Scientific Robustness Of The University Of Alabama At Huntsville MSU Data
By Dr. Roger Pielke Sr.
As a result of the persistent, but incorrect (often derogatory) blog posts and media reports on the robustness of the University of Alabama MSU temperature data, I want to summarize the history of this data analysis below. John Christy and Roy Spencer lead this climate research program.
The ad hominem presentations on this subject include those from the weblog Skeptical Science who have sections titled
Christy Crocks and Spencer Slip Ups
If this weblog intends, as they write, to contribute to
Explaining climate change science & rebutting global warming misinformation
they certainly have failed in this effort, with respect to the outstanding research that Christy and Spencer have accomplished.
To summarize specifically the UAH MSU dataset, it has gone through about 9 revisions (A, B, C, D, 5.0, 5.1, 5.2, 5.3, 5.4 – some listed in CCSP 1.1.) Two of the revisions involved changes Jim Wentz of RSS spotted, but the other seven were ones John Chrsity and Roy Spencer discovered (i.e. major ones like the spurious warming due to a change in the sensor when the satellite went in and out of sunlight).
Such corrections are what happens in the normal course of science when you are the first to build the data set and discover issues as time goes on, especially when a satellite goes through a calibration shift. Their data are publicly available and their methods published in a diverse range of peer-reviewed journals.
A example of their reporting on a correction and acknowledging who found it (in this case Jim Wentz) is written in the article
Christy, J. R. and R. W. Spencer, 2005: Correcting Temperature Data Sets. 11 November 2005: 972. Science. DOI:10.1126/science.310.5750.972
Text from their article includes [highlight added]
“We agree with C. Mears and F. J. Wentz (“The effect of diurnal correction on satellite-derived lower tropospheric temperature,” 2 Sept., p. 1548; published online 11 Aug.) that our University of Alabama in Huntsville (UAH) method of calculating a diurnal correction to our lower tropospheric (LT) temperature data (v5.1) introduced a spurious component. We are grateful that they spotted the error and have made the necessary adjustments. The new UAH LT trend (v5.2, December 1978 to July 2005) is +0.123 K/decade, or +0.035 K/decade warmer than v5.1. This adjustment is within our previously published error margin of ± 0.05 K/decade.”
I also reported on an independent check on the robustness of the UAH MSU analyses in my post
where I reported
In order to further examine the robustness of the Christy and Spencer analyses, in 2006 I asked Professor Ben Herman, who is an internationally well-respect expert in atmospheric remote sensing, to examine the Christy and Spencer UAH MSU and the Wentz and Mears RSS MSU data analyses. He worked with a student to do this and completed the following study
Randall, R. M., and B. M. Herman (2007), Using Limited Time Period Trends as a Means to Determine Attribution of Discrepancies in Microwave Sounding Unit Derived Tropospheric Temperature Time Series, J. Geophys. Res., doi:10.1029/2007JD008864
which includes the finding that
“Comparison of MSU data with the reduced Radiosonde Atmospheric Temperature Products for Assessing Climate radiosonde data set indicates that RSS’s method (use of climate model) of determining diurnal effects is likely overestimating the correction in the LT channel. Diurnal correction signatures still exist in the RSS LT time series and are likely affecting the long-term trend with a warm bias.”
The plot of the lower tropospheric temperature analysis, as obtained from the MSU data, for the RSS and UAH groups are shown below. The degree of correspondence between them is another check on the value of both data sets in assessing long-term variations in the global averaged lower tropospheric temperatures.
The bottom line conclusion is that the statements made by Kevin Trenberth, John Abraham, and Peter Gleick in
Opinion: The damaging impact of Roy Spencer’s science
that
Unfortunately this is not the first time the science conducted by Roy Spencer and colleagues has been found lacking. Spencer, a University of Alabama, Huntsville, climatologist, and his colleagues have a history of making serious technical errors in their effort to cast doubt on the seriousness of climate change. Their errors date to the mid-1990s, when their satellite temperature record reportedly showed the lower atmosphere was cooling. As obvious and serious errors in that analysis were made public, Spencer and Christy were forced to revise their work several times and, not surprisingly, their findings agree better with those of other scientists around the world…
are grossly incorrect and a retraction and a correction by Trenberth, Abraham and Gleick is appropriate. Similarly, weblogs such as Skeptical Science, if they want to move the debate on the climate issues forward, need to move towards a more constructive approach.
