02/28/2008 (11:24 am)
Warming? Debatable.
Those making the claim that we need to cut development drastically to stop human-caused global climate change have been saying “the debate is over” about three, basic items for about 20 years:
- The earth is warming
- Man is causing it
- We have to cut back development to stop it
That there’s no debate used to be almost true of the first point. It’s never been true of the other two, though it was closer about 10 years ago: a growing number of scientists dispute how much man is the cause of the warming, pointing to factors like solar effects on clouding, ocean currents, and the earth’s rhythmic cycles; and sensible people everywhere note that even the most drastic plans to cut back will not affect global temperature enough to matter, and we just have to get used to warmer weather. However, on the first point — that the earth is warming — there’s been very little serious disagreement.
Until recently.
Today, that’s on the table, too. Two things have happened:
- The earth’s measured temperature stopped rising around 1998, and this year it fell dramatically. There hasn’t been any warming for 10 years, so far as we can tell.
- It’s becoming apparent that much of the warming we’ve seen in the last 30 years has been because of bad measurements.
Temperatures Rising?
First item first. Yesterday, Hot Air noted a blog entry at DailyTech showing that all four of the global temperature tracking agencies had updated their 2007 data, and that global temperatures had fallen dramatically.
This is not surprising to some of us who have been following the reports, though, because in actual fact, global temperature readings have not risen since 1998. Take a look at this graph showing the readings of the four global tracking agencies (two by satellite, two from ground stations) since 1979. This was adjusted through a smoothing algorithm to eliminate some of the “noise,” but you can find the original graph at the link provided above:
All the measurements agree, there’s been no increase in global temperature since 1998 (all but one say we’ve cooled since 1998), and this year, we took a dive. Meanwhile, global CO2 continues to rise. Something other than CO2 seems to be driving the weather.
Stations Going Dark
Now look at the second item, that our measurements are broken. It’s not as though we can slip a thermometer under the globe’s tongue and wait five minutes in order to get a reading. Measuring the earth’s temperature is tricky. It involves taking temperature readings at thousands of weather stations around the globe, averaging them intelligently, and noting trends in the measurements. The temperatures have to be adjusted for factors that affect the readings, like urban heat island effects (cities get hotter), lighting, altitude, and such. The outcome of massaging the data depends a lot on who’s doing the massaging, and how.
When the Soviet Union collapsed between 1989 and 1992, a large number of weather stations from that nation stopped reporting altogether. The Soviet Union’s weather stations constituted a large percentage of the Arctic weather stations in the data sample, because Russia’s got a longer Arctic border than anybody else. No adjustment was made for the loss of those stations — and suddenly, around 1990, the earth’s temperature readings showed dramatic warming. The AGW crowd likes to speak of the 1990s as “the hottest decade ever.” but that might just be because of the loss of a large number of cold-weather stations. It would be like measuring the average height at the elementary school over a number of years, and suddenly one year the kindergarten class got cut.
Look at this graph(1) produced by Prof. Ross McKitrick of the University of Guelph, in Ontario:
That sharp decline in measuring stations corresponds closely to a sharp rise in temperature readings. Has anybody adjusted for the loss of stations in 1990? Not that I know of.
Millennium Bug
In August last year, science bloggers noted a sharp anomaly in NASA’s surface temperature readings around January of 2000. They had to reverse-engineer NASA’s graphing algorithm, since NASA refused to cooperate, but after reproducing NASA’s results, they notified NASA that they’d discovered a Y2K bug in the algorithm. NASA quietly adjusted their data the next posting cycle… and suddenly, 1998 was not the hottest year on record anymore.
As a result of the changes, four of the 10 hottest years on record now occur in the 1930s, whereas only 3 occur in the last 10 years.
Broken Thermometers
A recent comment on my blog by Evan Jones (you have to click on the link and then scroll down, it’s the long one at the bottom) reintroduced me to the work of Anthony Watts, whose blog site, Watts Up With That?, seems to come up every time I google climate change issues. Watts is a retired weatherman who does science for fun. Guys who do science for fun are responsible for a surprising number of epochal changes in science, things like genetics and relativity. I don’t know if Watts is in that league, but he’s making quite a stir in climate science these days.
