Here’s are interesting minutes of the IPCC WG1 leading up to the approval of the SPM.
Here’s something exemplar:
“Bordhag underscored the importance of consensus and certainty for policy makers. He stressed the role of IPCC in the climate change process and stated France’s support for a similar body for biodiversity.”
Overall, on the one hand I am surprised that they are being so open about how unscientific the whole thing is, but on the other hand, knowing the 5 second attention span and 200 word vocabulary of the average vidiot, I guess they figured that there would be very few people who would realize just how embarassing the true nature of these meetings actual is. Only a few people will ever read this, and fewer still will realize what a circus it has been and continues to be.
“I will therefore confine these comments to
the aspects of the “2007 Summary for Policymakers” which I find the most distasteful. They
come under the headings of unreliable data, inadequate statistical treatment and gross
exaggeration of model capacity.” — from an interesting commentary on the SPM by Dr. Vincent Gray here: http://www.climatescience.org.nz/assets/20072141112360.SPM07GrayCritique.pdf
Wow, that is an excellent excellent commentary by Dr. Gray! If you read only one commentary this year …. 10/10 — 2 thumbs, etc.
#2: One of the lead authors of 4AR/Chapter 11, Jouni Räisänen, is quoting Vincent Gray here (as an example of a “skeptic comment” to IPCC 4AR/Chapter 11 -draft):
This chapter is the most disappointing of the whole set. It contains hardly any actual regional climate information, based on observations. Where there are observations they are wrong…
RealClimate seems to have a different opinion about Dr.Gray.
“Gray and Muddy Thinking about Global Warming”:
The conclusion of the article was:
The Wall Street Journal has insinuated that there is some ageism involved in the reaction to Gray’s work (“Hurricane debate shatters civility of weather science,” by Valerie Bauerlein, Feb.2, 2006). The problem is not Gray’s age — we all revered Henry Stommel who did some of his finest work in his seventies. The problem is Gray’s failure to adapt to a modern era of meteorology, which demands hypotheses soundly grounded in quantitative and consistent physical formulations, not seat-of-the-pants flying. The WSJ also made much of the withdrawal of an invitation for Gray to join a debate on hurricane trends at an Atlanta tropical meteorology conference. We can’t speak for the organizers, but we find it easy to believe that their decision was guided more by the invalidity of Gray’s scientific reasoning than by any political or personal considerations.
RE: #2 and #5. They are different Grays – Vincent Gray and William Gray. Dr Vincent Gray seems persuasive.
…in quantitative and consistent physical formulations
Yeah? But how does that allow for “moving on” when your trial balloons have been shot down?
From Gray’s paper:
The mathematics that derived the 2001 curve have been comprehensively trashed by McIntyre
and McKitrick (2003). The curve does not appear in this Summary, and I understand it has
been withdrawn. Despite this embarrassment, this IPCC Summary claims:
“the warmth of the last half century is unusual in at least the previous 1300 years”
At least the curve has been withdrawn. Wonder what the HT thinks of this?
RE: #2 – I just had an “a-ha” moment. On page 5 of this critique, Dr. Gray mentions the highly inappropriate substitution of median daily temperature values for mean values. Consider the following. As of 1850, 90%+ of the measuring stations were located in Western Europe and the US east of the Mississippi. In the Marine West Coast climate zones which dominate the populated and most developed part of Europe and in the Humid Continential and Humid Subtropical zones of the Eastern US, the diurnal change between low and high temperatures is greater than only Equatorial and other Tropical lowland climates. But generally it is low.
Since 1850 and particularly since WW2 there has been a global population explosion in Arid and Semi Arid climate zones in both the Low and Low-Mid Latitudes. Here, we experience a much greater diurnal range, over the course of a typical year than in the climates where most people lived in 1850. And significantly, in terms of time spent at temperatures closest to the daily lows and highs, during there is substantially less time spent closer to the highs on an annualized basis. For an extreme example of this, look at the actual hourly plots of places like Palm Springs, Abu Dabi or Adelaide. For the most extreme examples look at Death Valley or Yuma. When the sun rises, the temperature goes through the roof and stays up there until sun down. During the Winter, after sundown it really plummets, relatively speaking. We simply do not have the RH in the atmosphere in these areas to keep the warmth in at night. Bottom line is, in arid and semi arid climates, annual median temperature is higher than annual mean temperature.
