It’s like a Whole New Post

If you have not visited my post lately on my son’s experiment on urban heat islands, go check it out, its like a whole new post.  Sixty comments and at least five updates.

I appologize to all the climate alarmist posters who have found my son’s project (to measure the Phoenix urban heat island) to be insufficiently rigorous.  I am sure all your baking-soda-and-vinegar volcanoes in 8th grade were much better done.

23 thoughts on “It’s like a Whole New Post”

  1. I will check it out.

    I thought it was a great science project for the kid. I said so too. Let me wax on a bit more. It was neat in conception. How it could be done with simple materials. How it incorparated a few things. How it was still simple enough that the boy could really have ownership and understanding. I really thought the camaredirie was touching, too. How you could work together on it. Good on you and the Longhorn.

    Just the comments about how showing that Phoenix has a UHI is anything new, that it moves the needle…those comments were silly, given that the work shows a UHI exists and that your opponents don’t contest that. I really worry about the Firm if it’s that hard for a consultant to dissaggregate issues.

    Of course they are busy getting all gaga over Hillarette and AGW themselves. And they were Enron silly billies. So maybe that’s a point in favor of skepticism. If McK is for AGW, it must be wrong. God knows half the crap done in late 90s, early 0s, seemed to be followed by scandal and bankruptcies!

  2. Ok, done checking it out. You’re walking back, but too slowly. Let me summarize your idiocy.

    A and B are connected by an AND gate to give output C.

    A is agreed by both sides to be ON. B is debated by the two sides. You do a great demonstration of effect A being on.

    You: Look how A is ON!!!!

    Us: We know that.

    You: But I also assert B is on, therefore proof of A being on should settle your hash.

    Us: Sigh.


    Lather, rinse, repeat.

  3. TCO, a metaphor is not an argument, however much you want it to be. You never specify what B is, or what conclusion he’s claiming to have reached. So, let’s take it from the top:

    The Null Hypothesis: Urban heat islands have had a negligible effect on surface temperature readings.

    The Test: To identify whether or not the UHI effect has a negligible effect in a city which has grown by an order of magnitude since the warming trend in surface temperature readings began.

    The Result: The UHI effect has a VERY non-negligible effect on this city. (The test itself isn’t perfect, but repetition in different environments would increase its accuracy)

    The breakdown: Given that the greater Phoenix area has gone from four hundred thousand to four million residents since the trend began, and given that the UHI effect will have a greater than linear growth coefficient (owing to heat spillage being reduced by a declining area-to-surface-area ratio), we can assume that all but a lesser order of magnitude of the total UHI will show up as a bias in the surface temperature readings. (I/e, if the final UHI effect turns out to be a 6 degree difference, then shifts in the UHI effect will result in a 5.4 degree temperature trend in Phoenix.)

    Now, I’m aware this isn’t perfect – changes to things like building materials will have changed the UHI dynamics; and other factors (such as humidity) will change the readings, hence the necessity for more experiments to reduce the uncertainty, which would presently be quite high. But the experiment itself is on perfectly fine footing, and your non-argument based on boolean logic doesn’t even make sense. (Particularly since this is a Bayesian field.)

  4. Climate skeptic – the experiment seems well done (although I’m curious as to why one of the graphs shows a continuous change in temperature while the other consists almost entirely of pairs of points of equal temperature separated by a small distance). But it doesn’t tell us anything new. Your claim that it was contributing something to science is sadly false.

    Adirian – disastrously wrong logic there. The null hypothesis you describe is so ambiguously worded as to be useless. The result doesn’t have any bearing on whether urban heat islands bias the temperature record. I don’t think you understand the difference between the existence of urban heat islands and their effect on the temperature record. Let’s make it clear one more time:

    1. Urban heat islands exist.
    2. They can be very significant (like in Phoenix, which is mentioned in GISS papers as a well known large urban heat island).
    3. Their effect can be corrected for, and you can see that this correction works by excluding urban stations entirely from the dataset, which gives you essentially the same warming trend.
    4. Many other independent lines of evidence also rule out urban heat islands as a significant bias, like satellite measurements agreeing with the surface record, sea surface temperatures also rising, glaciers melting, the greatest warming being seen in the Arctic, and there being no correlation at all between urbanisation and warming.
    5. Therefore, urban heat islands have not biased the surface temperature record.