Watts leads a citizen task force that’s evaluating just how sound our surface temperature readings have been, and what they’ve found so far is alarming: most of the weather stations were established in rural areas to avoid urban heat sources, but since around 1980, suburbia has crept out around them. They’re finding and photographing reading stations located just a few feet from the exhausts of air conditioners, located on asphalt roofs, located next to parking lots; this is in addition to the recognized urban heat island effect, which itself might be underestimated. None of Watts’ discovered problems are adjusted for by surface reading agencies. So far, the net bias of stations reviewed by the project suggests readings 2 degrees C high. Since the rise in global temperatures in the 20th century is less than 1 degree C, this bias could easily account for all the global warming we’ve been seeing… just because suburbia encroached on the thermometers.
Check out this photograph from Watt’s slide show. I’ve circled the sensor stand in the photo, or you might miss it.
The issues here are the concrete, the gravel, the cinder block building, and the boat. These apparently are all new additions since the placement of the sensor stand. As Evan Jones’ comment pointed out, all you have to do to raise the temperature of a greenhouse is to put a big rock in it; exposed masses with high specific heat, like concrete, steel, and stone, absorb solar radiation, hold it, and radiate it out slowly. Thus, building a concrete-and-steel cell phone tower and a cinder block building next to the sensor drives the temperatures from that sensor up, and keeps them artificially high.
So far, Watt’s project has surveyed only a fraction of all the weather stations, and only in the US; however, the US weather collection system is regarded as the Cadillac of the world’s systems (Australia’s may be just as good or better). Similar effects, or different biases, could be found elsewhere.
The verdict seems to be that the earth may not be warming at all. Global warming is, as of this writing, not a clearly established fact.
The past 20 years or so have seen activists attempting to shame and bully the entire world into turning political power over to them in order to stave off the destruction of the planet via the greenhouse effect. It’s a remarkable and powerful attempt, and if truth and learning survive the onslaught (which is still very much in doubt), will serve as a cautionary tale for centuries about how vigilant we must be to protect our liberty.
(1) Horner, Christopher, The Politically Incorrect Guide to Global Warming, Regnery Publishing, 2007, p. 112.
Update: Anthony Watts, whose project I highlighted in this post, responded in the Comments section and corrected a few items. Following the links in his comments, I found a much better illustration of the effect of suburban growth on weather sensor readings at the web site for his project. Here it is. The graphs are temperature readings over the past 100 years at each of the two stations, both of which have been reporting from the same location for the entire period, as reported by GISS (NASA). Notice the difference in the readings over time between the station that has been maintained properly, and the one around which growth was permitted:
« « William F. Buckley, RIP | Main | Cynical Obama Demagogues NAFTA » »
16 Comments »
Comment by Anthony Watts
You have misidentified the temperature sensor in the above photo, it is on the left side of the tower.
Here is a better photo to use:
http://gallery.surfacestations.org/main.php?g2_itemId=12970
This is the official USHCN climate monitoring station operated by the Atmospheric Science Department of the University of Arizona at Tucson.
These are people who should know how to correctly measure temperature.
Webmaster’s note: the photograph has been corrected per Mr. Watts’ suggestion. Thank you very much, sir, and my face is appropriately red. I’ve pasted his suggested photograph, which is clearer and just as dramatic, into a comment below. PW
Comment by Anthony Watts
“So far, Watt’s project has surveyed only a fraction of all the weather stations…”
We are at 500 now out of 1221. 41% If anyone wants to help us get the remaining ones, please go to http://www.surfacestations.org and sign up as a volunteer.
Thanks for the article here, it is much appreciated.
Webmaster’s note: Yep, 41% is a fraction. A big one. 
You’re very welcome for the article, and the plug for your project is welcome here.
Comment by Phil
Here’s the additional photo Anthony Watts suggested. It’s clearer than the one I used, and illustrates just as dramatic an error; note the asphalt, concrete, and cars surrounding the measuring station.
Comment by Evan Jones
urban heat island effect, which itself might be underestimated.
You see, they adjust urban stations based directly on the data of the surrounding rural stations. And if the rural stations themselves suffer from significant site violations (and, so far as observed around 6 in 7 do), well, the effect on UHI adjustment is obvious.
A true double-whammy bias.
Comment by Phil
If what you’re saying is true, Evan, this is a bias that is not being accounted for in Watts’ bias calculations. Watts’ calculation merely posits the minimum effect suggested by NOAA site violation metrics (e.g., a class 3 violation results in an error of at least 1 degree C). What you’re describing here is a separate bias occurring in the UHI adjustment.
