Measuring the Phoenix Urban Heat Island

Note Updates at the Bottom.  Could we please agree to actually read the whole post and the updates before commenting?  All commenters welcome, and I never delete comments except in the case of outright advertisement spam

This is a project my son did for Science Fair to measure the urban heat island effect in Phoenix.  The project could also be called "Disproving the IPCC is so easy, a child could do it."  The IPCC claims that the urban heat island effect has a negligible impact, even on surface temperature stations located within urban areas.  After seeing our data, this claim will be very hard to believe.

In doing the test, we tried to follow as closely as possible the process used in the Nyuk Hien Wong and Chen Yu study of Singapore as published in Habitat International, Volume 29, Issue 3 , September 2005, Pages 547-558.  We used a LogTag temperature data logger.  My son used a map and a watch to mark our times, after synchronizing clocks with the data logger, so he could match times to get temperature at each location.  I called out intersections as we passed them and he wrote down the times.  At the same time, I actually had a GPS data logger where I gathered GPS data for location vs. time, but I did not share this with him because he wanted to track locations himself on the map.  My data below uses the GPS data, which was matched with the temperature data in an Excel spreadsheet using simple Vlookup calls.

To protect the data logger from the 60mph wind  (we tried to drive at exactly 60 so my son could interpolate distances between intersections) we put the datalogger in a PVC Tee:

We added some insulation to reduce the effect of heat from the car's roof, and then strapped the assembly to the roof with the closed part of the Tee facing forward (the nose of the car is to the left in this picture).

We drove transects two nights in a row.  Both nights were cloudless with winds below 5 mph.  Ideally, we would have driven between midnight and 6 AM, but this was my kid's science project and he needs to get to bed so we did it from about 9PM to 11PM.  We were concerned that the air might still be cooling during the test, such that as we drove out from town, it might be easy to mix up cooling with time and cooling with location.  Our idea for correcting this was to drive and gather data on an entire loop, starting in the center of town, going about 30 miles out, and then returning to the starting point.  That way, with data taken in both directions, the results could be averaged and the cooling rate would cancel out.  As it turned out, we didn't even bother to do the averaging.  The two trips can be seen in the plots, but the urban heat island shows through pretty clearly in the data and the slope of the line between temperature and distance was about the same on the inbound and outbound legs.

I used the GPS lat/long points to calculate the distance (as the crow flies) from the center of town (My son did it the hard way, using a tool on Google maps).

The first night we went north (click to enlarge):

The second night we went south.  The urban profile going south is a little squirrellier, as the highway we were traveling tends to dip in and out of the urbanization.

Here is the total route over the two nights.  I'm still trying to figure out the best way to plot the temperatures on the map (again, click to enlarge)

You can see the results.  Even at the too-early time of 9-11PM, the temperature fell pretty linearly by about 0.2-0.3 degrees F per mile from the city center (as the crow flies).

I would really love to do is to go down to Tucson and run this same test starting at the USHCN weather station there and driving outwards.  That may have to wait a few weeks until my job calms down a bit.

Update:  Per some emails I have received, it is theoretically possible for the urban heat island effect to be real and to have integrity in the surface temperature record.  The first way this could happen is if the official measurement stations are well sited and outside of growing urban heat islands.  I know for a fact by direct observation that this is not the case.  A second way this might be the case is if one argues that urban heat islands exist but their effect is static over time, so that they may bias temperatures but not the warming signal.  I also don't think this is very credible, give growth of urban areas over the last 50 years.

A better argument might be that because most US temperature stations are arriving at daily temperature averages from just measuring daily min and max temperatures.  It might be arguable that while urban temperatures cool more slowly at night, they still reach the same Tmin in the early morning as the surrounding countryside.  Unfortunately, I do not think this is the case -- studies like this one taken at 5AM have seen the same results.  But this is something I may pursue later, redoing the results at whatever time of day Phoenix usually hits its minimum temperature.

A good argument for the integrity of the surface temperature measurement system is NOT that scientists blind to local station installation details can use statistical tools to correct for urban biases.  After looking at two stations in the Arizona area, one urban (Tucson) and one rural (Grand Canyon) it appears the GISS statistical method, whatever this double-secret process may be [insert rant about government-funded research by government employees being kept secret] it actually tends to average biased sites with non-biased sites, which does nothing to get the urban bias out of the measured surface warming signal - it just spreads it around a little.  It reminds me a lot of my kids spreading the food they don't like in a thin layer all over the plate, hoping that it will be less noticeable than when it sits in one place in a big pile. 

Again, I have not inspected their procedure, but looking at the results there seems to be a built-in assumption in the GISS algorithms that they expect an equal chance of a site being biased upwards vs. downwards.  In fact, I seem to see more GISS corrections fixing imagined downwards biases than upwards biases.  I just don't see how this is a valid assumption.  The reality is that biases in outdoor temperature measurement are much more likely to be upwards than downwards, particularly over the last 50 years of urbanization and even more particularly given the fact that the preferred measuremnt technology, the MMTS station, has a very very short cable length that nearly gaurantees an installation near buildings, pavement, etc.

Update #2:  To this last point, consider this situation:  Thermometer one in the city shows 2 degrees of warming.  Thermometer two a few hundred kilometers away shows no warming.  Someone aware of urban biases without a dog in the hunt would, without other data to guide them, likely put their money on the rural site being correct and the urban site exaggerated or biased.  The urban site should be thrown out, not averaged in.  However, the folks putting the GISS numbers together are strong global warming believers.  They EXPECT to find warming, so when looking at the same situation, absolutely sure in their hearts there should be warming, the site with the 2 degrees of warming looks correct to them and the no warming site looks anomalous.  It is for this reason that the GISS methodology should be as public as possible, subject to full criticism by everyone.

Update #3:  I know that many commenters see one line or even a title to a post and jump to the comment section to bang out their rebuttal without reading the post. I typcally do not respond to such folks, but there are just so many here I feel the need to say:  Yes, the IPCC knows urban heat islands exist.  What I said, and I think it is true, is that the IPCC does not believe urban heat islands substantially bias the surface temperature record, and, if they do, their effect can be statistically corrected by approaches like that used by the GISS and discussed above in Update #1.  I admit that this experiment alone, even if the quality was perfect, would not disprove that notion, but it has to make one suspicious (skeptical, even?)  By the way, if you want to yell "Peterson!" at this point, see here.  The volume of interest, pro and con, on this post I think is going to motivate me to go down to Tucson and run the same test with this USHCN station as the urban starting point, and then we'll see.

By the way, my point is clearly not, as some skeptical supporters might make out, that urban heat biases in surface temperature measurement account for all historical warming.  Clearly that is not true, as satellites, which do not have this urban bias problem, have measured real global warming, though at a lower rate than the surface temperature record.

Update #4:  To some of you commenters:  give me a break.  This is a junior high school science project funded with a $65 temperature logger and a half tank of gas.  I am sure the error bars are enormous and the R-squared probably has little meaning  (to tell the truth, Excel just put it there when I asked it to draw a trend line through the data).  Some of the data on the second run in particular looks weird to me and I would want to do a lot more work with it before I presented it to my PhD review board.  That being said, I would be happy to put it in front of said board next to the typical junior high baking soda and vinegar volcano project.

Given our constraints, I think we did a moderately thoughtful job of structuring the project-- better, in fact, than the published Singapore study we emulated.  In particular, the fact that we did the run both ways tends to help us weed out the evening cooling effect as well as any progressive heating effect from the car itself.  I honestly had zero idea what we would find when we downloaded the data to the computer.  I kind of thought it would be a mess -- remember, we were not really doing this at the right time of day.   It was not until my son did the charts using his position log he took by hand that I thoughy, "wow, there is a big effect here."   That is when I decanted the data from my GPS logger to check his results using a little more accurate position vs. time data and produced the charts here.  As I said, I really should have averaged position data for the forward and reverse runs, but I think the charts were fairly compelling.

Update #5:  The other half of my son's project was to participate in the SurfaceStations.org survey of USHCN temperature stations.  He did a photo survey of two sites.  Below is a picture from the USHCN station at Miami, AZ.  Left as an exercise to the commenters who are defending the virtue of the US surface temperature netork:  Explain how siting the temperature instrument within six feet of a reflective metal building that is perfectly positioned to reflect the afternoon sun from the SW onto the instrument does not introduce any measurement biases.  As extra credit, explain why the black gravel and asphalt road and the concrete building 6 feet away don't store heat in the day to then to warm up the air around the instrument at night as the heat re-radiates.