    Point out any flaw you can see in this reasoning, please.

  5. Scientist, could you provide a source for your #3? I am interested in seeing a proper analysis of warming trends using only rural station data. Could you comment on McIntyre’s re-analysis of the Peterson 2003 dataset? ( ) Using the data points Peterson classified as rural, he was unable to detect much of a warming trend at all from the late 1800s to present. Is McIntyre missing something?

    From McIntyre’s graphs, it is clear that there has been a recent warming trend, which would explain things like glaciers melting, and the disappearance of sea ice in the far north, but in the rural data, we’ve yet to warm back up to where we were in the 40s.

  6. Is it possible that an experiment such as yours would help to answer the question of whether or not a site should be designated as urban or rural. There has been an extraordinary amount of trafic at Climat Audit and at Tarmino’s blog Open Mind
    about the validity of certain adjustments to an urban station such as Puerto Maldon in Bolivia. Now it turns out that the site may be rural and should therefore not even be subject to the urban correction. I would think the Urban/rural designation should not depend on the size of the town whose name the station bears but whether or not it is subject to a UHI and if so when did it start (why else does GISS distinguish between urban and rural if not because of UHI?).
    Do you think your apparatus would detect a micro heat island around an otherwise well situated rural site where for example the mesuring equipment is situated over asphalt, next to a steel shed? And does GISS acknowledge any Micro Heat Island effect?

    Anyway, great project. I wish I had thought of it. I wonder what happens if you compare windy days to calm ones. Has anyone done this?

  7. The weird thing is that all you “side” people think that I am against the hypothesis of UHI biasing the results. And I’m not. Per se. I really don’t know one way or another. But showing a big UHI in Phoenix which is AGREED TO, does not move the needle. It’s just sophistry or confusion or something. Refer to the issue tree that I describe before (A, B, and C).

    A lot of the comments around “B” (impact of UHI on the average) ARE relevant to the discussion. They’re great. Let’s go deep. Let’s get it on mofos. Let’s do math. Let’s do science. Let’s do top-notch (non Houston office) issue analysis.

    But showing the UHI in Phoenix is not a debated point. It doesn’t affect understanding. It is a granted point. Dwelling on it is either showmanship or idiocy.

    Lather, rinse, repeat, lather, rinse.

    P.s. If the young man is reading this stuff…I really do think it’s a heck of a science project, your old man is a heck of a guy, and…umm UT has a lot of hottie chicas. Peace, bra.

  8. Scientist, the null hypothesis is quite clear if you understand each of the words. But words don’t seem to be your strong point

    And TCO, B is derivative of A. It’s not A && B -> C, it’s A->B->C. That is, if the UHI is non-negligible and at least proportionate to non-negligible populations, then non-negligible changes in population will result in a non-negligible change to the UHI, and thus to temperature trends as a whole. Phoenix has experienced an increase in population of one order of magnitude – unless you’re arguing that this will not have resulted in an order of magnitude increase in the UHI, that’s a very significant TREND. That is, that IS B. The ONLY reason you need additional evidence for B is that not all of the 6 degrees of variation may be a result of the UHI effect. B is proportionate to A*dP, where dP is urbanization or deurbanization trend at work.

  9. No, Adrian, it’s not at all clear. You haven’t even attempted to distinguish between the local and global effect, or the spatial and the temporal effects of heat islands, or between the existence and the effect of heat islands in your supposed ‘null hypothesis’. Very sloppy. Your description of a null hypothesis is useless as a result.

    joshv: have a look at Hansen et al 2001. Have a look at the figures on page 20. They show that after correction for urban and other effects, the warming trend seen in urban, peri-urban and rural stations is almost identical.

