Valmore Lamarche was perhaps the first person to suggest that temperature information could be extracted from bristlecone pine information and his early publications are often referenced. Lamarche et al. [1984] (with Fritts, Graybill and Rose) first postulated CO2 fertilization. As you know, I’m increasingly interested in changes in treeline elevation as a "low-frequency proxy". It turns out that Lamarche was as well. Lamarche [1973] contains an analysis of Holocene treeline variations at the key bristlecone sites of Sheep Mountain and Campito Mountain.
I’ve collected a large number of Lamarche articles over the past 18 months and would like to collect them all. His articles typically report on the sites which are sampled, with information on the geology, ecology, climate etc. By contrast, many more recent studies seem to jump into a discussion of the ring width or MXD "chronologies" and statistical manipulations thereon, while leaving the actual site undescribed. (These later articles all seem to end with a moral about global warming. When I read the genuflections of Zhu [1973] towards the Great Helmsman, I thought that the form of the genuflections were not dissimilar, if I may be permitted this observation without being accused of sarcasm.)
Lamarche’s relatively careful site descriptions undoubtedly reflect his training as a geologist. Geologists are trained to be careful observers and faithful reporters. They can theorize afterwards, but recording comes first. They are also trained to describe the context. In Canada, where there are many mineral exploration prospectuses, these requirements carry over even into qualifying report standards and prospectus standards for such offerings. I did not receive any formal geological training, but I’ve worked with geologists for many years and have learned to admire their systematic reports. These property reports are different formats from what geologists write in academic journals, which look like any other type of academic literature. However, their absence from dendroclimatic studies makes any efforts for third-party review and due diligence needlessly difficult.
As a passing thought, if I had a big policy job in which I needed to make a decision on climate policy, I would make up a commission of independent experts, which would include competent scientists, but which would not include any climate scientists actively involved in the debate. (They would be invited to make submissions to the panel.) While the people in the debate are smart and knowledgeable people, academics have a vested interest in showing that they are "right", which makes it impossible for them to be objective. Money is not the only form of bias in the world. I would make sure that some practicing geologists were heavily involved in the commission, because they have lots of relevant experience and, by and large, good judgement. (I’m told that they were very popular as intelligence officers in the Canadian armed forces in World WarII, because of their training to make sensible decisions on incomplete information.) I’d also have some medical people with statistical backgrounds and some statisticians. I’d probably ask one of the learned societies to nominate some candidates, but I wouldn’t leave it up to them to make their own panel.
I’ll get to Lamarche after one more digression. Lamarche’s protege, Donald Graybill, who collected the balance of the bristlecone ring width measurements in present-day use, did an undergraduate degree in sociology and a Ph.D. in anthropology at Tucson. I guess that he ran into tree ring studies in anthropological dating. He taught for a few years as an anthropologist, before returning to carry out tree ring research. We sometimes hear about the need to rely on proxy collections in the 1970s and 1980s because of the expense and difficulty of making the collections. I’ve someitmes made fun of this excuse, pointing out that exploration geologists also go all over the world and report in a timely basis. Is it coincidence that the main bristlecone collections were made by people who trained in geology and sociology/anthropology respectively?
Lamarche [1973] commences with a review of what was then known about Holocene treeline variations, which I excerpt at length below, together with (unusually for my posts here), a listing of some of the older literature, in case people are interested. I’ll post tomorrow on what he found at Sheep Mountain and Campito Mountain, two essential Hockey Team sites.