Comments

[Mitchell]

Great experiment.

[dreamin]

Seems to me that to "close the loop," you would need to dig up the GISS data for stations in your area and find out how much they have been "corrected" for UHI.

[Ari]

Fantastic. Run a regression analysis and see if those two slopes are statistically significant (though by their looks, they probably are). Also, though you did add insulation, it might not be enough to shield the logger fully from the heat island of the car. In fact, the car may, with time, be cooling down or heating up slightly. A vacillating car temperature could definitely affect your results. To be certain that what you note can be fully attributed to the urban heat island, you should attach a thermometer to the roof of the car to measure its temperature at every point at which you take the temperature with the logger. Then simply subtract Troof from Tair at every point and plot the points again. (Yes, I know that temperature probably can't be subtracted like that -- still, even if the actual temperatures are thus made flawed, the trends, at least, will be more accurate.)

[Tony Edwards]

As far as controlling for car effects, surely a better way would be to make some sort of pylon and mount it near the front of the roof. A foot or so should be enough to keep the sensor clear of any car bias.
Good job anyway, I hope he got an award for the efforts. Come to that, you should as well.

[Tony Edwards]

As far as controlling for car effects, surely a better way would be to make some sort of pylon and mount it near the front of the roof. A foot or so should be enough to keep the sensor clear of any car bias.
Good job anyway, I hope he got an award for the efforts. Come to that, you should as well.

[Scientist]

Wow, you really think the IPCC has said that heat islands don't exist? Please, try reading any IPCC document. You have proved nothing that is not already extremely well known.

[Richard]

How about time? Do you have to correct for time? I would think that if you wait 2 hrs in the evening it would cool too.

You did ride out of the city and back in, so that should give you some correction.

[Al Fin]

Good work! Since the weanies at the IPCC deny the significance of UHIs, when the reality gets rubbed in their faces they may have some splainin to do.

Here's a question: When the reality of global cooling hits the awareness of earth's billions, what do you think they are going to do to the obscurantists and would-be climate dictators? What happened to Mussolini and Ceausescu will be nothing compared to what happens to "Eli Rabbett" and the other rodents, I fear.

[DKN]

Nice work - your son deserves an A+++! I would, however, suggest that you do a multiple regression using both distance and elevation vs temp. Temps vary with elevation (higher = cooler usually, and nighttime air drainage can make lower = cooler sometimes) Also beware of irrigated areas (evaporation from them can create local cool spots).

[BillBodell]

Scientist,

I usually go out of my way to be as polite as possible, but... you're an idiot.

NO ONE IS SAYING THAT THE IPCC DENIES THE EXISTENCE OF URBAN HEAT ISLANDS!

The IPCC says that UHIs are adjusted for and their impact is small. We're not sure about that. They say that they adjust for it (and I'm sure they do try), we're just not sure that they're doing it right. The primitive code they use to do this has recently been released and people are working to reverse engineer the adjustments they make. It would help the government employees that collect this data would be more forthcoming with their methods.

By making the claim that we are saying that the IPCC denies the existence of UHIs, you're either dense or intentionally obfuscating.

[morganovich]

scientist-

Fourth Assessment Report from the IPCC (2007: p.244) says the following.

Studies that have looked at hemispheric and global scales conclude that any urban-related trend is an order of magnitude smaller than decadal and longer time-scale trends evident in the series (e.g., Jones et al., 1990; Peterson et al., 1999). This result could partly be attributed to the omission from the gridded data set of a small number of sites (<1%) with clear urban-related warming trends. In a worldwide set of about 270 stations, Parker (2004, 2006) noted that warming trends in night minimum temperatures over the period 1950 to 2000 were not enhanced on calm nights, which would be the time most likely to be affected by urban warming. Thus, the global land warming trend discussed is very unlikely to be influenced significantly by increasing urbanisation (Parker, 2006). ... Accordingly, this assessment adds the same level of urban warming uncertainty as in the TAR: 0.006°C per decade since 1900 for land, and 0.002°C per decade since 1900 for blended land with ocean, as ocean UHI is zero.

as you can see, the ipcc does in fact make the claim that the UHI effect is exceedingly small and need not be considered "significant". worse, they base the bulk of this opinion on min temps. what about max temps? look at a weather map around a city in the summer. where do the highest highs occur? asphalt can hold an awful lot more heat than grass. your straw man argument is a nice try, but holds no water.

[Mr X]

Wow. This is the first story on your home page? I hope that anyone visiting this website sees how ridiculous your pseudo-science is.

If you ask me, you are doing a terrible thing by trying to confuse everyone, even your own son. The real scientists have spoken. There is no "scientific" debate. The only debate right now regarding the science of the issue is between those who hold degrees in applicable fields, and idiots like yourself who have nothing better to do than to try to disprove a report that took thousands of people months and months of hard work.

[Larry Sheldon]

Some of these people a scary!

Seems like the teenagers are the ones getting to the truth with open minds.

Keep the good work and don't let them grind you down.

I don't know what got me to thinking about it, but last night I got to thinking about aqn uncle that died of modern medicine.

He had stomach ulcers and the settled science of the time was that the only treatment was surgery, so they kept cutting him up until they killed him.

[Adirian]

Mr X, given that he's done a pretty good job of poking a pretty big hole in a pretty significant part of the report, it is those "thousands" of people who should be ashamed of themselves.

And you fail so badly at "real" science it's not even worth making fun of. Just a hint, though - "real" science is skepticism. It's not blindly following what those in positions of authority say. Science isn't consensus, it isn't what "scientists" say, it is, in whole, the act of testing ideas. That's it. Science = Testing. If you aren't testing, you aren't doing science.

[TDK]

"Wow. This is the first story on your home page?"

Gentlemen, I think we're found a rare specimen of Neverheardofblogus.

[Papertiger]

"Wow. This is the first story on your home page?"

Got to love it, when a puedoscience believer proudly asserts his stupidity.

[Kyle Bennett]

I'm thinking maybe "X" isn't a pseudonym, it's how he really writes his name. Or, it's TCO again, too ashamed to show himself under his "real" name now that he's sobered up.

[TCO]

It's a wonderful science experiment. It just doesn't show bias over time. It shows heat islands. That's NOT CONTESTED.

[TCO]

If you think that the amount of UHI is larger than known in the literature, you should write a paper on it. I think you will find that the magnitude of the UHI has been well studied. It is a trickier issue to say how it is corrected for, how it changes over decades, the impact on the overall trend (which is some average of posited contaminated and uncontaminated sites). However, just showing "Phoenix has a big UHI" is not new, not contested.

[Kyle Bennett]

"...it's TCO again, too ashamed to show himself..."

"Posted by: TCO"

I guess I was wrong, he has no shame.

[TCO]

It's really sad with all the interesting things to think about in this world, that we have to deal with someone who thinks that he can argue a conclusion (requiring mutual on states of A and B), by citing evidence for A being on, and ASSUMING B on, when his opponents concede A, but not B.

It's just really sad. Is this lack of intelligence? Lack of honesty? Self deception?

[TCO]

I mean it's really bumming me out. Are skeptics stupid? I thought we were supposed to be smarter and more honest. More willing to let the chips fall how they may, than the professionals. But instead we get game playing and citing of earrly data, without final results, we get crap "pulbickation" in echo-chamber filled blogs. We get lack of statistical testing. We get Johnny lies too, so I will arguments.

It's just really, really sad. This is not the ethic that I learned in school. This is not cricial thinking. This is not how scientists should be. Dick Feynman would rip you all up the belly, for your self-serving ways.

[TCO]

I mean it's really bumming me out. Are skeptics stupid? I thought we were supposed to be smarter and more honest. More willing to let the chips fall how they may, than the professionals. But instead we get game playing and citing of earrly data, without final results, we get crap "pulbickation" in echo-chamber filled blogs. We get lack of statistical testing. We get Johnny lies too, so I will arguments.

It's just really, really sad. This is not the ethic that I learned in school. This is not cricial thinking. This is not how scientists should be. Dick Feynman would rip you all up the belly, for your self-serving ways.

[TCO]

In your update #2, you site a couple interesting issues.
A. Biases of the data analyzers.
B. Whether potentially biased thermometers should be included (with correction) or omitted.

On A: yes, perhaps your "opponents" could have the bias, you ascribe. However, even in your words themselves, you show your bias: you have an implicit assumption that the trending up station must be contaminated. But you never cited it as a predicate. I've seen a similar thing from Watts where he shows a station with an up trend and says that it must be classic urbanization or station micro-site: he doesn't even include as a possibility, that the station MIGHT be accurate and just be recording an up trend. Of course, it's easy to find your opponents wrong, if you can't even encompass a world view that maybe they are right.

