The Validity of SAT (Surface Air Temperature) Data and Problems with such Data
Warning:
Loooong Post
Note 2: Copy of a post that did not seem to appear in the forum list for some strange reason.
This is another new thread started from another thread because of my idea of wanted a single issue discussed, if possible, within one thread. The thread this came from was “Issues with Global Land Surface Temp Trends” and the specific topic was the reference made by Canuck to a research paper
http://blue.atmos.colostate.edu/publications/pdf/R-321.pdf.To Amaranth Rose. This paper is no where near 90 odd pages long. It is really only about 25 pages when the numbering, double-spacing, pages of citations and graphs at the end are taken into account. I do sympathise with anyone attempting to read it however. The language, in my view is obtuse, overly complex, and full of acronyms. The paper could easily have been written in English any reasonably well education person could understand without detracting from any of the technical nature of the paper. Unfortunately, this is true with a very large number of scientific papers and certainly is not limited to climate research. I sometimes wonder whether this is a hangover from the way the authors were taught that papers had to be presented or to some extent still part of the “exclusive club syndrome”. If it is written in dense enough language specific to a particular field of endeavour then it can only be deciphered by those expert in the field and thus it is “protected” from being, heaven forbid, available to the lay public who may misinterpret what is really being said.
The paper doesn’t seem to present anything new at all. It seems to be a reference to research done by a number of climate scientists and related experts over the past few years. Certainly, I didn’t find anything in it that I didn’t know before. The authors do seem to demonstrate a pro global warming bias even though the data they are presenting often does not support a general warming can be concluded from the data that is currently available unless adequately corrected. While the conclusions are somewhat vague, it does seem that the authors agree that it is really not possible to properly correct the data and that what has been previously spurned as irrelevant and very minor issues could actually have very large effects on the variation of the data between the real trend and that demonstrated by the data.
Somewhat disappointing is that this paper does not match the various problems with the data to the major variations between the reported SAT averages and those reported by Balloon and Satellite Data. Considering some of the figures used in magnitude levels greater than the enter alleged global warming over the last 120 years, the issue is relevant to what was being discussed.
The paper fails to mention averaging problems and just how biased these are, although I guess it could be argued that the paper was looking at physical reasons for data discrepancies. It also seems to whitewash the bias that the various errors introduced into the data collection is not random and therefore is not something that just simply averages out.
Examples are used such as the sealing of a car park, which then fades in colour before being resealed, or of a tree that grows before being cut down. Figures of some of the enormous effects these can have on the data are provided but the paper does not mention that pretty much every single problem to which the paper refers does not result in a random modification to the data but rather to an apparent warming trend when there may be no real warming trend at all. Sure, the sealing of a road near a weather station will result in a sudden jump in the recorded temperature and the fading of the road over time will lead to a slight decrease in that jump, however, overall the trend is for the recorded temperature to be higher. Same with the vegetation although this one is not as obvious. Weather stations were near vegetation very commonly until the 60s or the 70s and it was during this period that the science of meteorology really started to be a science, especially in the US. In other areas of the world, the process was delayed a bit. It became clear to those who were responsible for weather stations that having them in the shade is inconsistent and what generally happened was the shade was removed or the weather station was moved. If it was a big beautiful tree that shaded the post office then it was the weather station that moved. The recorded temperatures increased substantially because of all these moves but it is very rare to find any records of such a move. Often you need to actually locate photographs of the building at different times to determine what moves did occur.
The issue of airports is touched upon but not to the extent that such weather stations actually may have biased the world’s recorded weather. Airports changed over time and their recording systems went from rudimentary to often the most hi-tech of all weather stations for a country. These are locations where movement of the weather station occurs frequently, very rarely with any documentation and the movements, especially from the era of jet aircraft again introduced a major bias of the recorded temperatures being warmer.
Overall, what this research does demonstrate is just how appalling unreliable SAT data really is if it is to be used as a comparison of temperatures over time. The record coldest day is unlikely to be the day that the weather station has on record. The record hottest day likewise is unlikely to be the same. The wettest day has a better chance but unless in an open field even this would very much depend on the physical location of weather stations. For most people’s purposes, to say that a certain days was the hottest on record for 92 years, is “good enough”. But for the purposes of determining global climate trends “good enough” is nowhere near good enough.
The interesting point to all of this is I believed that we could manage to create a database of SAT that removed as much bias as possible and could be of value in general trend terms. The biggest issue to me is the consistency of data. That is, that the average being used is calculated the same way for all locations within the database. This is a big issue in itself.
But what do you do with the weather station that spent 25 years under a nice big deciduous tree and was then moved to a site near a building so that it was in shade for only a limited amount of months late in the afternoon. Do you, as the research suggests, use data from other surrounding areas so that the change can be “ironed out”. But here’s the rub. It is almost impossible to find any weather station that has had a consistent history in anything, whether it be the times of recording the temperatures, the equipment used, the location of the equipment, or changes to the direct environment around it or the more general environment in the case of urbanisation. So let’s take our 25 year under a tree weather station and we go and find the weather station 25 kms (about 15 miles) down the road. But this weather station is at an airport and during the period we want to correct that airport underwent a concrete apron expansion doubling the concrete surface of the airport. So we go 40 kms the other way, only to find that that town had a fire and it lost its weather station records for three years (if you think this is rare, go to the NASA site and look at pretty much any location graph – you will almost always find periods were data is missing).
I could go on for probably twenty or thirty pages myself on problems with the data, why such problems are rarely random, why they have especially in the 80s on been an accelerating bias towards increasing the reported temperature, with citations, data references, anecdotes etc but I think, as dry as the report is, it goes a long way to questioning the data everyone uses to “prove” global warming, with a huge amount of supporting research to back it up.
Regards
Richard