=================================================================
I’d like to add this from Dr. John Christy, from a comment he left on WUWT in this thread. – Anthony
Some clarifications are needed. The orbital decay effect was discovered by Wentz around 1997 which induced a spurious cooling effect on one of our microwave satellite products (lower troposphere) but not the others. However, most people forget that at the same time Roy and I discovered an “instrument body effect” in which the observed Earth-view temperature is affected by the temperature of the instrument itself, leading to spurious warming (Christy et al. 1998, 2000). This effect counteracted about 75 percent of the orbital decay cooling effect – so the net effect of the two together was almost a wash (a point rarely acknowledged.)
In 2005, Wentz and Mears discovered an error in the equation we used for the diurnal correction in one of our products (again, lower troposphere) which we quickly corrected and then published a “thank you” to Wentz and Mears in Science for their cleverness in spotting the error with an update on what the magnitude of the error was. Again, the magnitude of this error was small, being well within our previously published error estimates for the global trend. (Note that we were first to discover the diurnal drift problem back in the 1990s and initiated various corrections for it through the years.)
Roy and I were the first to build climate-type global temperature datasets from satellite microwave sensors, so we learned as we went – and were aided by others who read our papers and checked our methods. My latest papers continue to investigate error issues of our products and of the products of others.
The review of my one publication in Remote Sensing last year was done quite professionally and it was clear to me and my co-authors that the referees chosen to review the paper were specifically knowledgeable of the various satellite, radiosonde and statistical issues, leading to some substantial and useful revisions.
Kevin Trenberth was my MS and PhD graduate adviser at Univ. of Illinois.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
I’ve managed to get hold of a copy of the Santer paper which Pielke commented on and wanted to know whether I’m barking up the wrong tree. I looked in vane for the model for the climate noise they were using. Eventually it dawned on me that they weren’t modelling the noise. Instead it must go like this:
I want to know the signal to noise of my estimate of the number of apples. First. however tell me how many apples you have. Right, now my signal to noise is infinite because my estimate is precise. QED, I have a fantastic model which perfectly predicts the number of apples (after it knows how many to predict – and as it’s an exact prediction there is no noise).
Likewise I think this Santer paper is doing the same: they are creating a model to match the temperature signal and then assuming that the noise is any mismatch. Which is why the S/N gets increasingly better, the closer it matches the whole sample period because it is increasingly easy to get an exact fit to the temperature series over the whole period. This is not what the noise model suggests if you analyse it as it increases with increasing periods.
This simply doesn’t get around the real question which is: “how much is signal and how much is noise”. Instead it really says: “everything we can’t explain by tweaking the model is noise” and therefore the signal to noise is what we can’t post-factum explain. This is absurd nonsense.
The real signal to noise is the deviation between the model and actual signal. So, e.g. we need to take models produced prior to 2001 and compare their “signal” which was around 0.35C/decade warming with the “noise” which was … around 0.35C/decade less, all the “signal” could be explained as noise. Then, if we back project this, we find the potential limit at this trend (given that trends can last centuries even millennium) is that the noise could be as high as 3.5C/century. This therefore suggests that the actual signal to noise over a century is as low as 0.25 to 1
In other words, far from proving that we have to wait 30 years to assess climate, what the Santer type approach shows is that models are so bad that all the signal could easily be noise.
omnologos says:
September 13, 2011 at 1:10 am
Very well said. I will add that they are so ignorant of scientific method that they are not aware that there is a difference between Aristotle and Galileo on scientific method. Their arrogance makes their ignorance permanent.
G. Karst says:
September 12, 2011 at 7:05 pm
Make that “Pagan zombified dupes.” It is a religion after all.
It was SkepticalScience that made me a skeptic.