On B: This is the more interesting question. Very poorly addressed by the users of these adjustments or by the sniping critics. I honestly don't know what the statistical meaning of including such adjusted data to a trend is. How does it change the degrees of freedom and are we properly accounting for it? What is the meaning wrt long term trends of the geo average?

[Scientist]

BillBodell - are you illiterate or something? The man has clearly stated that he thinks that the IPCC doesn't believe in heat islands. He thinks that by measuring the urban heat island in Phoenix that he's 'disproved' the IPCC. Unfortunately he is making a fool of himself and also his son. He's got it into his head that GISS is evil, and nothing could possibly change his mind, even if GISS has specifically mentioned Phoenix as a good example of a strong urban heat island.

[Larry Sheldon]

Who (or what) is "TCO". Besides incoherent and annoying?

[TCO]

This is you buttplugs make me feel.

http://xkcd.com/386/

[BillBodell]

Scientist,

Let's try an exercise in communication. In the second sentence of his post, the author says "The IPCC claims that the urban heat island effect has a negligible impact". This makes it clear that the author knows that the IPCC is aware of urban heat islands. The author thinks that the IPCC believes that the UHI effect has a negligible impact. If you want to try and make a rational arguement you can debate whether the IPCC believes the UHI effect is negligible. But, since the author never said that the IPCC denies the existence of UHIs, it is not constructve to claim that he did. The author is arguing that the UHI effect is greater than the IPCC claims. Debate that, and then we'll have something to discuss.

[jim karlock]

Interesting, a little pack of Al Gore zombies that suddenly converged here. Time after time they have to shout down any challenges to their cash cow of global warming.

Here are my thoughts about the science:
A few years ago, I drove out of my town and watched a thermometer. I saw similar data.

At 60 mph, I’d be very surprised if any car body temperature affected the air temperature readings.

Who wants to set up a site like surfacestations.org to document this effect?

First, let me say: good work. Impressive.

You are measuring over an asphalt road. You probably should measure an open field at the end of the run out of the city to see if there is an asphalt heat effect form the road in the country.

Perhaps one should drive all major roads away from all of the weather stations. Some will go though the city center and some will go directly towards rural. All will add to the data.

If someone decides to do this on a national scale, how do you do it this using CHEAP instrumentation? (Some new cars have thermometers.)

Or, maybe, we just need to collect all of those area temperatures that TV news provides every day?

Thanks
JK

[Adirian]

TCO, let's put it simply: The conclusion of the runaway AGW proponents depends upon A, B, C, D, and E. A depends in turn upon F and G. The challenge here is to G - no, it doesn't disprove anything. It doesn't disprove the runaway AGW hypothesis. BUT it does poke a big hole in the assumptions that produced that hypothesis, in this case, that the UHI effect is negligible. (This is important because virtually all historic surface temperature readings have been situated in urban environments, particularly outside the United States.)

Does it prove that the data record is bad? No. But it does indicate some serious potential faults, and provides a reason for further study. It is, in essence, evidence for the claim that the science is not settled.

The skeptic questions the conclusion - skepticism is not arriving at another conclusion by the way.

[TCO]

Adrian: The existence of large UHIs is NOT CONTESTED. So how does demonstrating that they exist contribute to the examination of the trend lines? It's already accepted by your opponents that they exist. The question is if their changes over time have affected the overall record: to what extent has the factor changed with time and how well are time trends adjusted to deal with this. One can easily imagine that the UHIs are NOT being adequately adjusted for, but you have to look at that aspect. Just showing that they exist does nothing, because our opponents ALREADY agree that they exist.

CAN'T YOU GET THAT?

I mean this is child-like!!!

[Adirian]

TCO - the IPCC refers to the effect of UHI as negligible. The demonstration is not that they exist - can't YOU get THAT - but that they are significant. Yes, the argument is precisely that the UHIs aren't being adequately adjusted for - the opposition calls them negligible, and given that the UHI measured in this case is larger than the total change in temperature over the past century, I think we can agree they are anything but.

That is, the noise of the UHI is MUCH larger than the perceived signal of AGW. It's not the only noise that is being put forward as a problem, but it's one for which the evidence is readily presentable.

And the significant part of the efforts being discussed here are the examination of local CHANGES - that is, things which make differences to the trend, which obviously are not being well handled. For example, the measuring station being discussed in the other thread I'm arguing with you in. That's sort of the point of Watts' efforts - to show that the measuring stations, however well they may have originally been positioned, have with time been developed around, leading to a TREND of greater UHI noise.

And coming from the least polite individual on the forum, I can't take the accusation of child-like behavior seriously.

[TCO]

Well all that was done in this exercise was to show existence! You need to actually do some math, look at the whole network, look at the corrections, look at the nightlights, all that crap to make a call as to whether IPCC or GISS is citing a too high number because of UHI growth. I honestly don't know if they are or not. I have to see numbers and such to tell the difference.

P.s. I was just reading how the brighter someone was on a subject, the more likely he was to have a low certitude on an issue. And the less knowledgeable, the more certain and convinced of his foundation for said certainty. Listen to Rumsfeld...appreciate the philosophy of uncertainty.

[R.C.]

Hey, everyone; sorry to intrude on your verbal bar-brawl, but I have a question, for warming skeptics and true believers, alike:

As I see it, there are three ways to pooh-pooh the heat-island effect which may be adopted. You can either say...

(a.) Heat island effect has no possibility of affecting the warming conclusions of IPCC, no matter how big it is;

(b.) Heat islands would change the conclusions of IPCC if it were big enough, but it has been measured, and it actually isn't that big, so the conclusions are left unchanged;

(c.) Heat island effect was big enough to influence the IPCC data, but the IPCC data has used some method of adjusting the data to "correct" for its influence, which was sufficient to prevent it from skewing the results.

Now, it seems to me that one really can't argue (a.), and IPCC doesn't even try to. The quote given by Morganovich, above (Fourth Assessment Report from the IPCC, 2007, p.244) states, "...any urban-related trend is an order of magnitude smaller than decadal and longer time-scale trends....Thus, the global land warming trend discussed is very unlikely to be influenced significantly by increasing urbanisation." In short: If the urban heat island effect were big enough, it would make a difference, but it's "an order of magnitude" too small.

So, IPCC is arguing (b.), right?

Well, apparently not: To quote another post above: "The IPCC says that UHIs are adjusted for and their impact is small. We're not sure about that. They say that they adjust for it (and I'm sure they do try), we're just not sure that they're doing it right."

Okay, that sounds like option (c.): "The effect is there, it's big enough to make a difference, but we've corrected for it."

Which is it, (b.) or (c.)? Does IPCC correct for the heat-island effect, or just dismiss it as being too small to bother with?

If IPCC is actively adjusting for the effect, then they must have measured the effect in order to determine by how much to adjust, right? So, what was their methodology to measure it? Was it a better methodology than Climate Skeptic's little field trip described above?

Thanks for any clarification you folks can give.

[TCO]

I think their take is c, albiet your division into b and c is mildly confusing (partly me being tired).

To further complicate things: One of the points that the IPCC comment is trying to address is that they AGREE with skeptics that the impact of UHI growth is more likely to overstate temp in aggregate than to reflect total temp increase due to changed land use. Let me explain that.* If you could TRULY measure the temp at every point over land (and water) and integrate them, then the issue of UHI growth would be irrelevant. That is because land use itself is a cause of global warming. Pave the planet and it will warm. Everyone, warmer and skeptic agrees. But the issue is that only a very small portion of the planet is currently paved or increasingly being paved. Most of it really isn't. However temp instruments tend to be concentrated more towards areas where paving is going on (or exists already). So the potential problem is that localized paving could be falsely inflating the temps of the network imputed to be over all land (note that UHI has no appreciable impact on water temps).

There are various ways of dealing with this problem: essentially they mostly boil down to trying to use the parts of the network that are constantly rural (either by correction to all records, or just by using only rural records). In theory, what's really needed is to avoid records that are urbanIZING. However, that's not how they do it. One can easily imagine that this might be done properly or improperly. One has to dig into the details to see if the methods make sense. There are some interesting things for instance, like JohnV's charts where he shows similar temp patterns with all "best in class" stations as with the whole network. But bottom line, you have to dig into the details to make a call. Sorry Meyer, going with the ED's day one hypothesis is not good enough. You actually have to think this one through.