Wow. I think the correct, technical description of this is “f***ed up.”
Comment by Evan Jones
Right; it isn’t taken into account. No one even was considering the state the great majority of the sites were in before Anthony Watts went and spoiled everything. (I imagined he, himself, was taken aback.)
And while UHI effect does partially mask the effects of site violations, they are NOT taken into account. Thus, in turn, affecting UHI. (Take a page out of the IPCC handbook and call it “positive feedback”, if you will.)
We do have to be careful about those “estimated effects”. They are, after all, only estimated. A CRN3 violation (1°C), for example, is defined merely as having a heat source closer than 100′ (30m) but not within 10m. That can obviously vary a lot.
But the fact remains that at whatever times those estimates DON’T apply, when they DO apply (or at least are most likely to apply) are at TMax, and TMin. And, you guessed it, THOSE are the times used to calculate temperatures!
And Mr. Watts tends to err on the side of caution when doing his ratings. I hear LOTS of complaints about how his photograhs and ratings are “irrelevant”. All the dang time. (HOW they yell!) But never once do I hear that his ratings are “inaccurate”. (I also see stuff he rates as CRN-3 that I would have sworn was CRN-4 or even 5.)
The sites weren’t perfect to begin with. And there DOES have to be a heat increase in order for a heat sink to “kick in” and contribute. So I am guessing that there was a bit of 20th-century warming, but that it has been exaggerated, oh, perhaps twofold.
The story could be more gruesome, though, once we wade through the NOAA and GISS “adjustment” procedures! The semi-raw data in the US seems to show actual cooling, but the adjusted soup shows a very definite warming. This is an ongoing saga, to be determined at a future time: NASA is not very helpful in their explanations.
A fellow on another blog said he had discussed with two of the biggest wigs at GISS how they adjusted the data and it was just fine. I appealed to him to let us in on the details. He flatly refused. (Names withheld to protect the guilty.)
Comment by Phil
Evan, answer a couple of questions for me, will you?
1) Who are you?
2) You say, “And there DOES have to be a heat increase in order for a heat sink to “kick in” and contribute.”
Is this true? Consider a site where a black asphalt roof is placed 10 feet from a sensor. Isn’t the temperature of the asphalt a factor of the temperature of the radiant source (e.g. the sun), rather than of the ambient air? Can’t it therefore get hotter than the ambient air temp? If you’ve ever replaced a roof during the summer, you’ve experienced how hot the asphalt gets, and unless I’m mistaken, it’s as much as 20 or 30 degrees F hotter than the ambient air.
3) Can you explain briefly how the satellite temp readings from UAH and RSS are done? What, if anything, do they have to do with the ground temp stations Watts is examining?
Thanks for all your interaction here.
Comment by Old Mike
Can I sell anyone here a few only slightly used carbon credits?
Comment by Evan Jones
1) Who are you?
Just a layman observer. Going by real name. Not a scientist. That’s the point, really. I think it is very important for laymen to understand these issues. Unlike a lot of scientists involved in this, I do not think it beyond the ability for a layman to understand them.
Furthermore, it is laymen who must (and will) determine the policy, not the experts. The experts inform us, oh, yes. But they do NOT decide for us (and perish the thought)!
2) You say, “And there DOES have to be a heat increase in order for a heat sink to “kick in” and contribute. Is this true? Isn’t the temperature of the asphalt a factor of the temperature of the radiant source (e.g. the sun), rather than of the ambient air? Can’t it therefore get hotter than the ambient air temp? If you’ve ever replaced a roof during the summer, you’ve experienced how hot the asphalt gets, and unless I’m mistaken, it’s as much as 20 or 30 degrees F hotter than the ambient air.
In a word, yes. Such radiant heat is a direct offset as you describe. When a heat sink first appears on the scene, there is a direct bias a greater or lesser degree, the same as with waste heat.
And yes, this one-time offset appears in the record and biases the record to that extent. That is–supposed to be–adjusted for, whether it is a result of a site move or a change of site environment (i.e., they build the parking lot near the station). Usually it isn’t.
But a heat sink has a far more insidious effect. Namely that (after the direct offset) it increases the TREND of any subsequent temperature change. (Even if the offset has been adjusted out.)
