Are surface temperature records reliable?

Temp record is unreliable

U.S. weather stations have been located next to exhaust fans of air conditioning units, surrounded by asphalt parking lots, on blistering-hot rooftops, and near sidewalks and buildings that absorb and radiate heat. 89 percent of the stations fail to meet the National Weather Service’s own siting requirements that stations must be 30 metres away from an artificial heating or radiating/reflecting heat source. (Watts 2009)

1  2  Next

Comments 1 to 50 out of 91:

1. biocab at 13:03 PM on 13 September, 2007
   The error is 2.6 K: http://biocab.org/NOAA_vs._UAH.jpg
   Response: When comparing temperature anomalies, the trend is what you want to look at, not absolute values. Temperature anomaly is calculated as the difference from a baseline average. Often different temperature datasets use different baseline periods (eg - 1960 to 1990 or 1978 to 2000). However, the trend will be the same regardless of the period. In the case of the NOAA vs UAH graph, the trends are very similar. An additional complication is there is much uncertainty in satellite data regarding correcting for long term satellite drift (see our Satellite page for more info)
2. TMLutas at 05:50 AM on 15 September, 2007
   I strongly suspect that periodic physical checks are needed to avoid error. Poor ventilation, not painting when needed, allowing vents to get partially or totally blocked, etc. can have a heating effect on these stations independent of UHI. The abortive attempt to hide the station locations and an awful lot of hot air floating around the Internet about how physical checks aren't really necessary makes me think that the universal commitment to quality is more theoretical on some people's part than real.
3. Philippe Chantreau at 08:54 AM on 18 September, 2007
   John V on CA has results using "good" stations and "bad" stations (per Watts definitions) and comes to results extremely close to GISSTEMP. GISSTEMP is closest to the curve obtained withe the "good" stations. There is a post and a link on Rabett Run, you can also go directly to CA.
   Response: Phillipe, thanks for the comment. You can find Rabett's post here plus here's a direct link to John V's graphs on Climate Audit. What I find particularly interesting is Steve McIntyre's response:

   "...keep in mind that USHCN stations have already passed one cut of quality control. They are represented as “high quality” stations. No such representations have been made for stations in China - they may be good, they may be bad, they may have had accurate records throughout the turmoil of Chinese history, they may not. I don’t know how you’d even begin to place “confidence” in the Chinese record in the absence of such analysis."

   Eg - he concedes that in spite of all those photos of air conditioners and car parks, the US stations are actually good quality. So instead, let's go pick on China instead!
4. Philippe Chantreau at 07:46 AM on 19 September, 2007
   Obviously, it is beyond unthinkable at CA that the evidence pre-existing their "scrutiny" could have any validity.
   
   Let us not forget what exactly the CA/McIntyre effort basic drive consists of, in summary: we do not like what the scientific research concludes on this issue, so we are going to review every single detail, fishing for anything that could lead in the direction that we favor.
   
   On the other hand, the actual climate research follows this basic process: study climate, by considering the physical laws governing atmospheric dynamics and their interrelations, by modeling these on supercomputers, by gathering as much data as can be obtained and carefully sorting through and analyzing that data.
   
   It is not very suprising that when the CA folks actually get into a scientific way to analyze data, their conclusions confirm the prior ones from real researchers.
5. Wondering Aloud at 06:27 AM on 7 December, 2007
   If what you say is true than McIntyre's error is in conceding that the sites in the US are good. I believe what he is conceding is that they are likely to be better or at least no worse than elsewhere.
   http://wattsupwiththat.wordpress.com/
   
   So far, with 40% of this supposedly gold standard network surveyed, 85% of sites are showing errors in the site and operations that are likely to be > 1 degree C. In other words larger than the entire GW signal to date.
   
   People promoting catastrophic warming scenarios frequently refer to graphs from this very data set to support their claims. It is clear that we just can't make a useful reconstruction of surface temperatures using these sites.
   
   I am somewhat dismayed by the idea that modeling with supercomputers is somehow climate research. While going thru and evaluating actual data and methodology is apparently not what "real researchers" do. I could have saved so much time in grad school if I had only known that computer models were real research and the actual data wasn't.
   
   Modeling is a valuable tool in science but the models are not evidence in any way of what is happening in the climate. Adding the word supercomputer does not make it science, in fact quite the opposite. By the way since Cray isn't making them anymore what makes a computer super these days?
6. Wondering Aloud at 09:04 AM on 13 December, 2007
   How do NASA and GISS claim to remove UHI effects?
   
   The term Urban heat island is probably not a good one it is a land use issue not simply an urban issue. A station in NYC central park may be just fine while a station in the middle of nowhere can be bad if it is placed on asphalt next to an air conditioner exhaust.
   
   Last month, Energy and Environment 18:985-995, published a not very kind report by Douglass J Keenan. It shows that two well known and influential papers, that are still the basis for the IPCC claim that UHI has been removed from the global climate data sets, are in fact incorrect. In fact the word used is Fraudulent.
   