    As for McIntyre, I don’t trust him to do good analysis. He is, after all, a mining executive and not a scientist. Non-academics can do top class scientific work, of course (the discovery of the Silverpit crater by people searching for fossil fuel deposits is a fine example), but McIntyre’s work is notable by its absence from decent peer-reviewed journals, and his website is full of pejorative references to climate scientists – you can be in no doubt that he is thoroughly biased in his approach.

  10. Scientist,

    I took a look at Hansen et al 2001, page 20. What we have here are graphics showing no, or nearly no trend in the difference between adjusted rural and urban station data. I don’t know what “adjusted” means, and I don’t apparently get to look at the raw data, so I can’t judge for myself the effects of the adjustment.

    On the other hand McIntyre creates a graph of the raw data for rural stations using Peterson’s own station selection criteria, and produces a graph with no observable trend over the entire interval.

    So, if after adjustment, overall US average surface temperature has a significant uptrend (from page 21), and the difference between US average and rural sites has no uptrend (from page 20), this must mean that the adjusted rural data also has a warming trend.

    And yet McIntyre’s graph of raw rural data shows no warming trend.

    There are a some possible explanations for this:

    – The most obvious would be that the process of adjustment is introducing a warming trend into the rural data that is not present in the raw data.
    – McIntyre has made some sort of a mistake, or has faked his data. You might be a fan of this explanation, but his graphs are simple enough to check. It’s raw unadjusted data, and I am sure he will provide his data points if asked.
    – McIntyre’s (and therefore Peterson’s) selection criteria for rural stations is not representative of rural stations at large, and perhaps differs significantly from those used in Hansen 2001.
    – There is some sort of statistical mayhem occurring in the highly mathematically dubious process of “subtracting” two different averages of station data.

    Which is it? Or is there something I am missing?

  11. Have a look again at page 20. You can see that indeed, the raw rural data doesn’t show a warming trend. You can also see what happens at each stage of the data analysis to change that. It’s also discussed on page 7.

    Why did you think you couldn’t get access to the raw data? There is a section entitled ‘source data’, under which you can find this: The source of the monthly mean station temperatures for the GISS analysis is the Global Historical Climatology Network (GHCN) of Peterson and Vose [1997] and updates, available electronically, from the National Climatic Data Center (NCDC). You can also look at the reduced data if you want. From page 11: The current GISS analysis of surface air temperature change is available at The data set can also be obtained via ftp at The previous analysis [Hansen et al., 1999] continues to be available at the GISS web site, but it is not updated each month as the new analysis.

    Just a point to remember – this is all about US temperatures. Even if you decide you fundamentally disagree with the data processing, and you don’t believe they’ve shown that there is any warming, you need to look at the rest of the world for the whole story. Glaciers are melting, ice caps receding, sea levels rising and floods and droughts becoming more common. You don’t even need to measure the temperature to see that it’s rising.

  12. Sorry, I was off by one page due to the oddities of my PDF reader. Looking at the actual page 20, it seems that for the rural US, raw station data, we have an overall temperature anomaly of -0.05 degC from 1900-1999. Tiny, and negative.

    Good then we are in agreement, the raw rural data doesn’t show a warming trend.

    According to page 20, the adjusted rural data however somehow takes on a roughly 0.4 degC anomaly via the adjustment process, which closely matches the adjusted urban anomaly, which then allows Hansen to produce the following page of trend graphs showing no overall difference in the trends between rural and urban station data.

    This quite clearly demonstrates that warming in rural areas is purely an artifact of the adjustment process.

    Looking at the raw data, knowing that UHI is real, and that urban areas have been expanding, if I wanted to calculate the average US temperature over the last century I’d simply throw away the urban data points and never consider them in my analysis. Sure you’d have lower coverage, but certainly no worse than other areas on the planet.