Take the yucky scenario you describe. If the “real” temperature rises X degrees, the thermometer will go up MORE than X degrees. Likewise, if the temperature drops by X degrees, the temperature will go down by more than X degrees (but a certain chunk of the initial warming offset remains in any event).
This is not only because of the pumping up of the T-Max (which you describe), but also because when it gets cold at night, the accumulated joules come out of the asphalt and warm the air. LaDochy et al. (Dec. 2007) point out that T-Min is affected even more than T-Max insofar as urban heat Island effect is concerned. (T-Max and T-Min are averaged to get the daily temperature.)
I will therefore clarify my earlier statement:
AFTER THE HEAT SINK IS ALREADY IN PLACE (and we hope against hope its one-time effect is adjusted for), there must be a temperature increase past that point for a heat sink to exaggerate that trend.
I emphasize that what makes a heat sink such a vile monster is that it not only produces a one-time offset, but it exaggerates the TREND ITSELF.
3) Can you explain briefly how the satellite temp readings from UAH and RSS are done? What, if anything, do they have to do with the ground temp stations Watts is examining?
Fear not. Mr. Watts is not “going ASAT”. He is only measuring ground stations.
What is interesting about satellite records is that they do not measure temperature directly, they use a microwave proxy. Sometimes interferences occur and adjustments must be applied. (Orbital decay also intrudes.) Snow areas are also hard to measure for various reasons.
But they get a much smoother “overall looksee” than the scattered, poorly-distributed, homogenized (and, we suspect, pasteurized), long-suffering surface networks.
The fly in the ointment is that (according to Christy, if I understand him correctly) the troposphere has a bit of a heat sink about it: namely that the satellite TRENDS [sic] are said to be 1.2 times greater than the ground-level trends (and 1.4 in the tropics).
This takes us hideously into the realm of calculus: It is not enough to divide the satellite temperature by 1.2 (or 1.4), oh, no. One must actually do this to the slope of the trend.
On top of that, the AMOUNT the trend ITSELF varies depending on the baseline temperature, the rate of change increasing as the base temperature increases. As I pointed out before, the tropics use a 1.4 figure, not a the 1.2 for temperate zones.
Then throw in the humidity factor (which changes the density of the troposphere and screws with the equation. So deserts skew one way and jungles another.
At or before this point, I stand well off and radio for brain support. The only weaponry at my foot-slogging disposal is a lousy Masters in History. But what is important for the layman to know is that satellite data trends [sic! sic! SIC!] are NOT the same as surface temperature trends. So it is important to know if the satellite data has been adjusted to reflect this if you want to “compare” satellite and surface records.
There is also more than a little suspicion that when satellite measurements are “off”, they are adjusted to conform with surface data. Considering the variance in how much the trend should be adjusted (as per Christy), this should come as no surprise.
The only true antidote for this would be absolute openness in regard to data and adjustment methods so that the real scientists on both sides of the debate can have a crack at it. All Data, Algorithms, Code, And Operating Manuals, PLEASE! NASA, unfortunately, is notoriously reluctant to disclose (and is not alone in this).
Comment by NevadaDad
Layman have been playing a very important role in this debate. Anthony Watts and surfacestations.org is one example. Steve McIntyre and climateaudit.org is another.
While I fully appreciate the problems inherent in just measuring surface temperature accurately, and in properly accounting for all the variables that have biased these readings both episodically and systematically, the one topic that still seems elusive is the details of the climate models themselves.
Some questions:
- How many variables are input to these models? 5? 20? More?
- How sensitive is the output (I presume it is just a single variable – T) to small variations in input variables?
- What type of error bars exist around these model outputs based on uncertainties in the input variables?
- How non-linear are the various contributions of the input variables and how accurately are the inputs known, let alone their relationships to one another?
- How senstive are these models to boundary conditions? To initial conditions?
- How have these models been validated? Can they reproduce various historical data with a high degree of accuracy? For instance, if we load these models with data from 1950-1980, are they able to accurately “predict” temperatures in the period 1980-2000?
The models used are the entire basis for the “fact” of Global Warming and its relationship to increasing levels of CO2, yet those that wield the models are like some kind of secret order of priests that will not let the “laymen” understand what’s inside the black box.