   While Tom Wigley has sent me some references on sea temperature that seem pretty robust, (thank you sir) the land surface temperature measurements are in serious trouble. It looks to me like at least half of the late 20th century warming signal in this data is about to vanish. We really need a data set that is not badly contaminated, that uses sites that are properly placed and maintained, USHCN is not it.
7. Philippe Chantreau at 21:26 PM on 1 February, 2008
   It would be nice to give us a little more detail, W.A., especially considering the terminology employed (fraud). What are the papers criticized by E&E; (which is itself not a peer-review science publication)? Have the authors responded to it? This journal is far from being an objective reliable source:
   http://pubs.acs.org/subscribe/journals/esthag-w/2005/aug/policy/pt_skeptics.html
   
   This article treats of how UHI affects observations:
   http://www.ncdc.noaa.gov/oa/climate/research/population/article2abstract.pdf
   
   As mentioned higher, John V has plotted the data from the "good" sites (per Watts definition) and has found very good agreement with GISSTEMP, so they must be doing something right. It is worth emphasizing that Watts'effort concentrated on micro site effects, a different problem than UHI; nevertheless, agreement was still there in the data.
   
   I would not venture to say that climate science dispenses from going through and evaluating actual data. This RC post is of some interest as to how the UHI effect is accounted for:
   http://www.realclimate.org/index.php/archives/2007/07/no-man-is-an-urban-heat-island/#more-454
   
   I believe the post references this article:
   http://www.ncdc.noaa.gov/oa/climate/research/population/article3abstract.pdf
   
   Is that one of the 2 criticized by E&E;?
8. Philippe Chantreau at 23:38 PM on 1 February, 2008
   Another point, mentioned by John Cook on the "More on Urban Heat Island" thread, is that there is also good agreement with satellite data and weather balloon data, both immune to the micro site effects as well as UHI. If these were really that much of a factor, there would be significant discrepancies, but all the trends are consistentt.
9. Philippe Chantreau at 16:39 PM on 2 February, 2008
   I mistakenly included both UHI and microsite effects in the previous remark, UHI would still apply.
10. Wondering Aloud at 03:21 AM on 8 February, 2008
    Yes satellite nad balloon data have good agreement, but isn't the important thing here that neither of them correlate well with the surface record?
    
    I should have sited the actual papers but I thought you'd rather look yourself now I'll have to remember to dig it back out.
11. Philippe Chantreau at 15:53 PM on 9 February, 2008
    I don't see too much of a disagreement on this graph:
    http://tamino.files.wordpress.com/2007/08/global2.jpg
    The sources are:
    http://data.giss.nasa.gov/gistemp/
    http://www.remss.com/pub/msu/monthly_time_series/
12. Wondering Aloud at 07:31 AM on 13 February, 2008
    "Historic Climate Network surface stations only" That is the graphs and data sets that your post leads to Phillippe. (http://data.giss.nasa.gov.gistemp/) Exactly the portion of the data with which there is a problem. Compare to the Balloon data and the satellite data where the anomally is much smaller, and the trend much less pronounced.
    
    Yes the graphs from the surface data agree with the graphs from the surface data.
13. Philippe Chantreau at 17:37 PM on 13 February, 2008
    What are you talking about? The comparison is between GISS and MSU. You say there is a problem with USHCN and by extension GISS, this graph compares the "problem data" to satellite (MSU). Did you even look at REMSS.com? Tamino's graph puts them nicely together so as to compare the trends. I don't see any significant disagreement.
    
    You say:
    "Compare to the Balloon data and the satellite data where the anomally is much smaller, and the trend much less pronounced."
    Graphs, sources, data, links?
14. Wondering Aloud at 01:10 AM on 15 February, 2008
    I don't know what you want me to look at. The giss data is clearly surface only. From what I can actually open of the remss data which appears to be similar style data sets, it is clear that the anomaly numbers in the remss data are very much lower that the giss data. for 2007 for instance they appear to be more than a half a degree C lower!
    
    We have been discussing the surface station data. I am not familiar with all of the ways these different data sets are compiled. I am pointing out that the USHCN has clearly got problems in their data collection end.
    
    Is it your contention that the balloon and satellite data show the large anomaly that the surface stations data does? The satellite and ballon data match each other well but neither is nearly as dramatic as the "surface record".
15. Philippe Chantreau at 04:45 AM on 15 February, 2008
    That's because they are troposphere numbers. They should be like that. The trends are in agreement.
16. Wondering Aloud at 06:06 AM on 15 February, 2008
    Yeah but they are anomaly numbers, in other words how far they are from some mean. if surface anomaly is +.75 degrees are you saying it's ok that troposhere numbers are only +.25 degrees? Are you saying that it is ok that the surface is warming more than the atmosphere? That would directly falsify the entire greenhouse hypothesis.
    
    That can't be right.
17. Philippe Chantreau at 14:34 PM on 15 February, 2008
    Are you saying that it is ok that the surface is warming more than the atmosphere? That would directly falsify the entire greenhouse hypothesis.
    
    Are you really sure about that?
    
    Did you also look at what reference period is used to compute anomalies?
18. Wondering Aloud at 00:52 AM on 16 February, 2008
    Yes I am 100% sure about that... if the surface warms more than the atmosphere than the atmosphere cannot be the cause as this would violate the second law of thermodynamics, think in terms of entropy and in terms of what is known as zeroeth law.
    
