A Quiet Anniversary
(1 Oct 04)

Have you adjusted your temperature data lately?

Historical temperature records usually include effects of various changes in the locations, and kinds of, thermometers, and of observer procedures, including times of day of observations, all of which can have "non-climatic" effects on the temperatures that get recorded.

Such "non-climatic" effects on recorded temperatures get considerable attention by researchers, and various kinds of methods have been devised to adjust the "raw" data so as to compensate for those effects.

Many people may recognize the name John Christy due to his work on calculating atmospheric temperatures from microwave data collected by NOAA satellites. He is also the state climatologist of the state of Alabama, USA, and when he attempted to answer the question "when was the hottest summer in north central Alabama?", his attempt required a large amount of research. It is discussed in considerable detail in a paper in the Bulletin of the American Meteorological Society linked from the abstract (at that page, click on the "print version" link for a PDF of the paper), and it included attempts to compensate for such non-climatic effects.

For a data set called the United States Historical Climatology Network (USHCN), researchers compiled monthly data of about 1200 weather observation stations. They also compiled data about changes of equipment, location, observer, and time of day of observation. Summaries of such changes are included in a station history file, and various adjustments to "raw" temperature data are based on data about such changes.

Other collections of temperature data, such as the Global Historical Climatology Network have other methods of attempting to compensate for such "non-climatic" effects, but for the stations in the USHCN, use the USHCN adjustments. And recently, the NASA Goddard Institute for Space Studies (GISS) started using USHCN adjustments for temperature data from those stations instead of using the methods used by GISS for temperature data from other stations.

One may get the impression that the USHCN adjustments are regarded as relatively good adjustments, even by at least some other researchers who use other methods for temperature data from other stations.

That may lead one to wonder how good are the adjustments made by those people for temperature data from all of the other stations in the GHCN/GISS collections.

Oh yes, the title of this article alludes to an anniversary. October 1, 2004, is the tenth anniversary of the most recent update to the USHCN station history file.

About Nature
(14 Sep 04)

A magazine named Nature is said to be a "journal of science".

Its mission statement is said to be

First, to serve scientists through prompt publication of significant advances in any branch of science, and to provide a forum for the reporting and discussion of news and issues concerning science. Second, to ensure that the results of science are rapidly disseminated to the public throughout the world, in a fashion that conveys their significance for knowledge, culture and daily life.

CO2 Measurement Problems
(25 Aug 04)

When reading about extraction and analysis of atmospheric gases from ice cores, you may notice some things that seem rather odd. One favorite such oddity is a statement in the file Historical carbon dioxide record from the Vostok ice core:

Using semiempirical models of densification applied to past Vostok climate conditions, Barnola et al. (1991) reported that the age difference between air and ice may be ~6000 years during the coldest periods instead of ~4000 years, as previously assumed.

Hot and Cold Summers
(23 Aug 04)

A paper titled "European seasonal and annual temperature variability, trends, and extremes since 1500" was published in the issue dated March 5, 2004, of the AAAS magazine named Science. It seems that the authors concluded that the summer of 2003 was the hottest summer in Europe since 1500.

It also seems than Hans Erren does not share that conclusion, and in an article at A debatable European summer temperature since 1500 he suggests that the summer of 1540 rivals the summer of 2003 in the category of hot European summers.

Among the points of contention are the accuracy of tree ring, and other, proxies for temperatures, as well as the validity of extrapolation of small geographic area information to large geographic areas. You may find the discussion interesting.

Let me mention an aspect that particularly caught my attention. In the comparison of the Luterbacher et al. graph with a graph based on the Pfister index, the Pfister index indicates that the notorious summer of 1816 was particularly cold, which seems quite plausible to me; the Luterbacher et al. graph does not, which seems not quite so plausible to me.

Was there a tree ring problem, an extrapolation problem, or what?

Mauna Loa CO2 Measurements Updated
(5 Aug 04)

In March of this year a widely reported story about CO2 measurements at Mauna Loa was discussed in CO2 Report Makes the Rounds. Since then, the data on which the report presumably was based have been posted at CDIAC.

Here is a graphic summary of annual averages of the CO2 measurements at Mauna Loa since 1960, including the data added for 2003.