    But instead they attempt to correct for UHI by using nearby rural stations – fair enough. I would expect this process to lower the anomaly in urban areas, given the absence of anomaly in the rural raw data. But instead, on page 20, we see the adjustment process raising the anomaly. I would also expect such an adjustment to leave the rural data untouched, as after all, it’s serving as the gold standard for the nearby rural stations, but no, somehow the adjustment process raises the rural anomaly as well. This is fascinating mathematics.

    I do not comprehend this adjustment process. You take one set of data with a trend, another set of data without a trend, and somehow combine them to create two sets of data with almost identical trends that are larger than the original.

    You will have to forgive this layman for failing to understand why we shouldn’t just use the raw rural data.

    As for the rest of the world, I’d like to sort out the US first as we’ve got some of the best data on the planet.

    As for melting glaciers, etc. I will readily admit that the US, rural or urban, using raw or uncorrected data, is significantly warmer than it was 30 years ago. Thus I think it is reasonable to believe that the rest of the world has also significantly warmed, this may or may not be the cause of some of the effects you list.

  13. Scientist: “Surface temperature readings.” A reading is taken locally. You can’t have a regional reading – the concept doesn’t even make sense. The language is, indeed, perfectly clear.

    You’re correct, though, I didn’t make a temporal distinction – but, and you might gather this as important, this is wholly because the experiment in question didn’t cover the temporal direction. The question is wholly whether or not the UHI has a significant effect on individual readings.

    If it does, it is then either extrapolation and definition, or further experimentation, which renders the temporal case meaningful. This is “The Breakdown” – this was me extrapolating these results on a temporal scale using armchair logic, basic knowledge of thermodynamics, and the definition of the UHI effect. There’s a reason I made that a distinct statement from “The Results.”

    Am I done with your linguistic nonsense yet?

  14. The short form or Joshv’s posts: Hansen, et al, apparently “correct” for the urban heat island effect by averaging the corrupted data from urban locations with the good data from rural locations.

    Now, we might be misinterpreting Hansen’s calculations – but since he’s keeping them secret…

  15. adirian: Socratic dialogue: what is the difference between calculus with curly d’s and straight d’s? What is the relevance of this question to issue disaggregation?

  16. markm – you know he publishes papers? Have you tried reading them? Have you tried looking on the GISS website? If you think these things are kept secret, then you really haven’t even tried looking.

    Adrian – if by ‘linguistic nonsense’ you mean describing science carefully and accurately, no, you’re not done. Your result didn’t relate to your null hypothesis (which wasn’t worded carefully or accurately), and your ‘breakdown’ was supposition unsupported by the data. Sorry and all, but science does involve a bit more rigour than that.

    joshv – you can use just the rural data by all means but you can’t use the raw data. You do first need to correct for various effects, as described in the Hansen paper we’re talking about. The urban adjustment is only one of them. The process really is described in plenty of detail. Understanding scientific papers is not simple but the detail is all there if you persevere. And like I say, you can ignore every single word Hansen has ever written and you can still find extremely clear evidence of global warming and humanity’s influence on it.

  17. Thanks you Scientist, I’ve read the paper. They catalog the various adjustments, providing scant information on the details of the implementation of each adjustment. Oddly, all of the adjustments lead to an increase in warming trend.

    Looking again at page 20, I see that the time of day adjustment, apparently due to the fact that many of the stations have, in concert, conspired to change the time at which they record temperature, in such a way that decreases observed temperature, contributes a 0.17 degC warming delta to the raw rural data.

    Further then the “Max/Min & SHAP Data” adjustment contributes a whopping 0.22 degC warming trend delta. But wait, I read the rest of the paper. And I didn’t see anything about a “Max/Min & SHAP data” adjustment. Perhaps this is the Station History Adjustment – SHA? – guess I don’t know the lingo. Assuming it is, I find it amazing that the strongest component of the warming trend is based on an adjustment for stations histories for which the paper admits there is very sparse data. For many stations, we just don’t know where they were located. We don’t know when the moved, or where they moved from. But never fear, we can adjust, and on the basis of this adjustment rests the largest component of the rural data warming trend.