The position of the treeline is a valuable paleoclimatic indicator because temperature is an important factor in determining the altitudinal and latitudinal limits of tree growth…
There is evidence of broadly synchronous advances and retreats during the past several thousand years at many plaices in the northern hemisphere. In northern Europe, stumps and other remains of pine (P. sylvestris L.) beyond the present tree limit have long been known to mark earlier more northerly forest extensions. Hustich [1966] shows the distribution of such find in northern Finland and Norway. Citing dates from Andreev [1954] and Tikhomirov [1956], he correlates the advanced treeline position with the “post-glacial thermal maximum”. In the Calendonian Mountains in Sweden, pine stumps are found in bogs up to 150 m above the highest nearby living trees. Radiocarbon dates place most of these remnants between 8000 and 3000 radiocarbon years BP, within the same postglacial warm period. Lamb [1964] obtained a radiocarbon date of 4400 BP for a log from a bog in northwest Scotland. It …To the south the altitudinal treeline in the Alps may have been as much as 300 m higher than at present during this period [Ludi, 1955]…
McCullough and Hopkins [1966] found birch and spruce logs outside the present range of these species in the Seward Peninsula of northwest Alaska. Together with other evidences, these logs suggest a warm period beginning at least 10,000 BP and continuing to about 8300 years BP…Ritchie and Hare [1971] review the evidence for past changes in position of treeline in northwestern North America. .new pollen and megafossil evidence from Mackenzie Delta, together with radiocarbon dating of a rooted stump well north of the present treeline, suggest a period much warmer than present since about 4000 BP. A detailed treeline history has been worked out for the area west of Hudson Bay in northern Canada (Bryson et al, 1965; Sorensen et al 1971, Nichols, 1967)….forests moved northward following deglaciation and lake-draining about 5800 BP. Two separate northward advances, separated by periods of retreat followed, culminating about 2500 BP and 1000 BP respectively. …Jungerius [1969] also documented a major treeline advance in eastern Alberta between 4500 and 3600 BP.
Evidence of higher past treeline levels has been found at several localities in western United States, …In his work in the La Sal Mountains of southeastern Utah, Richmond [1962] related relict podzolic soils at high altitudes to a treeline advance up to 300 m above present levels during the “altithermal of postglacial optimum” prior to 2800 BP. A more recent advance, followed by retreat, is indicated by stands of large dead trunks that extend perhaps 100 m above present treeline. The death of the trees is attributed to cooler conditions during the historic Little Ice Age, which did not end until the 1800s. Curry 1968 also described dead trees at high elevations and interpreted these as evidence for previous higher treeline position at many points in the Sierra Nevada, California.
If anyone wishes to contribute a short summary of any of these articles,I’ll post it up.
References:
Lamarche, V., 1973. Holocene climatic variations inferred from treeline fluctuations in the White Mountains, California. Quat. Rese 2, 632, 660.
Hustich, 1966. Annals of the University of Turku A. II: 36 (rep. Kevo Subarctic Station 3) 7-47.
Lamb 1964. QJ Roy Met Soc. 90, 382-394.
Ludi, 1955. Ber Geobot Forschunginst Rubel (Zurich) 36-38
Richmond, 1962. US Geol Surv Prof Paper 324, 1-135.
Richmond, 1972. Quat Res 2, 315-322
McCoullough and Hopkins 1966. Geol Soc Am Bull 77, 1089-1108
Ritchie and Hare 1971. Quat Res 1, 331-342.
Bryson et al 1965. Science 147, 45-48
Bryson, 1966. Geog Bull 8, 228-269.
Nichols, 1967. Eiszeitalter and Gegenwart 18, 176-197
Sorensen et al 1971, Quat Res 1, 468-473
Jungerius 1969. Arc Alp Res 1, 235-245.
Climate Audit 
14 Comments
1. Good point on the better care shown in older papers.
2. I think it would be interesting if you could somehow get this point into more general circulation than a blog posting. Maybe a article on methods? This would require a bit more work (need to find some different examples, maybe similar themes.) I hesitate from urging that, as while you’ve obviously read a lot of literature and (should by this time) have professional level awareness, you don’t have greybeard level experience. Still…it’s an important issue and raising it, might benefit the community. (And I don’t quite agree with your worries of “making a mistake” and getting gigged for it (needing to be perfect)…I think you and the field are better off if you take more shots and in some cases make a mistake or two, but one that gets cleared up by the community and mnoves your own thinking also.) Also, the peer review and the reaction might increase your own understanding of issues on the experimental side. Hmm…what about combining forces with an experimentalist to write the note? I donno…
3. Anthropology/sociology: while one may not get the practical field work experience (or may if one is more towards archeology) of a geologist, the mathematics is pretty similar. However, I expect the really swinging guys in stats-heavy parts of either field come from mathematical physics background (as they do in economics): Mann’s background.
4. Nice summary of the elevational implications on climate. I wonder if any of the reconstructors use this type of info as a proxy. My intution would be that it would tend to be a more simple (less numbers) metric, but one that deserves higher weighting based on the better physical rationale. Not so good for fairing a curve, but great for understanding maxes and mins. Heck maybe there really is some stats way to make it numerical…
5. I’ll try and read one of those articles and summarize it. However, I warn you that I will give all the implications that I get (not just on your side). Also, you know the limitations of my style/skill (but will make a bit of an effort). Got one that is pdf, that you can post? Or do I need to go ILL it (I can).