    I suspect that there must be something else in there that we aren't seeing, if they are using different reference periods for their anomaly calculations then combining them in the graphs as they have; that would be amazing incompetence so I doubt that's what it is.
    
    It could be what I have suggested on other threads we ought to quit using the land surface record until we get a better handle on what the heck the problems with it are. This meets enormous resistance because the warming signal from balloon measures has been so much weaker and the satellite record is so short.
19. Wondering Aloud at 01:00 AM on 16 February, 2008
    Back to comment 3 John V's post actually suggests that the CRN5 stations are introducing a poitive bias in the surface results since 1960. He graphs it fartherr down the page.
20. Philippe Chantreau at 06:01 AM on 16 February, 2008
    How significant that actually is remains to be seen.
    
    About the satellite record:
    The T2 channel, used for troposphere measurements is influenced by the stratosphere. The T4 channel is all stratosphere.
    
    http://www.ncdc.noaa.gov/oa/climate/research/rss-msu.pdf
    http://climate.envsci.rutgers.edu/pdf/VinnikovGrody2003.pdf
    http://climate.envsci.rutgers.edu/pdf/TrendsJGRrevised3InPress.pdf
    http://www.ncdc.noaa.gov/oa/climate/research/nature02524-UW-MSU.pdf
    
    
    About balloons:
    http://www.sciencemag.org/cgi/content/short/309/5740/1556
21. Philippe Chantreau at 16:48 PM on 3 March, 2008
    Re-reading through this, it seems that there may be some confusion.
    "if the surface warms more than the atmosphere then the atmosphere cannot be the cause as this would violate the second law of thermodynamics,"
    The surface is not really the surface as in the surface of a spheroid. Surface temps measurements and estimates are rather the lowest troposheric temps and should be thought of that way. Sea surface temps would probably correspond better to the idea of surface as you use it in you thermodynamic view. But any AIR temperature can not be considered as surface that way, it is always atmospheric, even if it's 2 cm off the ground.
    
    "if they are using different reference periods for their anomaly calculations then combining them in the graphs as they have; that would be amazing incompetence so I doubt that's what it is."
    Actually that's exactly what it is and I don't know who you mean exactly by "they." This graph:
    http://tamino.files.wordpress.com/2007/08/global2.jpg Is a compilation by Tamino to show agreement in the trends and agreement does show, in spite of having different time periods for anomaly computation.
    The reference period for GISS is 1951-1980. Obviously, the satellite record can not use this same period. Satellite records use 1979-2000, during which average temps were already higher, making warm anomalies smaller than those seen on GISS. It is not incompetence to represent these on the same graph, so long as we know what we're looking at. In fact, it is a good test of the true trend. Incompetence would lie rather in the ignorance of the difference or using the graph for interpretations that ignore these differences. Putting these on the same graph to verify identical trends is not incompetence.
    
    One last thing: satellite measurements are, in fact, lower troposphere measurements, a sizable layer of atmosphere, and even the T2 channel includes a strong stratospheric influence. The papers cited higher give some details on that.
22. Wondering Aloud at 03:21 AM on 5 March, 2008
    Good job Phillipe that makes much more sense. That part is a baseline problem then and not a totally dumb one.
    
    Another way of saying 1979-2000 was already warming though is to say 1951-1980 was the coldest stretch in a century. Too bad GISS used this for a baseline but what can you do about that.
    
    Excellent first paragraph too. Though there is a lot of debate about where warming should be greatest vs where it is the greatest. I'm not sure who is right there. I think between the two things you took care of the problem I had with the anomaly numbers.
    
    Hurrah.
23. Philippe Chantreau at 03:53 AM on 5 March, 2008
    Glad to help.
24. Philippe Chantreau at 14:07 PM on 7 March, 2008
    And just as a reminder, it does not make any real difference what the ref period is.
25. Mizimi at 02:08 AM on 8 August, 2008
    According to WikiP. there are around 4000 stations around the world that are used for collecting data. Some are well maintained and calibrated. Some are not. From a fair number of stations the data does not arrive for incorporation at the right time, or sometimes not at all. So the data set is 'adjusted'.
    On a scale of 1 - 10 how would we rate the accuracy of this data source? And how reliable does this make any model we try to construct?
    Most of the stations are land-based and the sea based ones limited to particular sea routes; this means we have less data about sea temps then land temps...despite the sea being somewhat bigger.
    What skew does that put on any resultants?
26. Mizimi at 05:17 AM on 11 August, 2008
    On the subject of stations, I note that the number of stations used for data collection has dropped dramatically from 1990 'til present day. Coverage in (what was) the USSR and China has virtually disappeared.
    Coincident with the fall in stations the GMTemp has apparently risen.(???any connection here???)
    Look at a map of the current station locations and then tell me they are providing data that can be seriously used to construct a global model.
    Yes satellites provide additional cover but only during their overpass which is limited. Yes, their instrumentation is more accurate than land-based stations, but there are too few of them, so their 'correcting' effect on the overall dataset is diluted.
    The earth has around 510 million sqkm in surface; 150 land and 360 water. The vast majority of stations are land based and with around 4000 in use that works out to a station roughly every 38,000 sqkm. To try and model from that low level of distribution would be rejected by most reasonable people. The fact that most of these stations are actually concentrated in a much smaller area leaving HUGE areas un-monitored simply makes the data collected even more worthless for constructing any realistic model.
27. Mizimi at 01:29 AM on 12 August, 2008
    Sorry, forgot to post the sites....Look at:
    
    http://science.nasa.gov/headlines/y2000/ast21jul_1m.htm
    
    http://climate.geog.udel.edu/~climate/html_pages/ghcn_T_stn.html
    
    Note the closing comments......
    