I work in a company that makes mathematical models for process control so I have a cursory understanding of these matters. I also majored in Electrical Engineering; control theory and mathematical modeling of processes aren’t entirely foreign to me. With something as obviously complex as the climate, I am wondering just what makes us think that we have models that are even partially approaching robust and have any reliable predictive value whatsoever.
We have posited that the temperature is being driven predominantly by increased CO2, but how reliable are the models that purport to show this? Can they even model the historical relationship between CO2 and temperature accurately? What contribution to other input variables have?
My list of questions grows by the day. Anybody who can chime in and talk to me about these models and their veracity would be most welcome. But I’m becoming increasingly frustrated with “it’s complicated…you wouldn’t understand…leave it to the experts” hand waving.
Comment by Phil
Steve,
I believe the inability of the climate models to accurately predict historical temperatures based on earlier, historical temperatures was a feature of a court case tried a few years ago. A citizen watchdog organization called the Center for Regulator Effectiveness objected to the use of the National Assessment on Climate Change largely because the two data models on which it was based could not outperform a random number generator in predicting US ground temperatures.
If I recall correctly, the case was settled out of court with some modifications and disclaimers to the NACC.
By the way, do you recall the huge flap about a Bush administration flunky “rewriting scientific reports?” That was this case. The Bush administration was OBEYING A COURT ORDER, something the New York Times article mentioned only in a quote from one of the authors of the NACC, and only in a dismissive tone, as though it was an excuse.
Comment by NevadaDad
Phil,
This reminds me of the now infamous Mann “Hockey Stick” chart. Somebody (can’t recall who…Steve McIntyre?) took the model and showed that it would take random data and generate a hockey stick curve. The transfer function was essentially “anything in, hockey stick out”.
How convenient.
Having taken a lot of math as an undergrad on the way to an engineering degree, including a fair bit of control and modeling theory, it makes me wonder about the mathematical fluency of the people building these models. They may know climate science, but they seem to be absolutely abysmal at building models. And even worse at validating those models. You know, little things like running more than one data set through the model to see how it performs. Or trying it out on historical data before trying to give us an extrapolation out to the year 2100.
I work in the process industries where adaptive model-based control is used along with advanced models for plant optimization…things like running the plant in real time to minimize fuel costs, product quality variation, maximize output, etc. Maybe these climate scientists ought to get familiar with people whose models actually have to work in the real world, and work good enough to make money. Or understand the concept of model validation using historical data before claiming it is ready for prime time. But then again, these faulty models the climate scientists keep giving us are probably part of the full-employment act whereby continual fearmongering becomes its own industry. Models that don’t predict catastrophy probably don’t bring in grant monies or make for good headlines and soundbites.
Oh, and the NYT failing to get it right? That’s about as predictable as the sun rising tomorrow.
Comment by Phil
Steve,
You might be interested in reading this fellow’s experience with the Australian Carbon Office:
http://ncwatch.typepad.com/media/files/D-Evans2007.pdf
He’s a professional modeler like you, and among other things he describes the “gravy train” of AGW science. Six-figure salaries for scientists… not a bad gig, you don’t really want to mess that up, especially when it would mess up your friends as well. Check it out.
Comment by Mel Brown
The best argument to be made on the fallacy of man made global warning would be to publish the following data:
1. Actual carbon as a % of atmospheric gases
2. Carbon produced by water vapor.
3. Carbon produced from human and animal
exhaleing in the normal course of living.
4. Carbon produced from wetland biodegredation.
5. Carbon produced from avrage annual volcanoe eruptions.
6. Carbon produced by burning of fossil fuels.
a. a subset of USA carbon production
from burning of fossil fuels as a
% to world man made carbon production.
7. An analysis of the overall reduction of
man made carbon dioxide if the USA cut its
carbon production by 50%.
7. The cost to the USA economy in lowered
GDP if the 50% carbon reduction was mandated.
Pingback by Plumb Bob Blog » Looks Bad for Lieberman-Warner. (Good.)
[...] The earth may not have been warming at all, and has been cooling for at least a decade; [...]
Pingback by Plumb Bob Blog » Climate State-of-the-Science
[...] the temperature stations on which NASA relies for its surface temp readings (I reported on his work more than a year ago), and noting that urban heat elements had grown up around many of them since they were first [...]
RSS feed for comments on this post. TrackBack URI