Also (for comparison purposes) could you post a (modern) comparison paper on experimental gathering of cores (or first analysis). Doesn’t need to be the most egrigious example, but one that shows some of the modern tendancies, we posit as a problem.
Steve
Is the fact that the early Holocene was warmer than present in question?
Have you seen the paper
Sensitivity and response of northern hemisphere altitudinal and polar treelines to environmental change at landscape and local scales
Holtmeier FK, Broll G
GLOBAL ECOLOGY AND BIOGEOGRAPHY 14 (5): 395-410 SEP 2005
I paraphrase – at the landscape scale there is no scientific
justification for relating treeline position with only
one factor such as mean air or soil temperature Instead, many factors must be considered.
They’re likely not a silver bullet but well worth incorporating. (Of course there will be potential confounding factors for tree line movement also. Do we think they will be worse? Same? Better?)
In any case, more proxies->better.
Paul, a treeline derived temperature chronology will only establish a lower temperature boundary for a site. The actual temperature may be greater due to the many factors that will likely be less than optimal. Defining a lower boundary though not ideal is still quite useful.
If Wagner etal have done their Holocene CO2 work right, we could all have a lot of fun.
re#3,
Paul, have you consider that exactly the same complaint can be made about trying to use tree core ring widths or density as proxies for temperature? Steve’s main complaint, or at least one of the main ones, is that the proxyites have published tons of supposed temperature correlations without considering (except superficially) the other factors likely to affect the widths or density. Where is the heavy lifting which justifies tree rings as temperature proxies?
I say use both methods, just be sophisticated about deconvoluting confounding factors (where possible) and forthright to acknowledge dangers (where not).
Re: #3
Paul, your paraphrase sounds very reasonable.
Would not all proxies be influeneced by many factors?
I think that all proxies are influeneced by more than just temperatures. My question would be which of the many proxies which have been used for temperature should be weighted most strongly in creating a paleoclimate temperature approximation? Perhaps treelines are a better proxy than tree ring widths.
I have seen 4 proxies used for temperature: tree rings, tree lines, ice cores, and stalactites. Most likely there are others. Clearly each of these proxies is influenced by multiple variables.
Has anyone compared difference proxies to see which might be the “best” temperature proxy?
re #6,8, & 9: (Different Paul here.) I’ve made the point here several times that tree rings can’t be temperature proxies for the very reason that they depend on multiple variables. As we all learned in high school you need as many equations as unknowns to solve for the unknowns. Somewhere, maybe, a botanist has figured out how tree ring width varies with temperature, CO2, water, sunlight, soil fertility and whatever else. But that’s only one equation. You’re still at least four short of being able to solve for temperature. No amount of statistical analysis can get you that information, it’s mathematically impossible. End of story. The trouble is that people get caught up in statistics and forget the basic underlying mathematics. Statistics real value is to help you place errors on your measurements, it can’t recover nonexistent information.
The same holds true for any other proxy that depends on multiple variables. Unless there are independent measurements of multiple quantities that depend on the unknowns, as many as there are variables, you will not be able to extract temperature.
I don’t agree with 10 (the “it’s impossible” school) and the comment ignores recent content addressing some of this. Yes, there are potential confounding variables. That does not mean that they are prefectly confounded or that you can’t take RW and MXD (or even derivative of RW or fancy things like that) to get additional measurements. OR that you can’t insert moisture proxies or the like to help deconvolute.
yeah, maybe it’s not adequate now. But “it’s impossible, ever”? Don’t buy it…
#11: Yes it is impossible. McIntyre and McKitrick Mineral Co. owns five mines: diamond, gold, silver, tin, and aluminum. The total annual revenues of MMCO for the last five years are
2004: $2.1B
2003: $2.7B
2002: $2.4B
2001: $1.8B
2000: $2.5B
How much money did the diamond mine make in 2003 and 2000? Next, we have the following information on total revenue and the revenue from the diamond mine:
1999: $2.3B, $127M
1998: $1.7B, $202M
1997: $3.5B, $ 95M
1996: $2.7B, $307M
Find a statistical technique which will allow you to calculate diamond mine revenue for 2002.