    "The improved temperature record will guide efforts to refine computer models of the world's climate so that the behavior of the models more closely resembles the observed behavior of the atmosphere.
    
    Current models suffer from several shortcomings.
    
    For example, clouds are not well represented by the models. The resolution of current models is too coarse for features as small as clouds, Spencer said. Yet clouds clearly play a crucial role in climate due to their influence on humidity, precipitation and albedo (the percentage of solar energy reflected back into space as light).
    
    "The role of clouds is still regarded as one of the biggest uncertainties in global warming predictions," Spencer said.
    
    The ability of plants to remove carbon dioxide from the atmosphere and the role of soils have only recently been added to the models, and scientists aren't confident yet of how the models portray these factors, Spencer said.
    
    "While we know that vegetation takes up some of the carbon dioxide we generate from burning of fossil fuels, how that sink of carbon will change in the future is still pretty uncertain," Spencer said.
    
    Climate models are also limited by the computing power available.
    
    "The global models would be much better if computers were much faster," Spencer said. "Instead, a lot of approximations are made to make the models simple enough to do climate simulations over the whole globe.
    
    "Unfortunately," Spencer continued, "we know that many of the processes that are crudely represented are quite non-linear, and so have the potential to respond in unexpected ways."
28. theTree at 18:35 PM on 10 September, 2008
    Someone just suggested to me (from a book called 'Heat' I believe) that the earth has taken millions of years to store the potential energy of oil, gas, coal etc... and we're releasing it in a few hundred years so its bound to have an effect. Is this credible? How does this release compare to volcanic activity for instance? Opinions please! And apologies if this is the wrong thread for such a question.
    
    Just to tack another thought onto this: as people generally have an overriding opinion on AGW, do you think the multitude of factors and questions such as the one I've asked above are generally explained to support one's own 'overriding opinion'? And further, how many factors and questions would it take for someone educated in this field to 'change' their overriding opinion?
    
    I openly admit that I plead ignorance before I plead an opinion! The back and forth on this subject is dizzying.
29. Mizimi at 01:12 AM on 12 September, 2008
    theTree:
    Check out Wikipedia: It is (guess)timated that around 14 terawatts of heat is released from the earth's core through tectonic/vulcanic activities, around the same amount of energy that we currently consume.
    
    Science is about facts, not opinions. Opinions are shaped by the kind of person you are and you will find a lot of people will deny facts because they do not fit 'their' model of reality. That's why we need science, not opinion, not emotional hype, not fear induced reactions to an un-proven hypothesis.
    Science enables us to respond rather than react.
30. Quietman at 08:27 AM on 18 September, 2008
    theTree
    Yes it does have an effect, producing a false feedback through CO2 release which is a GHG. The argument on CO2 is climate sensitivity. Hansen claims a high sensitivity while Spencer claims a low one. The results thus far indicate Spencer is scientifically but not politically correct.
31. chris at 10:59 AM on 7 October, 2008
    Re #30-32
    
    These questions have been addressed fairly conclusively by the science.
    
    (i) You are correct (Tree) that the earth took many millions of years to sequester atmospheric CO2 in the form of fossil fuels (oil, gas, coal, shale and so on). Around 4000 billion tons of carbon is "stored" in this manner, and it's taken around 600 million years to do this.
    
    In the last 100 years we've released around 500 billion tons of this carbon back into the atmosphere, of which around 200 billion tons has remained there (around 300 billion tons has been absorbed by the oceans and terrestrial environment).
    
    see for example:
    
    http://www.physicalgeography.net/fundamentals/9r.html
    
    
    (ii) It's very clear that volcanic activity is on a miniscule scale with respect to our massive release of carbon dioxide. It's easy to demosstrate this. If one examines the high resolution atmospheric CO2 record over the last 1000 years, for example, one can see that the atmospheric CO2 levels remain rather constant over the period up to around the mid to late 19th century and rise massively in response to our emissions. The absence of significant activity from volcanoes can be observed by the absense of jumps in the atmospheric CO2 record as a result of the truly hummungous volcanoes of the last 1000 years (e.g. Santorini, Krakatoa and Tambora). Volcanic activity results in the release of something a ggod bit less than 1% of our current industrial emissions.
    
    see, for example, the high resolution atmospheric CO2 record compiled on page 3 of the IPCC summary for policymakers:
    
    http://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-spm.pdf (see page 3)
    
    
    (iii) Heat from tectonic activity is trivial with respec to greenhouse gas warming. This is one of those fallacious "arguments" that is doing the rounds! The practicioners avoid the three pertinent points. These are:
    
    (i) is there any evidence for enhanced tectonic activity during the period of very large warming (especially last 30-odd years)? After all tectonic activity has been occurring for millions of years. Has it suddenly intensified? Evidence please!
    