*sheesh, I feel like GAvin or something but I'm not. I'm a right to bear arms Republican...although pretty sick of Bush who is a Democrat when it comes to domestic policy. Rather have a Clinton in there and just have Congress stall than the crap...but I divert.

[Adirian]

TCO, I think you're more clear when you're tired.

Look at the experiment being described here as one item in two sets. The further experiments of this set should be to continue measuring the UHI effect of large cities, or even just Phoenix - the second set of experiments would be to measure the UHI effect of cities of approximate size to Phoenix at various points in time. This experiment demonstrates a non-negligible UHI effect, making the paired set of experiments worth doing, to establish exactly how much of the warming in Phoenix over the past, say, half century, have been due to urbanization.

Building materials have changed over the past half-century, so it's not going to be perfect, but we can get an idea. The population of the Phoenix area is ~4 million; I've found figures ranging from 100k to 270k in 1950 for the city itself (figures for the general area are harder to locate), but let's assume, say, 400k for the same geographic area. The next step would be for a reader who lives in a similar-environment and isolated 400k city to engage in the same experiment on a day with roughly similar temperatures (since UHI won't be linear with respect to "natural" temperature increases), and to take the difference, to find out approximately how much change the UHI effect has had in Phoenix, at least, over the past half century.

Anyone here live in a city whose general area houses ~400k people, with a roughly similar environment to Phoenix? New Mexico, Nevada, Arizona, and possibly Utah seem like likely states.

[TCO]

I fucking GET that this experiment shows evidence on point A. But point A is agreed to by both sides. Shit man, pull the dick out of your brain.

[TCO]

I'm almost physically sick with the butt-ass stupidity here.

There are so many important things to be done. Problems to be solved. Experiments to run. Analyses to do. And we end up having to explain and re-explain the most idiotic points like this. If this is skepticism, it needs a fucking enema. It needs soapy water rammed up its ass.

[Scientist]

Is the 'climate skeptic' aware that his 'update 2' is pure fiction? As is very well documented, urban sites showing a different trend to nearby rural sites are corrected to match the rural trend. Why does he think it's the other way around?

BillBodell - Don't you think it's a bit strange to appoint yourself as the interpreter of what the 'climate skeptic' is saying? Why do you think you know exactly what he means, when he himself refuses to clarify? If you think that somehow the difference between 'non-existent' and 'negligible' is important then I'm not much impressed at all. It's clear that the 'climate skeptic' has not read the many published papers on how GISS derives their surface temperature record. It's also very clear that he really does believe that he has shown something new and hasn't realised (although it has been pointed out to him) that Phoenix is a textbook example of an urban heat island. Why does he think he is 'disproving' the IPCC?

R.C. - you seem to think that the IPCC carries out scientific research. It does not. It describes the results of the scientific research carried out in the field of climate science regarding anthropogenic global warming. To find out which of your suppositions about urban heat islands is closest to the truth, try reading the literature. Don't rely on wilfully ignorant 'climate skeptics' to inform you - you might as well ask an Arsenal fan to commentate on a Tottenham game and expect to get an unbiased description.

[Pat]

Everyone knows that greenhouse gas concentrations increase at about 1% per year.

Although the exact sensitivity of global temperature to greehouse gasses is in dispute, it seems prudent to worry a bit over a phenomenon that is growing at 1% per year.

Consider then that in the US asphalt production is growing at 2% per year.

What is the effect of all this asphalt? No one denies that a blacktop parking lot has a surface temperature higher than a grass field. There is twice as much asphalt around today's temperature sensors as there was just a few years ago. When I look at Anthony Watts' pictures I don't see a lot of dirt roads.

Asphalt grows at twice the rate of CO2. Surface temperature sensors are increasingly likely to be mounted near asphalt. Even without this experiment or any other study, we should expect surface thermometers to show higher temperatures today than they did in the past.

the Urban Heat Island effect is prima facie plausible. Furthermore it can be directly experienced by almost anyone who drives into a city or just looks at a TV weather report. What is surprising is that the UN's IPPC has seemed to discount and minimalize UHI as an explanation for warming measurement trends.

The IPPC argues that adjusting for the small UHI effect is easy and they say they have already done so. Yet those who have looked closely at those adjustments are not encouraged. The adjusters don't keep very good records. In many cases their adjustments seem to be arbitrary and ad hoc. There is also a suspicion that the record has been "adjusted" to promote a warming agenda.

It is in the best spirit of science and freedom of inquiry to test the opinions of politiical authority. This kid is lucky to have a dad who encourages him to think for himself and test his ideas.

[Tony Edwards]

I think that several of you, in the heat of calling each other names, have forgotten one very important point. Not global warming, climate change or any of the science of adjustments, but the very simple fact that this whole effort was a schoolboy's science project. It may or may not have been suggested by his father, who, incidentally, has obviously gone a long way in helping his son, and as such, should be judged. School science projects are emphatically not Nobel prize candidates and are very often quite simple in form, design and objective and certainly not ground-breaking, nor are they expected to be.
So, just this once, leave off the mud-slinging, behave yourselves like gentlemen and apologise to the young man concerned for your behaviour. If you really want to act like... don't know, can't think of a descriptive word at the moment, but anyway go to some of the serious rough and tumble blogs.

[BillBodell]

R.C.

The answer is C.

Hansen at NASA (GISS) is responsible for putting out one of the two primary versions of the Surface Temperature record (the other is HADCru). He most definitely believes that there is such a thing as Urban Heat Islands. He adjusts the historical temperature to account for UHIs. What’s his method? He describes it in several places. Is this method accurate? Could it be better? What does his program actually do? These are all good questions. Hansen is not cooperating with efforts to address these questions. He was forced in September 2007 into releasing his code. There is a major effort underway at ClimateAudit to dig into this matter. If you want to know more, give it a visit.

Scientist

It is the policy of this blog’s author not to respond in the comment threads. I’m not sure that’s a good policy, but that’s what it is. How do I know? I read it on the blog. The difference between non-existent and negligible is significant. Non-existent means it DOESN”T EXIST. Negligible means that IT DOES EXIST, but has little effect. If you want to argue about how much effect UHIs have on the surface temp record or argue that the IPCC believes the effect of UHIs are more than negligible, then we’ll have something to discuss.

[Larry Sheldon]

Even if I agreed with "TCO" (I don't even understand most of what it says), I would still make this request:

Please ban "TCO" and its morphs for foul language. We want children to read this stuff.

[Adirian]

Sigh. We get rid of RW, and this "Scientist" individual appears following the exact same argument pattern - claim the post said something it didn't, then argue with everyone about the meaning of the word "is." Scientist, your pseudonym is laughable, and your argumentative technique makes your side look ridiculous.

And I think TCO has a good idea, basically, but his approach is acidic and therefore renders the effort useless. TCO, if you want people to listen to you, be polite and reasonable, and offer constructive criticism, rather than swearing wildly and making sullen comments about other people's uselessness.

The only use of acidity is to make somebody angry, which can on occasion be a useful approach - angry people tend to abandon facades of rationality, but only if they were only pretending at rationality to begin with; switching to rational modes of argument will then completely throw them off. (RW is a good case in point, as I utilized exactly that technique on him.) When arguing with rational people, it accomplishes little except to make you look bad.

[Tom V.]

If I recall, there was a couple of papers by some scientist named Parker who tried to prove that UHI does not exist.

I just Googled UHI and Parker together and found this story on Real Climate -

http://www.realclimate.org/index.php?p=43

"There are quite a few reasons to believe that the surface temperature record - which shows a warming of approximately 0.6°-0.8°C over the last century (depending on precisely how the warming trend is defined) - is essentially uncontaminated by the effects of urban growth and the Urban Heat Island (UHI) effect."

Seems like the RC scientists don't think it exists, yet here we have a student who can measure it. Maybe they never leave the office up there at RC.

Funny how stuff works out sometimes. No need to spread foul language, just give the kid an atta-boy and move along there TCO person.

TCO? What does that mean? Talking Completely Obtusely maybe?

[Alan Cheetham]

Here is an examination of Phoenix urban heat island effects using the NOAA temperature database: www.appinsys.com/GlobalWarming/Phoenix_UrbanHeat.htm
The nighttime minimum temperature has been increasing at Sky Harbor but not at the rural stations.

[Scientist]

Adirian - you must not be fully understanding the claims made by the 'climate skeptic'. He thinks that by measuring an urban heat island effect, he has proved the IPCC wrong. The experiment carried out only shows things which have been known for decades. And remember, the claim was that this would 'contribute something to science'.

