Problems in the Surface Temperature Record

Readers of this site won’t be surprised at reports of problems in the surface temperature record.  Joe D’Aleo and Anthony Watt have teamed up on a new paper published by SPPI analyzing the surface temperature record in depth.  I have only skimmed it, but it looks terrific  (and includes a few weather station site surveys and photos by yours truly).  From the summary:

1. Instrumental temperature data for the pre-satellite era (1850-1980) have been so widely, systematically, and unidirectionally tampered with that it cannot be credibly asserted there has been any significant “global warming” in the 20th century.

2. All terrestrial surface-temperature databases exhibit very serious problems that render them useless for determining accurate long-term temperature trends.

3. All of the problems have skewed the data so as greatly to overstate observed warming both regionally and globally.

  • given the huge percentage of measurements used to calculate the global average are “interpollated” as opposed to actually measured I decided to try a small experiment.
    I took 3 sites from GISS, Pittsburgh, State College and Allentown and did a little interperative dance of my own. I simply averaged Pittsburgh and Allentown and compared that average to the actual State College records for the same year, 1985. Pitt to State College is about 135 miles State College to Allentown in about 160 miles, State College is roughly mid point between Pitt and Allentown.

    Here are the Jan to Dec differences:


    In this case the State College actual temperatures averaged .9 degrees cooler than the simple interpollated records. It could easily have been .9 degrees warmer. The point is, interpollating temperature records of VERY, VERY inaccurate and more than 50% of the global grid is filled in with interpollated data.

  • Fred from Canuckistan

    Garbage in . . . Global Warming out.

  • NormD

    I have always been somewhat unsure of the exact nature of global warming theory.

    I assume:

    1. Because CO2 is a gas, at equilibrium, it is evenly throughout the atmosphere.
    2. The heat retaining properties of CO2 are also evenly spread.
    3. If some long running thermometers show no net warming (plenty fall into this category), then global warming is falsified

    There is no need to have hundreds or thousands of thermometers, since we are looking for a trend, a small number will do. For global warming theory to be correct all the thermometers must show a warming trend. Even one trendless thermometer dooms the theory.

    If its OK for warming to be localized then doesn’t this demand a much more complex theory that has CO2 sources, sinks, flows, differential warming by concentration, some areas warming, some cooling, etc. etc.

    Am I missing something?

  • DaveJ

    @NormD: I think you are missing something. The effects of an overall warming trend would not necessarily be expected to be evenly distributed. Climactic effects often skew some areas up and others down – for example, the Pacific La Nina and El Nino do this. So, even long-running thermometers don’t necessarily falsify the theory. Climate models, I believe, do attempt to incorporate the sources, sinks, etc.

    However, it’s not clear to me that the theory is even falsifiable. Given that climate (and weather) patterns are chaotic, any finite number of measurements will not really be sufficient unless the trend is very strong and clear. Furthermore, the attractor states generated by the chaotic system cannot be predicted, for the same reason they can barely predict tomorrow’s weather, so there is no way to really know where to measure. Using such logic one could make a case that Global Warming is “not even wrong.”

  • Hey Skipper

    On a recent trip to Denali NP in Alaska, I saw an (I think) ASOS II weather station near one of the visitor centers that appeared to be very well sited. What’s more, if there is a part of the world that need not worry about UHI, that is it.

    So when I got home, I did a little googling (Denali temperature 1950 1960 1970 1980 1990). At the top of the list was a study on central Alaskan temperatures based upon a very continuous and uncomplicated recorded history — roughly 80 years.

    The format isn’t cut and paste friendly, so I will summarize:

    — There has been about 3 deg F warming over the last 30 years.

    — However, taken over the entire 80 year record, warming is nearly invisible; about .5 deg F.

    — Once accounting for the PDO, it is essentially impossible to detect any climate warming.

    And this is from an organization that is looking at the consequences of AGW for central AK.

  • Anon y moose

    “Finally this month, a peer-reviewed analysis of the temperature data was published in the Journal of Geophysical Research. The paper used Watt’s station ratings to split all US weather stations into two categories: good (rating one or two) and bad (ratings three, four or five). The analysis then compared the raw, unadjusted data from the good and bad sites. In typical peer-reviewed understatement, the results were described as “counterintuitive”. They were in fact, a great surprise to many. Poorly sited weather stations actually show a cooler trend compared to the good sites.

    The cause of this cooling bias appears to have been a change in instruments. In the late 1980s, many sites converted from Cotton Region Shelters (CRS, otherwise known as Stevenson Screens) to electronic Maximum/Minimum Temperature Systems (MMTS). This had two effects. Firstly, MMTS sensors record lower daily maximums compared to their CRS counterparts. So the switch from CRS to MMTS sensors caused a cooling bias in certain stations.”

  • NormD

    @DaveJ: But the theory is that the sunlight comes down at frequencies unaffected by CO2, heats the surface which then re-radiates at frequencies that are absorbed by CO2 and thus the atmosphere heats up. Correct? CO2 will eventually evenly distribute via diffusion. Correct? So if CO2 went up at Mauna Loa then is also went up in the middle of England. So all points on the planet will heat up relative to when CO2 concentrations were lower. If I had a temperature reading over say 100 years, I should see a gradual increase in temperatures that track CO2 increases. If I see a flat temperature profile, such as:

    Then I have limited explanations:

    1. CO2 does not evenly distribute. This would require very complex models showing CO2 sources, sinks, flows and localized heating effects (which also distribute). Impossible to show “global warming”.

    2. CO2 does not cause local heating or is counteracted by other local effects.

    3. CO2 does cause local heating, but this heating is counteracted by other changes in the global climate which are also caused by CO2. I think you have an Occams Razor problem with this because you have to imagine a global change that just happens to counteract the minute amount of heating caused by the CO2 induced warming.

  • I just watched (and listened) to your video at ClimateGate. It was great! I’m definitely a layman, retired for some time now, but had a (mostly unused) education, undergraduate B.S.(1960) Tulane, physics/math, and about 30 hours or so of graduate level math, mostly at the Courant Institute of Mathematical Science.

    I imagine you’ve seen the 6 part youtube video on Henrik Svensmark?
    Have you seen the analysis by “dad & son” using NASA GISS data, comparing urban / rural pairs, temperature data since 1900 ?