Keeping in mind that the unit of measure is ppmv (parts per million by volume), it is clear that CO2 is a very small portion of the atmosphere. Among the so called greenhouse gases, it is a very distant second to H2O. And yet, some people portray CO2 as if it had almost mythical powers of dominating climate around the globe. A glance at some other measurements may help clarify the picture somewhat.


(3 month averages of lower troposphere temperature variances and inverted SOI)

While the atmospheric parts per million of CO2 have continually increased during recent decades, atmospheric temperatures have risen, and fallen, uninfluenced by any notions that some people may have about the powers of CO2. Atmospheric temperatures respond to realities, not myths.

As for the amount of the increase of atmospheric CO2 from 2002 to 2003, it was approximately two and one half parts per million, 2.54/1000000.00 to be more precise. Here is a graphic summary of the annual increases since 1960:



As he did in March, Miceal O'Ronain has provided, in both tabular , and graphical , forms, nice compilations of year to year differences of monthly, and annual, averages of Mauna Loa CO2 measurements.

Solar Activity and Terrestrial Climate
(28 July 04) updated (30 July 04)

On July 6, 2004, a BBC article, headlined Sunspots reaching 1000-year high, caught the attention of a number of readers. Related articles soon appeared in other publications, for example: Suspot activity hits 1000-year high at Swissinfo.org, The truth about global warming - it's the Sun that's to blame at the London Telegraph, and Hotter-burning sun warming the planet at the Washington Times.

It seems that at a conference of astronomers in Hamburg, Germany, Sami Solanki and colleagues presented a discussion of possible correlations between solar variability and terrestrial climate. The recent articles provide some clues to the content of the presentation, but a slightly closer look may be of interest.

Interestingly enough, the principal paper on which the presentation was based was published last year, and briefly discussed here by John Daly at Evidence for an Unusually Active Sun. An abstract of the paper is available at Millennium-Scale Sunspot Number Reconstruction: Evidence for an Unusually Active Sun since the 1940s. The paper is available in PDF form at I.G. Usoskin, S. Solanki, M. Schuessler, K. Mursula, K. Alanko, A millennium scale sunspot number reconstruction: Evidence for an unusually active sun since the 1940's, Phys. Rev. Lett., 91(21), 211101, 2003. (Hereinafter, Usoskin et al. 2003)

Of several closely related papers I will mention I.G.Usoskin, K. Mursula, S. Solanki, M. Schuessler, and K. Alanko, Reconstruction of solar activity for the last millennium using 10Be data, Astron. Astrophys., 413, 745-751, 2004.. (Hereinafter, Usoskin et al. 2004)

Using records of 10Be (Beryllium 10) concentration in polar ice, and some physical models for processes connecting 10Be concentration with sunspot numbers, the authors attempt to reconstruct average sunspot numbers for the period from the year 850 to the present. Their reconstruction indicates that the period of high solar activity during the last 60 years is unique throughout the past 1150 years.

The 10Be records they use include annual data from Greenland for the years 1424 to 1985, and eight year sampled data from Antarctica for the years 850 to 1900. They supplement those data with some 14C (Carbon 14) data, apparently for most of the period of the study, and with group sunspot numbers (GSN) from 1610 on. While their methodology has its problems, they claim it results in a better correlation of 11 year smoothed "reconstructed" sunspot numbers with historical GSN than a statistical regression would do, and they summarize their results in a graph that includes indicators of MM: Medieval maximum, Om: Oort Minimum, Wm: Wolf minimum, Sm: Spörer minimum, Mm: Maunder minimum, and Dm: Dalton minimum. It also includes estimated corrections of their calculated sunspot numbers in the solid red and green lines.

You can readily notice various differences between their Greenland and Antarctic results, as well as between either of those and GSN, particularly during the Maunder minimum. These, and some other difficulties, are briefly discussed in Usoskin et al. 2003.

Of the presentation in Hamburg there is available a file of slides, Solar Activity over the last 1150 years: does it correlate with Climate?. Among those slides is the above graph, as well as several other graphs not included in Usoskin et al. 2003. Among the latter are two on a slide labelled: Comparison: Sunspot number & Climate.