    (ii) how can it be that the areas of which major tectonic activity show little match to areas of temperature increase? For example Iceland is one of the most tectonically active regions on earth. However it is one of the few places on earth that has undergone a tiny bit of COOLING during the period of global warming:
    
    e.g. data on the scale and location of Arctic warming over the last 50 years from the Colorado University Arctic research center:
    
    http://arctic.atmos.uiuc.edu/CLIMATESUMMARY/2003/IMAGES/annual.1954-2003.tchange.png
    
    (iii) the heat released by undersea tectonic activity is around that of the geothermal background. This is around 0.1% of the heat energy from solar/greenhouse activity.
    
    e.g. according to Jeff Severinghaus of the Scripps Institution of Oceanography
    
    "... the average heat added from volcanoes to the ocean is of order 0.1 Watt per square meter. But the heat added (or removed) to the ocean from the sun and atmosphere is of order 100 Watt per square meter. So it is very hard for volcanoes to compete."
    
    So it's not just a question of showing that tectonic activity on the ocean bottom is significant with respect to warming (the evidence indicates it isn't), but of showing that this activity has increased in the last several decades to an extent that can have contributed to warming (the evidence indicates that it hasn't)...
32. Mizimi at 05:51 AM on 27 October, 2008
    Chris: "If one examines the high resolution atmospheric CO2 record over the last 1000 years,"......
    
    What highres records are we talking about please?
    Paleoproxies?
33. chris at 08:50 AM on 28 October, 2008
    Re #34
    
    No, not paleoproxies. That's clear from the data I linked to:
    
    http://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-spm.pdf (see page 3)
    
    
    The atmospheric CO2 record is the directly measured atmospheric CO2 either in the atmosphere (from the many sites around the world and the continuous record from Manua Loa since 1959), and trapped in bubbles in ice cores extending back many 100's of thousands of years, but at a high resolution extending back 1000 years:
    
    e.g. D. M. Etheridge et al (1996) "Natural and anthropogenic changes in atmospheric CO2 over the last 1000 years from air in Antarctic ice and firn" J. Geophys Res. 101, 4115 -4128.
    
    and later extended to 2000 years:
    
    CM Meure et al (2006) "Law Dome CO2, CH4 and N2O ice core records extended to 2000 years BP" Geophys Res. Lett. 33 Art. # L14810
34. Mizimi at 05:49 AM on 29 October, 2008
    Thanks Chris; somehow I missed the link.
    
    This thread is about temperature records and how reliable /accurate/representative are they.
    CO2 levels are assumed to vary only slightly due to effective atmospheric mixing, but this is very different from temperature which has much greater variation. Given the paucity of temperature recording stations I cannot accept that the data used for models is sufficiently representative of the global condition, and thus the resultant of the model is questionable.
    Even satellite records are questionable as recently demonstrated by the modification needed to the attitude correction algorithm.
35. chris at 08:56 AM on 1 November, 2008
    Re #36
    
    that's good...you agree that the atmospheric CO2 levels vary only slightly due to effective atmospheric mixing. You say that this is "assumed", but of course, as we both know [http://www.skepticalscience.com/co2-measurements-uncertainty.htm], this isn't an "assumption" at all..it's a real world observation [so long as we are careful to make CO2 measures in isolated locations and average over the relevant timescales for mixing (yearly averages are appropriate)].
    
    
    There's still a few problems with your post:
    
    (i) Temperature data isn't "used" for models
    of course. And so the "resultant" of the models isn't in any way "questionable" in relation to the temperature data which is an entirely independent data set. Model output (as predictions or hindcasts) might well be compared with the real world temperature....but that's another matter altogether.
    
    (ii) Notice that one doesn't need a huge number of "temperature recording stations" to assess changes in global temperature. Remember that the aim is not to determine the Earth's "average temperature" or "global temperature". These are terms with little meaning (after all the Earth's average sea level surface temperature will differ from the Earth's average 200 metres altitude temperature and so on). The Earth's temporal temperature evolution is determined as a change in the "temperature anomaly", which is the change in temperature in single locations averaged over a very large number of locations. Thus temperature stations at a whole range of locations and altitudes provide valid data sets. On similar lines, the fact that there is a strong correlation between temperature anomalies over large distances (100's of kilometres) means that the whole Earth doesn't need to be minutely sampled. Obviously we couldn't assess absolute global temperatures in this manner. But we're not assessing absolute global temperatures. We're assessing the change is absolute temperature at single locations and averaging these changes.
    
    So one needs to be clear about what the surface temperature anomaly means and how this is determined before attempting to trash it! [you might read the relevant descriptive papers here [*****]. Notice that in relation to the subject of this thread, the Earth's temperature anomaly progression under the influence of a marked 20th (and especially late-20th) century warming is essentially unchanged if the entire set of urban stations is omitted from the analysis.
    