The lack of ability of most people commenting here to understand the difference between the statements 'urban heat islands do not exist' and 'urban heat islands do not bias the temperature record' makes them all look ridiculous. And with your infantile talk of 'sides' I think you look rather ridiculous as well. Do you define your views according to which 'side' you see yourself as belonging to? Which 'side' do you think I am on? What are the 'sides'? How many 'sides' are there?

BillBodell - I can see that you are aware of the difference between the existence of urban heat islands and their effect on the surface temperature record. I think you can see that an experiment which shows that they exist says absolutely nothing about their effect on the temperature record. But the 'climate skeptic' cannot see that; how can you think that he can if you read what he says? For example: The IPCC claims that the urban heat island effect has a negligible impact, even on surface temperature stations located within urban areas. After seeing our data, this claim will be very hard to believe.

[Tuukka Simonen]

http://logicalscience.blogspot.com/2007/09/toughstations.html

Even if you compare the "best" temperature stations with the most contaminated ones you'll get very small differences in averages. And also note this:

"Since NASA uses satellite photos to remove stations that are near bright lights I will try and figure out which of these stations (if any) are a part of NASA's official temperature record."

NASA DOESN'T COUNT THE CONTAMINATED STATIONS IN ITS OFFICIAL TEMPERATURE RECORD!

[Wondering Aloud]

Speaking as a real scientists with the real degrees and all, and as a bit of a scientific method expert. I have to say wonderful work. People like morganovich are either entirely clueless about what science is or are pulling your leg.

[MIke Smith]

Want to know how stable these temps are for determining min/max? Let's do some science, a little experiment.

How about measurements in Phoenix at the same time spanning 9 F ? There's your science buddy.
http://www.wunderground.com/cgi-bin/findweather/getForecast?query=phoenix+az

Sunnyslope (KAZSCOTT10) for yesterday:
Min
44.8 F 85% y ~7 PM
Max
73.2 F 34% ~ 3 PM
Mean:
59 F

Woopdeedo. What's that tell us?

Personal Weather Stations in Arizona.
http://www.wunderground.com/weatherstation/ListStations.asp?selectedState=AZ&selectedCountry=United+States

[James Chamberlain]

I am a skeptic and believe the data. The only other thing that I can think of that I want to make certain is not an issue is altitude. I lived in Tucson for 5 years and know the altitude can change very quickly. For the milage that you did the experiment, is the altitude relatively flat? Again, I'm not trying to disprove, just trying to take out any artifacts like, "oops, we went partly up the Mogian rim" or even smaller bits.

Keep up the good work. Atta boy!!!

[Joe]

This is a fun experiment to show the urban heat island effect, but I do agree that it's pretty meaningless otherwise. One serious problem, though, is the geography of Phoenix itself likely affects the results. While not obvious to many, Phoenix is relatively wet compared to the surrounding desert. I suspect that, say, two hundred years ago, you would find a measurable drop in temperature between central Phoenix and thirty miles out when measured at night.

[Sam Urbinto]

TCO says: But the issue is that only a very small portion of the planet is currently paved or increasingly being paved.

That is true. It is also true that these huge paved areas impact the weather patterns, not just over and around the area, but because of wind and water vapor and clouds, other places too. If we put in big cities and roads, we change weather patterns and therefore the readings that construct the climate.

So we can all agree that paved surfaces (and buildings and cars and farmland and other things that absorb sunlight and therefore heat) are part of the phenomenum of global warming, regardless of what one may think about the accuracy of the measurements used to derive the data that is used to derive the anomaly and regardless of how one may feel about the anomaly trend's appropriateness.

How much do paved surfaces contribute to either the readings and/or the warming? 0%? 100%? The closer you are to the extremes will let you know to what extent you're a denier or an alarmist. Let us say you are smack dab in the center and think this is 50%. Then the next question becomes if there's evidence the levels of the antropogenic greenhouse gasses are saturated already or not. If so, what accounts for the other 50%? Non-anthropogenic ones, or non-greenhouse gas ones, or what mix? The sun? Cloud cover? If not, when does a saturation point get reached, or is the possibility that the rise in their IR absorbing effect is infinite?

But the question is one of measuring temperatures in whatever location. The fact is that where you measure it will determine what it is. And without knowing tmax and tmin and humidity and wind speed/direction and cloud cover, you only have some number that could be put together with an enormous number of combinations. For example, if the mean is 50, is that 20/80 or 30/70 or 40/60 or 50/50 and so on? Was it dry or moist, calm or windy, cloudy or sunny? Was the wind blowing in warm air from the South or cold air from the North or nothing? Was it raining at the time? Snowing? Dust storming?

The point of all this is not to argue about what's true and what's not, it's to point out that most of the information we have neither proves nor disproves anything, it simply brings up questions to be asked and hypothesis to be tested and then proven or disproven or neither. It doesn't mean we get an answer out of it, or that the answer is knowable.

When an answer is not knowable, people will argue about it that have the opinion the answer is a certain one, and as we've seen here, in a variety of odd ways. Unfortunate we can't prove or disprove the anomaly is anything but a number that just happens to be trending up less than 1 over a measly few decades of human existance at this point in history. Or that it's not reflecting energy levels in some manner. Just as we can do neither with what the mean temperature of the air was on a given day in a given place.

[Joe]

James has a good point. If you look at the southbound graph, note that it has two data cluster, one basically north of the 202 and one south. Anyone who has driven that road knows how quickly the geography changes around that point and, as I observed, how much the humidity drops--you can feel it. (Some of this humidity is due to irrigation, but some of it is natural.)

The elevation is also increasing steadily from 1117 feet in central Phoenix to 1398 feet in Casa Grande. Going northbound, the I-17 also increases steadily to 1245 feet in Anthem and 2014 feet in New River.

[aaron]

Very nice, but the warminsta nutjobs are correct that existance of UHI isn't an issue, it's how it changes over time that matters. The IPCC says it doesn't significantly change over time, however any reasonable person knows that both the extent and intensity of UHI should increase over time as both the number of sites and regions affected increase and the intensity of human activity increases over time.

[BillBodell]

Tuukka Simonen,

You said (loudly), "NASA DOESN'T COUNT THE CONTAMINATED STATIONS IN ITS OFFICIAL TEMPERATURE RECORD!"

You are wrong. They adjust sites with suspected UHI ("city lights" in the U.S.) effect by comparing them to nearby (within 1000 km) "rural" sites and then use the "adjusted" temperatures in the temperature record. Whether or not this is a good method is a matter for debate.

[Adirian]

Scientist - yes I regard the matter as divided into sides. It's a simple spectrum - on one end, the belief that humans are the primary drivers of modern climate. On the other, those who disagree with this belief. They may believe there is another primary driver, or they may believe - as I do - that the evidence is insufficient for either claim. On can slice the question in other ways, of course, but that is the way most people slice it. And yes, you make the people who claim that humans are the primary drivers look ridiculous, for the simple and expedient reason that you are utterly incapable of arguing the actual science, and rely instead on petty linguistic arguments claiming that the blog author said things which he has repeatedly gone out of his way throughout the life of the blog to make sure people don't interpret him as having said. Which, incidentally, is why I've chosen to pick on you, rather than anybody else in here, and why I usually pick on TCO - although he occasionally sullies his hands in more tangible debate.

[Sam Urbinto]

I've been told that water vapor (from the humidity aspect) holds far less energy than oxygen and nitrogen (as you might imagine, them being 99% of the dry air mass) The example was 77 F at 70% humidity versus 103 F at 2% humidity would be a ratio of air/water at 33:1 and 200:1 (roughly).

Of course, this doesn't take into account that it takes energy for water to change state from liquid/solid/gas and back, which might be considerably even more than what air is holding.

And yes, they don't use the questionable stations, but only if you qualify that with "as is". They adjust them (unless the quick check they're doing throws out the data as outside of the realms of reality) and use them adjusted. The question is, since they don't know what the stations actually are like, (as we've found out) how they figure out how to do those adjustments.

I am not commenting on their accuracy either, because I don't know how accurate they are.

For example, look at this scenario. t_min and t_max are 30 and 70 for a day, giving a mean of 50. Same thing tomorrow. Same thing the next day.