#9: That’s very interesting. So, even the anomalies are biased, if the number and locations of weather stations change. Fewer and fewer stations since 1950. I wonder if the relative number of stations in arid areas has increased. If so, that could explain part or all of the anomolies!
RE: #10 – Undoubtedly, arid areas such as the Persian Gulf, the western half of the USA’s Sun Belt, and Australia, have either added or retained stations while areas in dying industrial regions and at the margins of development at high latitudes or in interior intermountain areas, have been shut. The high growth regions get the money and the pull.
#9, from page 5
by halving the sum of the maximum and minimum
Isn’t that mid-range, not median. The Inferior Estimator of the mean of the normal population, when the sample size increases. Used somewhere?
Re mean temperatures – Haven’t the means always been computed as Max plus Min divided by 2? Of course the results would change if the stations changed to different climates.
The right way to calculate daily mean would be to average all the hourly readings (or whatever the period is, normally it’s hourly). That way if the overall curve is not sinusoidal, you’ll somewhat account for that. Of course, RMS is probably a better way to go than an average , but an average is good enough to account for non sinusoidality.
The optimal way to calculate mean daily temperatures is as you describe – from hourly, or even minute resolution data. Unfortunately hourly data isn’t available for much of the earlier station data.
So there is a choice: 1) discard all the early data, but this would severely restrict the length of the record; 2) estimate mean temperature from max/min where necessary and use all the data where possible, this is undesirable as there will be an inhomogeneity where the type of mean changes; 3) use max/min means throughout. Not perfect, but for many questions it is better than either of the alternatives.
RE: #15 – Which is why the surface data are, at the end of the day, essentially rubish. By lowering the bar to account for the early low quality data, we incur a positive bias due to the effect of growth in station coverage over the pasdt 150 years in arid low / low-mid latitude climates. Or we throw out everything before the hourly record era and end up with too short of a sampling period.
Back on #14 – RE: RMS – obviously, RMS value is going to be different from the “zero crossing” value of a perfect sinusoid. The beauty of RMS is that it somewhat mimics absolute heat in terms of trend. Absolute heat is where we need to end up to truly understand the thermal equation for the climate.
Perhaps this is a silly question, but couldn’t you use the new hourly data from each station to calculate a station-specific relationship between the mean and the max/min? For station with short hourly records, doing that would certainly be preferable to using a blanket (max+min)/2 formula. For stations with longer records, the change in the relationship of the mean to the max/min should also be a useful way of estimating the UHI effect, since the UHI effect typically changes the way that temperature declines after sunset and results in a increased min relative to the max. Such an affect should change where the mean falls since the rate of temperature increase in the morning would not show the same change.
estimate mean temperature from max/min where necessary and use all the data where possible, this is undesirable as there will be an inhomogeneity where the type of mean changes
Makes sense, need to think about it…
If 1 Hz sampled readings over a day are not symmetrical (e.g. true mid-range is not equal to true mean), switch from mid-range to hourly computed mean will be problematic. If they are symmetrical (on the average), no problem, the past data will just look more noisy (and it is!). Hmmm, computing the past-to-present global temperature is not that easy, lots of choices and adjustments 😉
I liked this comment about sceptics – search for this word in Steve’s original minutes:
On text noting high decadal variability in Arctic temperatures, Canada, supported by Norway, suggested removing a specific reference to a warm period observed from 1925 to 1945. The Coordinating Lead Authors explained that “climate sceptics” often point to this warm spell to question the IPCC for not acknowledging such warm spells. Participants agreed to keep the reference.
Canada and Norway defended a more intense censorship but the lead authors think that with such censorship, IPCC can also be criticized for censorship. That must be a hard time to be surrounded by all these obnoxious political big cats.
#19: I too noticed the comment when I read the minutes. It seemed so bizarre that I figured it the minutes were mistaken, or I misread it, or Pixie Dust* was involved.
#16: Your last sentence is absolutely right.
UHI effects, station relocations, recording practices, etc. certainly muddy our picture the near past. Even if we could fully untangle the last 200 years of observations the direction of absolute heat now that matters.
*Pixie Dust count correlates closely with bedtime.