    [e.g. Hansen et al (cited below) state in an analysis of urban heat effects that: “These examples illustrate that urban effects on temperature in specific cases can dominate over real climate trends. Fortunately, there are far more rural stations than urban stations, so it is not essential to employ the urban data in analysis of global temperature change.”]
    
    So the "urban heat island effect" is somewhat of a red herring (or a stalking horse) in the context of global temperature anomaly measures.
    
    [*****] Hansen et al (1999) GISS analysis of surface temperature changes J. Geophysical Res (Atmos) 104, 30997-31022
    
    or (for the Hadley analyses):
    
    Rayner NA et al (2003) Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century J. Geophys. Res. (Atmos) 108 (D14): Art. No. 4407 JUL 17 2003 etc. etc.
    
    
    (iii) Of course the proof is in the pudding. We've observed a large warming, especially of the high Northern latitudes (as predicted by models) with large attenuation of Arctic sea ice....we've observed large scale retreat of mountain glaciers....we've observed increased concentrations of atmospheric water vapor in response to atmospheric warming much as predicted ......we've observed widespread increases in the sea surface temperature...and so on.
    
    In fact it's possible to leave out direct surface temperature measures and construct a completely independent temperature scale by analysis of the record of mountain glacier retreat:
    
    e.g. J. Oerlemans (2005) "Extracting a climate signal from 169 glacier records" Science 308, 675-677.
    
    And as John Cook outlined in his top post, there are many other indicators of rising surface temperatures that are independent of direct temperature measures.
36. rob at 11:36 AM on 18 November, 2008
    Chris said,
    
    Notice that one doesn't need a huge number of "temperature recording stations" to assess changes in global temperature.
    Now I am glad she mentions this, there are enough rural sites certainly in the US that will give complete coverage, these rural sites show NO significant warming since 1900. Because of this problem (no warming) Hanson/Giss use over 1100 US weather stations many urban so they can then manipulate the raw data to push their cause.
    
    http://www.youtube.com/watch?v=LcsvaCPYgcI
    http://www.climateaudit.org/?p=1859
    http://www.climateaudit.org/wp-content/uploads/2007/08/peters27.gif
    
    Chris says,
    
    So one needs to be clear about what the surface temperature anomaly means and how this is determined before attempting to trash it! [you might read the relevant descriptive papers here [*****]. Notice that in relation to the subject of this thread, the Earth's temperature anomaly progression under the influence of a marked 20th (and especially late-20th) century warming is essentially UNCHANGED if the ENTIRE SET OF URBAN STATIONS IS OMITTED from the analysis.
    Fortunately, there are FAR MORE RURAL stations than urban stations, so it is NOT ESSENTIAL to employ the urban data in analysis of global temperature change.”]
    So the "urban heat island effect" is somewhat of a red herring (or a stalking horse) in the context of global temperature anomaly measures.
    
    I SAY PERHAPS CHRIS SHOULD TELL HANSON THAT.
    PERHAPS CHRIS CAN INFORM US WHY HANSON USES STATION PAIRS INSTEAD OF USING THE PRISTINE RURAL STATION DATA ALONE THAT IS READILY AVAILABLE OVER THE ENTIRE GLOBE.
    
    Station pairs disguise the actual temperature, by suggesting in flawed studies that there is little UHI effect, try London 9 degree C difference and every other major cities and towns on the planet.
    Google population growth and the UHI effect, 9 degrees C is NOT high.
37. chris at 08:33 AM on 21 November, 2008
    ummmmm Rob..
    
    Since I am quoting Hanson directly from one of his papers, I don't really need to tell him anything.
    
    Urban areas are generally warmer than the surrounds. Therefore one either eliminates urban areas from the record to establish the Earth's surface temperature evolution, or one corrects the data from urban stations by reference to local urban stations.
    
    However one does this (leaves out the urban stations or corrects these) the Earth's surface temperature anomaly is the pretty much the same.
    
    You would benefit from reading John Cook's article on urban heat island effect.
38. beam at 04:48 AM on 28 November, 2008
    I'm amazed at the rhetoric. If you lose the validity of the surface temperature record, your hypothesis, that warming is greater than expected, is invalidated. It is a logical fallacy to claim that it doesn't matter if one of your premises are proven false. It is impossible for the conclusion to be correct if the premises do not hold.
    
    Here in lies the crux of the problem. --------------------------------------------------------- So If I'm not mistaken this is the AGW Hypothesis:
    1. The world has been warming for a century, and this warming is beyond any cyclical variation we have seen over the last 1000 or more years, and beyond the range of what we might expect from natural climate variations.
    2. Almost all of the warming in the second half of the 20th century, perhaps a half a degree Celsius, is due to man-made greenhouse gases, particularly CO2
    3. In the next 100 years, CO2 produced by man will cause a lot more warming, from as low as three degrees C to as high as 8 or 10 degrees C.
    4. Positive feedbacks in the climate, like increased humidity, will act to triple the warming from CO2, leading to these higher forecasts and perhaps even a tipping point into climactic disaster
    5. The bad effects of warming greatly outweigh the positive effects, and we are already seeing the front end of these bad effects today (polar bears dying, glaciers melting, etc)
    6. These bad effects, or even a small risk of them, easily justify massive intervention today in reducing economic activity and greenhouse gas production [1] http://www.conservapedia.com/AGW_hypothesis ---------------------------------------------------------
    
    In order for this to be proven true at this point in time, the surface temperature record needs to be accurate, because the other forms of temperature data collection have not been around long enough to be relied on. We simply do not have upper atmospheric temperature measures for long enough to see any long term trends. Let alone trends that are not expected. This is also true of the surface temperature record, although it is slightly older.
    