However:

Day 1
Min 30 for 18 hours Max 70 for 1 hour

Day 2
Min 30 for 6 hours Max 70 for 6 hours

Day 3
Min 30 for 1 10 minute reading Max 70 for 1 10 minute reading

Now, what does each day tell us? For example, on day 3, we would have to know what the distribution time-wise is on the other 23 hours and 40 minutes to know if it was a nice mix between the two numbers, or if it spent more time in certain places than others. Say 90% of the day above 40 versus 10% of the day above 40.

So, ignoring that it may be the only point within hundreds of kilometers being sampled,how meaningful is using t_min and t_max to get the temperature for the day in that location in the first place?

[Larry Sheldon]

"but I think the charts were fairly compelling."

What may be most compelling is that we have another bright youngster interested in science and the method.

Good on both of you.

[dearieme]

In mathematical modelling, you'd like your approximations to be compelling, but sometimes you have to settle for their being just plausible. However, "They adjust sites with suspected UHI..effect by comparing them to nearby (within 1000 km) "rural" sites.,.": my golly I'd need some very strong evidence before I tried to persuade someone that 1000 km could ever have the physical implication of "nearby" in a geographical temperature field. Inverness is nearby to London? Oh aye!

[bill-tb]

Kudos for the experiment, it shows the facts as the facts need exposing. It's a wonder no one at the UN IPCC or USHCN thought about doing similar experiments, maybe even centered on station locations. I am sure that someone who wanted to, taking only a small portion of the billions of dollars invested in creating the hoax, could invest a few tens of millions getting to the bottom of this. It looks so hard, so maybe some alarmist researchers would be interested ... :-)

This little science demonstration has exposed the need to know, that is if you interested in real science.

I wonder if it is possible to construct a temperature profile using only properly sited stations, remove the ones that have significant heat island bias. I bet that would be telling.

[Adirian]

Warren - it says something that somebody is so critical of the experiment. If it weren't an effective statement, they'd have no reason to try to dispute it. Nobody starts arguing that volcanoes don't actually work that way regarding the vinegar "experiments," after all.

When the argument is bad, it doesn't require anything to refute it, and nobody will try.

It's when the argument is good, when it says something that people don't like, that people start trying to find faults. (Or, in Scientist's case, to fail entirely.)

[Uncle George]

Are you fed up with the Al Gore zombies?

Had enough of their sanctimonious preaching?

Tired of their "the science is settled" mantra?

Then help all the enviro-a**holes in your life put their butt where their mouth is.

Uncle George's Amazing Earth Savers

Because methane is 20x stronger than CO2 as a greenhouse gas.

If these a**holes believe we're all gonna die from global warming, help them help Gaia....get an Amazing Earth Saver for them.

Comes in a variety of sizes to fit every enviro-a**hole: From Ant/Termite tiny to Algore Huge.

www.unclegeorge4motherearth.com

[CT_Yankee]

This appears to be a high school project critiqued by people of two to five times the age of the kid. There appears to be room for improvement in technique, and it was certainly done on a limited budget.

I would like to see someone of greater age, education, and accomplishment repeat the general idea of the experiment with more refined methods. Certainly there is a college student out there looking for an idea for a thesis.

Until then, the high school student has shown his work, which is more than the IPCC has done with it's sensor error correction techniques. I can see what the boy might have done better, or differently, however I can't tell if the IPCC uses the best, or the worst, methods available, because they do not show their work.

NO legitimate scientist or scientific organization just says "trust me" and expects to be taken seriously.

I rate the boy's experiment as worthy of following up on, and the IPCC's reports as more politics than science until they are willing to release highly detailed data and methods that permit meaningful analysis of their processes. Certainly I would be more inclined to believe the word of this boy over the word of those who require profanity to express the level of hostility they feel towards an attempt by a father to spend a little time helping his son understand the scientific method.

[GeorgeM]

One extension of this experiment suggests itself. It has been postulated that the UHI bubble disappears at a certain wind speed threshold. Further runs when the wind is blowing (preferably several runs at different wind speeds) would be instructive. Also, there were some comments about the instrument location on the roof. Without a wind tunnel, it is difficult to predict where the eddies occur for a given vehicle geometry, but nearest the leading edge is usually the cleanest point. Auto manufacturers typically place their outside temperature sensor in the grille, but that data is badly contaminated when movement is slow or stopped. Reference to some model (or full size) aircraft airflow design data would be useful to guide the choice. You may know that Anthony of Surface Stations tried to make one of those runs with commercial equipment last fall, but the weather wouldn't cooperate. See his blog for details.

[tokay]

I didn't bother reading through all the comments, so forgive me if this has been mentioned before.

The temperature sensor requires some time before it reads the "real" temperature. I can't remember the term from process control right now, but with the sensors I used in the '90s, it was about 60 seconds. What that means is that you have to wait that amount of time before the sensor is in equilibrium with the system and gives you the proper reading. In other words, it's not instantaneous.

To get a more quantitative study, you have to sit on each site for the amount of time that the sensor needs to equilibrate. It might be better for a follow-up experiment to have less data points and better "quality" of data, i.e. "truer" temperature readings. That might account for some of the strange squirrelly behavior of the graphs, but it is hard to say until you get clean data.

Of course part of the science experiment is trying to figure out why the data appears the way it does and what the probable causes of errors are.

With that said, I think this is a very good experiment for a junior high student. The trend lines are clear. If there are random errors, they should cancel out, pretty much. If there are systematic errors, then all that means is that the absolute values may not be correct (for example, if the sensors are not properly calibrated) but the trend is still clear.

The main conclusion, I think, is that "urban heat effects are very likely not negligible." With a methodology in question, you can't conclude absolutely that UHI effects are necessarily significant, only that they may not be insignificant.

Right now, I'm just happy to see a critical-thinking young man getting his feet wet in experimental science. Good luck on the rest of his endeavors. :)

[Tuukka Simonen]

"By the way, my point is clearly not, as some skeptical supporters might make out, that urban heat biases in surface temperature measurement account for all historical warming. Clearly that is not true, as satellites, which do not have this urban bias problem, have measured real global warming, though at a lower rate than the surface temperature record."

At a lower rate? That's not true!

http://en.wikipedia.org/wiki/Satellite_temperature_measurements

Also a relevant point:

http://www.grida.no/climate/ipcc_tar/wg1/052.htm#2221

Full set of stations show less warming (0.65°C/century) than rural stations (0.70°C/century).

[The Loving Aunt]

I wish to thank some of you for teaching my nephew one of the best lessons to come of this discussion: how NOT to behave as an adult. I find people resort to curse words and name calling because their vocabulary and their reasoning are often quite limited. Reading through some of these comments brought to mind a preschoolers calling each other poopy heads because little joey just won't listen to little albert talk about E=MCsquared...

I'm all about a good curse word in the right situation. But please try to show my already smart and well behaved nephew that he does not need to dumb down his language to get his point across and in fact the opposite is true. The debater who can maintain his decorum without resorting to verbal fillers is more credible and less of well, a poopyhead.

[spinoneone]

I don't know if you have considered using two easily accessible examples: look at the differences between Washington Regan National Airport [KDCA} and Washington Dulles Airport [KIAD]. The data range here is aavailable for at least 25 years and will probably show an average difference of 5 degrees fahrenheit between night time lows. Dulles has been absorbed into the heat island over that period, which may also be useful for your investigation. It might also work to use Houston Hobby and Houston International. For a wider spread, New York La Guardia and New York Stewart.

[Magnus Andersson]

I have not read the whole post or all messages here, but let me just add a comment to
R.C. | February 24, 2008 at 03:54 PM.

I guess a comparison between the satellite data an GISS data between 1979 and present time is an important source.

We have a 0.1 degrees C higher increase of the GISS data this period of time. Some of that can have other sources than UHI, but a good part of it I guess is UHI.

Also the UHI correction made by the GISS team seems not to reduce the UHI effect, so the physicak UHI effect may be even smaller. Maybe 0.05 degreed C in 30 years on all stations. How many stations are city stations? If for exampel 1 station out of 20 are city stations it means an avarage 1 degrees C UHI effect for a city.

But more important in the AGW issue can be how large the UHI and adjustment effects are between 1880 and 1979. I think there are more station data adjustments which lower old data and thus increase global warming before 1979 than after 1979. Maybe we then we may have 90 years of a 0.15 degeres C UHI effect *plus* an adjustment effect from the ajustment mentioned e.g here:
http://www.climateaudit.org/?p=2793

The 0.7 degrees warming can very well be half as large, and that's what Patrick Michaels and Ross McKitrick found in this (peer reviewd) study:

http://www.spectator.org/dsp_article.asp?art_id=12492

http://www.uoguelph.ca/~rmckitri/research/jgr07/M&M.JGRDec07.pdf

This was my thoughts... Now I have to read this interresting post about this interesting experiment!!!