    The “Data fill & urban adjustment” then contribute the final 0.05 degC delta. I am not sure why “filling” in data would contribute to a trend in any statistically significant way, or why the rural stations require urban adjustment, but whatever, it’s a tiny adjustment, and probably not significant.

    And all of this is presented with absolute certainty, error bars are utterly lacking. I was trained as physicist. I’ve read my share of scientific papers. One thing always present in most serious scientific papers is a quantitative estimate of error. Error bars however are conspicuously lacking in this analysis. Is the time of day adjustment exactly +0.17 degC? Really? Wow, that’s pretty amazing.

    Is the station history adjustment, based on admittedly poor data, exactly +0.22 degC? Exactly? Again – wow.

    Perhaps you can point me to another analysis that includes an estimate of error.

  18. Why is it odd that adjustments lead to warming trends? Would you rather not see a warming trend?

    Have you looked for papers which discuss errors in a more quantitative way or are you just assuming that they don’t exist? I can tell you they do exist and they are not hard to find.

    Again it’s worth remarking that the paper we’re talking about only covers the US. There is plenty of empirical evidence of dramatic global warming – the US has hardly warmed at all in comparison to other parts of the world. You and others seem to be getting very hung up on minor aspects of the temperature record of 2% of the planet, when simple facts like rising sea levels and 95% of non-polar glaciers receding tell us that temperatures are rising rapidly. You can’t seem to see the wood for the trees.

  19. You might want to look more into glaciers. Receding usually is happens long after warming. They generally aren’t an indicator of rapid warming (and I would think that as they get smaller, they would recede faster–less ice per distance to melt).

    I’m not aware of a significant increase in the trend in sea level rise either.

  20. If you’re not aware of an increase in the rate of sea level rise, then you haven’t read the literature. Try looking up Church and White 2006 (Geophysical Research Letters 33, L1602). Glacier recession is also accelerating – look up Dyurgerov and Meier 2005, Glaciers and the Changing Earth System: A 2004 Snapshot. And contrary to your beliefs, glacier recession is an indicator of rapid warming. The lag behind temperature rises is of the order of a decade – perhaps not as long after warming as you think it is?

  21. aaro,

    Forget Gores film that shows glaciers receding from 1980.

    This relates to the link above:

    On her second visit to Glacier National Park in 1894, Mary Vaux (pronounced “vox”) was aghast at how the Illecillewaet Glacier had retreated since her previous visit seven years earlier. The lowest edge of the Great Glacier, as it was also known then, was clearly withdrawing upslope. We now know that most of the world’s glaciers were in retreat then as they are now. CO2 was I believe about 220ppm at that time.

    Obviously Gore was unaware of the fact that glaciers were on the retreat before 1850.

    Gore mentions lake Chad, suggesting global warming is the cause of the lakes loss of water, when in fact the main cause of the loss is that water from the rivers that flow into the lake are being diverted in increasing amounts for irrigation purposes. It is estimated that about one-third of the stream flow today is diverted from the Chari River before its flow reaches Lake Chad.

    Diversion of stream flow had been at a relatively low level, until the late 1970s when Lake Chad basin countries began to sharply intensify their food and fiber (e.g., cash crop) production efforts. According to UNEP GRID, “between 1953 and 1979, irrigation had only a modest impact on the Lake Chad ecosystem. Between 1983 and 1994, however, irrigation water use increased four-fold. About 50% of the decrease in the lake’s size since the 1960s is attributed to human water use, with the remainder attributed to shifting climate patterns.

    There is a long list of omissions, half truths and mistakes in this film, it is beyond me why the US and the UK governments have insisted the film should form part of the school curriculum.

    The slight warming is natural.

Comments are closed.