    Let's put it into perspective, if we scaled earth's total existence in time to a period of 1 year, the 50-100 years of data collection we now have would still be a fraction of a second on that time scale.
    
    So... claiming that you "Don't need" the temperature record is simply an act of hand waving by those too stubborn to admit defeat. At least for now, there is more work to be done.
39. chris at 21:33 PM on 29 November, 2008
    Yes, don't allow yourself to be taken in by "rhetoric" beam! In science it’s all about the evidence.
    
    I'm sure nobody would suggest that it doesn't matter if we "lose the validity of the surface temperature record". I've had a look through the thread and haven't found any post which claims that, let alone "claiming that you "Don't need" the temperature record"...that would be an odd claim indeed!
    
    Notice that in order to take action in response to real world observations we don't need "proof". Proof is a mathematical/philosophical concept. What we need is strong evidence.
    
    So the pertinent question is: "is there strong evidence that the temperature record is robust to the extent that we can reliably assess the Earth's temperature response in relation to our understanding/predictions of massive enhancement of greenhouse gas concentrations".
    
    The answer is yes I suspect we would agree for some of the reasons already outlined on this thread:
    
    (i) The record is independently assessed by three different organizations. Although there are differences in data compilation/analysis methods and some differences that relate to the nature of covering sparsely-monitored regions, the different compilations yield a consistent interpretation of the surface temperature evolution over the last 100 and a bit years.
    
    (ii) the surface record seems not to have significant contamination from the UHI since (a) a number of direct analyses indicate that the UHI isn’t significant [comparison of temperatures on windy days (with rapid excess heat dispersal) cf calm days, and other types of analysis, for example as described here: http://www.skepticalscience.com/Does-Urban-Heat-Island-effect-add-to-the-global-warming-trend.html, or in John Cooks introductory summary on this thread]; (b) one can remove all of the urban records from the analysis, and the temperature profile is pretty much unaffected; (c) those regions showing the largest warming are far, far away from urban centres and generally there is no correlation between local temperature evolution and local urban density [see for example: http://www.skepticalscience.com/urban-heat-island-effect.htm]
    
    (iii) completely independent records of the consequences of a warming Earth are consistent with the surface record [these include high latitude ice recession; independent temperature scales constructed from the record of high altitude glacier recession; tropospheric warming; enhanced tropospheric absolute humidity and so on].
    
    So the evidence supports the interpretation that the temperature record is robust.
    
    Your point about scaling of the record with respect to the Earth’s “total existence” isn’t an important comparison with respect to the question of the consequences of massive enhancement of the atmospheric greenhouse gas concentrations at this particular time in the Earth’s long history. In any case we have a huge amount of information about temperatures in the recent and much more distant geological past. This also informs our understanding and provides strong evidence in support of the expected surface warming response to enhanced greenhouse gas concentrations. For example there is a good correlation between atmospheric CO2 concentrations and the Earth’s “temperature” in proxyCO2 and proxytemperature data stretching back right through the Phanerozoic
    
    So in general, the paleorecord reinforces the data from our contemporary temperature record and all of the vast amount of information from understanding of basic atmospheric physics, to the spectroscopy of greenhouse gases, ice core records and so on and on, that informs us on the consequences of massive enhancement of greenhouse gas concentrations.
    
    That’s not to say that there isn’t much more work to be done!
40. Mizimi at 05:57 AM on 2 December, 2008
    Chris: we cannot adequately model climate right now, let alone model climate 'that was' millions of years ago. We assume land mass distribution, oceanic currents, atmospheric conditions and movement, biomass activity etc and then use proxy records to pin down climatic conditions. The best we can actually expect is an intelligent estimate of trends.
41. chris at 08:27 AM on 3 December, 2008
    We're not talking about modeling Mizimi. We're talking about measuring.
42. Mizimi at 04:31 AM on 7 December, 2008
    Your quite right Chris; apologies.
43. Bruce Frykman at 09:32 AM on 19 March, 2009
    I need clarification on what is meant be "average" in the context of temperature station readings. What are we interested in here daily averages, monthly averages, hourly averages etc. Ie is "average" simply some arbitrarily collected(high+low)/2 or are we talking about some fully intergrated average. Just what is the standard?
    
    Another question is what does science say about the type of average that is used to calculate the presnet state of the earth's "fever" (referring to your spokesman Al Gore's expression for it.)
    
    Do we do RMS averages, a simple arithemetic mean, or is a modal or median average most appropriate in determing the state of the this "fever"?
    
    Maybe can we mix them all up and wave some abra-cadabra (correcting the data) over it and voila - the current state of the earths fever is determined to 5 decimal points.
    
