes;
October 10, 2014 at 6:31 am
I also think tree ring data is fairly poor as a temperature proxy. Except in extreme cases. Super stratospheric events that prevent oceanic solar recharging via veiling and set up additional loss of heat via triggered El Nino conditions over months and years, especially around the tropical band, could lead to severe cases. Not only is oceanic heat lost, but no heat is gained. The oceanic currents would then distribute this condition globally, leading to evidence of severe cold, which IS seen in tree rings as killing freezes. And it takes a LOT of cold to cause extensive freeze damage in tree rings.

Walter Dnes builds a straw man. No dendrochronologist would assert that tree ring widths are "exclusively" influenced by temperatures. He writes (extract):
October 10, 2014 at 8:55 pm
A possible article for someone with an academic background in botany (especially dendrology) to run with; I seriously question the qualifications of most of the climatologists doing tree-ring analyses. How many of them have degrees in botany with specialization in dendrology? What qualifies them to assert that tree-ring widths are influenced EXCLUSIVELY by temperatures. As other comments have pointed out, major volcanic eruptions can influence precipitation and atmospheric transparency. Reduced/changed precipitation patterns can influence tree-ring width. Reduced atmospheric transparency cuts down photosynthesis, which cuts down tree-ring growth.

Mann, Michael E., Jose D. Fuentes, and Scott Rutherford. "Underestimation of volcanic cooling in tree-ring-based reconstructions of hemispheric temperatures." Nature Geoscience 5, no. 3 (2012): 202-205. doi:10.1038/ngeo1394

Martin P. Tingley, Alexander R. Stine and Peter Huybers. "Temperature reconstructions from tree-ring densities overestimate volcanic cooling" Geophysical Research Letters (2014) DOI: 10.1002/2014GL061268


  1. Nice work, Sou, although a suggestion: Posts that get this long need a summary up front.

  2. "They could check the wording used, and make sure it doesn't trigger denier attacks."

    I think you are being a bit unrealistic here. It does not matter what words are used deniers will read what they will into them. The Humpty Dumpty strategy.

    1. True, but we can make their work harder and their strategy even more obvious.

  3. Another important difference is that Mann et al. looked at Northern Hemisphere means, while Tingley et al. focused on latitudes poleward of 45 N.

    First, it's a simple fact of geometry that only 29% of the hemisphere's land area is poleward of 45 N. So a strong effect in the more northerly latitudes wouldn't necessarily dominate the hemispheric response.

    Second, it might well be that trees in the more northerly latitudes would be especially sensitive to radiative effects. I don't know if this is true.

    I can't see the full Tingley et al. paper yet so I don't know if these points were considered in their paper.

    1. Thanks, Raymond. I knew there was something else I meant to look at - it was latitude. There was some discussion in Tingley about the more open spaces in the higher latitudes, and diffuse light.

  4. Kind of makes sense to me. Latewood density is a function of summer temperature and the length of time that warm temperatures extend into autumn. Volcanic episodes (e.g. Pinatubo) are often described as 'the year without a summer' in temperate climes, so latewood growth is very suppressed.

  5. Re Willis Eschenbach being wrong in that he "... thinks that Michael Mann overestimates volcanic effects...."

    That depends upon whether Michael Mann favours the modeled result over the tree ring proxies underestimation of the effect. ie, He models that the effect was greater than the proxies indicate.

    Willis E would seem to be implying that Michael Mann favours the modeled result, and that may be an overestimation that effect. Willis E has also done some of his own analyses, concluding volcanic effects are often not discernible in the records, and he may be referring to that case.

    Meanwhile, MM perhaps has not stated his belief one way or the other:

    "This provides fairly clear, independent evidence that the tree-ring based paleoreconstructions do indeed show less cooling in response to the few largest pre-historic volcanic eruptions than indicated by state-of-the-art model simulations...."

    1. The whole reason that Mann and his colleagues investigated was because the extra large volcanic eruptions didn't show up as they should have in the tree ring growth reconstructions. So yes, Michael Mann and his colleagues are of the view that actual impact on surface temperature would have been greater than the tree ring growth reconstructions indicated.

      The temp drop did show up more in reconstructions that mixed tree ring proxies with other proxies, but still showed a "muted" response.

      I doubt you'll find any scientist who'd dispute that the volcanic eruptions Mann12 refers to, would have caused a large drop in surface temperature for a short period afterwards.

      Willis' mathturbations are not exactly scientific. In any case, as far as I know Willis has never looked at the particular volcanic eruptions that the Mann12 paper was referring to. I've only seen his articles on volcanoes in the instrumental era, not the VEI7 eruptions that Mann12 was discussing. If he's done other "work" I haven't seen it.


Instead of commenting as "Anonymous", please comment using "Name/URL" and your name, initials or pseudonym or whatever. You can leave the "URL" box blank. This isn't mandatory. You can also sign in using your Google ID, Wordpress ID etc as indicated. NOTE: Some Wordpress users are having trouble signing in. If that's you, try signing in using Name/URL. Details here.

Click here to read the HotWhopper comment policy.