/Magnus

[dreamin]

"However, the folks putting the GISS numbers together are strong global warming believers. They EXPECT to find warming, so when looking at the same situation, absolutely sure in their hearts there should be warming, the site with the 2 degrees of warming looks correct to them and the no warming site looks anomalous. It is for this reason that the GISS methodology should be as public as possible, subject to full criticism by everyone."

I agree 100%. And that's why your son's experiment alone doesn't prove anything.

What you need to do is find some sort of proxy for urbanization: For example, electricity use in each precinct. Then do the experiment to measure the relationship between current urbanization and current UHI effect. Then use historical records to extrapolate backwards and come up with an estimate for how much UHI has affected temperatures in the relevant areas. Then compare with the GISS adjustment.

[Magnus Andersson]

dreamin (04:16).

I agree with your points (but what GISS people think may be impossible to prove).

But city proxies? Why can't we just skip the urban stations? Of course that should be done based on data and excellent experiments like this!

The number of city stations I think shall be proportional to the area cities cover on the earth. I think something which is easy to argue for and should be done is that all lor almost all city stations are removed.

[Flatland]

One major point about R-squared in the graph.

It's completely irrelevent because you stated in the original post that you had to take a winding route to get out of town. Actually, you can't use the linear trend either. If you want to make people concerned about R-squared happy, just make your trend line something like a 8 degree polynominal.

Considering the way you collected your data, an R-squared of 0.837 is impressive and actually might be more than enough to prove linear correlation in any statistics package. To be sure, you'd have to make the trip a couple more times to increase degrees of freedom in an analysis.

[danbo]

I have to say. Your discussion of the method and limits of your research. As well as rational was great. I wish other "real" scientist would be so open.

If your son set up this study. I think we will see great things from him. Such care is usually lacking.

I'm also sorry your son has had to deal with people who believe they understand science and that science has spoken. Such small minds. Unfortunately, those who seek truth will have to deal with backlash of dogma.

[Chris Hill]

Has anyone noticed the correlation between people who call others idiots and people who have yet to master basic logic or the English language?

[Daniel O'Neill]

Some of the discussions are pretty 'heated', so I hate at this late date to introduce noise. The rounded off value for the adiabatic lapse rate is 3.5 degrees F per thousand feet. The Valley of the Sun is a valley. On the trip north to New River (Anthem?) there was a gain of 900 feet. The UHI thing is definitely bigger than the data has been adjusted for, but not quite as big as the raw data posted above presents.

[Tony K]

Ok I am going to try to paraphrase what I think is going on here.

Points of agreement:
1. UHI effect on temperature measurement exists.
2. UHI effect on temperature measurement in an area can be large. (The science experiment demonstrated a difference of possibly 12 degrees (assuming elevation differences are not relevant, etc.) In any case, big diffences are certainly plausible.
3. IPCC has made an attempt to correct for the UHI on its thermometers.

R.C. had a really good post up above in which he quoted something from the IPCC.

The quote given by Morganovich, above (Fourth Assessment Report from the IPCC, 2007, p.244) states, "...any urban-related trend is an order of magnitude smaller than decadal and longer time-scale trends....Thus, the global land warming trend discussed is very unlikely to be influenced significantly by increasing urbanisation."

As I read this, it is not inconsistent with anything that we agree on. If it were talking about individual thermometers, it would be in conflict, but i's talking about the overall trends. I think of it this way.

Suppose I were to put in a 1 square meter patch of blackest black asphalt in my backyard. If I put a thermometer on that patch it would read much higher than the thermometer in my neighbor's yard. I could probably apply a correction factor to that thermometer to get a reasonable estimate of what the actual temperature is. But I don't think that is the point that IPCC is trying to make.

Their point when I read the above quote is more like, "That one square meter patch of asphalt is not having a significant impact on the temperature of the entire neighborhood."

Does this clarify anything?

[aaron]

Tony, you are getting it. Now, have the patch grow with time fill it with people and power sources that also increase over time.

[koolbob]

Folks:

Chill! This is only a high school level experiment that challenges an international political body. Big deal. People do this all the time, and not all of them drink ideological Kool-Aid.

You can challenge the author's conclusion, but, if you do, you have explain all the UHI adjustments in various studies since most of the UHI adjustment methodology has not been revealed (Hansen), and in some cases, the data has not been released.

Yes, I am skeptical because I look for honesty, and not revealing the methodology is misleading. We can do without that kind of science from government employees.

[Dwight]

It seems like alot of people are missing the big picture. The Data clearly shows that there is a significant warming caused by the city environment. I run this same experiment at home and easily get a 5-10 degree change from city to country. Given that, the city data is corrupted and is not usable for world wide trend statistics. Someone saying they can correct for bad data by just looking at that data and putting in some adjustment without knowing the root cause of the bad data is smoking something. If I tried to do this at work here (Major Aero Space company with tons of simulations and data collection tools)I would get laughed out of the room. So what happens to the temp trends if only the rural sites are used. I am guessing that the temp history would show very little change over time.

[Scientist]

Dwight - wrong, sorry. If you take the temperature trend from rural sites only you get the same warming. Also, if you measure sea surface temperatures, you get the same warming trend. Also, if you look at satellite measurements, you get the same warming trend. And, if you look at a map of where the world's population is concentrated, and where the measured warming is greatest, you'll see that the greatest warming is happening not in the most populated places but in some of the least populated places, like the Arctic, and the Antarctic Peninsula. These observations disprove the hypothesis that urban heat islands significantly affect the surface temperature record.

koolbob - most of the UHI adjustment methodology has not been revealed (Hansen) - maybe you've never read a Hansen paper. There is extensive discussion of the correction for all biasing effects, including that for urban heat islands. Why did you think otherwise?

[Dwight]

Sorry I cannot agree with your assessment. I have looked at the rural data and much of it is flat. You are using examples of the artic which is being driven by a totaly different phenomina. The Satellite data does show some warming in one hemisphere but not the other. I will not disagree that there has been some warming, but am not convinced it is being driven largely by Co2 concentrations. As far as correcting for urban heat bias effects, it would have to be done by generating a very detailed model of each heat island and then validating those models under all the different conditions possible. This would have to be done for each unique urban environment. I am guessing that this is not being done, If it is, please give me specific examples.
This is how science is done in the real world. We do it all the time for complex systems (although not as complex as the weather system). The very thought of many climate people expounding that the science is settled and is not up for evaluation just indicates to me they are trying to hide something. Real science does not work this way

[Chris Huston]

As a chem student, I can say this is a very well thought out experiment. The data would likely look nicer if you did average your two routes, and that would also increase the Rsquared value because the averaged data would better represent the line.

Your results were also quite good. I know this because they were actually analogous to the very study you borrowed the method from. From the abstract of the Nyuk Hien Wong and Chen Yu study:

"This study has indicated a strong correlation between the decrease of temperature and the appearance of large green areas in the city. Although there is no distinct borderline between ‘urban’ and ‘rural’ areas in Singapore, maximum temperature difference of 4.01°C was observed."

a change of 4.01 Celcius is about a change of 7.218 degrees Farenheit. From a rough look at the graph you have, you observed a temperature change of around 8 degrees. You did a fantastic job replicating this paper with less sophisticated equipment. For a school project this is extremely impressive.

That being said, it's extremely silly to make a post about how this project calls the IPCC methodology into question. For one, your results trended in a similar way to the Nyuk Hien study, so this phenomena and its magnitude are not revelations. I do honestly think it's somewhat hypocritical of you to declare "look how this simple experiment has shown up the IPCC" and once making such a major pronouncement shun away when people question the methodology, saying now it's just a kid's science project. And a good project indeed.

So what we see from this, and from data provided by other commenters, is the UHI effect is well documented, and even well documented for Phoenix, and likely for a great multitude of other stations as well. So the argument comes down to whether the studies used by the IPCC are properly adjusting for this. And that is the argument the article doesn't actually address.

First, the idea that climate studies with hidden methodology are somehow making it past peer review in mass quantities because of pro climate change bias borders on paranoia and conspiracy theories. These studies would require rigorous methodology that would be constantly questioned.

If I think about it off the top of my head, it wouldn't be that hard. I can guess what they are doing just from the previous post talking about taking information from other sites up to 1000km away: They're doing a standard curve smoothing of the data. A process used to remove noise from a data set is to make sets of adjacent data points and average them together, giving you less noise but half the data points, and thus less resolution. There are several other methods that have different levels of noise reduction and resolution sacrificed, but they all work on the same principle. In some situations that lost resolution would matter, but with climate science and all the aggregate data, I can't see it being that significant.