The next slide from the presentation offers



I would guess that the trend that correlates at 0.7-0.8 level is somehow related to the very smooth lines that run through each of those two graphs, including the sunspot lines that continue climbing right through the Maunder minimum. If their models cannot get more realistic about a phenomenon such as the Maunder minimum, they would seem to need much more work before asking people to take their climate correlations seriously. I suspect that one can find a variety of statistical trends that appear to correlate incidental aspects of phenomena. As displayed in the presentation, the proffered trend seems to me to be a distraction from identifying correlations based on physical causation.

A subsequent slide of the presentation asserts SN runs ahead of climate by 10 years (SN presumably being their calculated sunspot numbers). I can imagine various lead times between sunspot numbers and various climatic effects, but I cannot see a 10 year lead time in either of those two graphs. (If you can, please let me know.)

While Usoskin et al. 2003 and related papers, including their difficulties, seem to me to be quite interesting, it also seems to me that the attempt in the presentation to portray a correlation with either version of the hockey stick, much less with both versions, was not conducive to resolving any of the difficulties.

Some incidental comments: Usoskin et al. 2003, and the presentation, emphasize the very much higher sunspot numbers in recent years than in earlier years, even within the historical series of group sunspot numbers. I estimate that the 11 year smoothed GSN of 1959 is about 57 per cent higher than that of 1790, which by my calculations was the highest before 1935. However, using Wolf sunspot numbers (WSN), the 11 year smoothed sunspot number for 1959 seems to be only about 20 per cent higher than that of 1790. Still higher, but not as strikingly so. The next two graphs present annual, and 11 year smoothed, comparisons between numbers based on each of the systems.


(Data for these two graphs are from: GSN and WSN ).

I am not aware of the relative merits of the two numbering systems. One obvious advantage of GSN is that they extend back to 1610, while WSN extend back only to about 1700, and some might suggest only to about 1750.

Update added 30 July 2004

A recent UPI article, Climate: The Vanishing Solar Factor, has been added to the mix. I will comment briefly on just a couple of parts of it.

In order for the sun to force the climate to the little ice age observed during the Maunder Minimum, the change in the solar constant had to be about twice what has been observed during modern, zero-sunspot periods.

1. Modern zero-sunspot periods have been very brief. The most recent calendar month with a zero WSN was June 1913; the most recent calendar year with a zero WSN was 1810. A zero sunspot day, or week, may not tell us much about what the so-called solar constant would do during a zero sunspot year, or a zero sunspot decade.

2. The statement implies an assumption that variations of solar wind and the magnetic heliosphere, and variations of the flux of cosmic rays, have no climatic effects, but without addressing research results indicating that they do. Raimund Muscheler probably could direct the author of the article to some of that research.

In the 2002 Harold Jeffreys Lecture to the Royal Astronomical Society in London, Solanki said: After 1980, however, the Earth's temperature exhibits a remarkably steep rise, while the sun's irradiance displays at the most a weak secular trend. Hence the sun cannot be the dominant source of this latest temperature increase, with man-made greenhouse gases being the likely dominant alternative.

1. That statement suggests that Solanki had not yet read Usoskin et al. 2003, which he co-authored. :-) Again, solar irradiance is not the only solar variable.

2. If the "latest temperature increase" refers to a globally averaged surface temperature increase which allegedly exceeds lower free troposphere temperature changes, then it refutes, rather than supports, notions of greehouse gases being a plausible, much less a likely dominant, alternative.

Corrigenda Anyone?
(2 July 04)

Steven McIntyre and Ross McKitrick have posted an update about the process of resolving issues raised by their assertions of errors in the Mann et al 1998 paper of Hockey Stick fame:

The Corrigendum in Nature today (July 1, 2004) by Professors Mann, Bradley and Hughes is a clear admission that the disclosure of data and methods behind MBH98 was materially inaccurate. The text acknowledges extensive errors in the description of the data set. Even more important is the new online Supplementary Information (SI) site, which reveals for the first time that key steps in the computations behind MBH98 were left out of (and indeed conflict with) the description of methods in the original paper.