    How does science deal with migrating weather stations? If I decide to place 500 weather stations in Arizona next year and call them "official" will the USA develop a strong "global warming" signal or just Arizona? or doesn't it matter at all? Conversely if I fund 10,000 weather stations in Siberia can I cool the planet's present fever?
44. Mizimi at 03:16 AM on 17 April, 2009
    #45
    
    Unfortunately most of the weather stations in Siberia have been shut down and if you look up the current distribution of weather stations globally they are distributed very unevenly...the highest density being in the USA. Many other parts of the world are not 'thermally' represented so any global mathematical average (however it is derived) is going to be wrong.
    
    Satellite measurement has been around now for only 30 years so whilst we have a more even distribution of data (not necessarily more accurate) the data series is too short for any predictive climate modelling.
    Interestingly, the current series of satellite temperature data shows a clear cooling trend since 2002 despite increasing CO2 levels.
45. HaydenK at 15:18 PM on 20 May, 2009
    That is a nice article!That's because they are troposphere numbers. They should be like that. The trends are in agreement.Well, it is not new to me. But aside from that breaking news, are you familiar with Victoria Gotti? Victoria Gotti is in trouble. She isn't getting arrested or anything, but the mortgage on the castle Victoria Gotti lives in is about to get foreclosed on – putting the Mafia princess out of her home. Daughter to criminal royalty, John Gotti, the onetime head of the Gambino crime family, she married a mobster herself (surprise!) and they were able to purchase a lavish home close to Long Island. She is $650,000 behind, so she isn't likely to be able to keep it, and that is definitely out of the reach of quick payday loans. However, with a mortgage that size there will probably be no mortgage loan modification in the future for Victoria Gotti
46. Frank Schnabel at 21:53 PM on 9 October, 2009
    Greeting all,
    
    When some outfit like Hadley or GISS offers an estimate of the global mean temp for a given year, do they present along with it an error estimate? e.g. In 2008 the average temp was 25 degrees C + or - 5 degrees. Seems like they would have to, given all that goes into coming up with an estimate. How do they assess the range of error, and how much confidence can we place on such estimates?
    
    cordially
    
    Frank
47. Marcus at 12:50 PM on 8 January, 2010
    Wondering Aloud, would Energy & Environment be the same "journal" that published that awful "paper" by Beck-the one claiming that, based on measurements of the day, CO2 levels in the 19th century were higher than in the 20th century? When you look at the results used by Beck, you see they have *massive* margins of error (around +/-200ppm) because most of the samplings were taken in urban environments, were measured without internal controls &/or were measured with equipment with sensitivities 10-100 times worse than modern equipment. Yet E&E; still accepted this paper, even when no other respectable journal would touch it, because it fit into their ideological agenda. I suspect the same is true with their "analysis" of monitoring stations.
48. Marcus at 13:28 PM on 8 January, 2010
    WA. 1951-1980 was not the *coldest stretch in a century*. The 30-year average of 1901-1930 was a good .24 degrees *colder* than 1951-1980-& is by far the coldest 30 years of the the 20th century. 1931-1960 average is almost identical to 1951-1980, & the 1961-1990 average is slightly warmer than either 1931-1960 or 1951-1980, so I don't see why you think GISS is in error for using this period as a base-line. The satellite data is from 1979-2000 because 1979 represents the start of the satellite measurement period.
49. Marcus at 13:59 PM on 8 January, 2010
    Mizimi, I have to tell you that your claims regarding the Siberian stations is just plain wrong. If you can provide *proof* of this wild allegation, I'd be happy to entertain it though. As for your claim regarding the satellite data post-2002. Even if what you said was true, 2003-2009 represents too small a sample group to be of any real benefit for determining trends-especially when that period is dominated by abnormally low sunspot numbers. That said, if we look at temperatures from AMSU-A for 1st July, we see that 2003 was -14.42, 2004 was -14.19, 2005 was -14.13, 2006 was -13.94, 2007 was -13.86, 2008 was -14.39 then 2009 was -14.11. This is only a single date, of course, but a look through the entire year shows that, if anything, the trend has been towards *warming* between 2003-2009, not cooling as you claim. As I said, though, picking on these dates-to prove either case-is the worst kind of cherry picking.
50. Marcus at 18:31 PM on 8 January, 2010
    What galls me is this-in spite of the fact that Dr Roy Spencer (of University of Alabama, Huntsville) has been shown that his adjustment for diurnal variations is in error, as far back as 2005, he still persists in using this method, which generates a long-term trend of +0.035 per decade lower than what is should be. If this error is accounted for, the long-term trend for UAH satellite data is around +0.2 degrees per decade (the same order of magnitude as RSS, which was around +0.23 degrees per decade). What's interesting is that *both* of the satellite data-sets give values which are *higher than for surface-based measurements (+0.187 degrees per decade).
    Another point is this, why do skeptics refuse to pillory Dr Spencer (a skeptic) for his errors, yet are quick to attack any apparent errors made by climatologists with CRU or GISS? That smacks of hypocrisy to me!

1  2  Next

Post a Comment

Political, off-topic or ad hominem comments will be deleted. Comments Policy...

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.