Trying to find info about the Medieval Warm Period I visited SkepticalScience. Here it had two images, one showing the heat anomalies during the MWP and one showing the heat anomalies for today. The image for the MWP was a lot colder than today. After reading several posts the reason was obvious, the first image was the average temperature over the entire MWP period, about 300 years, I guess he decided when it started and ended. The one for today was the average temp for a single decade. Apples, pears, anyone. It was also an obvious and deliberate deceit.
That for me was game over, I mean How do you know when a liar is telling the truth?
Skeptical Science is a joke, just ignore it.
Dr. Christy – People change, and I guess that is Trenberth’s problem. He stopped being a scientist and became the number one shill for Al Gore.
I call it Septical Science.
Matt says:
September 12, 2011 at 12:09 pm
Any “argument” that somehow implies that the target of the argument is the individual(s) who made it rather than the points made in the argument itself, is effectively ad hominem. Drawing attention to potetnially mistaken positions on unrelated subjects is also effectively an ad hominem attack. So asserting that Spencer and Christy have some sort of “history” of being mistaken a personal attack against the men themselves rather than what they have said. In fact, drawing Christy into the debate regarding a paper by Spencer and Braswell is outright irresponsible. Christy was not an author of the paper. These kinds of behaviours are “political” debate tactics and not scientific. Laymen readers may neither understand nor note the differences, which is the purpose. The arguments are not directed to fellow scientists but to the climate “laity.” The hope is to restore or reinforce “conviction” in the religious sense rather than clarify and inhance understanding.
There are no models in the vanes; your search would necessarily be in vain. 😉 ;p
Your conclusion echoes a critique I read long ago, that the climate models are skillful only at selecting their preferred noise. Is that an accurate précis?
Scotty (first comment)
“If there was a prize for cherry-picking, “Skeptical Science” would be Fruiterer of the Year.”
I love this term for the cherry picking award and it should become an annual prize “Fruiterer of the Year”. Has a resonance hasn’t it?
Just to clarify the total magnitude of the errors, in January 2009, I made a comparison of the current version of the analysis to that reported in earlier work for the period that they shared in common (January 1979–April 1997). Here is what I found:
Trend prior to 1998 paper by Spencer and Christy ( http://ams.allenpress.com/archive/1520-0442/11/8/pdf/i1520-0442-11-8-2016.pdf ): -0.076 C / decade.
Trend with correction in 1998 paper: -0.046 C / decade.
Trend using method of analysis current as of Jan 2009: +0.029 C / decade
In the 2008 paper, Spencer and Christy stated “We estimate the precision of the overall trend as +/- .05 C / decade.” However, the total change in the trend from immediately prior to the 1998 paper to the Jan 2009 analysis version was in fact +0.105 C / decade (with 0.075 of that occurring between the 1998 paper and the 2009 analysis.
Since the trend for the full data set through Dec 2008 was +0.127 C / decade, the change due to the longer time series is +0.098 C / decade. So, a little more than half the change in trend between the -0.076 C/decade trend prior to the 1998 paper and the +0.127 C / decade as of Dec 2008 was due to corrections and a little less than half was due to the longer time series.
[Since Jan 2009, there have been some changes in the UAH analysis, but as I understand the changes, they are mainly in the annual cycle and have very little effect on longer trends, so I haven’t bothered to redo the calculation with the latest version of the UAH analysis.]
The link I gave in my previous post appears to be out-of-date. Here is a working link to the 1998 paper of Spencer & Christy: http://journals.ametsoc.org/doi/pdf/10.1175/1520-0442-11.8.2016
Skeptical Science isn’t. Sadly it is nothing more than an echo chamber for pseudo science. Models are treated as science and data is treated as useless. Not worth reading anything there.
Davy123 said
After reading several posts the reason was obvious, the first image was the average temperature over the entire MWP period, about 300 years, I guess he decided when it started and ended. The one for today was the average temp for a single decade. Apples, pears, anyone. It was also an obvious and deliberate deceit.
This is interesting. Could you tell me which 10 year period from the MWP you would like to use in the comparison and how much data you have to establish the worldwide temperature for those 10 years?