[If you can answer these questions correctly, I’d like to let you share in an extremely rich gold deposit that was just discovered…but I digress.]
Now substitute ring width for total revenue and temperature for diamond revenue. The relationship for this toy example is simple, total revenue is just the sum of the individual mine revenues. In the case of tree rings the relationship of ring width to temperature, water, and the other growth factors is going to be much more complicated, but it doesn’t change the argument a bit. Using density doesn’t change things either, you’re just averaging over ring widths to arrive at the density. What’s more, it’s not independent of the ring width so you can’t use it as a second measurement.
Convinced or do I have to get technical with you?
I’m well aware that 2 equations, 3 unkowns doesn’t work. That does not change my point.
WRT your second example: An interesting thing is to look at TRS curves and discontinuities or changes in the trend. Sometimes, you can make connections to changes in policy or leadership and annotate the graph and extract lessons. You might take the different US (freight) railroads over the last 10 years and do that curve and then say if lessons can be extracted?
Some interesting information about the treelines and the upper limits of some crops reached in the past in the Italian Alps are available into old works of some Italian researchers (geographers, historians, agronomists, soil-scientists, geologists and so on). I think that this information should be used with an extreme prudence but, nevertheless, it could be useful in order to reconstruct the climate of the Middle age in Italy (it can be considered that the present value of the mean yearly temperature in the Po plain is about 12.5-13.5°C).
For example Umberto Monterin, in his book of 1937, spoke about archaeological finds that testify the presence, during the Medieval Optimum, of cultivated vineyards in Valle d’Aosta, at San Valentino, above Brousson – at an height of 1350 m asl, where the present limit for vineyards is about 850 m asl.
Now, if we hypothesize that the thermal limit for the commercial cultivation of grapevine is represented by an yearly mean temperature of 10.5°C (Lamb, 1966), we can deduce that in the Po plain (100 m asl) the yearly mean temperature was 10.5-1.3+1250*0.005=15.5°C (I’ve adopted a thermal gradient of 0.005°C and a gain for southern exposure of the vineyards in the Alpine area of 1.3°C).
During the Middle Age, olive trees for the production of olive-oil were present in the Po valley (Fabbri, 2006) where the cultivation of this crop is today impossible due to the sub-continental climate with the presence of severe frosts during winter. Moreover some historical works speak about the cultivation of olive trees for olive-oil into the Alpine area of Italy during Middle age, in Valle d’Aosta (Monterin, 1937) and in Valle Camonica, at Monno – about 1000 m asl (Berruti, 1998). About Monno site, if we hypothesize that the thermal limit for olive tree cultivation is represented by an yearly mean temperature of 12.5°C, we can deduce that in the Po plain (100 m asl) the yearly mean temperature was 12.5-1.3+900*0.005=15.7°C (I’ve adopted a thermal gradient of 0.005°C and a gain for southern exposure of the olive trees in the Alpine area of 1.3°C).
Furthermore Berruti (1998) wrote that until 1500 the beginning of the season of grazing in the upper pastures in the Alps (valle d’Avio) was about 60 days before the present beginning. An anticipation of 60 days can be the consequence of an increase of about 2.5°C in the yearly mean temperature (for the Po plain 13+2.5°C=15.5°C).
The same evaluation (temperature of 15-16°C for the Po plain during the medieval optimum, 2-3°C above present values) can be probably carried out on the base of other information about the old tree-lines that can be deduced from the presence above the present treeline of podzolised soils or fossil trunks.
I think that could be quite important to promote an effort in order to collect this multi-disciplinary information that in many cases isn’t written in English.
References
Berruti G., 1998. Clima e comunità alpine. L’alta Valle Camonica e l’alta Valle Trompia tra il XIV e il XIX secolo, Grafo, 78 pp.
Fabbri A., 2006. L’olivo nella storia dell’agricoltura dell’Italia settentrionale, Comunicazione presentata al convegno dell’Accademia dei Georgofili “L’olio di oliva nel Nord Est”, Padova, 17 maggio 2006 (olivicoltura.crpv.it/Olivo_storia.pdf).
Lamb H.H., 1966. The changing climate, Methuen, London, 236 pp.
Monterin U., 1937. Il clima della Alpi ha mutato in epoca storica?, CNR, Comitato Nazionale di Geografia, 54 pp.