Another point that was made was a direct reference to the Miami station, noting the detector was moved in 93 and 96. I would like to point out on those first and most easily refutable, that when I eyeball the heat record graph, the warming trend looks like it starts showing itself in the 70's

Secondly, if you moved the device from one place to another, and we assume that urbanization is the only cause of the heat increase, you would expect an immediate jump in temperatures then a new generally flat line as the detector is in the new environment. That's not what we see. We see an increase in slope, implying constantly increasing average temperature. Unless there has been constant yearly increases in the urbanization nearby, it wouldnt' account for that kind of increase. Looking at some of the other photos from the Miami site, it's a pretty open area minus the one building it's installed in. Certainly not a place where they are building new parking lots nearby at any appreciable rate.

[Magnus Andersson]

Scientist: "If you take the temperature trend from rural sites only you get the same warming."

This is no good argument. If there is a human urban warming effect which in total is much bigger than the differences we want to measure, then only small changes in human activity is likly to mess up the result!

Secondly I must comment your sentence. You presume there is a "warming" to measure. I assume you refere to the warming from CO2. Right? But that is the hypothesis which the measuring is meant to unfold as fact. So let's use the best measurement possible! (BTW there is no warming to measure now.)

I can't see any point in using urban stations, since there are possible positive effects and no negative effects from not using them.

[Zoro]

TCO is nuts, ignore him.

[Paul Biggs]

Scientist says:

R.C. - you seem to think that the IPCC carries out scientific research. It does not. It describes the results of the scientific research carried out in the field of climate science regarding anthropogenic global warming. To find out which of your suppositions about urban heat islands is closest to the truth, try reading the literature. Don't rely on wilfully ignorant 'climate skeptics' to inform you - you might as well ask an Arsenal fan to commentate on a Tottenham game and expect to get an unbiased description.

Don't rely on the IPCC bias to inform you either:

Documentation Of IPCC WG1 Bias by Roger A. Pielke Sr. and Dallas Staley - Part I

http://climatesci.org/2007/06/20/documentation-of-ipcc-wg1-bias-by-roger-a-pielke-sr-and-dallas-staley-part-i/

Additional Evidence On The Bias In The IPCC WG1 Report On The Assessment Of Near-Surface Air Temperature Trends

http://climatesci.org/2007/06/25/additional-evidence-on-the-bias-in-the-ipcc-wg1-report-on-the-assessment-of-near-surface-air-temperature-trends/

[VermontWeatherman]

I am surprised at some of the tone expressed here, and rather ashamed as a scientist. I am glad that a few people took the time to suggest additional methods to improve the experiment, which is truly how science has always worked and moved forward.

Unfortunately, we have a very vocal group that needs to go back and look at the scientific method. Hypotheses must be tested and re-tested and re-tested. It doesn't stop when you get the information or results that you want. Ptolemy had a pretty good theory for planetary motion and the structure of the Solar System. The science was "settled". There was "consensus". Of course Kepler and Copernicus didn't let that stop them, nor did it stop with Newton, or even Einstein. It doesn't stop, ever. All of us would be better served if we encouraged young people who are interested in science, rather than berated them. Give them an example of good science, rather than childish behavior.

[TCO]

Agreed.

[VG]

Actually I'm glad you did not nip the AGW's comments. Its probably the best proof that there ain't no serious scientific analytical AGW! (and they are digging their own grave). BTW... I am a scientist MSC, PhD. These comments by these people (and of course the data) have convinced me now that there is no AGW and I believe the skeptics and critics of AGW

[VG]

Actually I'm glad you did not nip the AGW's comments. Its probably the best proof that there ain't no serious scientific analytical AGW! (and they are digging their own grave). BTW... I am a scientist MSC, PhD. These comments by these people (and of course the data) have convinced me now that there is no AGW and I believe the skeptics and critics of AGW

[TCO]

VG: there are a lot of silly scientists with "union cards".

[Mike]

You and your son are to be congratulated. I suspect he has a bright future. Much brighter than for the members of the cult who have defiled this blog.

[Brian Macker]

It's possible the UHI is stronger in Pheonix than elsewhere and that many of the stations are located in non-urban areas. So that may be why it's discounted. If indeed it is discounted in the way claimed. Which I'm not certain of either. There are plenty of other reasons why this experiment may be wrong also. Changes in altitude, valley effects, etc. can cause temp changes. For instance if you were driving down hill even slightly and over time that path were in a valley then since cool air runs down hill and collects in valleys you would get a cooling effect, especially at night.

The nasty responses here are totally uneccessary. I found the one guys attitude funny. He's feels justified in being abusive about some kids experiment because he's tired of having to correct all the people on the internet who are wrong. I saw a cartoon once that exemplified this attitude. It showed a guy at his computer saying "Sorry honey, I can't come to sleep. Someone on the internet believes something that isn't true".

[Chicken Little]

Nice work. Glad to see you understand the scientific method. You have a theory. You come up with an experiment to disprove the theory. You run the experiment. When will the global warming enthusiasts do that? Answer: Never, they are afraid they will disprove their theory.

[bobclive]

The UHI effect does exist and many scientists believe it has had a significant bias upwards on the temperature trend from the early 20th century, possibly as much as 0.4c. Weather stations at that early time were mainly rural, an upwards bias has I believe occurred as the rural stations were either closed or slowly converted to urban sites by the ongoing development talking place around them. Once the vast majority of these sites became urban as now, surly any future surface temperature measurements taken from these sites would no longer be effected by the few rural sites that remained.
Is it possible that the UHI effect worked itself out in or around 1998, could it be that the level temperature trend seen since 1998 is partly due to the fact that the majority of the weather stations are now urban and are showing reasonably accurate temps with the UHI effect now being only marginal as the IPCC state, it appears Satellite temps now agree with ground temps.
Just a thought.

[Marty]

I am not a climate scientist, but have considerable experience in modelling complex systems, specifically urban transportation models, and I have a sense of how easily one's biases get built into the results. And if one WANTS the model to show a certain general answer, how easy it is to make that happen in a way that is almost impossible for an outsider to discover.

But speaking as someone who does public policy for a living, I suggest that the technical or scientific problem here is raising such heat (couldn't resist) because some people want to rework the entire economy, at great cost, based on data that from a policy point of view is woefully weak to support any such action.

The scientific questions are real and very very complicated, well beyond the current state of the science when it comes to making decisions that will cost trillions of dollars in coming decades. It is a shame that people like Hansen are not forthcoming with their data and assumptions, because transparncy is the only way this whole matter can ever be understood well enough to support intelligent policy.

[Tan Kia Wee]

To the writer of this article:

I think you have done a good study on a limited budget and a pretty tight set of constraints. However, your conclusions do not follow from the data and methodology. For one, since both sides agree Heat Islands exist, the only two questions that follow are:

1. Does the existance or growth of UHI bias already recorded data?
2. If current data is unbiased by UHI, how much of the phenomenon can be explained using the growth of UHI.

The language you use is frankly, terrible. If you wish to raise doubt by using your data, it would be much more prudent to use language that is neutral. The following is a inflammatory quote from your article:

"The project could also be called "Disproving the IPCC is so easy, a child could do it." The IPCC claims that the urban heat island effect has a negligible impact, even on surface temperature stations located within urban areas. After seeing our data, this claim will be very hard to believe."

What exactly is the IPCC claiming? That UHI do not bias their data by exclusion or correction? That UHI do not exist (this claim would be ridiculous)? You use "disprove" very lightly, without indicating that the study has serious limitations. You have also not examined the correction methods for climate change studies. At best, your low budget study calls for more effort into this effect. You have certainly not disproved the IPCC's conclusions. This claim (of disproving the IPCC's conclusions) is misleading at best and dishonest at worst.

Please consider rewording your article. You will gain more credibility by presenting your data and conclusions clearly, without including ridiculous and unfounded hyperbole. It is a great pity that your wording destroys the credibility of your well-thought-out experiment. I would suggest removing reference to the IPCC as you have not included a critique on the mothedology of the studies that the IPCC meta analysis considers.

[mcconnell]

Lighten up Tan Kia Wee. This was a school project (and a good one considering what was done at that level of inquiry and data collecting) and not some PhD level study. At least they were very transparent in their methodologies and that changes in how data are collected could be addressed better in future studies.